Methods in

Molecular Biology 2032

J. Philip McCoy, Jr
Editor

Immuno-
phenotyping
Methods and Protocols
METHODS IN MOLECULAR BIOLOGY

Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, UK

For further volumes:

http://www.springer.com/series/7651
For over 35 years, biological scientists have come to rely on the research protocols and
methodologies in the critically acclaimed Methods in Molecular Biology series. The series was
the first to introduce the step-by-step protocols approach that has become the standard in all
biomedical protocol publishing. Each protocol is provided in readily-reproducible step-by-
step fashion, opening with an introductory overview, a list of the materials and reagents
needed to complete the experiment, and followed by a detailed procedure that is supported
with a helpful notes section offering tips and tricks of the trade as well as troubleshooting
advice. These hallmark features were introduced by series editor Dr. John Walker and
constitute the key ingredient in each and every volume of the Methods in Molecular Biology
series. Tested and trusted, comprehensive and reliable, all protocols from the series are
indexed in PubMed.
Immunophenotyping

Methods and Protocols

Edited by

J. Philip McCoy, Jr
Frederick, MD, USA
Editor
J. Philip McCoy, Jr
Frederick, MD, USA

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-4939-9649-0 ISBN 978-1-4939-9650-6 (eBook)
https://doi.org/10.1007/978-1-4939-9650-6

© Springer Science+Business Media, LLC, part of Springer Nature 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been made.
The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer
Nature.
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.
Preface

Immunophenotyping has become more complex in recent years due to the increasing
availability of new fluorochromes with novel spectral characteristics and the development
of instrumentation with expanded capabilities. Propelled by technological advances in
cytometry as well as the development of commercially available monoclonal antibodies,
immunophenotyping of a multitude of cell types using flow cytometry has become com-
monplace in virtually all academic research centers, pharmaceutical companies, and clinical
laboratories. Thirty-five to forty years ago, immunophenotyping was limited to one or two
colors while today it is possible to construct panels consisting of several dozen fluorescent-
conjugated markers. Similarly, analysis of flow cytometric immunophenotyping has devel-
oped beyond real time, direct observation of electronic signals represented by one- or
two-parameter histograms viewed at the cytometer. Today, data acquired in list mode fcs
format files allows for remote, retrospective, highly complex, and sometimes automated,
data analysis performed independent of the cytometer and shared beyond the laboratory.
Additionally, newer cytometric technologies such as imaging flow cytometry and mass
cytometry permit immunophenotyping in conjunction with morphologic studies or with
an expanded number of parameters, respectively. The following chapters will present a
number of these topics as well as newer developments in the field of immunophenotyping.
Today, it is virtually impossible to present a complete and comprehensive compendium
of all immunophenotyping methods and applications in one volume. Therefore, this text will
present a representative collection of immunophenotypic methods and applications with
examples of newer technologies and reagents used in the research and clinical environments.
Basic methods in immunophenotyping such as construction of high-dimensional fluores-
cence and mass cytometry panels; fluorescence barcoding; and use of dried or lyophilized
reagents will be presented first, followed by immunophenotyping examples of specific cell
types often studied in research laboratories. The final chapters present immunophenotyping
topics useful in the clinical laboratory, including a chapter on the critical role of quality
control and immunophenotyping in the clinical environment.
I would like to thank each of the authors for volunteering to take the time to share their
expertise for this volume. Without their hard work and dedication, it would not have been
possible to share these methods with you.

Frederick, MD, USA J. Philip McCoy, Jr

v
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
1 High-Dimensional Immunophenotyping with Fluorescence-Based
Cytometry: A Practical Guidebook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Florian Mair and Aaron J. Tyznik
2 Immunophenotyping by Mass Cytometry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Gregory K. Behbehani
3 Fluorescent Cell Barcoding for Immunophenotyping. . . . . . . . . . . . . . . . . . . . . . . . 53
Valentina Giudice, Giovanna Fantoni, and Angélique Biancotto
4 Immunophenotyping Using Dried and Lyophilized Reagents . . . . . . . . . . . . . . . . 69
Marc Langweiler
5 Guidelines for Gating Flow Cytometry Data for Immunological Assays. . . . . . . . 81
Janet Staats, Anagha Divekar, J. Philip McCoy, Jr, and Holden T. Maecker
6 Strategies and Techniques for NK Cell Phenotyping. . . . . . . . . . . . . . . . . . . . . . . . . 105
Chen Ziqing, Andreas Lundqvist, and Kristina Witt
7 Immunophenotyping of Human B Lymphocytes in Blood
and in Adipose Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Alain Diaz, Maria Romero, Daniela Frasca, and Bonnie B. Blomberg
8 Multiparametric Flow Cytometry Analysis of Naı̈ve, Memory,
and Effector T Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Ankit Saxena, Pradeep K. Dagur, and Angélique Biancotto
9 Immunophenotyping of Human Regulatory T Cells . . . . . . . . . . . . . . . . . . . . . . . . 141
Janet Staats
10 Immunophenotyping of Human Innate Lymphoid Cells . . . . . . . . . . . . . . . . . . . . 179
Sara Trabanelli, Alejandra Gomez-Cadena, and Camilla Jandus
11 Enumeration of Plasmacytoid Dendritic Cells in Peripheral Blood
and Bone Marrow by Flow-Cytometric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Abdullah Alsuwaidan, Franklin Fuda, Weina Chen, and Mingyi Chen
12 Immunophenotyping of Circulating Endothelial Cells
and Endothelial Microparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Nicholas Wanner and Kewal Asosingh
13 Rare Event Phenotyping and Molecular Characterization:
Circulating Tumor Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Moen Sen, Ling Wang, Liping Yu, and Erica L. Carpenter
14 Quality Control of Immunophenotyping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Bruce Greig
15 Immunophenotyping of Acute Myeloid Leukemia . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Pallavi Kanwar Galera, Chunjie Jiang, and Raul Braylan

vii
viii Contents

16 Immunophenotyping of Acute Lymphoblastic Leukemia . . . . . . . . . . . . . . . . . . . . 297

Joseph A. DiGiuseppe and Jolene L. Cardinali
17 Clinical Flow-Cytometric Testing in Chronic Lymphocytic Leukemia . . . . . . . . . 311
Dalia A. Salem and Maryalice Stetler-Stevenson
18 Immunophenotyping of Paroxysmal Nocturnal Hemoglobinuria (PNH) . . . . . . 323
Andrea J. Illingworth, Iuri Marinov, and D. Robert Sutherland

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Contributors

ABDULLAH ALSUWAIDAN • Division of Molecular and Genomic Pathology, University of

Pittsburgh Medical Center, Pittsburgh, PA, USA
KEWAL ASOSINGH • Department of Inflammation and Immunity, Cleveland Clinic, Lerner
Research Institute, Cleveland, OH, USA
GREGORY K. BEHBEHANI • Division of Hematology, Department of Internal Medicine, James
Cancer Hospital, The Ohio State University, Columbus, OH, USA
ANGÉLIQUE BIANCOTTO • Precision Immunology, Immunology and Inflammation Research
Therapeutic Area, Cambridge, MA, USA
BONNIE B. BLOMBERG • Department of Microbiology and Immunology, University of Miami
Miller School of Medicine, Miami, FL, USA; Sylvester Comprehensive Cancer Center,
University of Miami Miller School of Medicine, Miami, FL, USA
RAUL BRAYLAN • Hematology Section, Department of Laboratory Medicine, Clinical Center,
National Institutes of Health (NIH), Bethesda, MD, USA
JOLENE L. CARDINALI • Department of Pathology and Laboratory Medicine, Hartford
Hospital, Hartford, CT, USA
ERICA L. CARPENTER • Division of Hematology and Oncology, Department of Medicine,
Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
MINGYI CHEN • Department of Pathology, University of Texas Southwestern Medical Center,
Dallas, TX, USA
WEINA CHEN • Department of Pathology, University of Texas Southwestern Medical Center,
Dallas, TX, USA
PRADEEP K. DAGUR • Flow Cytometry Core Facility, National Heart, Lung, and Blood
Institute, National Institutes of Health, Bethesda, MD, USA
ALAIN DIAZ • Department of Microbiology and Immunology, University of Miami Miller
School of Medicine, Miami, FL, USA
JOSEPH A. DIGIUSEPPE • Department of Pathology and Laboratory Medicine, Hartford
Hospital, Hartford, CT, USA
ANAGHA DIVEKAR • Department for Cellular Analysis, BioLegend, San Diego, CA, USA
GIOVANNA FANTONI • Center for Human Immunology, NIAID, NIH, Bethesda, MD, USA
DANIELA FRASCA • Department of Microbiology and Immunology, University of Miami Miller
School of Medicine, Miami, FL, USA
FRANKLIN FUDA • Department of Pathology, University of Texas Southwestern Medical
Center, Dallas, TX, USA
PALLAVI KANWAR GALERA • Hematology Section, Department of Laboratory Medicine,
Clinical Center, National Institutes of Health (NIH), Bethesda, MD, USA
VALENTINA GIUDICE • Department of Medicine, Surgery, and Dentistry, Scuola Medica
Salernitana, University of Salerno, Baronissi, Italy
ALEJANDRA GOMEZ-CADENA • Department of Oncology UNIL CHUV, University of
Lausanne, Epalinges, Switzerland
BRUCE GREIG • Hematopathology/Flow Cytometry, Department of Pathology, Microbiology,
and Immunology, Vanderbilt University Medical Center, Nashville, TN, USA
ANDREA J. ILLINGWORTH • Dahl-Chase Diagnostic Services, Bangor, ME, USA

ix
x Contributors

CAMILLA JANDUS • Department of Oncology UNIL CHUV, University of Lausanne,

Epalinges, Switzerland
CHUNJIE JIANG • Hematology Section, Department of Laboratory Medicine, Clinical Center,
National Institutes of Health (NIH), Bethesda, MD, USA
MARC LANGWEILER • Asheville, NC, USA
ANDREAS LUNDQVIST • Department of Oncology and Pathology, Karolinska Institutet,
Stockholm, Sweden
HOLDEN T. MAECKER • Institute for Immunity, Transplantation, and Infection, Stanford
University School of Medicine, Stanford, CA, USA
FLORIAN MAIR • Vaccine and Infectious Disease Division, Fred Hutchinson Cancer Research
Center, Seattle, WA, USA
IURI MARINOV • Institute of Hematology and Blood Transfusion, Prague 2, Czech Republic
J. PHILIP MCCOY, JR • Frederick, MD, USA
MARIA ROMERO • Department of Microbiology and Immunology, University of Miami Miller
School of Medicine, Miami, FL, USA
DALIA A. SALEM • Laboratory of Pathology, CCR, NCI, NIH, Bethesda, MD, USA; Clinical
Pathology Department, Faculty of Medicine, Mansoura University, Mansoura, Egypt
ANKIT SAXENA • Flow Cytometry Core Facility, National Heart, Lung, and Blood Institute,
National Institutes of Health, Bethesda, MD, USA
MOEN SEN • Division of Hematology and Oncology, Department of Medicine, Perelman
School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
JANET STAATS • Department of Surgery, Duke University Medical Center, Durham, NC,
USA; Duke Immune Profiling Core, Duke University Medical Center, Durham, NC, USA
MARYALICE STETLER-STEVENSON • Laboratory of Pathology, CCR, NCI, NIH, Bethesda, MD,
USA
D. ROBERT SUTHERLAND • University of Toronto, Toronto, ON, Canada
SARA TRABANELLI • Department of Oncology UNIL CHUV, University of Lausanne,
Epalinges, Switzerland
AARON J. TYZNIK • BD Biosciences, La Jolla, CA, USA
LING WANG • BD Technologies and Innovation, Research Triangle Park, NC, USA
NICHOLAS WANNER • Department of Inflammation and Immunity, Cleveland Clinic,
Lerner Research Institute, Cleveland, OH, USA
KRISTINA WITT • Department of Oncology and Pathology, Karolinska Institutet, Stockholm,
Sweden
LIPING YU • Applied Cells, Inc., Santa Clara, CA, USA
CHEN ZIQING • Department of Oncology and Pathology, Karolinska Institutet, Stockholm,
Sweden
 Chapter 1

High-Dimensional Immunophenotyping with Fluorescence-

Based Cytometry: A Practical Guidebook
Florian Mair and Aaron J. Tyznik

Abstract
Recent technological advances have greatly diversified the platforms that are available for high-dimensional
single-cell immunophenotyping, including mass cytometry, single-cell RNA sequencing, and fluorescent-
based flow cytometry. The latter is currently the most commonly used approach, and modern instrumenta-
tion allows for the measurement of up to 30 parameters, revealing deep insights into the complexity of the
immune system.
Here, we provide a practical guidebook for the successful design and execution of complex fluorescence-
based immunophenotyping panels. We address common misconceptions and caveats, and also discuss
challenges that are associated with the quality control and analysis of these data sets.

Key words Immunology, High-dimensional, Fluorescence, Flow cytometry, Polychromatic, Panel

design, Spillover spreading, SSM, Compensation, Staining

1 Introduction

Since its invention more than 50 years ago, flow cytometry has been
a key tool for novel discoveries in the field of immunology
[1, 2]. This has been driven primarily by the ability of flow cyto-
metry to qualitatively and (within certain constraints) quantitatively
measure multiple features on a single cell level. Given the diversity
of immune cell subsets and their functional complexity, there is an
increasing need to measure more targets per cell in order to accu-
rately classify a given subtype while still being able to assess addi-
tional features of interest.
This need has fueled the development of novel cytometry plat-
forms during the past decade, most notably mass cytometry (cyto-
metry by time-of-flight, CyTOF) [3–5] and single-cell RNA
sequencing (sc-RNAseq) [6–9]. Both of these approaches can pro-
vide previously unprecedented insight into the immune system, but
come with their own set of limitations. Mass cytometry currently
allows the measurement of up to 50 protein (or RNA) targets on

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_1, © Springer Science+Business Media, LLC, part of Springer Nature 2019

1
2 Florian Mair and Aaron J. Tyznik

millions of cells at a single cell level, but has a relatively slow

throughput and does not allow for recovery of living cells via
sorting [10]. Single-cell RNAseq increases the dimensionality of the
possible measurements by almost two orders of magnitude (depend-
ing on the platform used), but the associated reagent and sequencing
cost is currently prohibitive for analyzing more than tens of thousands
of cells [9, 11]. For further information on these technologies,
we direct the reader to some excellent reviews [5, 11, 12].
Within this chapter, we focus on fluorescent-based flow cyto-
metry as the most widely used platform for immunophenotyping.
Briefly, fluorescent flow cytometers rely on multiple lasers as light
sources to excite fluorescent molecules, which are usually coupled
to antibodies that are bound to antigenic targets of interest on the
samples being analyzed. In order to properly interrogate these cells,
a fluidics system utilizes hydrodynamic focusing to linearly align the
cells in a stream of liquid where they intercept the focused laser
beams one cell at a time. Excitation of the fluorochrome results in
the emission of photons that travel through an optical path con-
sisting of filters and mirrors and are detected by a set of photon
detectors, most commonly photon multiplier tubes (PMTs). In the
past decade, incremental technical improvements have resulted in
the latest generation of instruments, which allow the measurement
of up to 30 parameters [13]. Present efforts in the field include the
further development of spectral flow cytometry, which was origi-
nally developed at Purdue University and holds great potential for
the future [14–16]. Instruments based on this principle measure
the entire spectrum across the visible range and reconstruct indi-
vidual fluorophore signals by spectral deconvolution algorithms.
This approach will most likely increase the number of measurable
fluorescent parameters beyond 40 and promises several additional
advantages such as autofluorescence unmixing. However, spectral
flow cytometers are not very widely available yet, and thus within
the scope of this chapter we focus on “conventional” fluorescent
panel design.
Designing fluorescent panels to produce high-quality and reli-
able data for 15–30 parameters requires significant effort and thus
far has not been automated due to the wide range of available
instrumentation and the varying biological targets between differ-
ent research questions. Here, we provide a practical step-by-step
guidebook for designing high-dimensional fluorescent panels and
focus on common challenges. An overview of the process is out-
lined in Fig. 1, and from start to finish a 30-parameter panel can
successfully be established in a 3–6 week time frame (given that all
reagents are commercially available).
Of note, an important consideration for any high-dimensional
assay is the subsequent analysis of the resulting data. Manual gating,
while still the gold standard in the field and required for quality
control of the data, is often insufficient to fully explore high-
 High-Dimensional Immunophenotyping 3

Fig. 1 Overview of the typical steps required for a high-dimensional immunophenotyping experiment. Each of
the six steps is described in detail in Subheadings 3.1 through 3.7

dimensional datasets [17]. An exhaustive discussion of this topic is

beyond the scope of this chapter, and for further information we
direct the reader to the following reviews on computational analysis
methods for flow cytometry data [18–20], and some key papers
describing the comparison of computational methods to traditional
manual gating [21, 22]. Furthermore, for a comprehensive review
of the current state-of-the art on many cytometry-related topics, we
recommend the “Guidelines on the use of flow cytometry and cell
sorting in immunological studies” [23].

2 Materials

2.1 Instrument and Flow cytometer.

QC Particles Computer with analysis software capable of reading FCS files.
Hard-dyed reference beads (e.g., Spherotech Ultra Rainbow beads,
#URCP-38-2F).
Instrument QC-particles if required (see instrument-specific
instructions).

2.2 Reagents and Fluorochrome-conjugated antibodies (various vendors).

Plastic Ware Fc-blocking reagent (various vendors).
Fixable live/dead reagent (depending on the panel, various
vendors).
Compensation particles (e.g., BD Biosciences, anti-rat compensa-
tion beads #552844 and anti-mouse compensation beads
#552843, or anti-mouse Plus beads #560497).
4 Florian Mair and Aaron J. Tyznik

Brilliant staining buffer (various vendors).

PFA-based fixative (various vendors, or freshly prepared 2% PFA
solution).
Optional: fix/perm reagent for intracellular or intranuclear targets.
PBS.
Fetal bovine serum (FBS).
96-well round-bottom plates.
5-ml polystyrene tubes.

2.3 Solutions to Stain/wash buffer: add 10 ml of heat-inactivated FBS to a 500 ml

Prepare bottle PBS.
Fc-Block/viability dye solution: add 25 μl of Fc-blocking reagent
and 1 μl of reconstituted live/dead reagent to 500 μl of PBS
immediately prior to use. Titration of the live/dead reagent is
recommended.

3 Methods

3.1 Generating Your Any immunophenotyping experiment should start with a biological
Hypothesis and List of hypothesis and a list of targets that are to be included in the
Biological Targets experiment. Typically, this will comprise two types of target anti-
gens: the so-called lineage markers, which are used to delineate
canonical immune cell populations (such as CD3 for T cells), and
the markers of interest, which can either have a known expression
pattern that changes under experimental conditions (such as CD69
as an early activation marker for T cells), or a potentially unknown
expression pattern.
As a first step, we recommend compiling a table that lists all
target antigens, their expected expression pattern (continual or
variable), expression type (bimodal, continuum), and expression
level (low to high) as well as expected coexpression patterns, that
is, which markers are expected to be expressed on the same cell
types. An example for such a table with some human T cell-centric
markers is given in Fig. 2. Furthermore, it is helpful to include
reagent availability in this table, that is, which fluorophore–anti-
body conjugates are available from which vendor for a given target.
Having this information at hand simplifies the actual panel design
process, which will be discussed in detail in Subheading 3.4. The
antigen targets listed within this chapter are based on human
leukocytes, but the same principles apply for any other cell type
such as murine leukocytes.
1. Define the cellular populations that are relevant for the planned
experiment (e.g., CD4+ T cells, CD8+ T cells, and CD56+ NK
cells).
 High-Dimensional Immunophenotyping 5

Fig. 2 Example table with target antigen information. This table lists a selection of antigens for a human T cell
centric panel and the anticipated coexpression patterns. Researchers should adopt this table to fit their own
needs. Alternatively, one can draw a gating tree instead of using a table. If available, it is helpful to assess
common expression patterns in preliminary experiments or in public data sets such as those included in
published OMIPs

2. List the lineage markers required for the reliable identification

of all canonical cell populations (e.g., CD45/CD3/CD4 for
CD4+ T cells).
3. Define the target antigens of interest (e.g., for measuring the
phenotype of naı̈ve vs. effector T cells: CCR7, CD45RO,
CD45RA, CD25, CD127, etc.).
4. List the expression patterns according to the following cate-
gories: bimodal, continuous, unknown.
5. List the expected expression level for each target antigen
according to the following categories: low, medium, high.
6. Optional: perform an Internet search for reagent availability for
all targets and list the corresponding catalog numbers and
vendors.
7. Save and print this table for future reference.

3.2 Instrument For most researchers, the instrument available is an institute-

Characterization and specific fixed constant and not necessarily subject to change. How-
Detector Gain ever, it is critical to familiarize oneself with the given instrument
Optimization configuration and the available optical detectors [24], and perform
an optimization of the detector gains. The latter is a necessity that
6 Florian Mair and Aaron J. Tyznik

stems from the fact that photon multiplier tubes (PMTs), which are
still the most commonly used detector type, show significant
model-to-model variation, and thus each detector will require a
different voltage gain to deliver optimal resolution (i.e., optimal
signal-to-noise ratio). A more extensive discussion on the factors
describing the performance of a given PMT can be found in the
following reference [25]. Of note, avalanche photodiodes (APDs)
which are better suited for the detection of long-wavelength signals
[26] are becoming more popular but are beyond the scope of this
chapter.
There are several methods available to perform detector opti-
mization: for instruments from BD Biosciences, the Cytometry
Setup and Tracking (CS&T) module is commonly used to generate
consistent settings and track cytometer performance (see vendor
specific CS&T application guide). This system relies on a set of
hard-dyed dim, medium, and bright beads (CS&T beads). During
initial setup, voltages are determined such that the signal of the dim
beads is placed 10 times above the robust standard deviation (rSD)
of the electronic noise. Later on, the signal of the bright beads is
used to set target values, which are being matched during daily
quality control (QC) by modification of the detector voltages.
Often times, users utilize PMT voltage settings from the CS&T
application for setting and tracking detector gains. However, as
CS&T was developed at a time when the selection of available
fluorophores was limiting, the CS&T routine might deliver unnec-
essarily high baseline voltages for some detectors excited off the
violet, blue, and ultraviolet laser, thereby pushing signals from cells
off scale. Additionally, the rSD of the dim bead may not be the same
as the autofluorescence of a biological sample and correlation to
optimal PMT voltage settings may not be accurate. As a result,
researchers may need to lower or, in some instances, increase their
detector gains manually, making the CS&T method of setting
baseline voltages suboptimal.
Another approach is based on using unstained cells and setting
the voltages in such a way that the rSD of the cells equals 2.5 times
the rSD of the electronic noise, which ensures that dim signals are
pulled out of the electronic noise [27]. Finally, there are two
approaches that utilize a voltage titration experiment using either
the peak 2 of 8-peak rainbow beads (called the second peak
method, mimicking a dim signal) or single-stained samples (using
cells or antibody capture beads).
Here, we will describe a simplified voltage titration (voltration)
experiment using cells stained at saturating levels with a single
antibody clone (e.g., anti-CD4 or anti-CD8) conjugated to all
fluorophores that can be measured on a given instrument
[28]. Of note, the same data can be used to generate a relative
fluorochrome brightness chart, which will be useful in Subheading
3.4 for panel design. This experiment needs to be carried out only
 High-Dimensional Immunophenotyping 7

once for any cytometer unless the optical configuration, that is,
lasers, filters, or detectors are changed. A representative analysis of
two typical voltage titrations is shown in Fig. 3a, b, and two
fluorochrome brightness charts with a classification of fluoro-
chromes into four broad categories is shown in Fig. 3c.
Performing the voltage titration experiment (unique to
instrument):
1. Obtain antibodies for a given clone (typically targeting a line-
age marker such as CD4) conjugated to all different fluoro-
phores that can be measured on the available instrument.
2. Obtain cells (typically cryopreserved PBMCs or murine
splenocytes).
3. Prepare antibody dilutions of every conjugate at a saturating
titer in stain buffer (see Subheading 3.5).
4. Aliquot 500,000 cells in as many wells as required into a
96-well plate.
5. Spin plate at 400
g for 5 min. Flick supernatant, dry the top
of the plate by carefully placing it on a paper towel and draining
remaining liquid.
6. Add 100 μl of a single antibody dilution per well, resuspend,
and incubate for 20 min at RT in the dark.
7. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel. Repeat this wash
for a total of two washes.
8. Optional: add 100 μl of fixative to each well, resuspend, and
incubate for 20 min at RT in the dark (see Note 1).
9. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel.
10. Resuspend in 200 μl of stain buffer and store in the dark at 4  C
until acquisition.
11. Perform flow cytometer startup and daily QC as required per
instrument-specific instructions.
12. Load one of the single-stained samples, and adjust the voltage
of the respective detector in 30 V increments from 200 V to
700 V. Record 10,000 events for every given detector voltage
setting and name the tubes accordingly (BV421_200V,
BV421_230V, BV421_260V, etc.).
13. Disregard samples where the MFI of the positive signal has
exceeded detector linearity or is no longer on scale as errone-
ous values may be reported.
14. Repeat this step for all single-stained samples.
Fig. 3 Examples for a voltage titration experiment and a fluorochrome brightness chart. (a) and (b) Two
representative voltage titration experiments are shown as concatenated pseudocolor plots (left) and as
histogram overlays (right). Staining indices have been calculated as described in Subheading 3.2. For
CD4-PE the plateau of the stain index is reached around 420 V, indicating that this is the minimal voltage
required to deliver optimal resolution. If the positive population exceeds the upper end of the scale as is the
case for CD4-BB515, see Note 2. (c) Two representative fluorochrome brightness charts generated on two
independent commercially available high dimensional flow cytometers. As each instrument is unique in laser
power and detection optics, it is best to characterize the relative fluorochrome brightness for each individual
instrument. However, assigning fluorochromes to four broad categories will allow some generalization for
relative brightness (e.g., APC-H7 tandems being at the lower end of the scale, or BV421 and PE-tandem dyes
being at the upper end of the scale)
 High-Dimensional Immunophenotyping 9

15. Export data in FCS3.1 file format and load into a flow cyto-
metry data analysis program.
16. For a given single-stained control, draw gates delineating sing-
lets, and subsequently the positive peak and the negative peak
for the corresponding fluorophore.
17. Optional: concatenate the individual FCS files for a given
single-stained control into a single FCS-file, or generate histo-
gram overlays to obtain a visual representation of the voltration
experiment.
18. Repeat this step for all single-stained samples.
19. Calculate the median fluorescence intensity (MFI) of both the
positive and the negative peak, as well as the rSD for the
negative peak.
20. Calculate the staining index (SI) for each sample by using the
following formula: SI ¼ MFI(pos)  MFI(neg)/(2
rSD).
21. Using a spreadsheet program, generate a plot showing the SI
on the y-axis relative to the voltage on the x-axis.
22. For each detector, pick the voltage on which the SI starts to
reach a plateau, which is the minimal voltage delivering the best
possible resolution for a given detector. As an alternative visu-
alization, plot the SI as a percentage of maximum SI on the y-
axis relative to the voltage on the x-axis (also see Note 2).
23. Once the optimal baseline PMT voltages have been deter-
mined, record a sample of appropriate reference beads (e.g.,
Spherotech Ultra Rainbow Beads) as described in steps 29–34.
See Note 3 for examples of reference particles.
Create relative fluorochrome brightness ranking chart unique to
the instrument
24. Identify the single-stained samples acquired to reach the opti-
mal PMT voltage for any given detector.
25. Calculate the median fluorescence intensity (MFI) of both the
positive and the negative peak, as well as the rSD for the
negative peak.
26. Calculate the staining index (SI) for each fluorophore by
using the following formula: SI ¼ MFI(pos)  MFI(neg)/
(2
rSD).
27. Using a spreadsheet program, generate a plot showing the SI
for each fluorophore and arrange them by decreasing values.
Alternatively, plot the SI as percentage of maximum SI.
28. Assign the fluorochromes into broad categories based on their
SI, that is, bright, intermediate-high, intermediate-low,
and dim.
10 Florian Mair and Aaron J. Tyznik

Setting and maintaining target MFI Values

29. Set optimal PMT voltages as determined in steps 1–23.
30. Load a sample with reference beads (see Note 3) that provide
an adequate positive signal in each channel being utilized on
your flow cytometer and acquire 20,000 events.
31. Optional: Repeat acquisition of reference particles ten inde-
pendent times to determine baseline MFI and standard devia-
tion (SD) of signal.
32. Create histograms for all detectors and place an interval gate
on the positive peak(s) to determine median fluorescent
intensity (MFI).
33. Using an appropriate analysis program, export the MFI for all
baseline runs and calculate average MFI and standard
deviation.
34. After the daily instrument-specific start-up and QC proce-
dures and prior to every experiment, acquire reference beads
and adjust PMT voltage for each detector to set signals to
match the target MFI  5% (or  acceptable variability based
off the optional standard deviation calculations).

3.3 Characterization Conventional fluorescent flow cytometry instruments use optical

of Spreading Error and band pass filters placed in front of each PMT, balancing between
Fluorochrome catching the main peak of a given fluorescent signal and simulta-
Brightness neously avoiding signals coming from other fluorophores. Despite
instrument manufacturers’ best efforts, there will be spectral over-
lap between different fluorophores measured (Fig. 4a). One usually
refers to the primary detector as the one intended to detect that
fluorophore (e.g., G575 for PE, excited by green 532 nm laser) and
the secondary detector as the one that collects spillover, that is, the
remaining portion of the spectrum that is detected but does not
represent the primary signal in that detector (e.g., G610, intended
for PE-CF594 detection but also collecting signal from PE). It is
important to note that spillover can not only occur on neighboring
detectors off the same laser line but also on detectors off a different
laser line if the fluorophore shows some excitation by that wave-
length (e.g., U570, intended for BUV563 but will also collect
signal from PE which is excited primarily by the 488/532 nm
laser but to a certain extent by the ultraviolet 355 nm laser)
(Fig. 4a, right panel).
Spectral overlap is corrected by a process termed compensation,
a mathematical manipulation of the acquired raw data
[29, 30]. The compensation calculation is based on single-stained
controls, which are discussed in more detail in Subheading 3.6.
After compensation, typically a percent value is reported that will
describe the relative fluorescence detected in the secondary detec-
tor compared to the primary detector. During compensation, this
 High-Dimensional Immunophenotyping 11

Fig. 4 Fluorochrome overlap, compensation, and spreading error (SE). (a) Fluorochrome spectra depicting
excitation (dotted curve) and emission (solid filled curve) of Brilliant Violet 650 (BV650, left) and Phycoerythrin
12 Florian Mair and Aaron J. Tyznik

signal portion is subtracted from the total signal detected in the

secondary detector.
A very common historic misconception relates to which mag-
nitude of a compensation value is considered “acceptable.” When-
ever the compensation wizard reports high values (especially
approaching 100% or more), some users adjust PMT voltages to
lower this value, assuming that this will improve their instrument
setup. However, this is usually not the case and rather poses the
danger of reducing the PMT voltage below the minimum value
required to deliver optimal resolution (see Subheading 3.2). It is
important to note that in high-dimensional fluorescent cytometry,
some fluorochrome spectra will be so close, that at optimal detector
settings compensation values can often approach or even exceed
100%, which does not necessarily negatively impact assay perfor-
mance. An excellent description and sample data demonstrating
this effect has been published by Ashhurst and colleagues [31].
The correct metric to use in this context is the so-called spread-
ing error (SE), which is an extremely useful metric that was only
recently developed by the Roederer lab at NIH [32]. In a nutshell,
the SE describes the spreading of the negatives of any given com-
pensated population in the secondary detector. Often this increase
in the spread (measured by robust standard deviation, rSD) is
attributed to compensation, which is incorrect—compensation
does not generate the SE but merely makes it visible at the low
end of a logarithmic/biexponential scale. The underlying reason
for the SE is the fact that the actual measurement of the

Fig. 4 (continued) (PE, right curve) (adapted from http://www.bdbiosciences.com/us/s/spectrumviewer). The

shaded boxes indicate typical detection windows that are used to collect the emission peaks of these
fluorophores. Spectral overlap will occur into the neighboring detectors, as well as the same detection
window on different laser lines (indicated for PE, which is not only excited by the 488 nm and 532 nm laser,
but also by the 355 nm UV laser). (b) Connection between signal intensity and spreading error (SE). Three
different titers of single-stained anti-human CD3-BV786 PBMCs are shown as an overlay contour plot, with the
U-780 detector (excitation laser: 355 nm, detection window: 785/62) on the y-axis. The limit of detection on
CD3+ events is indicating by a solid line, highlighting that the SE increases with signal intensity. (c) Examples
showing the disconnect between the compensation value and SE. Compensation beads were stained with
anti-human CD4-PE (upper panels) and CD4-BV650 (lower panels). PE will show significant spillover into the
U-570 (excitation laser: 355 nm, detection window: 570/40) and G-610 (excitation laser: 532 nm, detection
window: 610/20) detector, but no spillover into the B-515 detector (excitation laser: 488 nm, detection
window: 515/20). BV650 will show significant spillover into the U-660 (excitation laser: 355 nm, detection
window: 660/40) and V-710 (excitation laser: 405 nm, detection window: 710/40) detector. Note that the
relative extent of SE does not correlate with the compensation value (numbers above plots). (d) N
N view of
some selected compensated single-stained controls from a 28-color experiment. Fluorophore detector pairs
with relevant SE have been marked with arrows. (e) The same view as in the first row of (d), but with single-
stained samples derived from human whole blood. Using an anti CD4-antibody, this will deliver a bright
population (T cells) as well as an intermediate population (monocytes, colored in blue) and allows to visually
assess the extent of SE at different signal intensities
 High-Dimensional Immunophenotyping 13

fluorescence signal is imprecise and will carry some variance due to

the Poisson error in photon counting [32]. This measurement
error will depend on many factors (such as the PMT model and
the wavelength of the detected signal) and is an intrinsic property of
the measurement process.
There are three key aspects of SE that will be relevant for panel
design as discussed in Subheading 3.4. First, SE is proportional to
signal intensity, that is, the brighter a signal in the primary detector,
the more pronounced the SE in the secondary detector will be
(Fig. 4b). Second, SE is reducing the resolution in the secondary
detector, that is, the detector that is collecting spillover (Fig. 4c).
Third, SE is additive, that is, if a detector collects SE from multiple
different fluorophores, the overall increase of the rSD of the nega-
tives (and thus the loss in sensitivity) will be larger.
As a result, an essential tool for modern panel design is the
spillover spreading error matrix (SSM), which is different from the
compensation matrix. The SSM will provide a comprehensive over-
view for a given instrument on (1) the relative contribution of any
fluorophore to SE in secondary detectors and (2) the relative loss of
resolution in any secondary detector due to SE collected from all
other fluorophores. The mathematical formula for calculating the
SSM is provided in the original publication [32] and has also been
made available online by Ashhurst and colleagues, but the SSM can
be calculated more easily in a common data analysis package,
FlowJo (version 10.4 and higher).
It is important to note that the extent of SE cannot be pre-
dicted from the corresponding value in the compensation matrix,
which is exemplified in the plots displayed in Fig. 4c. Below, we
describe the steps for obtaining an SSM for a given flow cytometer
using antibody capture beads (compensation beads). Alternatively,
one can directly utilize the data that has been generated during
PMT voltage baseline setup in Subheading 3.2. Of note, these data
sets can not only be used to calculate the SSM, but can also serve as
a visual reference to judge the extent of SE. After appropriate
compensation, visualizing the single-stained control samples in
the form of N
N plots (i.e., each fluorophore by each detector)
will provide a quick visual reference of the amount of SE into all
corresponding detectors to identify potential “watch-out” channels
for coexpressed markers (Fig. 4d). If these plots are based on
human whole-blood samples utilizing CD4 as an antigen, the
added advantage is that the impact of signal intensity on SE can
be visualized, as monocytes will show as CD4-intermediate cells,
while T cells will show a CD4-bright population (Fig. 4e). This
approach reiterates the fact that SE is proportional to signal inten-
sity, and can provide a useful “generic” biological reference as you
plan your fluorochrome assignments.
Overall, the concept of SE and the SSM is invaluable for panel
design, and an example SSM and how to interpret it is shown in
14 Florian Mair and Aaron J. Tyznik

Fig. 5 The spillover spreading matrix (SSM). (a) A representative spillover spreading matrix (SSM) calculated
from a 28-color experiment is shown. Color coding is from no SE (white) to high SE (red). Note that the SSM is
instrument specific and does not provide generalizable conclusions about the extent of SE, and SE is a
unitless, relative number. For this particular instrument, the three fluorophores contributing the least SE are
BB515, AF647 and BUV395. Thus, these three dyes would be optimal to assign for lineage markers as they are
contributing minimal SE to all the other detectors. In turn, the three detectors receiving the least SE are B-515,
V-510 and V-450. Since BB515 as well as BV421 are typically two very bright fluorophores (see Fig. 3c), on
this particular instrument these two detectors would be optimal for dimly expressed targets. Note that when
assessing the row-sums, that is, the total SE contributed by a given fluorophore, the case can be that this sum
is driven mostly by one particular value, for example BUV661, where SE into R-660 is the major contributor to
the sum. In this case, BUV661 can still readily be used without major impact to all other detectors except
R-660

Fig. 5. The use of the SSM is more extensively discussed in the next
section.
1. Obtain antibodies carrying all fluorochromes that are to be
included in the new panel (the antigen targets do not necessar-
ily matter at this point in time, but should match the fluoro-
chromes that will be used) (also see Note 4).
2. Aliquot antibody-capture beads (i.e., compensation particles)
into as many wells as required into a 96-well plate.
3. Resuspend in 100 μl of staining buffer.
4. Add 1 μl of a single antibody into each well and resuspend
using a 100 μl pipette. Incubate for a minimum of 15 min at RT
in the dark.
5. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel. Repeat this
wash once.
6. Optional: add 100 μl of fixative to each well, resuspend, and
incubate for 20 min at RT in the dark.
7. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel.
 High-Dimensional Immunophenotyping 15

8. Resuspend in 200 μl of stain buffer and store at 4  C until

acquisition.
9. Perform flow cytometer startup and daily QC as required per
manufacturer’s instructions.
10. Acquire the single-stained control tubes at the optimal voltages
determined in Subheading 3.2.
11. Export data in FCS3.1 file format and load into a flow cyto-
metry data analysis program capable of computing the SSM
(FlowJo).
12. Open the compensation wizard, calculate the compensation
matrix and then select “compute SSM” (see http://docs.
flowjo.com/d2/experiment-based-platforms/plat-comp-over
view/spillover-spreading-matrix/).
13. Export the SSM as a CSV-file and color-code it based on SE
using a spreadsheet program (for example low SE no color,
large SE red) (see Fig. 5).
14. Apply the calculated compensation matrix to the single-stained
control samples.
15. Arrange the samples in the form of an N
N plot and visually
assess the extent of SE.

3.4 In Silico Panel The two key considerations for successful panel design are (1) fluo-
Design Using the rochrome brightness, and (2) dealing with spreading error (SE),
Spillover Spreading whereby the latter is the more relevant issue. There are several ways
Matrix (SSM) that SE and the resulting loss of resolution can be controlled. The
most straightforward way is to use fluorophore–detector pairs with
high SE for mutually exclusive markers (e.g., CD3 and CD19)
because in this case the SE will not interfere with detection of either
signal. However, this tactic is only applicable to a limited extent as
most researchers’ experimental aim is to analyze as many target
antigens as possible on one or two cell types.
Thus, if fluorophore–detector pairs with significant SE are to be
used for molecules expressed on the same cell type, one has to assess
the expression level of the target antigens. This can be done either
by an internet search, or by preliminary experiments assessing
antigen density on the target population. An excellent resource in
this context are optimized multicolor immunophenotyping panels
(OMIPs) [33, 34], which typically show raw data in their supple-
ments and cover many of the commonly used targets in immunol-
ogy [13, 35, 36]. If the target expression level is known, one can
use a fluorophore generating a large SE for low-expression anti-
gens, which will thus generate only weak signals, which in turn will
reduce the amount of SE (as SE is proportional to the signal
intensity). Alternatively, one can use a highly expressed antigen in
the secondary detector, and assess whether the signal of the target
16 Florian Mair and Aaron J. Tyznik

in the secondary detector is above the spread of the negatives

generated by the SE.
Another useful approach for dealing with SE is to purposefully
titer down signals of lineage markers, which typically show bimodal
expression patterns and are only used to delineate positive from
negative cells. Again, lowering the overall signal by lowering the
used antibody titer will also decrease the SE generated (see example
in Fig. 4b). Of note, working at nonsaturating staining conditions
has important drawbacks that will be discussed in more detail in
Subheading 3.5 below.
Finally, one has to consider relative fluorochrome brightness
(see fluorochrome brightness chart, Fig. 3c) such that
low-expression antigens are paired preferably with bright fluoro-
chromes, while high-expression antigens can be paired with dim
fluorochromes. However, it is important that this assignment takes
SE into account. Below, we describe the steps that we typically
perform for the in silico design of a new panel. The example SSM
shown in Fig. 5 summarizes some of the key aspects that are listed
below.
1. Print out the color-coded SSM as well as the fluorochrome
(reagent) brightness chart generated in Subheading 3.4.
2. Identify the three highest values in the SSM and assign these to
mutually exclusive antigen targets.
3. Identify the fluorophores in the SSM with the lowest row sums,
that is, the ones that contribute the least amount of overall SE
to all the remaining detectors.
4. Assign these fluorophores to your most relevant lineage mar-
kers (e.g. CD3 for a T cell-centric analysis).
5. Identify the detectors in the SSM which have the lowest col-
umn sums, that is, collect the least amount of SE and thus will
not have any increased spreading of the negatives introduced.
6. Assign the most critical/dimly expressed antigen targets to the
fluorophores used in these detectors. Make sure that the
corresponding fluorophores fall into the “bright” or “interme-
diate” category of brightness.
7. Identify the detectors in the SSM which have the highest
column sums, that is, those receiving the most SE. For these
detectors it is essential that the assigned antigen–fluorophore
pairs deliver a signal that is above the SE. This has to be assessed
in preliminary experiments using fluorescent-minus-one
(FMO) controls compared to a full staining panel (see Subhead-
ing 3.6).
8. For the remaining antigens and fluorochromes, pair the anti-
gens with higher expression levels to the fluorochromes gen-
erating less SE, and the antigens with lower expression levels to
the fluorochromes generating higher SE.
 High-Dimensional Immunophenotyping 17

9. As there will be numerous possibilities in step 7, design up to

three different panel combinations. Note that in most cases,
there will not be one single optimal panel, but several possible
combinations.
10. Order the necessary reagents in small test-size aliquots.

3.5 Titration of After selection of reagents for the panel run, titration is an essential
Reagents step that is often omitted. The primary reason for titrating anti-
bodies is twofold: first, to minimize background from nonspecific
binding and second, to ensure that staining is performed at satur-
ating concentration (if necessary) [37]. For a typical titration exper-
iment, it is recommended to do 8–12 twofold serial dilutions
starting from a highest concentration of 2 μg per staining reaction.
After titration, the stain index (SI) can be plotted relative to the
dilution factor. This will reveal the saturating titer, which is the
concentration at which further addition of antibody does not
increase the signal intensity, that is, all target binding sites have
been occupied. If any quantitative measurements (i.e., extracting
MFI as a means of estimating target antigen density) are intended,
antibodies have to be used at the saturating concentration. How-
ever, for some lineage markers (e.g., CD3 or CD4) it is sufficient to
work at nonsaturating concentration. The benefits are that it saves
reagent and also minimizes spreading error into other detectors. If
one chooses to do so, it is important to note that signal intensity at
nonsaturating concentration can change with the number of cells
stained, incubation length, and temperature. Thus, one has to work
with standardized cell numbers and staining conditions. For differ-
ent examples of antibody titrations, we direct the reader to the
supplementary materials of recent OMIPs [13, 35] and the infor-
mation found in the “Guidelines” [23].
1. Obtain cells (typically cryopreserved PBMCs or murine sple-
nocytes). See Note 5 for targets that are not readily detected on
steady-state cells.
2. Aliquot cells in as many wells as required into a separate 96-well
plate.
3. Spin plate at 400
g for 5 min. Flick supernatant and dry the
top of the plate by carefully placing it on a paper towel and
draining remaining liquid.
4. For a given titration experiment, add 96 μl of stain buffer to the
empty well 1 of a 96-well plate, and 60 μl of stain buffer to wells
2–12.
5. Add 24 μl of antibody stock to well 1 and mix. If the specific
antibody concentration is known, use 2 μg of antibody stock
solution and add it to well 1 with additional stain buffer as
required to achieve a final volume of 120 μl.
18 Florian Mair and Aaron J. Tyznik

6. Transfer 60 μl of the solution in well 1 into well 2, mix, and

then transfer 60 μl from well 2 into well 3 and so on. From the
last well (well 12), remove 60 μl and discard.
7. Using a multichannel pipette, take up 50 μl from the prepared
antibody dilutions and use this to resuspend the cells and
incubate for 20 min at RT in the dark.
8. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel. Repeat this
wash once.
9. Optional: add 100 μl of fixative to each well, resuspend, and
incubate for 20 min at RT in the dark.
10. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel.
11. Resuspend in 200 μl of stain buffer and store at 4  C until
acquisition.
12. Start up the flow cytometer and perform daily QC as required
per instrument-specific instructions.
13. Acquire all samples at the optimal voltages determined in
Subheading 3.2.
14. Export data in FCS3.1 file format and load into a flow cyto-
metry data analysis program.
15. For a given titration sample, draw gates delineating singlets,
and subsequently the positive peak, and the negative peak.
16. Calculate the median fluorescence intensity (MFI) of both the
positive and the negative peak (gates are typically assigned by
the compensation wizard), as well as the rSD for the
negative peak.
17. Calculate the staining index (SI) for each titration sample by
using the following formula: SI ¼ MFI(pos)  MFI(neg)/
(2
rSD).
18. Using a spreadsheet program, plot the SI relative to the
dilution.
19. Determine the saturating titer by identifying the dilution at
which the SI does not show any further increase (i.e., plateaus).
20. If there is no saturating concentration, identify the dilution at
which the SI is highest, while maintaining a low background
signal of the negative population.
21. Record the optimal as well as the saturating concentration
together with the antibody clone, lot, and target in a
spreadsheet.
 High-Dimensional Immunophenotyping 19

3.6 Single-Stained Once all reagents have been titrated, the full panel can be tested.
Controls and Testing of The most essential aspect of this validation experiment is not nec-
Full Panel essarily the full stain but rather the inclusion of the appropriate
controls, specifically fluorescence-minus-one controls (FMOs)
[38]. An FMO contains the full antibody mix except one antibody,
and thus allows assessing problems that can potentially arise from
unforeseen spillover or unexpected interaction between dyes
[39]. While it is beneficial to prepare FMOs for every single reagent
in a panel, this can be tedious and might not always be necessary.
Instead, one should include FMO controls for every target where
the expected signal is either not known or includes potentially
problematic fluorochrome–detector pairs (also see Note 6).
The quintessential prerequisite (and most common source of
error) for the success of any polychromatic experiment is the quality
of single-stained controls [30]. This is an area where several mis-
conceptions are commonly found as to what constitutes a “good”
single-stained control, in particular whether one should use cells or
antibody-capture beads (compensation particles). There are four
generalizable rules that apply for single-stained controls, and as
long as these are met one can use either cells or beads (see also
Note 7): (1) Within one single-stained control, the positive and
negative particles must be of the same kind, that is, have the same
autofluorescence. This means that both populations must be beads
(of the same batch and type), or both populations must be the same
type of cells (and not one population comprising T cells and the
other one monocytes, as would be the case if using human PBMCs
and staining them with CD3). The reason for this is that during
compensation, the compensation value is calculated such that MFIs
of positive and negative populations are equal [30]. If the auto-
fluorescence, that is, the background fluorescence of these two
populations are different, this will introduce an error into the
compensation matrix. (2) The positive signal of the single-stained
control must be as bright as or brighter than the experimental
sample, but remain within the linear range of a given detector.
The reason for this rule is that the error that is associated with a
measurement is proportional to the square root of the signal inten-
sity, such that a dimmer signal will carry a proportionally larger
measurement error. Therefore, it is better if the single-stained
control is bright but does not exceed the linear range of the
detector (instrument-dependent), which would again introduce
errors into the compensation matrix. (3) The single-stained control
must use the exact same fluorophore as the real experimental sam-
ple. Though this is an obvious rule, it is sometimes not followed
because researchers use different antibodies for their single-stained
control relative to their experiment. However, this can be detri-
mental in particular for tandem dyes because different product lots
will show different conjugation efficiencies. Therefore, it is required
that the antibody–fluorochrome molecule used to prepare the
20 Florian Mair and Aaron J. Tyznik

single-stained control is the same type and lot as the one used to
stain the actual sample. (4) The single-stained control should be
treated exactly the same way as the real experimental sample,
including fixation. The reason for this is that some fixation proto-
cols can impact fluorophore characteristics, and similar to the pre-
vious point, the single-stained control should carry the exact same
fluorophore as is present in the sample.
1. Prepare a spreadsheet showing antibody volumes to be
pipetted in the planned experiment.
2. Obtain cells (typically cryopreserved PBMCs or murine
splenocytes).
3. Resuspend cells in the appropriate volume and dispense into a
96-well plate.
4. Centrifuge plate at 400
g for 5 min. Flick supernatant and
dry the top of the plate by carefully placing it on a paper towel.
5. Add 100 μl of freshly prepared Fc-Block/Viability dye solution
and incubate for 15 min at room temperature (RT) in the dark
(see Note 8). In the meantime, start preparing the antibody
master mix (and the required FMO mixes) using Brilliant
Staining Buffer (see Note 9).
6. Wash by adding 200 μl of stain buffer, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel.
7. Add 50 μl of antibody staining mix containing the correct final
dilutions of all antibodies, resuspend, and incubate for 20 min
at RT in the dark.
8. In parallel, dilute 7–15 drops of compensation beads with stain
buffer to a total volume of 1.5 ml, aliquot 50 μl into the wells of
the 96-well plate, add 0.7 μl of a single antibody into each well.
Incubate for a minimum of 15 min at RT in the dark.
9. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel. Repeat this wash
once for a total of two washes.
10. Optional: add 100 μl of fixative to each well, resuspend, and
incubate for 20 min at RT in the dark.
11. Add 200 μl of stain buffer to all wells, centrifuge plate at
400
g for 5 min, flick supernatant, and dry the top of the
plate by carefully placing it on a paper towel.
12. Resuspend the cells in 100–200 μl of stain buffer and store at
4  C in the dark until acquisition.
13. Startup flow cytometer and perform daily QC as required per
manufacturer’s instructions.
 High-Dimensional Immunophenotyping 21

14. Adjust the PMT voltages to match your MFI target values
of your reference particles as described in Subheading 3.2,
steps 29–34.
15. Record 10,000 events from each single-stained control, and
perform compensation using instrument-specific acquisition
software.
16. Record as many events as desired from the full stain as well as
the FMO controls.
17. Export data in FCS3.1 file format and load into a flow cyto-
metry data analysis program. Follow the analysis steps outlined
in Subheading 3.7.

3.7 Analysis and Within this section we focus on manual analysis and quality control
Troubleshooting of the newly established panel. This can be performed using any
software that is able to read and visualize FCS files. Among the
commonly used stand-alone packages are FlowJo (BD Biosciences),
and a powerful cloud-based solution is offered by Cytobank (www.
cytobank.com), but there are several other options available on the
market (e.g., FCS Express from De Novo Software).
For fluorescent data an often-undervalued aspect is the appro-
priate transformation and visualization of the data, which is essen-
tial for successful manual analysis. Historically, flow cytometric data
has often been displayed using a purely logarithmic scale. This can
introduce visual artifacts at the lower end of the scale, which is why
the logicle/biexponential transformation was developed in the
mid-2000s [40–42]. These so-called hybrid scales will be logarith-
mic at the high end of the scale and gradually approach a linear scale
around 0, after which they would transition back to a logarithmic
scale below 0. In the biexponential implementation, the “width”
basis will determine how much space around 0 is displayed in a
linear domain, thus either “compressing” or “widening” the
appearance of the negative populations. Currently, there is no
reliable automated solution available, and thus researchers need to
adjust the width basis as well as the amount of negative decades
manually. This should be done in a way that (1) all events are visible
on a plot and (2) the negative population appears as a normally
distributed (i.e., round) population. In particular for complex
polychromatic data sets, it is essential to use logicle/biexponential
(or any other appropriate version such as arcsin) transformation of
the data in order to visualize spreading error (SE) accordingly.
After the first run of a newly established panel, validation and
troubleshooting is critical, as even with the best possible planning,
unexpected issues can arise from SE or improper compensation. To
assess this, two different strategies should be employed: (1) Prepare
an N
N view of all parameters against each other and visually
screen for erroneous patterns. This include either “leaning,” trian-
gular populations or the so-called “super-negative” populations
that appear below 0 (Fig. 6a). Of note, the negative population
22 Florian Mair and Aaron J. Tyznik

Fig. 6 Examples of erroneous staining patterns and (SE) spreading error. Example data are from the
development of a 28-color immunophenotyping panel centered on myeloid cells (published in OMIP-044).
 High-Dimensional Immunophenotyping 23

for a highly compensated fluorochrome pair can sometimes appear

as slightly tilted oval, which is not an issue, but a result of the
compensation calculation [43]. (2) Compare the fully stained
panel with all FMO controls. If there is a positive signal present in
the controlled detector of the FMO control, this usually can be
traced back to either erroneous compensation, or SE that has not
been anticipated. The source of this SE can usually be found by
visualizing the detector in question against all other fluorophores
present in the experiment.
An example where SE was dealt with by assigning fluorophores
to mutually exclusive markers is shown in Fig. 6b. Figure 6c depicts
an example where spreading error from one antigen (CX3CR1-PE-
Cy7) was lowering the resolution in the B790 detector. The solu-
tion to this problem was to move the highly expressed CX3CR1 to
another fluorochrome (BV421) that did not generate SE in the
B790 detector, and instead assigning PE-Cy7 to a target generating
less intense signal (and thus less SE).
1. Load all the data in your preferred flow cytometry data analysis
program and make sure that compensation has been applied.
2. Assess signal/time for all parameters by either displaying a
signal off each laser versus time, or use the automated function
of your analysis program (in FlowJo: press cmd + I). Signal over
time should be constant. Major, recurrent fluctuations in signal
over time could indicate issues with the stability of instrument
fluidics.
3. Do a cleanup gating based on FSC-A vs SSC-A to exclude
debris.
4. Exclude doublets by gating on FSC-A vs FSC-H and FSC-A vs
FSC-W.
5. Exclude dead cells by gating on live/dead stain negative events.
This population should be used for all further analysis and
validation.

Fig. 6 (continued) (a) Upper panel shows correctly compensated data. If overcompensated (middle panel), this
would be visible either as a population of “super-negative” events (if displaying two unrelated detectors) or as
a leaning, banana-shaped population pointing below 0 (if displaying the fluorophore–detector pair that is
wrongly compensated). If the population is undercompensated, this would be very hard to diagnose solely
from the full stain, but the FMO control clearly identifies the issue by showing a leaning, banana-shaped
population pointing toward the upper right. (b) Two examples where spreading error (SE) was dealt with by
assigning the corresponding fluorophores to mutually exclusive antigens. (c) Example for improving a
suboptimal panel after initial testing. Left plot shows CX3CR1 on PE-Cy7, which will show significant SE
into the B-780 detector (excitation laser: 488 nm, detection window: 780/60) and makes it hard to discern
CD38+ events in this detector. To improve this, CX3CR1 was moved to BV421, a fluorophore that does not
show any SE into the B-780 detector. PE-Cy7 should instead be used for a target that delivers a dimmer signal,
which will in turn diminish the SE generated
24 Florian Mair and Aaron J. Tyznik

6. Use the fully stained sample and generate an N
N view of all

parameters against each other.
7. Visually screen for “leaning” population patterns as displayed
in Fig. 6a (lower panel).
8. Visually screen for “super-negative” populations as displayed in
Fig. 6a (middle panel).
9. Prepare an overview gating delineating your populations of
interest, and at least one plot of every marker in your panel.
10. Using this gating hierarchy, compare the fully stained sample to
all FMO controls. Assess whether there is any spreading of the
negatives in the respective channel of interest in the FMO
control. Note that the FMO control can be used to set reliable
gates during manual analysis.
11. Assess whether all the targets in the panel are resolved.
12. If this is not the case, usually this is due to spreading error
(SE) from another fluorophore. Trace poorly resolving pairs
and assess whether the signal is too dim or whether there is
unexpected SE from another signal.
13. If the signal is too dim, assess whether a brighter fluorophore
can be used instead.
14. If SE is the problem, assess the problematic pair and try to
switch it to another fluorophore combination, keeping the
aspects in mind that were discussed during panel design in
Subheading 3.5 (see example in Fig. 6c).

4 Notes

1. Fixation is not a prerequisite for every experiment, but can be

considered optional. If processing human samples, fixation
might be required by shared resource laboratories prior to
acquisition for biosafety reasons, while for murine samples
this is not the case. In general, unfixed samples should be
acquired within 6–12 h after processing. If samples have been
fixed, they can be stored in the dark at 4  C for several days.
However, as we discuss in Subheading 3.6 it is imperative that
all single-stained controls are treated the same way as the
sample (i.e., fixation).
2. If in a voltration experiment maximal resolution (i.e., plateau of
SI) has not been reached before the CD4 population has
exceeded the linear range of the detector, an antibody with a
lower antigen density can be chosen for that detector and the
voltage titration repeated. Alternatively, a PMT voltage can be
chosen that sets the median fluorescence intensity (MFI) of the
CD4 positive peak no higher than 60,000. As a rule of thumb,
 High-Dimensional Immunophenotyping 25

at this MFI most common higher density antigens will remain

on scale for that detector while maintaining sufficient resolu-
tion of dim antigen targets. Finally, as discussed in more detail
in the following ref. 28, one can use quantum simply cellular
beads (QSC beads, available from Bangs laboratories) for the
voltration experiment, which will provide multiple signal inten-
sities within one sample.
3. The key requirements for fluorescent reference particles are
(1) long-term stability, (2) an appropriately bright signal in all
detectors and (3) an intrinsically small coefficient of variation
(CV). Commonly used are single-peak reference beads such as
Spherotech Supra Rainbow MidRange Fluorescent particles
(#SRCP-35-2A) or MidRange Rainbow Fluorescent particles
(#RFP-30-5A). However, since instrument configuration and
detection optics vary widely, it might be required to use a
combination of multiple fluorescent particles to obtain a suffi-
ciently bright signal in all detectors. In this case, Spherotech
Ultra Rainbow Particles (#URCP-38-2F) have been proven
useful, containing six peaks of varying fluorescent intensity
which are also suitable for detectors excited with a UV laser.
4. As discussed in Subheading 3.6, it is important that the positive
signal of your compensation control is as bright or brighter
than your experimental sample. The underlying reason for this
is again the Poisson error during photon counting at the detec-
tor. This error is proportional to the square root of the signal
intensity, that is, the error will be relatively larger for dimmer
signals. As the error is propagated throughout the compensa-
tion calculation, the compensation matrix will only be accurate
up to this signal intensity, and will be increasingly imprecise for
signals that are brighter than the compensation control. In the
author’s experience, using 1 μl of antibody with standard com-
pensation particles will deliver an appropriately bright signal,
though it might be subsaturation stain of the binding sites on
the compensation bead. If the signal is not sufficiently bright,
one has to either use more antibody to saturate all binding sites
or switch to the so-called Plus-compensation particles available
by various vendors which are larger in size and deliver a
brighter signal. In turn, if the signal from the compensations
beads is off-scale (i.e., too bright), one should use appropriately
diluted antibody solution to achieve a dimmer signal.
5. Depending on your biological targets, antigen expression eval-
uation and corresponding reagent titration may require either
perturbation to your target cells (stimulation, etc.) or the
addition of costains to identify small subsets of cells that express
your antigen of interest. In the case of variable or inducible
antigen expression, it is advisable to use a similar sample source
to what will be used in your final assay as background staining
26 Florian Mair and Aaron J. Tyznik

and receptor expression may vary and impact the choice of an

appropriate reagent titer. Whenever possible, a primary cell that
represents your target antigen expression level should be pre-
ferred over cell lines or other surrogates. In addition, your
titration experiments should be performed on cells that are
treated the same way as your experimental cells including any
digestions, stimulations, permeabilization, or fixation steps.
If the target antigen is expressed on a small subset of
your bulk cell population and you require costains, it is impor-
tant to make sure that the primary antibody–fluorochrome
needed to identify your targets do not show any SE into the
detectors that are used for titrated antibodies. Appropriate
fluorophores can be determined by utilizing the SSM as
described in Subheading 3.3.
6. Commonly, FMOs are used to control only for a single fluoro-
chrome. However, there are two situations where one could
use FM2 or FM3 controls: First, if the detector to be controlled
for is collecting SE from multiple fluorochromes, the FM2/3
can be used during panel optimization to assess the cumulative
SE that is contributed from the two or three omitted fluoro-
chromes relative to an FMO. Second, if FMO controls are to be
included in repeated experiments, it can save pipetting effort
(and sample) if FM2 controls are made in a way that the two
omitted fluorochromes are explicitly tested to not have any
cross talk between each other.
7. While compensation beads are the preferred way of preparing
single-stained control samples, in the authors’ experience, not
all compensation particles work equally well with all fluoro-
chromes, in particular some of the Brilliant-Violet and Brilliant
Ultraviolet tandem dyes (e.g., BUV737, or BV786). For these
dyes, the resulting compensation value might have variability
compared to cells that is not appropriate for your specific assay
requirements. Thus, users should always assess whether the
chosen type of compensation particle delivers appropriate com-
pensation values by utilizing appropriate FMO controls.
8. As shown by multiple labs, utilizing Fc-blocking reagents is
critical to prevent possible artifacts arising from unspecific anti-
body binding to the Fc-receptors of various myeloid cells
[44]. To save time, in the authors’ experience it is possible to
combine the Fc-blocking and live/dead staining step. This will
result in a slight decrease in the intensity of the live/dead
staining signal, as the amino-reactive dye will also react with
the Fc-blocking reagent, but this does not negatively impact
assay performance.
9. Polymer-based dyes of the Brilliant-Violet or Brilliant-
Ultraviolet family are commonly used for high-dimensional
 High-Dimensional Immunophenotyping 27

panels, and will likely become even more prevalent in the future
[45]. Due to their chemistry, these polymer-based dyes can
potentially interact with each other, causing artifacts in flow
cytometry experiments that appear as inappropriately compen-
sated data. To prevent this, it is advisable to use commercially
available staining buffers preventing this interaction every time
one is using more than two polymer-based dyes in an experi-
ment. Since different vendors use different proprietary formu-
lations for these buffers, users might need to test which buffer
works best for their experimental setup.

Acknowledgments

The authors thank all members of the Prlic lab and Dr. Sabine Spath
for critical reading of the manuscript. F.M. and A.J.T. are supported
by a Marylou scholarship from the International Society for
Advancement of Cytometry (ISAC).

References
1. Fulwyler MJ (1965) Electronic separation of 8. Klein AM, Mazutis L, Akartuna I et al (2015)
biological cells by volume. Science Droplet barcoding for single-cell transcrip-
150:910–911 tomics applied to embryonic stem cells. Cell
2. Robinson JP, Roederer M (2015) History of 161:1187–1201. https://doi.org/10.1016/j.
science flow cytometry strikes gold. Science cell.2015.04.044
350:739–740. https://doi.org/10.1126/sci 9. Zheng GXY, Terry JM, Belgrader P et al
ence.aad6770 (2017) Massively parallel digital transcriptional
3. Bandura DR, Baranov VI, Ornatsky OI et al profiling of single cells. Nat Commun
(2009) Mass cytometry: technique for real time 8:14049. https://doi.org/10.1038/
single cell multitarget immunoassay based on ncomms14049
inductively coupled plasma time-of-flight mass 10. Brodie TM, Tosevski V (2017) High-
spectrometry. Anal Chem 81:6813–6822. dimensional single-cell analysis with mass cyto-
https://doi.org/10.1021/ac901049w metry. Curr Protoc Immunol
4. Bendall SC, Simonds EF, Qiu P et al (2011) 118:5.11.1–5.11.25. https://doi.org/10.
Single-cell mass cytometry of differential 1002/cpim.31
immune and drug responses across a human 11. Papalexi E, Satija R (2017) Single-cell RNA
hematopoietic continuum. Science sequencing to explore immune cell heteroge-
332:687–696. https://doi.org/10.1126/sci neity. Nat Rev Immunol 510:363. https://doi.
ence.1198704 org/10.1038/nri.2017.76
5. Spitzer MH, Nolan GP (2016) Mass cytome- 12. Chattopadhyay PK, Roederer M (2012) Cyto-
try: single cells, many features. Cell metry: today’s technology and tomorrow’s
165:780–791. https://doi.org/10.1016/j. horizons. Methods 57:251–258. https://doi.
cell.2016.04.019 org/10.1016/j.ymeth.2012.02.009
6. Tang F, Barbacioru C, Wang Y et al (2009) 13. Mair F, Prlic M (2018) OMIP-044: 28-color
mRNA-Seq whole-transcriptome analysis of a immunophenotyping of the human dendritic
single cell. Nat Methods 6:377–382. https:// cell compartment. Cytometry A 106:255.
doi.org/10.1038/nmeth.1315 https://doi.org/10.1002/cyto.a.23331
7. Shalek AK, Satija R, Adiconis X et al (2013) 14. Grégori G, Patsekin V, Rajwa B et al (2012)
Single-cell transcriptomics reveals bimodality Hyperspectral cytometry at the single-cell level
in expression and splicing in immune cells. using a 32-channel photodetector. Cytometry
Nature 498:236–240. https://doi.org/10. A 81:35–44. https://doi.org/10.1002/cyto.
1038/nature12172 a.21120
28 Florian Mair and Aaron J. Tyznik

15. Futamura K, Sekino M, Hata A et al (2015) Cytometry A 73:767–776. https://doi.org/

Novel full-spectral flow cytometry with multi- 10.1002/cyto.a.20595
ple spectrally-adjacent fluorescent proteins and 27. Meinelt E, Reunanen M, Edinger M et al Stan-
fluorochromes and visualization of in vivo cel- dardizing application setup across multiple
lular movement. Cytometry A 87:830–842. flow cytometers using BD FACSDiva™ Ver-
https://doi.org/10.1002/cyto.a.22725 sion 6 Software: technical bulletin
16. Feher K, Volkmann von K, Kirsch J et al (2016) 28. Perfetto SP, Ambrozak D, Nguyen R et al
Multispectral flow cytometry: the conse- (2012) Quality assurance for polychromatic
quences of increased light collection. Cytome- flow cytometry using a suite of calibration
try A 89:681–689. https://doi.org/10.1002/ beads. Nat Protoc 7:2067–2079. https://doi.
cyto.a.22888 org/10.1038/nprot.2012.126
17. Kvistborg P, Gouttefangeas C, Aghaeepour N 29. Bagwell CB, Adams EG (1993) Fluorescence
et al (2015) Thinking outside the gate: single- spectral overlap compensation for any number
cell assessments in multiple dimensions. Immu- of flow cytometry parameters. Ann N Y Acad
nity 42:591–592. https://doi.org/10.1016/j. Sci 677:167–184
immuni.2015.04.006 30. Roederer M (2001) Spectral compensation for
18. Saeys Y, Gassen SV, Lambrecht BN (2016) flow cytometry: visualization artifacts, limita-
Computational flow cytometry: helping to tions, and caveats. Cytometry 45:194–205
make sense of high-dimensional immunology 31. Ashhurst TM, Smith AL, King NJC (2017)
data. Nat Rev Immunol 16:449–462. https:// High-dimensional fluorescence cytometry.
doi.org/10.1038/nri.2016.56 Curr Protoc Immunol 10:5.8.1–5.8.38.
19. Mair F, Hartmann FJ, Mrdjen D et al (2016) https://doi.org/10.1002/cpim.37
The end of gating? An introduction to auto- 32. Nguyen R, Perfetto S, Mahnke YD et al (2013)
mated analysis of high dimensional cytometry Quantifying spillover spreading for comparing
data. Eur J Immunol 46:34–43. https://doi. instrument performance and aiding in multi-
org/10.1002/eji.201545774 color panel design. Cytometry A 83:306–315.
20. Chester C, Maecker HT (2015) Algorithmic https://doi.org/10.1002/cyto.a.22251
tools for mining high-dimensional cytometry 33. Roederer M, Tárnok A (2010) OMIPs—
data. J Immunol 195:773–779. https://doi. orchestrating multiplexity in polychromatic sci-
org/10.4049/jimmunol.1500633 ence. Cytometry A 77:811–812. https://doi.
21. Aghaeepour N, Finak G, FlowCAP Consor- org/10.1002/cyto.a.20959
tium et al (2013) Critical assessment of auto- 34. Mahnke Y, Chattopadhyay P, Roederer M
mated flow cytometry data analysis techniques. (2010) Publication of optimized multicolor
Nat Methods 10:228–238. https://doi.org/ immunofluorescence panels. Cytometry A
10.1038/nmeth.2365 77:814–818. https://doi.org/10.1002/cyto.
22. Brinkman RR, Aghaeepour N, Finak G et al a.20916
(2016) Automated analysis of flow cytometry 35. Liechti T, Günthard HF, Trkola A (2018)
data comes of age. Cytometry A 89:13–15. OMIP-047: high-dimensional phenotypic
https://doi.org/10.1002/cyto.a.22810 characterization of B cells. Cytometry A
23. Cossarizza A, Chang H-D, Radbruch A et al 103:2262–2596. https://doi.org/10.1002/
(2017) Guidelines for the use of flow cytome- cyto.a.23488
try and cell sorting in immunological studies. 36. Moncunill G, Han H, Dobaño C et al (2014)
Eur J Immunol 47:1584–1797. https://doi. OMIP-024: pan-leukocyte immunophenoty-
org/10.1002/eji.201646632 pic characterization of PBMC subsets in
24. Roederer M (2015) A proposal for unified flow human samples. Cytometry A 85:995–998.
cytometer parameter naming. Cytometry https://doi.org/10.1002/cyto.a.22580
A. https://doi.org/10.1002/cyto.a.22670 37. Hulspas R (2010) Titration of fluorochrome-
25. Perfetto SP, Chattopadhyay PK, Wood J et al conjugated antibodies for labeling cell surface
(2014) Q and B values are critical measure- markers on live cells. Curr Protoc Cytom
ments required for inter-instrument standardi- Chapter 6:Unit 6.29–6.29.9. https://doi.
zation and development of multicolor flow org/10.1002/0471142956.cy0629s54
cytometry staining panels. Cytometry A 38. Perfetto SP, Chattopadhyay PK, Roederer M
85:1037–1048. https://doi.org/10.1002/ (2004) Seventeen-colour flow cytometry:
cyto.a.22579 unravelling the immune system. Nat Rev
26. Lawrence WG, Varadi G, Entine G et al (2008) Immunol 4:648–655. https://doi.org/10.
Enhanced red and near infrared detection in 1038/nri1416
flow cytometry using avalanche photodiodes.
 High-Dimensional Immunophenotyping 29

39. Mahnke YD, Roederer M (2007) Optimizing a 43. Roederer M (2016) Distributions of autofluor-
multicolor immunophenotyping assay. Clin escence after compensation: be panglossian,
Lab Med 27:469–485. v. https://doi.org/10. fret not. Cytometry A 89:398–402. https://
1016/j.cll.2007.05.002 doi.org/10.1002/cyto.a.22820
40. Parks DR, Roederer M, Moore WA (2006) A 44. Andersen MN, Al-Karradi SNH, Kragstrup
new “Logicle” display method avoids deceptive TW, Hokland M (2016) Elimination of erro-
effects of logarithmic scaling for low signals neous results in flow cytometry caused by anti-
and compensated data. Cytometry A body binding to Fc receptors on human
69:541–551. https://doi.org/10.1002/cyto. monocytes and macrophages. Cytometry A
a.20258 89:1001–1009. https://doi.org/10.1002/
41. Trotter J (2007) Alternatives to log-scale data cyto.a.22995
display. Curr Protoc Cytom Chapter 10:Unit 45. Chattopadhyay PK, Gaylord B, Palmer A et al
10.16–10.16.11. https://doi.org/10.1002/ (2012) Brilliant violet fluorophores: a new class
0471142956.cy1016s42 of ultrabright fluorescent compounds for
42. Herzenberg LA, Tung J, Moore WA et al immunofluorescence experiments. Cytometry
(2006) Interpreting flow cytometry data: a A 81A:456–466. https://doi.org/10.1002/
guide for the perplexed. Nat Immunol cyto.a.22043
7:681–685. https://doi.org/10.1038/
ni0706-681
 Chapter 2

Immunophenotyping by Mass Cytometry

Gregory K. Behbehani

Abstract
Mass cytometry is a novel technology similar to flow cytometry in which antibodies are tagged with heavy
metal molecules rather than fluorophores and then detected with time-of-flight mass spectrometry. This
enables measurement of up to 50 simultaneous parameters with no autofluorescent background and little or
no spillover or required compensation. Mass cytometry has tremendous potential for the analysis of highly
complex research or clinical samples and can measure 40–50 immunophenotypic markers at a time. This
chapter describes most of the commonly used methods for performing basic immunophenotyping experi-
ments by mass cytometry, and how this can be combined with measurement of cellular functional
properties.

Key words Immunophenotyping, Mass cytometry, CyTOF, Flow cytometry, Minimal residual dis-
ease, Acute leukemia, Aberrant marker expression

1 Introduction and General Considerations

Flow cytometry is one of the most used technologies in the basic

science and clinical research laboratory. It is now an essential tool in
the study of malignancies, infectious diseases, and immune system
function. Its utility stems from its ability to analyze single cells for
their expression of virtually any antigen to which an antibody can be
specifically bound. This enables the identification and quantifica-
tion of immunophenotypically different cell types in complex bio-
logic samples even if the cells of interest are extremely rare. Despite
its utility, however, most current clinical flow cytometry technolo-
gies are limited to analysis of 4–15 simultaneous measurement
parameters, primarily due to the dependence of the technology
on fluorescent reporter molecules (with relatively broad absorption
and emission spectra), which also makes flow cytometry susceptible
to various artifacts due to the fluorescent properties of the cells
being studied and interactions between the fluorescent reagents
being used to study them. While the technology of fluorescent
cytometry continues to advance rapidly with newer machines and

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_2, © Springer Science+Business Media, LLC, part of Springer Nature 2019

31
32 Gregory K. Behbehani

newer reagents, the most transformative advance in cytometry has

been the development of mass cytometry, which allows for immu-
nophenotypic measurements without the problems created by fluo-
rescent reporters and optic measurement [1].

1.1 Mass Cytometry Mass cytometry (MC) is a cytometry technique similar to fluores-
Versus Fluorescent cent flow cytometry that detects cellular antigens through the use
Flow Cytometry of mass spectrometry to detect binding of heavy-metal conjugated
antibodies to the cells being studied. Mass cytometry experiments
are typically performed using the same antigen-specific antibodies
used in conventional flow cytometry, but the binding of these
antibodies is measured by vaporizing and ionizing the cells (with
the bound metal-conjugated antibodies) followed by the
subsequent detection of the molecules of isotopically purified
heavy metal atoms by time-of-flight mass spectrometry
(ICP-MS). The primary advantage of this approach stems from
the ability of ICP-MS to distinguish ions of different atomic weight
with less than 0.5% signal spillover between adjacent masses (which
are each conjugated to a particular antibody). The data generated
from the analysis of each antigen is very similar to the data gener-
ated when the same antigen is analyzed by flow cytometry, and
currently up to 40–50 parameters can be simultaneously recorded
per cell with minimal or no signal compensation and no back-
ground due to autofluorescence. The greatest advantage of mass
cytometry is the large increase in the number of simultaneous
parameters that can be recorded. This is particularly useful in the
analysis of highly complex or heterogeneous cell populations in
which the additional parameters enable simultaneous identification
of numerous cell types or functional states to be uniquely identified
and characterized, even when some of these cell types are very rare.
There are now numerous publications utilizing this technology and
many excellent reviews [2–6]; this chapter will provide a basic
protocol for immunophenotyping by mass cytometry, how this
can be combined with measurement of intracellular antigens for
functional studies, and focus on the unique practical considerations
that are important in designing mass cytometry experiments.
Since mass cytometry experiments can typically be performed
using the same exact antibody clones as fluorescent flow cytometry
experiments, the immunophenotypes derived from the two techni-
ques should be nearly identical. However, labeling of antibodies
with chelated heavy metal and its detection by time-of-flight mass
spectrometry creates several unique considerations that must be
accounted for in the design of a mass cytometry experiment.
These can be grouped into issues related to the antibody labeling
polymer and the ion detector.
Antibody considerations: Mass cytometry antibodies are pro-
duced by conjugation of polymer of DTPA chelators (functiona-
lized using a maleimide group) to available sulfhydryl groups of the
 Immunophenotyping by Mass Cytometry 33

antibody of interest. The majority of antibodies are produced using

a polymer called Maxpar® X8, which researchers can purchase in a
kit from the manufacturer (Fluidigm) to custom-label their own
antibodies (note that Fluidigm also uses a second polymer for a
small subset of their preconjugated antibodies that has slightly
different properties; see Note 1). The vast majority of commercially
available antibodies can be labeled in this way (except for antibodies
sensitive to reducing conditions, such as IgM clones), provided that
the antibody can be obtained (from the manufacturer or by purifi-
cation) without any other proteins in the antibody solution.
Regardless of whether the antibodies are custom-conjugated or
purchased preconjugated, most mass cytometry antibodies have a
staining intensity similar to what would be achieved if that of the
same antibody was conjugated to FITC and analyzed by flow
cytometry. This can be advantageous in that most of the reagents
will have similar measurement sensitivity, but this also means that
antigens of very low abundance that typically show changes of only
two- to threefold when measured by flow cytometry using antibo-
dies conjugated to bright fluorochromes (e.g., APC or PE) can be
difficult to detect with mass cytometry. Such low signals can poten-
tially be increased by detecting the antigen of interest with multiple
metal-labeled antibodies that are specific for the antigen (provided
that the antibodies do not sterically interfere with one another).
Additionally, the molecular structure of the metal chelating poly-
mers appears to cause them to adhere nonspecifically to certain rare
cell types such as eosinophils (likely due to interactions with intra-
cellular cations), but this effect can generally be blocked using
heparin, as described by Rahman et al. [7] Another advantage of
mass-tagged antibodies is that they are generally stable to the
majority of cellular processing and permeabilization methods,
enabling staining of surface antigens followed by permeabilization
or other additional processing steps without significant disruption
of the bound antibodies or their attached metal. By using post-
staining fixation, mass cytometry surface antigen staining has been
combined with hybridization intracellular RNA [8], staining for
nucleotide incorporation [9], and staining for hypoxia [10], along
with detection of other routine intracellular antigens. Given this
advantage, even in experiments that combine immunophenotypic
assessment with measurement of intracellular functional markers,
surface markers can be stained fresh or after only minimal cell
manipulation (e.g., gentle fixation).
As with all antibodies, mass-tagged antibodies should always be
titrated for optimal staining concentration. It is worth noting that
most metal-labeled antibodies do exhibit very small amounts of
nonspecific cell binding at high concentrations, meaning that (like
fluorescent flow cytometry) saturation of the antigen of interest and
optimal signal to noise ratio can occur at antibody concentrations
below the maximal achievable staining intensity. It is thus
34 Gregory K. Behbehani

important to include good negative controls in all antibody

titrations. Fortunately, the relative lack of spillover between
mass channels allows numerous antibody reagents to be titrated
simultaneously.
Detector considerations: The mass cytometer’s detector is
another obvious difference from a traditional fluorescent flow cyt-
ometer. Most notably, the mass cytometer only has a single ion
detector. The ion detector is automatically tuned daily, greatly
simplifying machine setup relative to fluorescent systems with mul-
tiple (often instrument-specific) filter sets and photomultipliers.
This fact also creates a drawback, however, in that it is not possible
to adjust detector sensitivity separately for each measurement chan-
nel. This makes it important to consider the mass channels used for
antigens expected to be at high abundance, as discussed in Sub-
heading 2.3. Additionally, unlike fluorescent flow cytometry detec-
tors, the mass cytometer’s detector has a finite lifetime and will
typically need to be replaced approximately every 12–18 months
depending on usage. Staining of samples with high levels of metal
can greatly accelerate the wear of the detector and, since its sensi-
tivity cannot be adjusted separately for each channel, it is important
to ensure samples are not stained with excessively high amounts of
metal.

1.2 General Method In this chapter the most common approaches to mass cytometry
Design cell processing will be described. However, if converting a previ-
ously established fluorescent flow cytometry protocol to a mass
cytometry protocol, the same sample processing and antibody
staining conditions can typically be used. The only absolutely
required modifications needed to convert a flow cytometry immu-
nophenotyping protocol to mass cytometry protocol are as
follows: (1) use of metal-conjugated antibodies instead of the
fluorochrome-conjugated ones, (2) the use of a fixative-containing
metal intercalator solution after the completion of antibody stain-
ing (Subheading 3.6, steps 1–3), and (3) running the sample into
the mass cytometer after dilution in either pure water or a
specialized running buffer free from nonorganic salts (Subheading
3.6, steps 4–9). If using commercially prepared solutions for fixa-
tion and permeabilization, all solutions should be checked for heavy
metal contamination on a regular basis (unless the solutions are
specifically made for mass cytometry).
The minimum number of cells required for mass cytometry
analysis varies based on the exact number of centrifugation steps,
the type of centrifuge tubes employed, and the number of washes
performed. For the protocol described here (utilizing standard
5 mL polystyrene FACS tubes), 500,000 cells is a good minimum
number as the mass cytometer generates data from 30% to 50% of
the cells that are injected (after staining and washing), necessitating
a proportional increase in the number of starting cells (as compared
 Immunophenotyping by Mass Cytometry 35

to fluorescent flow cytometry). As a general rule, we typically

expect to get data events from 10% to 20% of the cells present at
the beginning of the experiment, thus 500,000 cells will typically
yield 50–100,000 final cell events in the final data. Utilizing smal-
ler, polypropylene tubes, and reducing the number of staining and
wash steps can allow for the analysis to be preformed with lower
starting cell numbers (as low as 100,000). A good approach to
determining the number of starting cells, is to estimate the
expected frequency of the rarest cell population of interest and
the absolute number of these events desired, the total number of
starting cells necessary can then be calculated based on these num-
bers and the 10–20% rate at which starting cells result in data
events. Cells fixed according to this protocol (Subheading 3.4)
will be stable for at least several months at
80  C, and we have
successfully analyzed cells 2–3 years after fixation.

2 Materials

2.1 Precautions To avoid background due to heavy metal contamination, all solu-
tions for mass cytometry should be prepared using ultrapure water
and all reagent solutions (when first prepared from new lots of stock
reagents) should be tested for heavy metal contamination (see
Note 2). Note that cisplatin, paraformaldehyde, and the Smart
Tube buffer are all toxic and potentially mutagenic. They should
be handled with appropriate protections (gloves, eye and respira-
tory protection as appropriate). It is also important to mention that
standard laboratory dishwashers, autoclaves, and dishwashing
detergents are commonly contaminated with heavy metals (partic-
ularly lead and barium) that can disrupt experiments or damage the
mass cytometer. We use either disposable plastic containers or
glassware that we wash by hand for producing and storing all
mass cytometry solutions and reagents.

2.2 Fixatives, 1. Paraformaldehyde (PFA), 16% solution: This must be

Buffers, and Tubes methanol-free (i.e., not formalin). We purchase this in 10 mL
ampules (Electron Microscopy Sciences) and transfer it to foil-
wrapped tubes, as it will lose activity over several weeks upon
exposure to air and light. We will throw out any PFA that has
been open for more than 1 month.
2. Smart Tube proteomic stabilizer (STPS; alternate fixative): can
be purchased from Smart Tube incorporated (San Carlos, CA).
This comes as a working solution that can be added directly to
cell samples at a sample–buffer ratio of 1:1.4.
3. Pure methanol (optional): This should be kept cold (
20  C to
4  C) in a sealed bottle. This should not have any drying agents
added, as these may contain heavy metals.
36 Gregory K. Behbehani

4. Cell staining media (CSM): Standard phosphate buffered saline

(PBS), plus 0.5% bovine serum albumin (BSA), and 0.02%
sodium azide, pH ¼ 7.4. This is typically made this in 4 L
batches by adding 20 g of BSA and 800 mg of sodium azide.
Start with 3 L of sterile PBS and mix in the dry ingredients until
dissolved. Then add additional PBS to a total volume of 4 L.
Sterile-filter this using a 0.2 μM bottle-top filter into rinsed,
500 mL glass bottles. Rinse and hand-wash these bottles as
necessary, but do not place them into standard laboratory dish-
washers. Premade Maxpar® cell staining buffer can be substi-
tuted for CSM. Other premade PBS and blocking protein
solutions could also likely be substituted for CSM; however,
they should be checked for heavy metal contamination
before use.
5. Intercalator solution is made by addition of the iridium-based
intercalator solution (Fluidigm, 201192A or 201192B) to PBS
at a final concentration of approximately 125 nM. To this add
1/10th volume of 16% PFA to achieve a final concentration of
1.5% PFA. This solution should be made fresh just before
addition to the cells. The failure to use adequate amounts of
fresh (i.e., active) fixative will cause cells to degrade while being
analyzed (since analysis is typically performed in pure water).
6. Cisplatin: This can be purchased from a variety of manufac-
turers or can be purchased from Fluidigm as a natural isotopic
mixture or in the form of its purified 194 or 198 isotopes. We
typically prepare this ourselves as a concentrated 5 mM stock in
DMSO and aliquot it into single-use tubes.
7. This protocol is written for use with standard 5 mL polystyrene
FACS tubes, but a variety of other tubes can be used. We have
successfully performed the protocol in 1.5 mL Eppendorf-style
microcentrifuge tubes, and 1.1 mL polypropylene “cluster”
microcentrifuge tubes, though additional washes will be
required if the tube size is small relative to the residual volumes
left after each wash step (see Note 3).
8. Cells: Essentially any cell that can be placed into a single-cell
suspension (without disrupting the biology of interest) can be
analyzed by mass cytometry; however, cells that normally grow
or exist in a single-cell suspension tend to yield the best results.
Note that each mass cytometry sample typically requires at least
500,000 cells per sample (we typically collect a minimum of
two million cells to allow for one million cells to be stained and
analyzed twice if necessary; see Subheading 3).

2.3 Antibodies 1. As described above in the introduction, antibodies can be

and Antibody Panel purchased preconjugated from the manufacturer (Fluidigm)
Design Considerations or can be conjugated with metal using a kit from Fluidigm in
 Immunophenotyping by Mass Cytometry 37

accordance with the provided protocol. Additionally, other

non-Fluidigm conjugation protocols are available [11].
2. This protocol is largely based on our experience using cells
fixed using PFA or STPS fixation and methanol cell permeabi-
lization, an approach used by a large fraction of other MC
researchers. We have additionally performed similar experi-
ments on fresh (unfixed) cells with good results (notably,
most antigens exhibited similar staining regardless of fixation).
We have not extensively tested staining under other permeabi-
lization conditions (e.g., saponin), though other permeabiliza-
tion methods would likely be compatible with this method as
these same antigens have been successfully analyzed with other
protocols and are supported by Fluidigm [12].
3. The relative lack of spillover in mass cytometry allows greater
flexibility in panel design; however, many of the same consid-
erations that are important in fluorescent panel design are still
important for mass cytometry. Specifically, it is still important
to pay close attention to situations in which high abundance
antigens are on mass channels near very low abundance anti-
gens. The mass cytometer itself has a spillover of only 0.3% into
adjacent channels (technically termed abundance sensitivity);
however, most of the metals used for metal conjugation are
isotopically purified from natural populations of mixed isotopes
(e.g., natural neodymium is purified into its seven isotopes,
each of which is a separate mass channel) and this purification
is usually not complete. Thus, most mass cytometry metals are
contaminated with 0.5–2% of one or more other isotopes of
that same metal (see Note 4). Additionally, all metals oxidize to
some extent after ionization and before the TOF analysis; this
effectively creates a 0.1–2.5% spillover into the channel 16 Da
above (due to the mass of an oxygen atom; the exact percentage
of oxide signal is dependent on the chemical properties of each
element; see Note 5).
4. While generally not needed, compensation can be applied to
mass cytometry experiments. We have performed manual com-
pensation in rare cases where it is required, by either compen-
sating for the known isotopic impurities of the relevant isotope,
by using cells known to be negative for the antigen receiving
the spillover, or by using a “minus one” control cell. In addi-
tion, Chevrier et al. [13] have recently developed an automated
methodology for applying compensation to mass cytometry.
While both approaches can effectively compensate spillovers
(achieving a median spillover signal of 0 counts), signals of
greater than several thousand counts can create spillovers that
cannot be compensated with a linear matrix. Additionally, even
lower intensity signals 1000–5000 counts can lead to widened
population distributions (in channels receiving spillover) of
38 Gregory K. Behbehani

10–20 counts around the correctly compensated median inten-

sity that can overwhelm changes in very low expression anti-
gens being measured on the channel receiving spillover. Such
spillovers issues should be avoided by designing the antibody
panel so that high expression antigens are placed onto channels
with minimal spillovers into channels measuring antigens with
low expression levels. Alternatively, high abundance antigens
can be placed onto mass channels with low sensitivity (e.g.,
89Y), which can also facilitate measurement of more markers
by saving the higher sensitivity channels for the low abundance
markers that require the additional sensitivity (see Note 6).
5. Fluidigm does provide an online tool to assist with these panel
design issues (available on their website). Alternatively, the use
of premade commercial antibody cocktails, or antibody designs
previously published by other researchers can alleviate many of
the concerns around panel design.

3 Methods

3.1 Viability Staining If a significant fraction of the cells of interest could be dead at the
(Optional) time of collection, it is generally advisable to perform a viability
(live–dead) stain prior to antibody staining or cell fixation. Dead or
dying cells typically display different size and surface marker expres-
sion characteristics and almost universally will have altered levels of
intracellular functional markers. It is thus critical to have a means of
excluding dead cells (if expected to be present) as part of most
experimental designs. This is particularly important for analysis of
cryopreserved cells, in which the freezing and thawing process will
reliably kill a fraction of cells by necrosis. The simplest method of
live–dead staining is use of cisplatin (which nonspecifically and
covalently binds to protein), based on the publication from Fien-
berg et al. [14], which will be described here. Alternative
approaches include staining with bifunctional chelators loaded
with heavy metal, such as barcoding reagents [15], or use of
metal-containing DNA intercalators such as rhodium [16]
(although in our experience the latter stain is generally not stable
enough to be suitable if intracellular antigens will be measured). In
all cases, the basic concept is to incubate cells in a reagent that does
not rapidly cross the cell membrane of live cells but can rapidly
enter the disrupted membrane of a dead cell where it will bind to
intracellular DNA or protein, resulting in higher signal in the dead
cells.
1. Fresh cells (from primary tissue or in vitro culture media, prior
to any fixation) should first be washed one or two times with
serum/protein-free PBS (see Notes 7 and 8).
 Immunophenotyping by Mass Cytometry 39

2. The cells should then be resuspended in serum/protein free

PBS at a concentration of one to ten million cells per mL and
mixed to achieve an even single cell suspension.
3. Cisplatin should be added from the concentrated stock to the
cell suspension to achieve a final concentration of 1–10 μM (the
manufacturer recommends 5 μM; see Note 9).
4. Mix the cell suspension well and maintain at room temperature
for 5 min. (The time of incubation can also be titrated if
necessary.)
5. Following incubation with cisplatin, the reaction should be
quenched with 5 volume of PBS with either BSA (0.5%) or
serum (10–50%) added to it; alternatively pure human or ani-
mal serum or CSM could be used to quench the staining
reaction (note that the sodium azide in CSM could interfere
with subsequent functional studies). The protein in this solu-
tion will bind any remaining cisplatin and stop the staining
reaction.
6. Wash cells once in either: PBS + BSA, serum, or CSM.
7. If properly titrated, this staining step will not cause significant
short term (minutes to a few hours) toxicity to live cells [14], so
subsequent stimulations or other brief functional assessments
can be performed after cisplatin staining, if desired.
8. Platinum signal from natural platinum is typically recorded on
the 195 Dalton channel; if using an isotopically purified cis-
platin, measurement should be performed on the relevant
channel (194Pt or 198Pt). When the Pt staining reaction has
been properly titrated, there will be low levels of Pt staining on
live cells and dead cells will demonstrate a 1–2 log increase in Pt
incorporation.

3.2 Cell Stimulation While beyond the scope of this chapter, if the planned experiment
(Optional) calls for a cell stimulation prior to antibody staining (e.g., cytokine
stimulation [17] or IdU incorporation for cell cycle analysis [18])
this should be performed after viability staining and before anti-
body surface staining. Typically, if a stimulation is performed, the
cells will need to be fixed with PFA or STPS (or another fixative
product) before antibody staining to preserve the intracellular
response to the stimulation.

3.3 Fresh Sample If the primary objective of an experiment is measurement of immu-

Surface Staining nophenotype (rather than functional cell properties), cells can be
stained live, provided that the relevant antibodies have been titrated
on live cells using the staining time and temperatures that will be
employed in the final experiment. It is worth noting that many of
the original mass cytometry publications utilized cells that had been
previously been fixed before staining and stained for longer
40 Gregory K. Behbehani

(45–60 min) incubations at room temperature, thus the published

antibody concentrations from such publications may not perform
the same using the protocol described below. Note that traditional
barcoding methodologies [15, 19] will not perform well on live
cells; however, antibody-based barcoding methods [20] can be
used if barcoding is needed. (If this is desired, the relevant pub-
lished protocol should be followed for antibody barcoding.) Live
cell staining will provide the most accurate representation cell
surface immunophenotype; however, the cell processing of the
live cells is very likely to disrupt the functional state of the cells
(intracellular signaling, cell cycle state, etc.) so if such measure-
ments are important, gentle fixation prior to staining can be used
to “freeze” the intracellular state of the cells at the expense of
potentially decreasing staining of a subset of fixation-sensitive
antigens.
1. The desired antibody cocktail should be prepared just prior to
the start of cell processing (though we have prepared cocktails
the day before with success). Each antibody should be diluted
such that it will be at its empirically determined optimal stain-
ing concentration after the cocktail is added to the sample (see
Note 10). For live cells, the cocktail can be made in CSM or
alternatively (if the sodium azide will interfere with subse-
quently planned functional assessment) in PBS with added
albumin or serum. We typically add 400 units per mL of
sodium heparin to the antibody cocktail to prevent nonspecific
antibody binding [7]. Typically, we will aspirate cells after each
wash down to a final volume of 50 μL and make our antibody
cocktail at a 2 concentration, so that we can add 50 μL of
antibody cocktail to the residual 50 μL to achieve a final 1
antibody concentration in 100 μL.
2. Cells should be washed (see Note 7) in the desired staining
buffer one or two times prior to the start of the staining
reaction. It is desirable, but not absolutely required, that pro-
tein be present in this solution to block nonspecific antibody
binding. Typically for live cell staining, serum from the relevant
species is used directly or added to PBS. (CSM could also be
used, but the sodium azide could create cellular toxicity and
may need to be removed when staining live cells.) Cells can also
be stained in whole blood or the primary in vivo body fluid the
cells came from. We typically add sodium heparin to a final
concentration of 400 Units per mL into the final wash of the
cells. Typically, cells are cooled to 4  C during this process (see
Note 11). If blood is used, we typically use sodium heparin as
the anticoagulant when the blood is initially collected.
 Immunophenotyping by Mass Cytometry 41

3. (Optional) Add an Fc blocking reagent at the manufacturer’s

recommended concentration and incubate for 10 min (see
Note 12).
4. Add the antibody cocktail at the previously determined optimal
concentrations. Typically for staining of live cells, staining
should be performed for 30 min at 4  C or on ice. Cells should
be gently mixed periodically or placed on a low speed rotator or
shaker to keep the cells evenly suspended.
5. Wash twice in CSM or PBS.
6. If performing intracellular staining skip to Subheading 3.5 on
intracellular staining; if only performing a surface stain skip to
Subheading 3.6 on DNA intercalation.

3.4 Fixed Sample 1. Fixation: If the goal of the experiment is to study the intracel-
Surface Staining lular properties of the cells of interest, it is typically advanta-
geous to gently fix the cells prior to antibody staining.
Commonly used fixation methods utilize PFA at 1.5–4% for
10–15 min at room temperature. For analysis of fresh primary
human cell suspensions (typically, human peripheral blood or
bone marrow aspirate), we will typically utilize Smart Tube
proteomic stabilizer (STPS) solution in accordance with the
manufacturer’s instructions. A variety of other commercial
fixative solutions are available and would likely be compatible
with this protocol. This solution utilizes a proprietary fixative
cocktail, but is sufficiently gentle to allow subsequent lysis of
red blood cells as well as detection of antigens known to be
disrupted by standard PFA fixation. An alternative fixation
procedure developed by Chow et al. [21] utilizing PFA fixation
followed by Triton X-100 can also work well for ex vivo sample
fixation; however, this method tends to be slightly more dis-
ruptive to fixation-sensitive antigens in our experience. With
either approach, cells can be frozen at
80  C after fixation but
before red cell lysis. If freezing other cell types after fixation
with PFA, cells should be washed twice in CSM and then
resuspended in CSM + 10% DMSO before snap-freezing at

80  C (see Note 13). As previously noted, some surface
antigens can be disrupted by fixation, so the effect of fixation
should be empirically tested for each antigen–antibody combi-
nation of interest.
2. Red cell lysis (optional): If blood is present in the sample, it
should be lysed before proceeding with antibody staining. If
using STPS, we lyse the red cells in accordance with the man-
ufacturer’s instructions, if using the protocol of Chow et al.
[21] lysis is performed with a detergent containing lysis buffer.
After the completion of red cell lysis the cells should be washed
twice in CSM.
42 Gregory K. Behbehani

3. Barcoding (optional): If performing cellular barcoding [15],

this should be performed prior to antibody staining.
4. The desired antibody cocktail should be prepared just prior to
the start of cell processing (though we have prepared cocktails
the day before with success). Each antibody should be diluted
such that it will be at its empirically determined optimal stain-
ing concentration after the cocktail is added to the sample (see
Note 10). We typically add 400 units per mL of sodium
heparin to the antibody cocktail to prevent nonspecific anti-
body binding [7]. Typically, we will aspirate cells after each
wash down to a final volume of 50 μL and make our antibody
cocktail at a 2 concentration, so that we can add 50 μL of
antibody cocktail to the residual 50 μL to achieve a final 1
antibody concentration in 100 μL. (If using a saponin permea-
bilization to stain surface and intracellular antigens simulta-
neously, both should be added to this cocktail; see Note 14.)
5. Cells should be washed in CSM one or two times prior to the
start of the staining reaction. We typically add sodium heparin
to a final concentration of 400 Units per mL into the final wash
of the cells. When using fixed cells, all steps can be performed at
room temperature if desired, but antibody titration and stain-
ing should always be performed under the same conditions.
(If using a saponin-based staining protocol, saponin should be
present in this wash solution.)
6. (Optional) Add an Fc blocking reagent at the manufacturer’s
recommended concentration and incubate for 10 min (see
Note 12).
7. Add the antibody cocktail at the previously determined optimal
concentrations. Stain for 30–60 min at room temperature.
Cells should be gently mixed periodically or placed on a low
speed rotator or shaker to keep the cells evenly suspended.
8. Wash twice in CSM.
9. If performing intracellular staining, continue to Subheading
3.5 on intracellular staining; if only performing a surface stain
(or using saponin to simultaneously stain surface and intracel-
lular antigens), skip to Subheading 3.6 on DNA intercalation.

3.5 Intracellular 1. Surface Antibody Fixation (optional): While not absolutely

Staining required, we typically fix surface antibodies in place after sur-
face staining. After the last wash of the surface stain protocol
(Subheading 3.3, step 5 or Subheading 3.4, step 8), add 1 mL
of PBS plus 100 μL of 16% of PFA (1.5% final PFA concentra-
tion) and incubate for 15 min at room temperature.
2. Pellet cells by centrifugation and aspirate supernate down
to a minimal volume (typically ~50 μL; see Note 15). Vortex
cell pellet to resuspend it to an even single cell suspension
 Immunophenotyping by Mass Cytometry 43

(see Note 16). Rapidly add 1 mL of ice-cold methanol imme-

diately after vortexing the cell pellet then place the cells onto
ice. It may be helpful to vortex the cell suspension at low speed
during methanol addition to ensure that an even suspension
is maintained. Incubate cells on ice for 10 min. If desired, cells
can be stored at
80  C overnight (or for up to a week) in
methanol.
3. Add 1 mL of PBS to the cells in methanol. Then fill the
remainder of the tube with CSM (see Notes 17 and 18).
4. Centrifuge cells for 5 min at 600  g. Aspirate supernate to a
minimal volume. Vortex to resuspend the cell pellet.
5. Wash 2 times in CSM.
6. Add a staining cocktail containing with the desired intracellular
antibodies. We typically stain one to two million cells in 100 μL
staining reaction for 40–50 min at room temperature, but
other staining volumes, incubation times, or incubation tem-
peratures can be used provided the antibodies are titrated for
optimal staining under the same conditions.
7. After the completion of staining, wash cells at least twice
with CSM.

3.6 DNA 1. After the final wash following the surface or intracellular stain-
Intercalation and Data ing reaction, resuspend the cell pellet by vortexing thoroughly.
Acquisition 2. Poststaining fixation (optional): If the cells of interest are par-
ticularly fragile, or if clogging of the mass cytometer has been
experienced when running similar cell samples, it may be help-
ful to perform an additional fixation step, by resuspending the
cell pellet in PBS plus 1.5% PFA and incubating for 10–15 min
at room temperature. This additional fixation will reduce lysis
of particularly fragile cells during sample acquisition, which can
lead to clogging of the mass cytometer’s fluidics. Ensure that
that the PFA is fresh. After this fixation, pellet the cells by
centrifugation and resuspend the cell pellet by vortexing
thoroughly.
3. Add at least 100–200 μL of intercalator solution for each
million cells in the cell pellet. This can also be titrated for the
specific cell types being stained. Mix or vortex cells gently to
resuspend them evenly in the intercalator solution. Place the
cells at 4  C for at least 20 min. The cells will be stable in the
intercalator solution at 4  C for at least a week.
4. After intercalation, cells should be washed twice in either ultra-
pure (heavy metal free) water or in a cell acquisition solution
(see Note 19). Cells should then be resuspended at a density of
approximately one million cells per mL in either water or cell
acquisition solution (if used). Four element EQ beads should
44 Gregory K. Behbehani

be added to the solution at the manufacturer’s recommended

concentration (which should result in analysis of
4–10 beads per second).
5. Begin analysis of cells on CyTOF mass cytometer. Exact acqui-
sition settings are dependent on which CyTOF version is being
used and the cell types being studied. Startup procedures and
cell injection protocol will depend on the exact CyTOF instru-
ment version being used; follow the manufacturer’s recom-
mended procedures.
6. Once acquisition has begun, a preview of the data should be
performed with the cytometer in TOF mode to ensure that
there is not significant metal contamination (see Note 20).
7. Once acquisition is started, the frequency of cell events per
second should be verified. An ideal rate is typically 200–400
cells per second (though the Helios machines may be able to
run slightly faster). Rates higher than this will lead to larger
numbers of cell doublet events and commonly leads to cell
clogs in the cytometer fluidics. If the sample acquisition rate
is too high, samples should be diluted with a mixture of water
and EQ beads (or cell acquisition solution mixed with EQ
beads) until the cell event rate reaches the desired level.
8. The number of that needs to be collected will be experiment-
specific; however, it is advisable to predetermine the expected
frequency of the most rare cell of interest and the minimum
number of such cell events that will be needed for the experi-
mental objectives, then back calculate the number of total
events that will be required (see Note 21).
9. Water or cell acquisition solution should be run for several
minutes between each sample. The data preview function
should be used to ensure that no further cell events are visible
when running water before starting the next cell sample.
Should cell events persist for more than several minutes, the
cytometer’s fluidics should be cleaned with either wash solu-
tion (Fluidigm) or with 3% Nitric acid (trace metal free), fol-
lowed by several minutes of water (see Note 22).

3.7 Initial Data There are now a large number of data analysis approaches for high
Processing dimensional mass cytometry data ranging from large scale analysis
of hundreds of standard biaxial plots to complex high-dimensional
analysis strategies. A thorough discussion of the pros and cons of
each approach is beyond the scope of this chapter, but some guid-
ance can be found in several recent publications [1, 3, 6]. Regardless
of the final analysis approach, several steps should always be per-
formed to ensure the data quality before more complex analysis is
undertaken. It is always ideal to still view the data in the relevant key
biaxial plots that have been traditionally used to define relevant cell
 Immunophenotyping by Mass Cytometry 45

populations (either before or after more complex analyses), to

ensure that the cells of interest would have been similarly gated by
the majority of other researchers.
1. Normalization: Upon completion of cell acquisition, the
resulting data should be normalized, using either the mass
cytometers built-in software or the Matlab application devel-
oped by Finck et al. [22] This normalization step is critical for
experiments in which total cell acquisition requires more than
an hour, as a significant loss of measurement sensitivity can
occur in as little as 1–2 h of sample analysis. Without use of
EQ beads and postanalysis data normalization, samples run
later in the experiment acquisition will appear to have lower
signal across all channels as compared to samples run earlier
during data acquisition. The normalization process will effec-
tively correct this effect in final data files. Note that samples that
have not been adequately fixed will frequently lose antibody
signal during prolonged incubations in water (see Note 23).
2. Debarcoding (optional): If barcoding was performed, debar-
coding should be performed after data normalization and
before any subsequent data analysis.
3. Singlet gating: Once data has been uploaded into the flow
cytometry analysis software of choice, a singlet gate should be
applied to the remove cell doublets and debris. This gate is
created in a biaxial plot of event length (i.e., the number of
“pushes” of the TOF analysis that ions from each cell event are
spread across) vs. DNA intercalator signal (measured on the
191 or 193 channel). In this plot, doublet events will appear as
a tail of cells with high DNA signal and a sharply increased
event length. The gate on singlets should exclude this tail of
event length high cell events as well as any cells with low
amounts of DNA intercalator signal, that are likely to represent
debris. It is best to determine the exact borders of these gates
empirically whenever possible. This can be done by identifying
events that are clearly doublets (e.g., cell events positive for
CD3 and CD15), then plot these doublet events on the event
length vs. DNA biaxial plot. The same can be done for events
that are clearly debris. When determined empirically and drawn
properly, this singlet gate will remove 50–60% of cell doublets
without removing a significant amount of real cell events. To
more completely exclude doublets, cells need to be run very
slowly (75–150 cells per second) or barcoded with a nonde-
generate code. Under these conditions, more than 95% of
doublet events can be excluded.
4. Viability gating (optional): After normalization, debarcoding,
and singlet gating, a viability gate should then be applied to
remove dead cells from experimental analysis (unless
46 Gregory K. Behbehani

measurement of cell viability is an objective of the experiment).

This gate is created on a plot of DNA vs. platinum (195, 195,
or 198 Pt). If a significant fraction of dead cells was present at
the time of viability staining, those dead cells should form a
population with an approximate 1–2 log increase in Pt signal,
compared to the background staining in the remaining live
cells.
5. After these steps, the data are ready for more complex analyses
as appropriate for the given experimental objectives.

4 Notes

1. The several Fluidigm-manufactured, preconjugated antibodies

utilize a different polymer for metal chelation and attachment
to antibodies. This polymer enables conjugation of more metal
atoms to each antibody but also has slightly higher nonspecific
binding and can exhibit very high background on cells that
have been permeabilized with certain saponin preparations.
(This is lot specific, and saponin-containing buffers purchased
directly from Fluidigm generally do not exhibit this issue.) As
of December 2018, the specific antibodies labeled with this
polymer are anti-human CD19/b4 (clone: hib19) on 142nd,
165ho, and 169tm; anti-human CD25 (clone: 2a3) on 149sm
and 169tm; anti-human CD45ra (clone: hi100) on 143nd,
153eu, 155gd, 169tm, and 170er; anti-human CD127
(clone: a019d5) on 143nd, 149sm, 165ho, 168er, and
176yb; anti-human CD45 (clone: hi30) on 89y; anti-human
CD196/ccr6 (clone: g034e3) on 141pr and 176yb; anti-
human CD196/ccr6 (clone: 11a9) on 141pr; anti-human
CD49d/α4 integrin (clone: 9f10) on 141pr and 174yb; anti-
human tgfbeta/lap (clone: tw4–6h10) on 163dy; anti-human
CD41/gpiib (clone: hip8) on 89y; anti-human CD16 (clone:
3g8) on 209bi; anti-human CD47 (clone: cc2c6) on 209bi;
anti-human CD61 (clone: vi-pl2) on 209bi; anti-human
CD11b/mac-1 (clone: icrf44) on 209bi; and anti-mouse
CD45 (clone: 30-F1130-F11) on 89Y.
2. Reagents can be tested using solution mode on the mass cyt-
ometer (recording all metal channels) or by collecting data in
the preview mode using the TOF setting that shows the entire
mass range. Do not inject any concentrated solution that may
contain significant amounts of heavy metal directly into the
mass cytometer without first testing a 1/10th or 1/100th
dilution of the solution. Note that during regular data acquisi-
tion, only the metal masses selected in the reagent panel for
that experiment will be shown, as a result, if there is contami-
nation from a metal not being used for an antibody reagent
 Immunophenotyping by Mass Cytometry 47

(even if at very high levels) it will not be visible unless TOF

mode is used.
3. After each antibody staining step, sufficient washes should be
performed such that residual unbound antibody is diluted by at
least 1000- to 5000-fold. Note that washing away unbound
antibody is more important in mass cytometry than in flow
cytometry as there is no sheath fluid for excess reagents to
dissipate into and excess unbound reagents will increase wear
on the mass cytometer’s detector.
4. While almost all mass channels have less than 4% spillover into
any channel, the mass channels with the highest spillovers
(>1%) are In113; Nd(143 and 145); Sm(147 and 149); Gd
(156, 157, and 158); Dy(161, 162, 163, and 164); Er
(166, 167, and 168); and Yb(171, 172, and 173).
5. The mass cytometer’s plasma temperature is tuned to ensure
oxide formation is always less than 2.5%; however, individual
metals have different propensities to form oxides. The metals
with the highest levels of oxide formation (1–2.5%) are La, Ce,
Pr, Nd, and Gd.
6. Current mass cytometers are significantly less sensitive to
metals with masses below 139 Daltons or above 176 Daltons
(approximately three- to tenfold less sensitive than the other
metal channels). Across the remainder of the mass range, sensi-
tivity generally only varies by threefold with the most sensitive
channels in the range of 159Tb to 169Tm).
7. In this method, a “wash” refers to addition of CSM or PBS to
resuspend the cell pellet in 50–100 volumes (~5 mL for stan-
dard FACS tubes), followed by a 5 min centrifugation, then
aspiration of the supernate down to a minimal volume (typi-
cally about 50 μL for a standard FACS tube). The cell pellet
should then be resuspended to an even single cell suspension in
the residual volume (by tapping, or gentle vortexing). All
centrifugation steps prior to methanol permeabilization can
be performed at 300–600  g (or whatever centrifugal force
is known to be optimal for the cells of interest) for 5 min. After
methanol permeabilization, centrifugation should be per-
formed at a minimum of 600  g. We recommend use of a
refrigerated swinging bucket centrifuge (e.g., Sorvall Legend
XTR [Thermo/Fisher], or similar) set to 4  C or 20 C. If the
volume of the tube being used is not large enough to allow for
50–100 volumes of CSM to be added, perform additional
washes until the total dilution is greater that 1000-fold. Note
that washing is much more important in mass cytometry than
fluorescent flow cytometry, since residual metal reagents in
solution can create significant background, measurement
error, and may increase wear on the mass cytometer’s detector.
48 Gregory K. Behbehani

8. We have successfully optimized this cisplatin staining protocol

for staining in cells in protein-containing media; however, this
requires much larger concentrations of cisplatin and addition of
concentrated protein solutions to quench the reaction. The
protocol can also be optimized for use with different incuba-
tion periods or temperatures, provided that these conditions
will be very consistent during the final experiment.
9. The exact concentration that provides optimal live–dead dis-
crimination will depend on the exact cell numbers, the amount
of residual protein that is not washed away in the previous
washes, the fraction of dead cells, and the length of incubation.
This should be empirically titrated for the planned experimen-
tal protocol and cells of interest.
10. Titrations should be performed for all antibodies (with the
possible exception of preoptimized antibody cocktails devel-
oped for a specific cell type and staining protocol). It is impor-
tant that titrations be performed under the conditions that will
be used to stain the cells, particularly similar staining volume,
cell number, staining time, and staining temperature. Once the
optimal conditions have been established, if staining a larger
number of cells we will increase the total staining volume to
achieve the same ratio of cells per volume of staining reaction.
We typically titrate our antibodies for staining one to two
million cells in a 100 μL staining reaction, and scale up or
down the staining volume (without changing antibody con-
centration) to maintain a constant ratio of cells to volume of
staining cocktail (e.g., 100 μL per two million cells).
11. Live cells are typically stained at 4  C to prevent antigen inter-
nalization; however, it is not clear if antigen internalization
would actually alter mass cytometry staining (since it would
likely take many minutes for a cell to break down and excrete
the chelated heavy metal).
12. We typically use a commercial Fc block from Biolegend, but a
variety of other Fc blocking reagents are available. Care should
be taken to use a blocking agent that does not disrupt mea-
surement of Fc receptors (e.g., CD64, CD32, CD16) if these
antigens will be measured. We have also successfully used total
mouse IgG for blocking. For samples being stained in whole
blood or whole serum, blocking may not be necessary due to
the serum IgG already present.
13. At the desired time of sample analysis, the fixed cells should be
thawed at 4  C; this can be done rapidly in a circulating 4  C
water bath or by placing cells on ice.
14. If using a saponin-based permeabilization to allow for simulta-
neous staining of surface and intracellular antigens, cells should
be fixed (using PFA, STPS, or a commercial fixative
 Immunophenotyping by Mass Cytometry 49

preparation) and the protocol in Subheading 3.4 should be

used. It is important to use the appropriate saponin-containing
buffer for the washes before staining and include saponin at the
appropriate concentration (typically 0.2%) in the staining reac-
tion. After staining, skip to Subheading 3.6; the other steps of
the protocol are typically unchanged. See also Note 1 regarding
interactions between certain Fluidigm reagents and some lots
of saponin.
15. When aspirating the CSM (or PBS + PFA) from the cells,
remove as much as possible (without aspirating the cell pellet),
preferably to a pellet of 50 μL or less. (This will enable a final
methanol concentration of 90–95% to be achieved.)
16. It is absolutely critical that cells be in an even single-cell sus-
pension prior to the addition of methanol. If cells are not
evenly resuspended, they will reliably form large clumps and
aggregates.
17. The BSA in the CSM will precipitate if added to a solution with
a high concentration of methanol, so it is important to reduce
the methanol concentration with PBS before adding the CSM
to the cell suspension.
18. If the volume of methanol added in step 1 was more than 20%
the volume of the tube being used, then additional CSM
washes should be performed so that the final methanol con-
centration of the residual buffer and cell pellet (just before
adding the antibody staining cocktail) is less than 0.5%. If
using 1.1 mL polypropylene “cluster” microcentrifuge tubes,
the cells will first have to be centrifuged in methanol, the
methanol aspirated, then add 0.5 mL of PBS followed by
0.5 mL of CSM and centrifuge again. Then perform two
additional CSM washes to remove residual methanol.
19. The purpose of using a running buffer rather than pure water
for analysis of cells is to reduce the chance of cell breakdown or
antibody disassociation while samples are being run or waiting
to be run (after having been washed into pure water). This
comes at the cost of increased mass cytometer sensitivity deg-
radation due to solute buildup on the mass cytometer’s cones.
Typically, this step is unnecessary if cells are well fixed; however,
we and others have observed this behavior in certain cell types
(or with certain antibody–antigen pairs) in situations when
fixation would have been expected to have been adequate. If
a running buffer is used, inorganic metal salts (e.g., sodium
chloride) should be avoided as inorganic (mineral) salts seem to
cause the mass cytometer’s signal to degrade much faster,
requiring more frequent machine cleaning. Fluidigm sells a
ready-made cell acquisition solution for this purpose.
50 Gregory K. Behbehani

20. Note that in regular data collection mode, only signal within
the time-of-flight windows of the masses set in the panel will be
recorded or displayed on the preview window of the computer.
As a result, if the sample is contaminated with a metal not being
measured in the experiment’s antibody panel, this contamina-
tion (even if at relatively high levels) can be invisible to the user.
A quick preview in TOF mode will show all of the ion species
striking the detector, and if the preview shows cell events
without significant amounts of metal contamination, the
chance of problematic contamination is quite low.
21. When calculating the number of required cell events, it is
important to remember that up to 10% of events will be from
EQ beads and will not be present in the final data file. Addi-
tionally, if significant amounts of debris are present in the
sample, these debris particles will often be counted as cell
events. Thus, if debris is observable in preview screen of the
cytometer during sample acquisition, additional events will
need to be collected in proportion with the approximate fre-
quency of debris relative to true cell events. It is thus not
uncommon to need to collect 10–50% more events than
desired final number of cell events.
22. If running wash buffer (which contains acid) or nitric acid
between samples, it is important to ensure that all acid has
been washed away before starting the next sample. Persistent
acid will strip metal from the antibodies attached to the cells in
subsequent samples until the acid is neutralized.
23. If samples have not been adequately fixed before being placed
into pure water the antibodies can disassociate from their anti-
gens and/or cells may lyse during prolonged cell acquisition.
This effect is directly dependent on the time cells are in water
and cannot be normalized by the EQ bead-based normaliza-
tion algorithms. This problem can be minimized by ensuring
that the PFA used in initial sample fixation and in the inter-
calator solution is sufficiently fresh, by minimizing the time
cells are kept in water prior to being loaded in the cytometer,
and by keeping cells cold between water washes and analyses.
Additionally, use of the optional poststaining fixation steps
(Subheading 3.5, step 1; Subheading 3.6, step 2), or use of a
cell acquisition solution can minimize this problem.

References

1. Bendall SC, Nolan GP, Roederer M, Chatto- applications. Cancer Immunol Immunother
padhyay PK (2012) A deep profiler’s guide to 62(5):955–965
cytometry. Trends Immunol 33(7):323–332 3. Spitzer MH, Nolan GP (2016) Mass cytome-
2. Tanner SD, Baranov VI, Ornatsky OI, Bandura try: single cells, many features. Cell 165
DR, George TC (2013) An introduction to (4):780–791
mass cytometry: fundamentals and
 Immunophenotyping by Mass Cytometry 51

4. Bjornson ZB, Nolan GP, Fantl WJ (2013) 14. Fienberg HG, Simonds EF, Fantl WJ, Nolan
Single-cell mass cytometry for analysis of GP, Bodenmiller B (2012) A platinum-based
immune system functional states. Curr Opin covalent viability reagent for single-cell mass
Immunol 25(4):484–494 cytometry. Cytometry A 81(6):467–475
5. Di Palma S, Bodenmiller B (2015) Unraveling 15. Zunder ER, Finck R, Behbehani GK et al
cell populations in tumors by single-cell mass (2015) Palladium-based mass tag cell barcod-
cytometry. Curr Opin Biotechnol 31:122–129 ing with a doublet-filtering scheme and single-
6. Behbehani GK (2017) Applications of mass cell deconvolution algorithm. Nat Protocols
cytometry in clinical medicine: the promise 10(2):316–333
and perils of clinical CyTOF. Clin Lab Med 16. Ornatsky OI, Lou X, Nitz M et al (2008) Study
37(4):945–964 of cell antigens and intracellular DNA by iden-
7. Rahman AH, Tordesillas L, Berin MC (2016) tification of element-containing labels and
Heparin reduces nonspecific eosinophil stain- metallointercalators using inductively coupled
ing artifacts in mass cytometry experiments. plasma mass spectrometry. Anal Chem 80
Cytometry A 89(6):601–607 (7):2539–2547
8. Frei AP, Bava F-A, Zunder ER et al (2016) 17. Bendall SC, Simonds EF, Qiu P et al (2011)
Highly multiplexed simultaneous detection of Single-cell mass cytometry of differential
RNAs and proteins in single cells. Nat Methods immune and drug responses across a human
13:269 hematopoietic continuum. Science 332
9. Behbehani GK, Bendall SC, Clutter MR, Fantl (6030):687–696
WJ, Nolan GP (2012) Single-cell mass cytome- 18. Behbehani GK (2018) Cell cycle analysis by
try adapted to measurements of the cell cycle. mass cytometry. In: Lacorazza HD
Cytometry A 81(7):552–566 (ed) Cellular quiescence: methods and proto-
10. Edgar LJ, Vellanki RN, Halupa A, Hedley D, cols. Springer New York, New York, NY, pp
Wouters BG, Nitz M (2014) Identification of 105–124
hypoxic cells using an organotellurium tag 19. Behbehani GK, Thom C, Zunder ER et al
compatible with mass cytometry. Angew (2014) Transient partial permeabilization with
Chem Int Ed Engl 53(43):11473–11477 saponin enables cellular barcoding prior to sur-
11. Han G, Spitzer MH, Bendall SC, Fantl WJ, face marker staining. Cytometry A 85
Nolan GP (2018) Metal-isotope-tagged (12):1011–1019
monoclonal antibodies for high-dimensional 20. Hartmann FJ, Simonds EF, Bendall SC (2018)
mass cytometry. Nat Protoc 13 A universal live cell barcoding-platform for
(10):2121–2148 multiplexed human single cell analysis. Sci
12. Chang Q, Ornatsky OI, Koch CJ et al (2015) Rep 8(1):10770
Single-cell measurement of the uptake, intratu- 21. Chow S, Hedley D, Grom P, Magari R, Jacob-
moral distribution and cell cycle effects of cis- berger JW, Shankey TV (2005) Whole blood
platin using mass cytometry. Int J Cancer 136 fixation and permeabilization protocol with red
(5):1202–1209 blood cell lysis for flow cytometry of intracellu-
13. Chevrier S, Crowell HL, Zanotelli VRT, lar phosphorylated epitopes in leukocyte sub-
Engler S, Robinson MD, Bodenmiller B populations. Cytometry A 67(1):4–17
(2018) Compensation of signal spillover in sus- 22. Finck R, Simonds EF, Jager A et al (2013)
pension and imaging mass cytometry. Cell Syst Normalization of mass cytometry data with
6(5):612–620 e615 bead standards. Cytometry A 83(5):483–494
 Chapter 3

Fluorescent Cell Barcoding for Immunophenotyping

Valentina Giudice, Giovanna Fantoni, and Angélique Biancotto

Abstract
Immunophenotyping using flow cytometry highly benefits from multiplexing samples for generation of
more robust data, because of reduction of antibody consumption, batch effect and technical variations. One
way to multiplex is via fluorescent cell barcoding (FCB) prior to staining procedure.
FCB is a high-throughput multiplexed assay using various concentrations of different fluorescent dyes.
Individual samples are uniquely labeled, then mixed together, stained and analyzed as a single sample,
decreasing technical variations and increasing throughput and speed of acquisition. In addition, FCB
simplifies implementation of normalization using a bridge control sample.
In this chapter, we illustrate the protocol for FCB and recommendations for choosing barcoding dyes and
concentrations among other technical considerations.

Key words Fluorescent cell barcoding, Immunophenotyping, Multiplexing

1 Introduction

Standardization of flow cytometric analysis process is critical in

order to have reproducible high-quality data that are comparable
at intralaboratory and interlaboratory levels over time [1–3]. Fluo-
rescent cell barcoding (FCB) enables high-throughput multipa-
rameter flow cytometry by mixing multiple samples in a single
tube for further staining and analysis [4]. Using covalently bound
fluorescent dyes at various concentrations for labeling, samples are
“barcoded” and distinguished based on fluorescence emission
wavelengths and intensities. Multiplexed barcoded samples can be
mixed together and subsequently stained with antibodies in a single
tube. Thus, FCB minimizes staining variability, antibody consump-
tion and decreases total sample volume needed. However, to be
implemented FCB requires optimization [5].
In this chapter we describe FCB procedure for immune cell
phenotyping that can be performed on whole fresh blood, periph-
eral blood mononuclear cells (PBMCs), or cryopreserved PBMCs.

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_3, © Springer Science+Business Media, LLC, part of Springer Nature 2019

53
54 Valentina Giudice et al.

FCB protocol for a 3x3 matrix – 9 samples

DyLight 350 vs DyLight 800
1. FCB dye preparation 2. Sample preparation
0.6-1 x 106 cells/well of 3 mL of 1x
Fresh PBMCs or Lyse/Fix Buffer
Frozen PBMCs or
Fresh whole blood after
RBC lysing
5μL of each dye/well
15 min incubation at RT
250
DyLight 350

10μL of DyLight 350 only

13

10μL of DyLight 800 only Centrifuge at 400g for 5 min

3 mL of cold 1x
Unstained sample
0

Perm Buffer II
DyLight 800

0 13 250

20 min incubation on ice

Keep on ice
until FCB
Centrifuge at 400g for 5 min
3. FCB
Add 30μL of sample for each well Resuspend cells in
cold 1x Perm Buffer
II to have 0.6-1 x 106
DyLight 350

cells/30μL

DyLight 800
15-25 min incubation on ice

Add 3 mL of 1x Barcoding Wash Buffer and centrifuge at 400g for 5 min

Resuspend in 300 μL of Wash Buffer

4. Antibody staining Keep single-color controls

and unstained sample on
Y YYYY ice until acquisition
Y YY
Y
20-30 min incubation at RT
Add 3 mL of 1x Barcoding Wash Buffer and centrifuge at 400g for 5 min
Resuspend in 300 μL of Wash Buffer for cytometer acquisition

Fig. 1 Illustration of FCB protocol for a 3
3 matrix using DyLight 350 and DyLight 800. In (1) FCB dye
preparation is shown how to perform dye dilution (Subheading 3.1); then, 5 μL of each dye at various
 Fluorescent Cell Barcoding 55

In this protocol, FCB is obtained using two amine-reactive

fluorescent dyes with active esters (succinimidyl ester, NHS)
(DyLight 350 and DyLight 800) because of their efficiency in
tethering molecules [6].
In this chapter, we describe barcoding of nine or 12 samples
combined with a five-color staining, by listing reagents needed and
a detailed protocol procedure (Fig. 1). FCB is performed using
96 round-bottom well plate; however, 1.5 mL Eppendorf tubes can
also be used. In addition, gating strategy to analyze samples is
described for manual and semiautomated analyses by using avail-
able software. Finally, Notes describe technical recommendations
based on immunophenotyping of interest (phospho-flow, fluoro-
chromes and instrument used).

2 Materials

2.1 FCB Buffers Barcoding dyes from Thermo Fisher Scientific (Waltham, MA,
and Dyes USA): DyLight 350—NHS; DyLight 800—NHS; and Pacific
Orange—NHS (see Note 1).
Barcoding buffers from BD Biosciences (San Jose, CA, USA):
BD Perm Buffer II; 5
BD Phosflow Lyse/Fix Buffer; 4
BD
Phosflow Barcoding Wash Buffer.
Other reagents: Ficoll-Paque gradient centrifugation (MP Bio-
medicals, LLC, Santa Ana, CA, USA), according to the manufac-
turer’s instructions for PBMC isolation; Ack Lysing buffer (Quality
Biological, Gaithersburg, MD, USA) for red blood cell lysis; Phos-
phate buffer saline (PBS) and dimethylsulfoxide (DMSO) Sigma-
Aldrich (St. Louis, MO, USA).

Fig. 1 (continued) concentrations are pipetted in appropriate wells (according to colors). In two separate
wells, 10 μL of DyLight 350 or DyLight 800 are added and will be used as single-color controls. In another well
no dye is added and 40 μL of unstained sample will be pipetted, will be considered unstained control for
cytometer setting. Once prepared, dyes can be kept on ice until use. In (2) Sample preparation,
0.6–1
106 cells/well per sample of fresh PBMCs or frozen PBMCs or fresh whole blood after RBC lysing
are incubated for 15 min at RT with 3 mL of BD Lyse/Fix Buffer. After centrifugation, supernatants are removed
and cell pellets loosen; then, 3 mL of cold 1
Perm Buffer II are added and samples incubated for 20 min on
ice. After incubation, cells are washed with 2 mL of PBS and centrifuge at 400
g for 5 min. Cells are
subsequently resuspended in cold 1
Perm Buffer II to have 0.6–1
106 cells/30 μL. In (3) FCB, samples are
added in appropriate wells prepared with FCB dyes and incubated on ice for 15–25 min. Samples from the
3
3 matrix are combined together, while single-color controls and unstained sample are transferred to
separate tube. Then cells are washed with 3 mL of 1
Barcoding Wash Buffer and centrifuge at 400
g for
5 min. In (4) Antibody staining, after centrifugation, cells are resuspended in 300 μL of Wash Buffer and
antibodies are added according to the manufacturer’s instructions. After 20–30 min incubation at RT, samples
are washed with 3 mL of Barcoding Wash Buffer, centrifuged, and resuspended in 300 μL of the same buffer
for cytometer acquisition
56 Valentina Giudice et al.

2.2 Buffer and Dye All buffers and dyes are prepared using PBS and stored as indicated.
Preparation
1. Prepare 1
BD Perm Buffer II by diluting 1:2. Add 10 mL of
BD Perm Buffer II at room temperature (RT) and 10 mL of
cold PBS to a 50 mL Falcon Conical tube (Fisher Scientific,
Loughborough, UK), mix by inversion and store at +4  C.
2. Prepare 1
BD Phosflow Lyse/Fix Buffer from the 5
stock
solution. Add 10 mL of 5
BD Phosflow Lyse/Fix Buffer and
40 mL of PBS at RT to a 50 mL Falcon Conical tube, mix by
inversion and store at RT.
3. Prepare 1
Barcoding Wash Buffer by diluting 1:4. Add 5 mL
of cold 4
Barcoding Wash Buffer and 35 mL of PBS at RT to a
50 mL Falcon Conical tube, gently mix by inversion and store
at RT (use the solution within 1–2 days).
4. Dyes used in this protocol are DyLight 350 and DyLight 800.
Upon arrival dry powder dyes are reconstituted at 500 μg/mL
concentration in DMSO in 15 mL Falcon Conical tubes
(Fisher Scientific, Loughborough, UK). For 1 mg of dry pow-
der, add 1 mL of DMSO to the tube containing the powder
and transfer to a 15 mL Falcon Conical tube; then rinse with
1 mL of DMSO the tube and transfer to the 15 mL Falcon
tube. Mix well by pipetting and aliquot in 1.5 mL Eppendorf
tubes. Store at 80  C protected from light (see Note 2).

3 Methods

3.1 Sample FCB for lymphocyte phenotyping can be performed on fresh sam-
Preparation ples, as whole blood or peripheral blood mononuclear cells
(PBMCs) or cryopreserved PBMCs. If frozen PBMCs are used,
thaw cells in a water bath at +37  C for 1–2 min. Add 2 mL of PBS
and centrifuge at 400
g for 5 min. Discard supernatant and
vortex to loosen cell pellets. If fresh PBMCs are used for barcoding,
start from step 5.
If whole fresh blood is employed in the experiment:
1. Collect at least 0.2 mL of whole blood in the presence of
anticoagulant (EDTA or sodium heparin).
2. Lyse with 1:10 Ack Lysing buffer. For 0.2 mL of fresh blood,
add 2 mL of Ack Lysing buffer and incubate for 10 min on ice
(see Note 3).
3. Centrifuge at 400
g for 5 min and discard supernatant.
4. Repeat steps 2 and 3.
5. Add 3 mL of 1
Lyse/Fix Buffer, gently mix by pipetting and
incubate at RT for 15 min.
6. Centrifuge at 400
g for 5 min. Discard supernatant and
vortex to loosen the cell pellets.
 Fluorescent Cell Barcoding 57

7. Add 3 mL of cold 1
Perm Buffer II. Incubate on ice for

20 min (see Note 4).
8. Add 2 mL of PBS and centrifuge at 400
g for 5 min. Discard
supernatant and vortex to loosen cell pellets.
9. Resuspend cells in cold 1
Perm Buffer II and keep cells on ice
until use (see Subheading 3.4).
Calculate the volume of cold 1
Perm Buffer II for resus-
pension in order to have 0.6–1
106 cells/30 μL. In addition,
in order to have controls for cytometer set up, for one sample
calculate: 30 μL for unstained sample; and 30 μL for each
single-dye control (e.g., DyLight 800 and DyLight 350).

3.2 Nine-Sample Prepare dye working concentrations with DMSO from stock con-
Matrix Preparation centration immediately before starting the experiment, as DMSO
(3
3) tends to precipitate.
1. Thaw dye stock solution (500 μg/mL) at RT protected from
light. After thawing, vortex vigorously.
2. Prepare dye working concentrations with DMSO in round
bottom 96-well plate and label wells as “DyL350-0”,
“DyL350-13”, “DyL350-250” and “DyL800-0”, “DyL800-
13”, “DyL800-250” (see Note 5).
For each dye, prepare the following dilutions:
250-dilution ¼ 25 μL of 500 μg/mL dilution + 25 μL of
DMSO
50-dilution ¼ 10 μL of 250-dilution + 40 μL of DMSO
13-dilution ¼ 13 μL of 50-dilution + 37 μL of DMSO
0-dilution ¼ 30 μL of DMSO
The 50-dilution is used only for preparing the 13-dilution
and not for FCB matrix preparation (see Note 6).
3. Add 5 μL of diluted dye to appropriate well in a round bottom
96-well plate as shown in Fig. 1. For a final volume of 40 μL/
well (10 μL of dyes +30 μL of sample), final concentrations will
be 0, 3.25, and 62.5 μg/mL.
4. Seal or cover the plate and keep at RT protected from light.
5. Proceed to Subheading 3.4.

3.3 Twelve-Sample Prepare dye working concentrations from stock concentration right
Matrix Preparation before starting the experiment, as DMSO tends to precipitate.
(4
3)
1. Thaw dye stock solution (500 μg/mL) at RT protected from
light. After thawing, vortex vigorously. In a round bottom
96-well plates, prepare dye working concentrations with
DMSO as follows (Fig. 2).
2. For DyLight 350, name tubes/wells as follows: “DyL350-0”,
“DyL350-30”, “DyL350-150” and “DyL350-500”. Prepare
dilutions as follows:
58 Valentina Giudice et al.

4x3 matrix – 12 samples

DyLight 350 vs Pacific Orange Pacific Orange
25 μ
μL 10 μ
μL

Pacific Orange

500
50
25 μL 40 μL 40 μL
DMSO DMSO DMSO

0
500 250 50 0
DyLight 350
DyLight 350
0 30 150 500
15 μ
μL 10 μL
μ

35 μL 40 μL 40 μL
DMSO DMSO DMSO
500 150 30 0

Fig. 2 Illustration of 4
3 matrix preparation. Using four concentrations of

DyLight 350 and three of Pacific Orange, a 4
3 matrix is prepared for
12-sample FCB. Detailed information on how to prepare dye dilutions is
described in Subheading 3.3

500-dilution ¼ 50 μL of 500 μg/mL dilution

150-dilution ¼ 15 μL of 500 μg/mL dilution +35 μL of
DMSO
30-dilution ¼ 10 μL of 150-dilution +40 μL of DMSO
0-dilution ¼ 40 μL of DMSO
For a volume of 40 μL/well combining dye and samples,
DyLight 350 final concentrations are 0, 3.8, 18.8, and
125 μg/mL.
3. For Pacific Orange, label wells as “PO-0”, “PO-50”, and
“PO-500”.
Prepare dilutions as follows:
500-dilution ¼ 50 μL of 500 μg/mL dilution
250-dilution ¼ 10 μL of 500 μg/mL dilution + 10 μL of
DMSO
50-dilution ¼ 10 μL of 150-dilution + 40 μL of DMSO
0-dilution ¼ 40 μL of DMSO
The 250-dilution is made only for preparing the
50-dilution and not for barcoding. For a volume of 40 μL/
well combining dye and samples in a 12-sample matrix with PO
final concentrations are 0, 6.3, and 125 μg/mL.
4. Add 5 μL of diluted dye to appropriate well in a round bottom
96-well plate.
5. Seal or cover the plate and keep at RT protected from light.
6. Proceed to Subheading 3.4.
 Fluorescent Cell Barcoding 59

3.4 Fluorescent Cell Once matrix and samples are prepared, proceed to FCB as follows.
Barcoding Perform all steps at RT unless otherwise specified.
1. Add 30 μL of each sample into appropriate well, according to
their matrix position. Mix well the solution as dyes tend to
precipitate.
2. Add 30 μL of samples into one well with 10 μL of DyLight
350 and other 30 μL of sample into a different well with 10 μL
of DyLight 800. These samples were prepared in step 9 of
Subheading 3.1 and will be used as single-color controls for
cytometer setting.
3. Seal or cover the plate and incubate for 15–25 min on ice.
Longer incubation time is preferable for whole blood barcod-
ing (see Note 7).
4. Transfer all nine barcoded samples into one FACS tube named
“Combo”. Transfer single-color controls prepared in step 2 in
two separate FACS tubes labeled “DyL350” and “DyL800.”
5. Wash with 3 mL of 1
Barcoding Wash Buffer (see Note 8).
Centrifuge at 400
g for 5 min and discard supernatant.
Vortex to loosen cell pellets. Repeat step 5 (see Note 9).
6. Resuspend “Combo”, “DyL350”, and “DyL800” in 300 μL of
1
Barcoding Wash Buffer. Keep “DyL350” and “DyL800”
on ice until cytometer setting.
7. Add antibodies according to the manufacturer’s instructions to
the appropriate tube and incubate at RT in the dark for
20–30 min.
8. Wash with 3 mL of 1
Barcoding Wash Buffer. Centrifuge at
400
g for 5 min, then discard supernatant and vortex to
loosen cell pellet. Repeat washing step one more time.
9. Resuspend cells in 300–500 μL of 1
Barcoding Wash Buffer
and keep on ice or +4  C until acquisition.
10. Acquisition can be carried out using appropriate lasers
(355, 407, or 630 nm) on cytometers. Preacquisition compen-
sation can be calculated using bead standards for each fluoro-
chrome (anti-mouse Ig, κ/Negative Control Compensation
Particles Set, BD Biosciences) and prepared barcoded cells
with the highest concentration of each dye. For each fluoro-
chrome, cells as well as single-color controls can be used.
A minimum of 50,000 lymphocytes needs to be recorded.

3.5 Control Samples During protocol optimization, a bridge sample within one matrix is
for Protocol used in order to measure intra-assay variability. Indeed, percent of
Optimization positive cells and MFI values should not significantly differ across
all 9/12 barcoded populations within the same matrix.
Control samples are (1) “No fix/perm” control, without fixa-
tion, permeabilization, and barcoding; and (2) “No FCB” control,
60 Valentina Giudice et al.

a fixed and permeabilized sample without barcoding. These two

controls are used to compare percentages of positive cells before
and after barcoding.
After optimization, it is preferable to include the same bridge
control used for protocol optimization across all experiments in
order to measure interassay variability.

3.5.1 “No fix/perm” 1. Resuspend cells from step 4 of Subheading 3.1 in PBS.
Control Preparation 2. Add antibodies according to the manufacturer’s instructions
and incubate at RT for 20–30 min in the dark.
3. Wash with 3 mL of PBS, centrifuge at 400
g for 5 min and
discard supernatant. Vortex to loosen cell pellets.
4. If intracellular staining is required, add 200 μL of CytoFix/
CytoPerm BD buffer and incubate at +4  C for 30 min. Other-
wise skip to step 7.
5. Wash with 2 mL of 1
Perm Wash, centrifuge at 400
g for
5 min, discard supernatant, and vortex to loosen cell pellets.
6. Resuspend cells in 175 μL of 1
Perm Wash and add intracel-
lular markers antibodies according to the manufacturer’s
instructions. Incubate at +4  C for 30–40 min in the dark.
7. Wash with 2 mL of 1
Perm Wash. Centrifuge at 400
g for
5 min, discard supernatant and vortex to loosen cell pellets.
8. Resuspend cells in 300 μL of 1
Perm Wash for acquisition.

3.5.2 “No FCB” Control 1. Prepare cells as in Subheading 3.1 (steps 1–4).
Preparation 2. After centrifugation, add 3 mL of 1
Lyse/Fix Buffer and
incubate at RT for 15 min.
3. Centrifuge at 400
g for 5 min. Discard supernatant and
vortex to loosen the cell pellets.
4. Add 3 mL of cold 1
Perm Buffer II. Incubate on ice for
20 min.
5. Add 2 mL of PBS and centrifuge at 400
g for 5 min. Discard
supernatant and vortex to loosen cell pellets.
6. Resuspend cells in 200 μL of 1
Barcoding Wash Buffer.
7. Add antibodies according to the manufacturer’s instructions to
the appropriate tube and incubate at RT in the dark for
20–30 min.
8. Wash with 3 mL of 1
Barcoding Wash Buffer. Centrifuge at
400
g for 5 min, then discard supernatant and vortex to
loosen cell pellet. Repeat washing step one more time.
9. Resuspend cells in 300–500 μL of 1
Barcoding Wash Buffer
and keep on ice or +4  C until acquisition.
 Fluorescent Cell Barcoding 61

For measurement of inter-assay variability across matrixes and

over time, we suggest including in each matrix one of the donors
used for FCB optimization. This sample will be used as internal
control and bridge through different experiments facilitating lon-
gitudinal studies. This bridge sample should always be barcoded
with the same dye concentrations, thus the same position in the
matrix. By comparing percentages of positive cells and/or MFI
values from bridge internal controls across matrixes, inter-assay
variability can be measured and outliers can identify experiments
that should be repeated.

3.6 Post-Acquisition Post-acquisition analysis can be carried out using FlowJo or RStu-
Analysis dio software. Recommended analysis steps using both software are
described below.

3.6.1 FlowJo Flow cytometry analysis of barcoded samples, or deconvolution,

can be performed in FlowJo software as follows (Fig. 3):
1. Make post-acquisition compensation matrix and apply it to
barcoded samples. Export the matrix as .csv file.
2. Identify lymphocytes and/or monocytes based on forward
scatter area (FSC-A) and side scatter area (SSC-A).
3. Exclude doublets based on FSC-A vs FSC-H.
4. On single cells, identify barcoded populations based on FCB
dyes (e.g., DyLight 800 vs DyLight 350).
5. On each barcoded population, perform your gating strategy
according to antibody staining.

3.6.2 RStudio RStudio also can be used using the following packages (flowCore,
flowClust, flowViz, flowWorkspace, ggcyto, and flowType)
(Fig. 4a). In RStudio, antibodies are identified by cytometer chan-
nel used during acquisition (e.g., on LSR Fortessa cytometer,
DyLight 350 was acquired using the channel BUV396-A and
DyLight 800 with APC-H7-A channel).
1. Remove debris by filtering the data using rectangleGate func-
tion and FSC-A and SSC-A parameters. Events with FSC values
lower than 20 k are removed by using the Subset method.
2. Compensate data using the matrix made in step 1 of FlowJo
analysis and compensate function of flowWorkspace.
3. Transform compensated data using arcsinh function.
4. Filter data for FCB dye channels (e.g., BUV.396.A and APC.
Cy7.A) by rectangleGate function.
5. Identify barcoded populations using flowClust function. Clus-
tering is performed by setting varNames parameter on FCB dye
channels (e.g., BUV.396.A and APC.Cy7.A) and K is set based
on the number of barcoded samples in the matrix (9 or 12).
62 Valentina Giudice et al.

250K
105

200K
4
10
150K
SSC-A

FSC-H
103 100K

0
50K

0
0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K
FSC-A FSC-A

105
250
DyLight 350

DyLight 350

104
13

103
0

102
DyLight 800 0

0 103 104 105

0 13 250
DyLight 800

Fig. 3 Manual deconvolution using FlowJo. In order to gate the nine barcoded populations using FlowJo
software, first lymphocytes are identified based on linear parameters (FSC-A vs SSC-A) and double cells
excluded (FSC-H vs FSC-A). On single cells, barcoded populations are shown using FCB dye parameters
(DyLight 350 vs DyLight 800) and samples are visualized on a dot plot according to dye concentrations (bottom
left panel)

6. After clustering, data can be shown using ggcyto package.

7. Barcoded populations are identified individually and a flowset
created for further analysis.
8. Gating strategy can be performed employing rectangleGate
function using parameters of interest (e.g., CD3 vs SSC-A or
CD4 vs CD8) and barcoded populations included in one flow-
set shown simultaneously using ggcyto.
Fig. 4 Automated deconvolution using RStudio. (a) Using RStudio software, first debris (events <20 k in FSC)
are removed by filtering for linear parameters (FSC-A vs SSC-A) with rectangleGate function. After compen-
sation and arcsinh transformation, data are filtered for FCB dye parameters by rectangleGate and subse-
quently, nine populations are identified using flowClust function. Barcoded populations are identified
individually and a flowset created for further analysis. (b) Gate boundaries for unfiltered and filtered data. A
detailed description is provided in Subheading 3.6.2
64 Valentina Giudice et al.

For standardization of gate boundaries in RStudio, Minimum

(Min), first quartile (1st), Median, third quartile (3rd), and maxi-
mum (Max) values of individual antibody can be extracted from
each barcoded population and fluorochrome. Variations in fluores-
cence intensity among populations are considered, identifying the
lowest and highest values for each parameter. On unfiltered data,
below the lowest first values are removed from the analysis, because
considered as outliers. Gates (Q1–Q4) are then defined as follow-
ing: Q1, Parameter 1 (lowest 1st, highest 3rd) and Parameter
2 (highest 3rd, highest Max); Q2, Parameter 1 (highest 3rd, high-
est Max) and Parameter 2 (highest 3rd, highest Max); Q3, Param-
eter 1 (highest 3rd, highest Max) and Parameter 2 (lowest 1st,
highest 3rd; and Q4, Parameter 1 (lowest 1st, highest 3rd) and
Parameter 2 (lowest 1st, highest 3rd). For filtered data, gates were
defined as follows. Gates (Q1–Q4): Q1, Parameter 1 (lowest min,
highest 3rd) and Parameter 2 (highest Median, highest Max); Q2,
Parameter 1 (highest 3rd, highest Max) and Parameter 2 (highest
Median, highest Max); Q3, Parameter 1 (highest 3rd, highest Max)
and Parameter 2 (lowest Min, highest Median; and Q4, Parameter
1 (lowest Min, highest 3rd) and Parameter 2 (lowest Min, highest
3rd) (Fig. 4b). Based on FCS data, the first quartile can be sub-
stituted with Min and Median with the Mean in order to adjust gate
boundary. However, we suggest keeping the same gate boundary
strategy during analysis of samples within the same experiment.

3.7 Recommended Purity of deconvolution is defined as the distance between MFIs

Assessment FCB and CVs to obtain 70% of the respective barcoded population with
Efficacy and Intra- 95% purity. Others reported that MFIs should be separated by a
assay/Interassay three-fold increase for a good deconvolution (Fig. 5a).
Variability

3.7.1 MFI Fold Increase 1. Make post-acquisition compensation matrix in FlowJo soft-
Calculation ware and apply it to barcoded samples.
2. Identify lymphocytes and/or monocytes based on forward
scatter area (FSC-A) and side scatter area (SSC-A).
3. Exclude doublets based on FSC-A vs FSC-H.
4. On single cells, shown each FCB dye as histogram and identify
each dye concentration used and name populations accordingly
(e.g., “DyL350-0”, “DyL350-13”, “DyL350-250” and
“DyL800-0”, “DyL800-13”, “DyL800-250” when a nine-
sample matrix is made).
5. Calculate MFI and CV for each FCB dye dilution. Export data
as .csv or .xls file.
6. In a spreadsheet, calculate MFI fold change
 as follows: MFI

fold increase ¼ MFIpeak2  CV peak2 = MFIpeak1 þ CV peak1 :
 Fluorescent Cell Barcoding 65

100 100
MFI fold increase
13μg/ml / 0μg/ml
Normalized Barcoded

80 80
[MFIpeak2 – CVpeak2]
Cell Number

60 60
[MFIpeak1 + CVpeak1]
40 40
MFI fold increase
20 20 250μg/ml / 13μg/ml
[MFIpeak2 – CVpeak2]
0 0
0 102 103 104 105 0 102 103 104 105
[MFIpeak1 + CVpeak1]

DyLight 350

B DyLight 350 (μg/ml) 1.56 50 500

DyLight 800 (μg/ml) 0 500 0 500 0 500
HC-1 92.2 95.5 93.7 95.8 97 98.9

Control
Mean SD M+2SD M-2SD
w/o FCB
95.5 2.4 100.3 90.8 95.7

Mean and SD using samples

within the same matrix

Variability Ratio
LOW Mean
M-2D < Control w/o FCB < M+2D 0.8 < < 1.2
Control w/o FCB
HIGH
Control w/o FCB < M-2D
or
Control w/o FCB > M+2D

Fig. 5 FCB efficacy and inter-assay/intra-assay variability. (a) FCB efficacy is calculated using the MFI fold
change. For each dye, MFI values are calculated for each of the three concentrations  and MFI fold 
change
 is determined
 according to the following formula: MFI fold increase ¼ MFI peak 2
 CV peak 2
=
MFIpeak1 þ CVpeak1 : (b) Inter-assay/intra-assay variability is assessed for each fluorochrome and percent
of cell population (Subheading 3.7). For intra-assay variability, using a 3
3 matrix with the same donor in all
nine positions and matched No-FCB controls, range of variability is calculated as < mean in controls –2SD or
> mean in controls 2SD and percent of barcoded populations after deconvolution should be within this
range. For inter-assay variability, the average of all barcoded samples across all experiments is calculated and
compared to the average of all matched control samples used. Values should be within mean of controls
2SD. The ratio of variability is calculated as the ratio between the mean of barcoded samples and the mean
of matched No-FCB controls. The ratio should be within 0.8–1.2
66 Valentina Giudice et al.


For example, MFI fold increase ¼ MFIDyL350‐13 CV DyL350‐
  
13= MFIDyL350‐0 þ CV DyL350‐0 MFI fold increase ¼ MFIDyL

350‐250  CV DyL350‐250 = MFIDyL350‐13 þ CV DyL350‐13 .

3.7.2 Intra-assay/Inter- 1. Calculate percent of positive cells (e.g., CD3+ cells, CD4+
assay Variability of Percent cells) in control samples using FlowJo software and export
of Positive Cells Calculation data as .csv or .xls file.
2. Calculate the range of variability as <mean in controls –2SD or
>mean in controls +2SD (Fig. 5b).
3. Compare percent of barcoded populations after deconvolution
to matched control samples. Values should be within 2SD.
4. For intra-assay variability, for each cell population (e.g., CD3+
cells), calculate the average across all barcoded samples within
one matrix and compare to the average of all matched control
samples used in that matrix. Values should be within mean of
controls 2SD.
5. For inter-assay variability, of each cell population (e.g., CD3+
cells), calculate the average of all barcoded samples with the
same FCB dye dilution (e.g., 0 μg/mL) across all matrix and
compare to the average of all matched control samples used.
Values should be within mean of controls 2SD.

3.7.3 Intra-assay/Inter- 1. Calculate MFI for each antibody used in barcoded samples
assay Variability of MFI using FlowJo software and export data as .csv or .xls file.
Values for Each 2. Calculate the range of variability as <mean in controls –2SD or
Fluorochrome Calculation >mean in controls +2SD.
3. Compare MFI values of barcoded populations after deconvolu-
tion to population barcoded with lowest concentration (e.g.,
0 μg/mL) (see Notes 10 and 11). Values should be within 2SD.
4. For intra-assay variability, for each cell population (e.g., CD3+
cells), calculate the average across all barcoded samples within
one matrix and compare to the average of all matched control
samples used in that matrix. Values should be within mean of
controls 2SD.
5. For inter-assay variability, for each cell population (e.g., CD3+
cells), calculate the average of all barcoded samples with the
same FCB dye dilution (e.g., 0 μg/mL) across all matrix and
compare to the average of all matched control samples used.
Values should be within mean of controls 2SD.

4 Notes

1. We describe combination of FCB dyes using DyLight 350 and

DyLight 800 but the user can choose a different dye (e.g.,
Pacific Orange) depending on cytometer setting and
 Fluorescent Cell Barcoding 67

fluorochromes used. Indeed, if DyLight 350 and DyLight

800 are chosen, cytometer needs to be equipped with ultravio-
let (355 nm) and red (633 nm) lasers. If ultraviolet laser is not
available, we recommend the use of Pacific Orange and
DyLight 800 for barcoding, as they are excited by violet
(407 nm) and red (633 nm) lasers. However, Pacific Orange
limits the use of other fluorochromes excited by violet laser
because of the spillover in other violet channels.
2. Aliquots of 500 μg/mL stock solution can be prepared in order
to avoid repeated thawing cycles. However, do not make small
volume aliquots as solutions could not be as homogenous and
stable as the ones in larger volumes.
3. Gently mix samples with Ack lysing buffer by pipetting and
repeat occasionally during the incubation time.
4. If pSTATs are included in the antibody staining, for permeabi-
lization, use 3 mL of cold Perm Buffer III instead of 1
Perm
Buffer II. Avoid sample incubation below +4  C, because
DMSO can freeze.
5. If 0- and 13-dilutions do not achieve an MFI fold changes 3,
instead of preparing a 13-dilutions, make a 15-dilution by
mixing 15 μL of 50-dilution and 35 μL of DMSO.
6. Higher concentrations of FCB dyes might change MFI values
of same fluorochromes because of the increase in autofluores-
cence of barcoded samples. For this reason, prefer higher dye
concentrations or 12-sample matrix concentrations for surface
marker staining when MFI values are not required (e.g., immu-
nophenotyping or T cell subset characterization).
7. After incubation, it is preferable to combine samples in a tube
already filled with 3 mL of 1
Barcoding Wash Buffer, because
residual unbound dyes are quickly diluted in a larger volume of
wash buffer upon transfer.
8. To get rid of residual unbound dyes that might react with cells
by the time of acquisition repeating the wash step twice, espe-
cially if barcoded samples are not acquired in the same day.
9. Barcoded “Combo” cells can be used for up to 3 different
antibody-staining by transferring 100 μL of resuspended cells
in each of three separate tube. Label tubes accordingly as
“Staining 1,” “Staining 2,” or “Staining 3.” If only one stain-
ing and/or the acquisition of more events are required, cells
can be resuspended in 200 μL of 1
Barcoding Wash Buffer
and antibodies directly added to the sample.
10. Deconvolution can be carried out at any step of gating strategy;
however, percent of positive cells may differ for two reasons:
(a) If deconvolution is performed later in the analysis, not
barcoded cells are included in all gates before identifica-
tion of barcoded populations.
68 Valentina Giudice et al.

(b) Percent of positive cells is calculated using a given popu-

lation as total (parent, 100%). For example, if we want to
know the percent of CD4+/CD8+ cells:
l What if we do deconvolution first? Each barcoded
population is our parent population on which all per-
centages will be calculated.
l What if we do deconvolution later? If we identify first
CD4+ and CD8+ cells, these subsets become our par-
ent populations. Then, if we do deconvolution of CD4
+ cells for example, we will identify nine barcoded
populations but percentages represent the proportion
of each of them within the CD4+ subset.
11. During optimization of protocols for phosphoFlow in combi-
nation with FCB, in addition to percent of positive cells and
MFIs, MFI fold change values for pSTATs (MFI of stimulated
samples/MFI of unstimulated samples) need to be calculated
for each barcoded population and compared to matched
controls.

References
1. Finak G, Langweiler M, Jaimes M et al (2016) throughput drug screening and signaling
Standardizing flow cytometry immunopheno- profiling. Nat Methods 3:361–368
typing analysis from the human immunopheno- 5. Giudice V, Feng X, Kajigaya S, Young NS, Bian-
typing consortium. Sci Rep 6:20686 cotto A (2017) Optimization and standardiza-
2. Maecker HT, JP MC Jr, FOCIS Human Immu- tion of fluorescent cell barcoding for
nophenotyping Consortium et al (2010) A multiplexed flow cytometric phenotyping. Cyto-
model for harmonizing flow cytometry in clini- metry A 91:694–703
cal trials. Nat Immunol 11(11):975–978 6. Lim CY, Owens NA, Wampler RD et al (2014)
3. Maecker HT, McCoy JP, Nussenblatt R (2012) Succinimidyl ester surface chemistry: implica-
Standardizing immunophenotyping for the tions of the competition between aminolysis
human immunology project. Nat Rev Immunol and hydrolysis on covalent protein immobiliza-
12(3):191–200 tion. Langmuir 30(43):12868–12878
4. Krutzik PO, Nolan GP (2006) Fluorescent cell
barcoding in flow cytometry allows high-
 Chapter 4

Immunophenotyping Using Dried and Lyophilized Reagents

Marc Langweiler

Abstract
Antibody reagents that are used for flow cytometry immunophenotyping have traditionally been prepared
by combining individual liquid antibody conjugates into mixtures. These cocktails have limited shelf-life,
and their preparation is time-consuming and prone to laboratory error. Manufacturers of these reagents, in
collaboration with several clinical and research centers, have made advances in constructing dried antibody
cocktails which have addressed many of the problems inherent in preparing the liquid cocktails on the lab
bench. This chapter discusses methods for the use of dried reagents.

Key words Flow cytometry, Immunophenotyping, Harmonization, Dried antibody cocktail,

Lyophilization

1 Introduction

From the outset, one of the fundamental uses for flow cytometry
has been immunophenotyping of cell suspensions using antibody
cocktails. Throughout the evolution of this methodology, the par-
adigm for the procedure has been and continues to consist of
(1) preanalytical steps including processing of a cellular substrate
to be stained and preparation of antibody cocktails, (2) analytical
steps including cell staining and data acquisition using cytometers,
and (3) postanalytical steps including off-line data analyses. There
have been marked advances in each of these steps, including the
number of markers contained in cocktails, the variety of fluoro-
chromes available to be conjugated with these markers, the number
of parameters available on cytometers that are manufactured with
various technologies for creating flow streams, and the progression
of the data analysis strategies.
Throughout this process, there have always been the goals of
maximizing accuracy, precision, and reproducibility, while minimiz-
ing variability and cost. These goals are essential in performing
good science, but they also have a financial impact on the opera-
tions in the research and clinical laboratory realm [1]. One of, if not

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_4, © Springer Science+Business Media, LLC, part of Springer Nature 2019

69
70 Marc Langweiler

the most difficult, stages of the process outlined above is the

construction, preparation, and storage of liquid antibody cocktails.
The challenges in preparing liquid cocktails are varied and include
the need for repetitive pipetting of small volumes of reagent and
degradation of cocktail constituents due to aging and exposure to
light and temperature variation.
Antibody cocktails that have been prepared by drying them
into lyophilized buttons or, more recently and effectively, by drying
directly to the plastic of staining vessels have addressed these issues.
Dried reagents have been utilized for what has been referred to as
“multicenter harmonization” studies [2] in clinical centers investi-
gating immunodeficiency syndromes and hematologic malignan-
cies. For example, members of the EuroFlow Consortium have
developed exhaustive strategies for standardization of immunophe-
notyping panels to study normal, reactive, and malignant human
leukocytes [3]. Several of the panels are composed of dried formu-
lations of antibody cocktails manufactured by BD Biosciences and
Cytognos [4, 5].
Similar products manufactured by Beckman-Coulter have been
used to develop a screening tube for lymphocyte subsets in periph-
eral blood, bone marrow aspirates and body fluids [6, 7]. More
recently, the Laboratory of Immunomonitoring in Oncology has used
dried cocktails and liquid “drop-in” reagents in an initiative to
characterize the immune status of cancer patients [8].
Another multicenter harmonization project, organized by the
Human Immunophenotyping Consortium (HIPC), under the aus-
pices of the Federation of Clinical Immunology Societies (FOCIS)
attempted to minimize several contributory variables in multicenter
clinical trials [9]. The markers used for implementation of this
study were manufactured in a lyophilized form by BD Biosciences
[10] utilizing a set of standard markers for the phenotyping of
T-cells, B-cells, NK-cells, monocytes, and dendritic cells to enu-
merate the basic subsets of the above cell types and their activation
status [11]. The methodology described in this chapter developed
following the culmination of this study, as a means to augment the
functionality of the antibody cocktails and staining protocol
reported in that study [12].

2 Materials

2.1 Assay Lyoplate, Each of the five rows contains 12 replicates of the five cocktails (see
BD Biosciences Note 1).
Designate 175 (Fig. 1)
A: T-cell.
B: Treg.
C: B-cell.
 Dried Antibody Cocktails 71

Fig. 1 Assay Lyoplate (number 175) from BD Biosciences is a 96-well plate with rows A through F and columns
1 through 12 containing lyophilized antibody cocktail buttons. Each row includes 12 replicates of the complete
cocktail as follows: T-cell (A), Treg (B), B-cell (C), DC-Mono-NK (D), and T-helper (E)

D: DC-Mono-NK.
E: T-helper.

2.2 Compensation Each row contains wells with individual antibodies that are consti-
Lyoplate, BD tuents of one of the five cocktails, as well as negative and positive
Biosciences Designate compensation bead particles. Beads are not used to incorporate
177 (Fig. 2) viability dyes into the compensation matrix. Instead, cellular mate-
rial is stained separately, in tubes with cells, and these are acquired
on the cytometer at the same time as bead acquisition. Several
strategies are available for calculating the compensation matrices
for each cocktail (see Note 2).

2.3 Fluorescence- Each row contains antibodies specific for each of the five cocktails in
Minus-One (FMO) a similar fashion as the compensation Lyoplate. Columns A and B
Lyoplate, BD include all the cocktail antibodies. Cells that have been stained with
Biosciences Designate viability dye are added to column A which represents the “com-
176 (Fig. 3) plete” cocktail. Cells that have not been stained with viability dye
are added to column B which represents the viability dye FMO. The
remaining columns contain the antibody cocktail minus one of the
antibody conjugate constituents (see Note 3).

2.4 Phosphate- 1, pH 7.4 (Quality Biological, catalog number 114-058-101).

Buffered Saline (PBS)

2.5 Supplemented Staining Buffer: PBS supplemented with 1% fetal calf serum, 1%
PBS normal mouse serum, 0.02% sodium azide (see Note 4).
72 Marc Langweiler

Fig. 2 Compensation Lyoplate (number 177) from BD Biosciences is a 96-well plate with rows A through F and
columns 2 through 10 containing lyophilized buttons of a single antibody from each cocktail per well along
with both negative and positive beads that are included in BD Biosciences Anti-Mouse Ig Compensation
Particles Kit (part number 552843). Each row contains the constituents of one of the five cocktails as follows:
T-cell (A), Treg (B), B-cell (C), DC-Mono-NK (D), and T-helper (E). The antibody in well E3 is a rat monoclonal
anti-CXCR5 conjugate; therefore, the compensation bead set for this well is derived from the BD Biosciences
Anti-Rat Ig Compensation Particle Kit (catalog number 552844)

Fig. 3 FMO Lyoplate (number 176) from BD Biosciences is a 96-well plate with rows A through F and columns
1 through 11 containing lyophilized antibody buttons. Columns 1 and 2 are complete cocktails as follows:
T-cell (A), Treg (B), B-cell (C), DC-Mono-NK (D), and T-helper (E). Columns 3 through 11 are “minus-one” of
the constituents of the cocktail for that row
 Dried Antibody Cocktails 73

2.6 Fixation Medium 2% formalin solution in PBS, pH 7.4: Prepared fresh from a 10%
methanol-free buffered formalin stock (Polysciences, catalog num-
ber 04018-1).

2.7 Viability Dye Fixable Blue Live/Dead Cell Stain Kit (ThermoFisher Scientific,
catalog number L23105): The kit includes four vials of frozen dye
and one vial of dimethyl sulfoxide (DMSO) diluent. Stock solutions
are prepared by addition of 50 μl of DMSO to each of the four vials
of dye, which are pooled. Aliquots of 16 μl are placed in screw-
capped freezing vials which are stored at 80  C. One of these vials
is sufficient to stain cells from five different samples.

2.8 Cell Samples PBMC: peripheral blood mononuclear cells that had been prepared
using density gradient centrifugation and preservation in liquid
nitrogen for long-term storage (see Note 5).

3 Methods

The methods outlined below refer to the overall protocol used to

perform staining of cells using three different lyoplates. These
plates include—in the order of use—Compensation, FMO, and
Assay. The compensation plate is handled in a slightly different
manner than the latter two plates, as described below.
An adherent foil wrapping covers all three types of lyoplates,
beneath the lid of the plate. The foil is removed from the plate at
the time of use, and buttons within the wells are inspected, with any
anomalies noted, in the event unexpected results are obtained that
might be associated with any defects. The Compensation and FMO
plates are used in their entirety once they are unpackaged. How-
ever, the Assay plates may have unused wells, if there are less than
12 samples to be assayed. In these cases, the foil wrapping is only
partially removed so that it remains adhered to unused plate wells.
If the plate is used again for a subsequent staining procedure, the
wells are evaluated for evidence of hydration of the remaining
lyophilized buttons. If this is noted in any well the plate is
discarded.
Once the substrate is added to the lyophilized antibody cocktail
button, the workflow for surface-marker immunophenotyping
using dried reagents is not markedly different from that used with
liquid reagents. Hence, this chapter does not proceed past the
preparation of samples to be presented for data acquisition on a
flow cytometer; these steps, as well as data analysis strategies, are
particular to individual flow cytometry laboratories that utilize
liquid antibody cocktails. Depending on the vendor of the dried
cocktails, they may be manufactured in tubes or plates. Hence, the
particulars for the procedure following admixture are contingent
on the staining vessel. There are advantages and disadvantages
74 Marc Langweiler

inherent in each type of containment, regarding staining versus

cytometer data acquisition (see Note 6).

3.1 Processing 1. PBS is added to each well in 100 μl volumes, starting in well
the Compensation A1, proceeding down the column then moving to the top of
Lyoplate successive columns.
2. After addition of PBS to all wells, the plate is incubated for
30 min at the ambient temperature, protected from light.
There is a 5-min period during which the added buffer dis-
solves the lyophilized button; the plate is placed on a rotary
plate mixer at low speed for the remaining 25 min.
3. At the end of this interval, the plate is centrifuged (ambient
temperature, 400  g, for 5 min) and flicked to remove super-
natants, followed by addition of 100 μl of Staining Buffer. The
contents of each well are transferred into labeled polystyrene
tubes and held at 4  C until cytometer acquisition, which is
completed within 30 min of staining. Beyond this interval, the
beads will begin to deteriorate, which may compromise the
validity of the compensation matrices.

3.2 Processing Cell 1. After cells have been thawed and washed, 1 ml of PBS is added
Samples to the residual volume remaining in each tube.
2. Before the addition of 1 μl of the viability dye working solution,
a 50 μl aliquot of cell suspension is removed and added to
another labeled tube containing 100 μl of Staining Buffer, to
be acquired on the cytometer as unstained samples.
3. The tubes with cells and viability dye are incubated, at the
ambient temperature, protected from light for 5 min, followed
by the remaining 25 min on a rotating plate mixer.
4. At the end of this interval, the tubes are centrifuged (ambient
temperature, 400  g, for 5 min), decanted one-by-one, and
the pellets resuspended by gentle repipetting in their residual
supernatant with addition of 450 μl of Staining Buffer, to
ensure sufficient volume for staining the five wells for each
sample (100 μl per well).

3.3 Processing 1. For either plate, once removed from its pouch, and peeling off
the FMO and Assay of the foil cover, 100 μl of cells from Subheading 3.2 is added
Lyoplates into rows A through E in the column assigned to each sample.
Pipetting of sample starts in column 1 and proceeds through
column 10, using a new pipette tip for each well.
2. Once all wells are filled, a 30-min incubation at the ambient
temperature ensues, with the plate protected from light. After
the first 5 min, the cells are resuspended, column-by-column,
using an eight-channel multipipettor set at 80 μl. The suspen-
sions are mixed by gentle repetitive pipetting. After addition to
each column, the tips are discarded and replaced with new tips
 Dried Antibody Cocktails 75

before proceeding to the next column. Following the resus-

pension of cells, there is an additional 25-min incubation,
protected from light, at the ambient temperature on the rotary
plate mixer.
3. Washing steps (see Note 7).
Wash 1: 100 μl of PBS is added to each well, mixed by gentle
repetitive pipetting, then centrifuged and flicked to remove
supernatants.
Wash 2: 200 μl of PBS is added to each well, mixed by gentle
repetitive pipetting, then centrifuged and flicked to remove
supernatants.
Wash 3: 200 μl of 2% formalin is added to each well, mixed by
gentle repetitive pipetting. The plate is then incubated at
the ambient temperature, protected from light, for 30 min,
then centrifuged and flicked to remove supernatants.
Wash 4: 200 μl of Staining Buffer is added to each well, mixed
by gentle repetitive pipetting, then centrifuged and flicked
to remove supernatants.
4. The final suspension volume of 200 μl of Staining Buffer is
added, followed by gentle repetitive pipetting. The contents of
each well are transferred to polystyrene tubes labeled with the
row/column identifier, already containing 250 μl of Staining
Buffer. Samples are collected on the cytometer immediately
following the staining procedure. In the interval prior to
acquisition on the cytometer sample tubes are protected from
light and held at 4  C until ready for acquisition, either refri-
gerated or on wet ice.

4 Notes

1. The original iteration of the Lyoplate product for the HIPC

project [11] did not include the utilization of a 355 nm laser
since not all instruments from participating centers were
equipped with one. With our desire to use this laser, a
CD45∗BUV conjugate (BD Biosciences, catalog number
563792) was included in each lyophilized cocktail. The original
iteration directed usage of a viability dye excited by the 488 nm
laser; this was modified to use the 355 nm laser-excited viability
dye discussed above in Subheading 2.6. This made the 488 nm
laser-excited conjugates available for incorporation of an addi-
tional conjugate - either FITC (rows B and D) or
BB515 (rows A, C and E) - in each cocktail:
Row A (T-cell) <viability dye/CD45∗BUV/CD3∗V450/
HLA-DR∗V500/CD28∗BB515/CD4∗PerCP-Cy5.5/
76 Marc Langweiler

CCR7∗PE/CD45RA∗PE-Cy7/CD38∗APC/
CD8∗APC-H7>
Row B (Treg) <viability dye/CD45∗BUV/CD3∗V450/
HLA-DR∗V500/CD39∗FITC/CD4∗PerCP-Cy5.5/
CD25∗PE/CCR4∗PE-Cy7/CD127∗AF647/
CD45RO∗APC-H7>
Row C (B-cell) <viability dye/CD45∗BUV/CD3∗V450/
IgD∗V500/CD10∗BB515/CD19∗PerCP-Cy5.5/
CD24∗PE/CD27∗PE-Cy7/CD38∗APC/
CD20∗APC-H7>
Row D (DC-Mono-NK) <viability dye/CD45∗BUV/
CD14∗V450/HLA-DR∗V500/CD163∗FITC/
CD123∗PerCP-Cy5.5/CD56∗PE/CD11c∗PE-Cy7/
CD16∗APC/CD3 + CD19 + CD20∗APC-H7>
Row E (T-helper) <viability dye/CD45∗BUV/CD3∗V450/
HLA-DR∗V500/CXCR5∗BB515/CD4∗PerCP-
Cy5.5/CXCR3∗PE/CCR6∗PE-Cy7/CD38∗APC/
CD8∗APC-H7>
2. The original iteration of the Lyoplate product for the HIPC
project [11] used a single plate that contained the compensa-
tion beads and the assay cocktails. Combining the assay and
compensation wells in one plate results in the compensation
beads being washed more aggressively and incubating longer
than necessary, leading to inconsistencies when calculating
matrices. An alternative strategy is to use separate plates for
compensation and assay. Using a separate compensation
plate—with more wells available—allows the use of the full
complement of the conjugates for each cocktail. This offers
the opportunity for two strategies to construct the matrices.
In the first scenario, separate matrices are constructed for each
cocktail, using the full complement of wells in each row. In the
second scenario, only one well of conjugates with replicates
present in more than one cocktail is acquired. Following data
acquisition, the experiment in the cytometer software is repeat-
edly duplicated resulting in five copies, one for each cocktail.
The data files that are not relevant for a given cocktail are then
deleted before the matrix is constructed. With regards to the
frequency of preparation/data acquisition/data analysis of the
Compensation Lyoplate, the expectation at the outset, when
the compensation wells and antibody conjugate wells were
included on a single plate, was that new matrices would be
constructed for every plate. In the protocol described herein,
matrices were constructed with much less frequency—only
when there were significant changes to the cytometer instru-
ment optics/fluidics/electronics during servicing. The rela-
tively low number of parameters associated with the panels
 Dried Antibody Cocktails 77

described above makes it possible to construct N  N plots of

all fluorochrome combinations within the cytometer software
experiment template before exporting files for off-line analysis.
With the use of the experiment template copied from the
previous acquisition session, changes to matrix values are easily
accomplished to address minor changes inherent in day-to-day
variability in cytometer performance.
3. In the original iteration of the BD Lyoplate used in the HIPC
study [10–12], there was no provision for performing FMO
analysis. Given its importance of in immunophenotyping [13],
the use of dried reagents for this procedure is a significant
addition to the protocol. Therefore, a limited number of
FMO lyoplates were prepared during the production run.
The use of dried FMO cocktails in plates is a significant time
and cost-saving step when compared with using liquid cocktails
in tubes. That said, once the initial quadrant gate placements
are derived from the initial FMO files analyses, subsequent
FMO plates are only run following significant alterations
following servicing of the cytometer. The strategy for confirm-
ing proper quadrant gate placement, after initial placement
from analysis of FMO data files, involves the use of Boolean
gating strategies to derive “NOT” gates in order generate
internal “negative” cell populations in the absence of
markers that would definitively identify contextual
leukocyte subpopulations.
4. Several stages of the procedure are carried out using PBS
without any additives, while others utilized this buffer supple-
mented with 1% fetal calf serum, 1% normal mouse serum and
0.02% sodium azide (Staining Buffer). PBS, instead of Staining
Buffer, is used during cell processing before addition of viabil-
ity dye to prevent diminution of viability dye uptake in the
presence of protein. PBS is used during processing of the
Compensation lyoplate because the presence of normal
mouse serum in the Staining Buffer would block binding of
murine antibodies to the compensation particles. PBS is used in
the first two wash steps following staining, before the formalin
fixation step. All other steps utilize Staining Buffer.
5. Though several of the cited references report on the use of
dried antibody cocktails to study bone marrow aspirates and
body fluids, the primary substrate used is peripheral blood,
which is prepared using lysis techniques (before or after addi-
tion to antibodies), or density gradient separation. In most
cases, the latter procedure is an antecedent to long-term stor-
age in liquid nitrogen. The samples used in this procedure were
not processed in the same laboratory that performed the stain-
ing; hence, the methodology is not described. There are
recommendations for using certain types of cell processing,
78 Marc Langweiler

depending on the nature of the study being conducted. In

studies that involve single-time point determinations at a single
testing site, the use of lysed blood would be preferable to
minimize cell loss and possible phenotypic modification that
might occur with cell separation methods and long-term stor-
age. In studies that involve temporal sampling and/or testing
at multiple sites, as described herein, the use of frozen periph-
eral blood mononuclear cell preparations is a necessity.
6. The choice of whether to use dried antibodies prepared in
microtiter plates or 12  75 mm polystyrene tubes has
depended on the format chosen by the vendor manufacturing
the product. For example, the original Lyoplate product from
BD Biosciences was only available in plates, and the original
Duraclone product from Beckman-Coulter was only available
in tubes. It is an open question as to whether there will be the
opportunity to request one or the other format. That being the
case, regarding cell preparation, the use of microtiter plates is
an advantage compared to tubes concerning processing time
and nonantibody reagent costs. The same applies to data acqui-
sition on instruments with high-throughput acquisition mod-
ules. The techniques described herein used staining in plates,
with transfer to tubes before cytometer acquisition. The reason
for transferring to tubes was due to a preponderance of pauci-
cellular samples, and a desire to collect at low flow rates. Since
the plates take several hours to run to completion, and there is
no mechanism to cool the plate during collection, it is neces-
sary to keep the samples chilled until the time that they would
be acquired, to prevent cell and conjugate degradation.
7. Lyoplate washing procedures, including centrifugations
(400  g for 5 min), are carried out at the ambient tempera-
ture. Supernatants are removed by flicking the plates into a
sink. The sink is disinfected with a 10% bleach solution follow-
ing each wash step. Centrifugations in tubes, carried out during
the preparation of cellular material before staining, are per-
formed at the ambient temperature at 400  g for 5 min,
with supernatant removal by decanting.

References
1. Hammerling JA (2012) A review of medical 3. van Dongen JJM, Lhermitte L, Böttcher S et al
errors in laboratory diagnostics and where we (2012) EuroFlow antibody panels for standar-
are today. Lab Med 43(2):41–44 dized n-dimensional flow cytometric immuno-
2. Jamin C, Lann L, Alvarez-Errico D et al (2016) phenotyping of normal, reactive and malignant
Multi-center harmonization of flow cytometers leukocytes. Leukemia 26(9):1908–1975
in the context of the European “PRECI- 4. Van der Velden VHJ, Flores-Montero J, Perez-
SEADS” project. Autoimmun Rev 15 Andres M et al (2017) Optimization and test-
(11):1038–1045 ing of dried antibody tube: the EuroFlow LST
and PIDOT tubes as examples. J Immunol
 Dried Antibody Cocktails 79

Methods. https://doi.org/10.1016/j.jim. 9. Maecker HT, McCoy JP, FOCIS Human

2017.03.011 Immunophenotyping Consortium (2010) A
5. Glier H, Heijnen I, Hauwel M et al (2017) model for harmonizing flow cytometry in
Standardization of 8-color flow cytometry clinical trials. Nat Immunol 11(11):975–978
across different flow cytometer instruments: a 10. Chan RC, Kotner JS, Chuang CM et al (2017)
feasibility study in clinical laboratories in Swit- Stabilization of pre-optimized multicolor cock-
zerland. J Immunol Methods. https://doi. tails for flow cytometry applications. Cytome-
org/10.1016/j.jim.2017.07.013 try B Clin Cytom 92(6):508–524
6. Hedley BD, Keeney M, Popma J et al (2015) 11. Maecker HT, McCoy JP, Nussenblatt R (2012)
Novel lymphocyte screening tube using dried Standardizing immunophenotyping for the
monoclonal antibody reagents. Cytometry B human immunology project. Nat Rev Immu-
Clin Cytom 88(6):361–370 nol 12(3):191–200
7. Rajab A, Axler O, Leung J et al (2017) 12. Finak G, Langweiler M, Jaimes M et al (2016)
Ten-color 15-antibody flow cytometry panel Standardizing flow cytometry immunopheno-
for immunophenotyping of lymphocyte popu- typing analysis from the human immunophe-
lation. Int J Lab Hem 39(Suppl. 1):76–85 notyping consortium. Sci Rep. https://doi.
8. Pitoiset F, Cassard L, Soufi E et al (2018) Deep org/10.1038/srep20686
phenotyping of immune cell populations by 13. Feher K, Kirsch J, Radbruch A et al (2014) Cell
optimized and standardized flow cytometry population identification using fluorescence-
analyses. Cytometry A. https://doi.org/10. minus-one controls with a one-class classifying
1002/cyto.a/23570 algorithm. Bioinformatics 30(23):3372–3378
 Chapter 5

Guidelines for Gating Flow Cytometry Data

for Immunological Assays
Janet Staats, Anagha Divekar, J. Philip McCoy, Jr,
and Holden T. Maecker

Abstract
“Gating” refers to the selection of successive subpopulations of cells for analysis in flow cytometry. It is
usually performed manually, based on expert knowledge of cell characteristics. However, there can be
considerable disagreement in how gates should be applied, even between individuals experienced in the
field. While clinical software often automates gating, and some guidelines do exist (especially for clinical
assays), there are no comprehensive guidelines across the various types of immunological assays performed
using flow cytometry. Here we attempt to provide such guidelines, focused on the most general and
pervasive types of gates, why they are important, and what recommendations can be made regarding
their use. We do so through the display of example data, collected by academic, government, and industry
representatives. These guidelines should be of value to both novice and experienced flow cytometrists
analyzing a wide variety of immunological assays.

Key words Flow cytometry, Gating, Analysis

1 Introduction

The analysis of flow or mass cytometry data uses a process com-

monly called “gating,” whereby populations of cells are virtually
separated from one another. This is accomplished through the use of
specialized flow cytometry software programs to identify subsets
within a sample of particles or cells, and to assess the expression
levels of specific markers of interest on those selected subsets, with-
out confounding data from other cellular subsets. In this context,
some authors may refer to an “analysis region” rather than a “gate,”
which formally refers to physical sorting of cells; but we will not
distinguish the two here. As widely practiced today, gates are set by
human input based on visual examination of the flow cytometry data.
Subsets of interest can range in magnitude from very abundant
to exceptionally rare. The relative abundance of a subset, as well as
its separation from other populations, affects the ease with which

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_5, © Springer Science+Business Media, LLC, part of Springer Nature 2019

81
82 Janet Staats et al.

that subset can be accurately and reproducibly gated, as well as the

purity of the gate.
Most software programs used for flow cytometry analysis allow
for gates to be drawn in one or two dimensions (i.e., on histograms
or on two-parameter plots). Two-dimensional gates may be rectan-
gular, elliptical, or irregular polygons; and the plots on which they
are drawn may depict contours, density, or individual events (dot
plots). In theory, gates could be drawn in more than two dimen-
sions at once, but this is practically not feasible for humans. Hence,
subsets that are distinguished by a series of markers are usually
identified by sequential one- or two-dimensional gates (known as
Boolean hierarchical gating).
An important aspect of hierarchical gating is knowledge of the
expression profile of a population of interest across many different
markers. Hierarchical gates can include both positive and negative
markers; and the sequential order of the hierarchical gates can
influence the ability to clearly resolve the final population. There-
fore, accurate and robust gating should be based on a strategy that
involves an optimal sequence of marker-based gates, rather than
just a single terminal gate based on one or two markers.
Gating is subjective; substantial variations in gating can be
observed between the gate set by different analysts. In addition to
operator expertise, this is in part dependent on the quality of
staining. Many factors influence the observed fluorescence intensity
of the samples, as summarized in Table 1. While most of these
factors have been reviewed previously, it is generally assumed that
the operator will make appropriate adjustments for these factors in
terms of gating, although, ideally, these factors should be first

Table 1
Factors affecting population resolution in flow cytometry

Root level Factor References

Biological Donor variability in expression levels, background [1]
Sample integrity (shipping, cryopreservation/thawing, viability, etc.) [2–4]
Aggregates [5]
Instrument Optical configuration (lasers, laser power, filters) [6]
Daily performance, including laser delays, voltage calibration [7]
Hydrodynamics (air bubbles, clogs) [8, 9]
Reagents Antibody clones, titers, and lot-to-lot variability [10, 11]
Panel design (fluorochrome brightness and spillover) [12]
Reagent degradation, especially tandem dyes [12]
Staining protocol, including activation, fixation, permeabilization, [11]
intracellular vs. cell-surface staining, washing
Data transformation Accuracy of compensation for fluorescence spillover [13, 14]
Scaling (logarithmic vs. biexponential, valley artifact) [15, 16]
 Flow Cytometry Gating Guidelines 83

minimized through technical means (e.g., experimental design and

careful adherence to protocol and to best practices).
Because of the vast numbers of different assays that can be
performed, flow cytometry has developed into a widely used tech-
nology in a number of scientific endeavors. These applications can
be roughly divided into four major areas: clinical/diagnostic, clini-
cal/translational research, biopharma/drug discovery, and basic
research. While these four areas use a common technology, each
has its nuances that relate to the specific applications. Thus, when
guidelines for flow cytometry are established, some elements might
be common to all applications while others might apply to one or a
few of the applications. Perhaps for this reason, there are few
general guidelines on how to place gates that are applicable to the
vast majority of flow cytometry analyses.
As one might expect, the clinical environment has led the way
in efforts to establish practice guidelines for flow cytometry, in large
part due to the implementation of CLIA-88 (Clinical Laboratory
Improvement Act of 1988). These laws were enacted to assure
reliable, high quality, accurate, and safe testing in all clinical labora-
tories. The laws encouraged a standard of practice, as well as profi-
ciency testing and inspections. CLIA, in turn, led directly to the
issuance of guidelines for the performance of flow cytometry in the
clinical environment by the College of American Pathologists
(CAP) and the Clinical Laboratory Standards Institute (CLSI)
(formerly NCCLS). These guidelines dealt with a variety of issues
ranging from lot testing to instrument set up to establishing stan-
dard operating procedures (SOPs) for nearly every aspect of the
laboratory.
In the clinical and translational research environments, there
have been a number of assay-specific guidelines that have been
established. Perhaps the best known are the guidelines for CD4 T
cell counting in the context of HIV disease. One of the first of these
were the guidelines published by the US Centers for Disease Con-
trol and Prevention (CDC) in 1992, with supplements and revi-
sions to these guidelines issued in 1994, 1997, and 2003 [17–20].
Similar guidelines for enumeration of CD4 T lymphocytes have
been put forward by the World Health Organization [21].
Guidelines for additional clinical tests have also been published
and gained wide acceptance. Among these are guidelines for enu-
meration of CD34+ hematopoietic stem cells (proposed by the
International Society of Hematotherapy and Graft Engineering
[22]), and guidelines for detecting paroxysmal nocturnal hemoglo-
binuria (PNH) [23]. These included specific directions for how the
gating is to be performed for these assays.
While the clinical arena is heavily regulated with enforceable
guidelines, the remaining areas in which flow cytometry is used are
far less regulated. Except for some network clinical trials work, for
which GCLP compliance is required, most often adherence to
84 Janet Staats et al.

guidelines in research environments is voluntary rather than man-

dated. For biopharmaceutical and translational research cytometry,
there have been guidelines established to promote high-quality
practices, although these pertain more to general laboratory prac-
tice rather than being specific for flow cytometry. In the realm of
translational research, several guidelines have been proposed which
would also be applicable to basic research. Examples of these
include guidelines to intracellular cytokine staining, proposed by
the Cancer Immunotherapy Consortium [24] and guidelines for
HLA-peptide multimer assays proposed by the Cancer Vaccine
Consortium [25]. Of these, only the former included details
concerning how the gating should be performed.
Basic research using flow cytometry is the area in which the
least has been done to standardize or harmonize how assays are
performed. This is due to a number of factors including the vast
array of assays that are performed by flow cytometry in basic
research, the fiercely independent nature of many researchers, and
the lack of the ability to enforce compliance with guidelines in this
area. Nomura and colleagues demonstrated that gating of flow
cytometric data was a significant factor in interlaboratory variations
[26], a finding that was later confirmed by Finak et al. [27]. The
latter study also explored advantages offered by an automated
gating strategy.
A number of automated gating algorithms have been intro-
duced to reduce the subjectivity of manual gating, decrease the
time and labor of manual analysis, and/or discover populations
not anticipated by manual gating (reviewed in [28, 29]). These
range from tools that help to define the edges of a selected cell
population, to supervised approaches for analyzing a complete data
set, and even fully unsupervised clustering algorithms. In general,
with increasing complexity and decreasing supervision comes the
prospect that the automated method may produce spurious or
undesirable results; or at least, will not match the manual gating
results of recognized experts. As such, automated gating tends to
still require some manual confirmation. It is not our purpose here
to debate the relative merits of automated approaches; but rather,
recognizing that the vast majority of gating is still done manually, to
propose guidelines that could improve the consistency and accuracy
of such manual gating. Certainly, some of these guidelines might be
useful to designers of automated algorithms as well.

2 Guidelines

2.1 Acquisition Definition: When acquiring samples using a flow cytometer, the
Threshold or cells of interest are usually in a suspension that contains many other
Trigger Gate cells or particles. Each cell or particle is measured by the cytometer
as an event in the form of a pulse; some events represent signals of
 Flow Cytometry Gating Guidelines 85

interest and some are noise. An operator may set a limit for the type
of pulse that is acceptable to be included in the resulting data file;
this is called an acquisition (detection) threshold or trigger gate.
Trigger gates are set prior to acquisition as a means of excluding
events of noninterest or noise. The most common type of trigger
gate is used to exclude debris based on size, measured using for-
ward scatter. Using forward scatter, the trigger gate is set to exclude
events that are too small to be considered cells. Another type of
commonly used trigger gate is based on CD45 expression, where
CD45 is a panleukocyte marker (see Fig. 1). Setting the trigger gate
on CD45+ cells will essentially exclude all nonleukocyte events.
Note that on some cytometers, a storage gate can be set in
addition to an acquisition threshold or trigger. Care should be
taken that the storage gate as well as the acquisition gate includes
all populations of potential interest for later analysis. For example, a
storage gate set on lymphocytes (using forward vs. side scatter)
would preclude downstream analysis of monocytes.

Fig. 1 A CD45 trigger gate (threshold) is used to exclude nonleukocytes. In a sample heavily contaminated with
erythrocytes, a CD45 threshold can help to exclude these cells and “clean up” the downstream gates. In this
example, the same erythrocyte-contaminated sample was run with a high, medium, or low threshold on CD45.
Left panels: With no further gating on CD45, the initial FSC vs. SSC gate for lymphocytes (first column) is
uncontaminated with a high CD45 threshold (top row), but progressively more contaminated as the CD45
threshold is reduced. This results in an artificially low observed percentage of B cells (second column), due to
dilution of lymphocytes by contaminating erythrocytes. Right panels: The same data can be effectively rescued
by applying a postacquisition gate on CD45high events (first column), prior to lymphocyte gating (second
column). This results in similar observed percentages of B cells (third column), albeit at the expense of greatly
diminished cell numbers when there was a low CD45 threshold on acquisition
86 Janet Staats et al.

How data is affected: Trigger gates have a “Goldilocks” zone. A

low or no trigger gate leads to excess noise and reduces overall
signal-to-noise ratio. But when trigger gates are set too high, they
partially or completely exclude the target population of interest and
decrease the overall signal. When applied properly, trigger gates
effectively remove unwanted events or cells, thus decreasing noise
and improving the overall signal-to-noise ratio.
Conclusion: Threshold gating is a very effective tool for reduc-
ing noise. To ensure the removal of noise and not signal, setting
trigger gates properly requires both prior knowledge of the sample
and the target population of interest.

2.2 Time Gate Definition: Time was first introduced as a flow cytometry parameter
to measure kinetic responses [30]. Subsequently, in 1987, James
Watson first described the concept of using time as a quality control
(QC) parameter [8]. The principle of flow cytometry is based on
regulating the differential pressure between sample and sheath fluid
to establish a laminar flow, enabling a single cell to pass in front of
the laser for interrogation. The steps involved with acquiring a
sample influence the fluid dynamics throughout the course of
acquisition. When a sample is first introduced into the cytometer,
a burst of pressure may cause a high density of events to pass in
front of the laser during the first few seconds. After the initial
sample introduction, the flow rate stabilizes and becomes laminar.
Nearing the end of acquisition, the sample volume is greatly
reduced, which may influence the fluid dynamics again. During
acquisition, fluidic changes may be observed as either an increase
or decrease in the event rate. The most frequent culprit is air
bubbles or clogs. For these reasons, the time parameter may be
used during analysis to exclude those events that were acquired
during periods of nonlaminar flow. This is called a time gate. For
manual analysis, a time gate is created by visualizing time versus
scatter or fluorescence. Alternatively, there are automated software
options, such as FlowClean [9], that may be used to effectively
remove anomalies, such as those caused by changes in the event
rate, from an FCS file. Of note, the time parameter may be dis-
played differently based on the software being used for acquisition
or analysis.
How data is affected: Anomalies in the flow rate result in wider
CVs and/or shifts in scatter and fluorescence parameters (see
Fig. 2). The effect of nonlaminar flow on the resulting data may
be modest or severe, depending upon the degree of fluidic instabil-
ity and the integrity of any subsequent quality control gates, such as
singlet and fluorescence gates. In the most severe case, an entire
data file may be unacceptable for analysis due to unstable fluidics
that occurred throughout the time of acquisition.
Conclusion: The time parameter is an effective QC tool for
identifying temporal changes in fluid dynamics that occurred
 Flow Cytometry Gating Guidelines 87

Fig. 2 Use of time gates. (a) Time as a QC parameter. PBMCs were thawed, permeabilized, stained, and
acquired. Time versus SSC-A is used to visualize the temporal changes in fluid dynamics that occur
throughout the sample acquisition. Left plot illustrates an ideal time parameter plot. The burst of sample at
the beginning, observed as the red or higher density lower scatter population, has an initially wider SSC CV
that rapidly resolves into a tight SSC CV. The SSC CV remains the same throughout the remainder of
acquisition and slightly decreases near the end of acquisition, as the sample rate decreases (red turns to
orange). There are no skips or jumps in the number of events relative to time and no shifting of SSC over time.
Middle plot: moderate nonlaminar flow. A more noticeable sample burst at the beginning of acquisition with
wide SSC-A CV. SSC-A CV changes throughout the acquisition in an oscillating pattern, caused by swirling
fluidics. The time gate, indicated in pink, is set to include the most stable portion of the events acquired. Right
plot: illustrates nonstable fluidics that occurred throughout the entire sample acquisition. Clogs are indicated
by skips or jumps in time, where the event rate is greatly reduced followed by bursts of events when the clog
breaks. Changes in SSC-A CV and shifting of SSC-A are prominent throughout the acquisition. The time gate,
in pink, is placed around the most stable area, although no area in this example is ideal. While a significant
portion of events are excluded due to unstable fluidics, the data included in the time gate represents those
events with the highest integrity. (b) Time gate used to exclude shifting scatter created by nonlaminar flow.
LyoCell PBMC control product was stained with CD14. The gating strategy shown was used to identify CD14+
monocytes by first parsing through time, then viability, singlets (FSC and SSC), and scatter gates. Both rows
represent analyses using the same FCS file. Top row: the time gate is set around the region with best laminar
flow, identifying approximately 71% of monocytes as CD14+. Bottom row: using a time gate set on the region
with nonlaminar flow there are no CD14+ monocytes. (c) Time gating affects CV of downstream gates. Human
PBMC were stained with CD4 PE-Cy7. The leftmost plot is ungated, while the middle and the right histograms
are gated on FSC/SSC Singlets. If placement of time gate includes sections of the plot where nonlaminar flow
exists then the CV as well as median of the CD4+ cell peak and the frequency of CD4+ cells is affected.
Inclusion of events from the unstable fluidics results in a wider CV and higher percent positive cells and
exclusion of these events results in a tighter peak with lower CV and higher fluorescence intensity
88 Janet Staats et al.

throughout sample acquisition. Gating on “good” acquisition

interval(s) can improve data quality.

2.3 Singlet Gates Definition: Aggregates consist of one or more cells or particles; they
form for many reasons, including cells clumping together due to
dying/death, adherent properties, antigen presentation, and also
high sample concentration, high flow rate, and even dye–dye aggre-
gation. When running clumped samples with many aggregates, it is
important to filter the samples prior to acquisition to remove as
many aggregates as possible. When filtering is insufficient to
remove all aggregates, they may be further removed using singlet
gating. To understand how singlet gates work, it is important to
understand how signals are detected by a flow cytometer. When a
particle passes in front of the laser, the cytometer measures the
resulting photons of light in the form of a pulse. Digital cytometers
are capable of collecting three values that define a pulse: Width is a
measurement that represents the time a cell passes in front of the
laser. Width is proportional to the size of the cell or particle and is
unaffected by PMT voltage settings [31]. Height is the measure-
ment of the intensity of the signal (maximum fluorescence, size, or
scatter) and is affected by PMT voltages. Area is not measured; it is
derived using both width and height measurements, and is only
slightly affected by PMT voltage, owing to its relationship to both
width and height. Area scaling is used in some digital instruments
to equalize height and area. Pulses can be distorted by coincidence,
when two events (a doublet) pass in front of the laser simulta-
neously. This distortion results in a twofold stretching of the area
and width, while the height remains unchanged. During analysis,
singlet gating uses the distortion aggregates impose on pulse width
and area to distinguish singlets from aggregates. Singlet gating is
performed using either forward scatter (FSC) or side scatter (SSC)
parameters. It is important to remember that FSC is collected using
a diode, unless you have an instrument specially equipped with FSC
PMT, and SSC is collected using a PMT. Since PMTs are more
sensitive than diodes, a better resolution between singlets and
aggregates may be visualized using the SSC over FSC parameter.
As a best practice, ideally both FSC and SSC singlet gates would be
used sequentially to remove coincidental events occurring in differ-
ent orientations. There are two common approaches to singlet
gating; one is dependent upon area scaling and one is not: area
versus height and width versus height. For area versus height singlet
gating, coincidence is identified by the events for which there is an
increase in area without a corresponding increase in height. Thus,
cells along the diagonal are singlets and should be included in the
singlet gate. Use of area versus height singlet gating requires the
area scaling to be properly adjusted on your cytometer using cells;
otherwise, there is a risk for gating errors. Maximum resolution
between singlets and aggregates is observed using width versus
 Flow Cytometry Gating Guidelines 89

height, as both independently measured values. Another advantage

to using width versus height singlet gating is that neither value is
affected by area scaling [32]. For width versus height singlet gating,
coincidence is identified by an increase in width, while height
remains unchanged. Irrespective of the pulse value used to establish
singlet gates, debris may be readily distinguished from cells in the
lower corner and may be excluded using the same gate as the
aggregates. Examples of singlet gates are shown in Fig. 3.
How data is affected: It is important to remove aggregates
during data analysis of immunophenotpying data, as these can
give false double positive events. As an example, if a cluster of
cells includes B (CD19+ CD3 ) and T (CD19 CD3+) cells and
if these are not removed from the data analysis, then the operator
may report the presence of a cell population that is CD3+ CD19+
(a T-B cell), which does not exist. Applying doublet discrimination
helps remove these oddities from data.
Conclusion: In the context of immunophenotyping, singlet
gates are necessary for accurate identification of cells that are coex-
pressing two markers or double positives. Once singlet gating is
applied, in some instances, this will “clean up” data and thus result
in loss of a certain cell population, whereas in other cases it may
result in enrichment of a cell type due to loss of nonspecific popula-
tions. There are, however, circumstances under which one might
want to evaluate aggregates, and for those assays the use of singlet
gating would not be applicable [33].

2.4 Exclusion Gates Definition: In the most general sense, an exclusion gate is any gate
that removes unwanted cell populations, before positive gating on
markers of interest. Practically, this is often useful to prevent the
interference of these excluded populations on the downstream
analysis. Besides time and singlet gates already discussed above,
two common variations of such exclusion gates are viability gates
and lineage marker cocktails (often called a “dump channel”).
Viability is most frequently assessed by the fact that live cells have
intact membranes and thus exclude dyes that would otherwise
enter the cell and bind, for example, DNA. Propidium iodide and
7-AAD are examples of such dyes. Alternately, agents that react
with free amine [4] or thiol groups can be used, taking advantage of
the fact that relatively few such reactive groups are present on the
cell surface, but many more are available in the interior of cells.
Many dyes are available in this category, with different fluorescent
properties, such that they can be combined with almost any com-
bination of fluorochrome-labeled antibodies. A major advantage of
these dyes is that they form a covalent bond with cellular proteins,
and thus do not diffuse away if cells are fixed, permeabilized, or
otherwise processed after application of the viability stain. For this
reason, they are often referred to as “fixable” viability dyes.
90 Janet Staats et al.

a WvH AvH
250K

200K

150K
FSC-H

FSC-H
Forward
scatter 75.7
100K

50K

0
FSC-W FSC-A
250K

96.6 200K Single cells SSC

72.8
150K
SSC-H

Side SSC-H
scatter 100K

50K

0
SSC-W SSC-A

b Without singlet gating With singlet gating

Q1 Q2 Q1 Q2
105 1.01 31.7 105 0.96 36.0

104 104
FoxP3 PE

FoxP3 PE

103 103

0 0
Q4 Q3 Q4 Q3
103 26.7 40.6 103 20.4 42.6

-103 0 103 104 105 -103 0 103 104 105

CD25 APC CD25 APC

Fig. 3 Singlet gating. (a) Width vs height and area versus height singlet gates using FSC and SSC. Whole blood
was subjected to erythrocyte lysis, stained, and acquired by flow cytometry. Top row: Forward scatter
parameter is used to create singlet gates. Bottom row: side scatter is used to create singlet gates. Left
column: width (-W) versus height (-H) plots are used for singlet gating and to exclude debris in the lower
corner. Aggregates are those events for which there is an increase in -W relative to -H. Right column: area (-A)
versus -H is used for gating singlets; aggregates are those events for which -A increases disproportionately to
-H. (b) Exclusion of doublets can alter the frequency of gated populations. Whole blood was stained with a
cocktail containing CD3, CD4, CD127, CD25, and Foxp3. Left image depicts frequency of Treg cells without
doublet exclusion and right panel shows frequency of Treg cells with singlet gating or doublet exclusion. Note
increase in frequency of CD25+ Foxp3+ cells following gating on singlets. Singlet gating was performed using
SSC-A vs SSC-H followed by FSC-A vs FSC-H
 Flow Cytometry Gating Guidelines 91

How data is affected: Dead cells tend to bind antibodies and/or

fluorochromes nonspecifically and as such interfere with the quan-
titation of cells identified with positive markers alone. This is par-
ticularly important when quantifying rare populations like antigen-
responsive T cells (Fig. 4). In such assays, dead cells may add to the
“background” or unstimulated response, as well as potentially
obscuring the distinction between positive and negative
populations.
A lineage cocktail uses markers for one or more cell lineages
that are not of interest for downstream analysis. For example, B, T,
NK, and monocyte markers are commonly combined in a lineage
cocktail for the downstream gating of dendritic cells (DCs), which
bear none of these markers. This exclusion improves the ability to
perform positive gating on DCs (commonly gated as lineage-HLA-
DR+). But other applications are also possible. Identification of
antigen-specific T cells with certain reagents may be improved by
first gating out monocytes and B cells, which express Fc receptors.
Samples that have erythrocyte contamination may benefit from
exclusion gating on CD235a, an erythrocyte-specific marker. And
use of a platelet-specific antibody like CD62P can be used to
remove platelet-bound lymphocytes and thereby clean up intracel-
lular cytokine staining of antigen-responsive T cells [34].
To some extent, lineage cocktails also help to remove dead cells
via their nonspecific binding; but they are not a perfect surrogate
for a viability dye and can serve other functions as noted above. Of
course, a viability dye and a lineage cocktail can also be combined in
a single channel.
Conclusion: Exclusion gates reduce noise, thus enhancing the
overall signal-to-noise ratio. Rare populations are more accurately
quantitated with exclusion of dead cells and/or irrelevant cell
lineages.

2.5 Scatter Gates Definition: Fluorescence flow cytometry typically measures forward
and right-angle (side) scatter of the laser light used to excite fluor-
ochromes on the particles of interest. These two parameters contain
information on cell size and granularity, respectively. As such, they
are frequently used as initial gates on immune cell populations, such
as lymphocytes, monocytes, and granulocytes.
How data is affected: While a population such as lymphocytes
may appear to have a relatively homogeneous scatter profile, there
are in fact subtle differences in the scatter profile of different
lymphocyte subsets. Most dramatic is the increased forward scatter
of proliferating lymphocytes (blasts); a tight lymphocyte gate thus
misses most of the proliferating cells. But even populations such as
B cells or CD8+ T cells can have subtle shifts in their scatter relative
to the rest of the lymphocytes. The result is that the placement of
the lymphocyte gate can have small but potentially significant
92 Janet Staats et al.

Fig. 4 Excluding dying/dead cells reduces noise. PBMCs were thawed, stimu-
lated with either no specific antigen or CEF peptide mix (optimal 8–9mers for
CMV, EBV, and influenza) in the presence of protein transport inhibitors, per-
meabilized, stained, and acquired. For all plots, CD3 versus IFN-γ + IL-2 are
shown after gating on CD8+ T cells. For this ICS assay, a positive response is
defined as a response that is at least 2 background and greater than 0.10%
and values are reported after background subtraction. Left column: unstimu-
lated. Right column: CEF stimulated. Top row: The background subtracted CEF
response in the absence of a dead cell exclusion gate is 0.058% or no response.
Bottom row: When the same FCS file as was used in the top row is analyzed
using a dead cell exclusion gate, the background is reduced, and the net
response is 0.109% and thus positive

effects on the relative percentages of different subsets reported

within that gate (Fig. 5).
Conclusion: Scatter gates should be used with caution and
checked by backgating to avoid unintended exclusion of relevant
cells (recovery) and inclusion of irrelevant cells (purity). Backgating
refers to the visualization of a final gated population in all prior
gating steps (generally with those prior gates removed). This allows
one to see the effect of each parent gate in the hierarchy, and can
quickly identify parental gates that create undesirable effects (e.g., a
too-tight scatter gate that cuts through a specific downstream
subpopulation of interest).

2.6 Specific Definition: A specific population gate defines a cell population such
Population Gates as CD4+ T cells, B cells, or monocytes, focusing further subset
identification within those populations. The size, placement, and
 Flow Cytometry Gating Guidelines 93

250 250

200 200

150 150
SSC-H

SSC-H
100 100
NK cells B cells

50 50

0 0
0 50 100 150 200 250 0 50 100 150 200 250
FSC-H FSC-H

Fig. 5 Differential scatter profiles of specific lymphocyte subsets. NK cells (left panel) and B cells (right panel)
were gated separately using specific markers from the green, liberal scatter gate for lymphocytes and then
projected back into the scatter dimensions as red dots. Note that the pink (tight) scatter gate for lymphocytes
would exclude a subset of both NK and B cells because of their biased and/or diffuse scatter profile relative to
most lymphocytes

order of such gates can influence the reported results in ways that
may not be obvious to an inexperienced operator.
How data is affected: Knowledge of the expression properties of
specific subsets of interest can be useful for setting a parental
population gate. For example, CD3 may be used as a parental
gate prior to analysis of T cell subsets such as ɣδC5; T cells. By
examining a dot plot of CD3 vs. SSC, one can identify the ɣδC5; T
cells as CD3high relative to the rest of the CD3+ population
(Fig. 6). Knowing this, care should be taken not to exclude
CD3high cells when enumerating ɣδC5; T cells as a subset thereof.
Another example in T cell gating concerns the simple discrimi-
nation of CD4+ and CD8+ populations. Since these are largely
mutually exclusive populations, it may seem irrelevant whether
one gates on CD3+ CD4+ and CD3+ CD8+ populations, or rather
sequentially gates on CD3+ and then CD4 vs. CD8. However, the
presence of a small but variable population of CD4+ CD8+ T cells
means that these two strategies will not give identical results. In
fact, the former strategy will double-count these CD4+ CD8+
events as belonging to both populations, while the latter method
allows their true discrimination.
Still another example with regard to T cell gating is the down-
modulation of CD3, CD4, and CD8 on activated T cells. When
performing stimulation assays, the gates for these markers need to
be sufficiently “loose” to avoid cutting off potentially small sub-
populations of activated (and thus responding) T cells that have
partially down-modulated these markers.
94 Janet Staats et al.

Specimen_001-Tube_004 Specimen_001-Tube_004

(x 1.000)
150 200 250
105
104
APC-A

SSC-A
P3
103

100
50
102
TCR γδ

102 103 104 105 102 103 104 105

PE-A PE-A
CD3

Fig. 6 ɣǖC5;T cells are CD3high. Human lysed wash blood was stained with CD3 and TCRɣǖC5; antibodies.
Orange population indicates TCRɣǖC5; cells on a plot of CD3 vs TCRɣǖC5; (left) and CD3 bright staining on
the TCRɣǖC5; population depicted in CD3 vs SSc plot (right image)

Rare event gates are another source of potential variability. For

example, gates on cytokine+ cells responding to specific antigen, or
MHC-tetramer+ T cells, require consistent placement relative to
the negative population. If the gate is drawn too close to the
negative population, there will be increased and variable “back-
ground” created, from cells that are on the edge of the negative
population but counted as positive. Conversely, if the gate is drawn
too far away from the negative population, dim positives will be
missed. In this regard, a gating control (e.g., unstimulated cells for
cytokine assays or an FMO for MHC-tetramer staining) may be
useful for setting the boundary of the positive gate.
Conclusion: Knowledge of the staining properties of subsets
within an initial population gate is important for accurate and
consistent downstream enumeration. Backgating is a useful tech-
nique for checking these initial population gates to be sure they are
not unwittingly excluding populations of interest. A gating control
can also be useful, particularly for rare and/or dim populations.

2.7 Data Display Definition: The premise of flow cytometry is that one marker is
tagged with one fluorescent dye and the photons generated by the
2.7.1 Compensation
dye are measured in a single detector. Spectral overlap occurs when
the fluorescence emission from one dye emits photons into the
fluorescence emission spectrum of a neighboring detector. In a
multicolor panel, spectral overlap from multiple dyes may be
observed in a single detector. Electronic subtraction is applied to
correct the interference caused by the spectral overlap in each
channel; this is called compensation. A compensation matrix is
generated to correct the spectral overlap for each dye versus all of
 Flow Cytometry Gating Guidelines 95

Fig. 7 Undercompensation leads to a false positive response in an ICS assay. Undercompensation between PE
(IFNγ + IL-2) and FITC (CD4) created the appearance of CD4+ IFNγ + IL-2+ cells, when in fact with correct
compensation these cells were CD4 . In this case, the operator used software to create a compensation
matrix but neglected to verify the accuracy of the compensation matrix by visual inspection the data (e.g.,
using an N  N plot matrix as described in the text)

the other dyes in the panel, respectively. The goal of compensation

is to specifically detect fluorescence from each dye in its on respec-
tive detector. Compensation does not change the raw data, but it
merely changes how data are displayed. By reducing the apparent
interference from neighboring fluorophores using compensation,
the markers tagged with those fluorophores are displayed with
more separation, enhancing the appearance of distinct populations.
How data is affected: During analysis, where visualizing a clear
separation of populations is essential, an error in compensation may
lead to an erroneous classification of cells, including false positives
or false negatives (Fig. 7). It is therefore important to assess the
accuracy of compensation by examining biaxial plots of all combi-
nations of fluorescence channels, before the application of gating
on fluorescence channels. Typically, an N  N set of plots is
generated, with appropriate transformation (see Subheading
2.7.2), where N is the number of fluorescence channels in the
experiment. Some flow cytometry software allows the automatic
creation of such a plot layout. The user can then judge the accuracy
of compensation by the degree to which populations that should be
orthogonal to each other for particular markers do in fact align
orthogonally. Undercompensation or overcompensation can be
easily identified as an artificial curved edge to such populations.
96 Janet Staats et al.

Conclusion: It is critical to appropriately compensate cytome-

try data prior to analysis. To ensure proper compensation, opera-
tors should manually verify compensation for accuracy, looking for
“hooks” created by overcompensation or undercompensation, as
well as subnegative populations resulting from overcompensation
with a neighboring detector. An easy way to review compensation
is by visualizing a compensation layout containing N  N dot
plots.

2.7.2 Transformation Definition: The range of signals in flow cytometry can span very
negative values (particularly after compensation) to very high values
(104 or greater). Given the approximately log-normal distribution
of many populations in flow data, a linear scale will severely distort
populations and make inefficient use of the display space. However,
a simple logarithmic scale is unable to display zero or negative
values. Therefore, transformations have emerged that preserve the
ability to display negative and zero values while approximating a
logarithmic scale in the positive space. These have variably been
referred to as logicle, biexponential, or VLog scales [15, 35, 36].
Some version of these is available in all modern flow cytometry
software, often with automated application in a data-dependent
fashion.
How data is affected: Choosing an optimal transformation is
important for a number of reasons. First, visualizing the effects of
compensation requires a scale that smoothly transitions through
zero and displays negative data. Undercompensation or overcom-
pensation can then be accurately judged based on how orthogo-
nally the edges of populations align in the negative to low positive
range. Second, the choice of transformation is important to effi-
cient use of the graphical display space. If a majority of the axis
length is devoted to high values (e.g., >102), and if the data
consists entirely of dim events, then the transformation is not
appropriate. Conversely, if a majority of the axis length is devoted
to negative values, and high positive populations are compressed
and difficult to distinguish, the transformation is not appropriate.
Finally, an inappropriate transformation can create artifacts such
as “splitting” of a negative population into what appears to be
two populations (“valley artifact”). This particular artifact is a
result of displaying too many decades of logarithmic scaling at
the low end of the log scale (usually below 10 or 1). Since
judgement of positive and negative expression is often a goal of
flow cytometry, such artifacts can clearly be an impediment to
interpretation.
Conclusion: Choosing the right transformation is critical
for accurate visualization of compensation, for efficient use of
display space, and for complete visualization of negative popula-
tions without introducing the artificial appearance of breaks in
the data.
 Flow Cytometry Gating Guidelines 97

2.8 Analytical An unstained cell sample is often used for calculation of compensa-
Controls tion, relative to single positive samples for each fluorochrome.
However, this is not strictly required for performing compensation,
2.8.1 Unstained Cells
if negative populations are included in each single-color control. An
unstained control is, however, still useful for judging the PMT
voltage gain in each sample, since it represents the position of the
most negative possible population. It is generally not useful for
determining gate placement on stained samples, since it does not
account for either spillover from other fluorescent channels, or
nonspecific binding in the channel of interest.

2.8.2 Isotype Controls Isotype controls are sometimes useful when deciding gate place-
ment in flow cytometry. These controls are made of the same
isotype as the antibody of interest and have the same fluorescent
dye attached to them. In specific instances, such as when monocytes
are being stained with an antibody tagged with tandem dyes such as
PE/Cy7, APC/Cy7, APC/Fire 750, APC-H7, or PE-CF594, the
use of an isotype control antibody is extremely helpful. Monocytes
tend to nonspecifically bind to dyes such as Cy7, Fire 750, and H7.
Figure 8 shows the effect on tandem dye binding to isotype con-
trols. Some manufacturers formulate their antibodies with reagents
to reduce this background and others may sell buffers to overcome
these issues (e.g., Monocyte Blocker from BioLegend).

2.8.3 Fluorescence In the absence of situations like those shown in Fig. 8, the major
Minus One Controls (FMOs) background signals in high-dimensional flow cytometry tend to
come from spillover rather than nonspecific binding. As such, an
FMO control, which leaves out a single antibody conjugate but
maintains all other stains in the panel, is often more useful for
setting gates than an isotype control in the absence of other stains.

2.8.4 Combined Isotype A control which leaves out one antibody conjugate, but replaces it
and FMO Controls with an isotype control antibody, can be considered a “combined”
control. With such a control, one attempts to combine the benefits
of FMO (accounting for spillover) and isotype controls (accounting
for nonspecific binding). While this in theory may be the “best of
both worlds,” it is worth mentioning that there are still drawbacks
to the use of isotype controls that are not alleviated by combining it
with FMO. Specifically, the particular isotype control antibody may
or may not have similar nonspecific binding properties as the test
antibody [37]; and it will likely require titration of the isotype
control to at least approximate those properties.

2.8.5 Internal Controls For functional readouts like cytokines in intracellular cytokine
(e.g., Unstimulated Cells) staining (ICS), the best gating control may be an internal control
(e.g., unstimulated cells from the same donor). Even so, it can be
difficult to decide exactly where to place the cytokine or functional
98 Janet Staats et al.

unstained Isotype PE-Dazzle 594 CD3 PE-Dazzle 594

54 36 41

41 27 31
Count

Count

Count
27 18 21
M3 M4 M3 M4
98.58% 0.00% 95.78% 0.29%
14 9 10

0 0 0
-101 102 103 104 -101 102 103 104 -101 102 103 104
PE-DZL-A PE-DZL-A PE-DZL-A
unstained Isotype APC-Cy7 CD14 APC-Cy7
30 41 79
M4 M3 M4
99.21% 0.79% 26.94%
23 31 59

M1 M1 M1
Count

Count

Count
15 0.00% 21 M2 61.66% 40 M2 82.13%
40.03% 18.20%

8 10 20

0 0 0
-102 103 104 105 -102 103 104 105 -102 103 104 105
APC-Cy7-A APC-Cy7-A APC-Cy7-A

Fig. 8 Tandem dyes nonspecifically stain monocytes and thus isotype should be used to set marker gate
(s) appropriately. Top row: Human lysed washed blood was not stained or stained with isotype control
PE/Dazzle 594 or CD3 PE/Dazzle 594. Histograms are gated on FSC and SSC Singlets followed by monocyte
gate (scatter based). Red marker gates (M1 and M2) indicate gate placement using unstained cells (leftmost
panel) and blue marker gates (M3 and M4) depict frequency of CD3 positive cells based on gate placement
with isotype control (middle panel). If care is not taken interpreting data based on unstained cells, then users
may report monocytes to be CD3+. Bottom row: A similar case using APC/Cy7. Here we show that gates based
on the isotype control are more appropriate for the positive monocyte stain, CD14

marker gate relative to the negative population. Placing the gate

too close to the negative population increases the background of
false positive events; but placing it too far away can reduce the
readout of true positives. This becomes especially critical when
dealing with rare events such as antigen-specific responses. Guide-
lines for this have been published, both for manual gating [24] and
for gate selection based on a statistical measure [38]. For manual
gating, it is generally advised to set the gate above the “halo” of
cells at the edge of the negative population, since increased back-
ground, which may be sample-specific, will be the most detrimental
to detection of low-level responses.
 Flow Cytometry Gating Guidelines 99

2.9 Flow Cytometry It is worth noting that the readout of a gated population can be
Units expressed in different units. Most gating software by default reports
percent-of-parent statistics, that is, the percentage of cells in the
parent population that are contained in the gate of interest. How-
ever, this may not always be the best readout to report.
Consider the case of lymphocytes, which have major subpopu-
lations of B, T, and NK cells. If an individual’s T cell count
increases, the percentage of lymphocytes that are B and NK cells
will necessarily decrease, even if their absolute frequency is
unchanged. In order to avoid the effect of such reciprocal relation-
ships, one can report absolute cell counts (per microliter of blood).
However, this requires a volumetric measurement or use of count-
ing beads. It can be done most accurately by single-platform meth-
ods (where the counting beads or volumetric measurement is done
in the same flow cytometry assay as the cell population staining).
Alternately, a dual platform method can be used, relating the results
of a counting assay to the flow cytometry staining. For example,
total lymphocyte count can be measured by a complete blood count
with differential (CBCD). The total lymphocyte count can then be
multiplied by the percentage of lymphocytes gated as B, T, and NK
cells, to derive absolute counts of these three populations. Because
of variations in the way lymphocytes are defined in the CBCD
versus flow cytometry platform, however, this will necessarily be
less accurate than single platform counting. Still, both methods can
be made relatively reproducible [39].
Short of reporting absolute counts, one can also report popu-
lation percentages but using a higher-level ancestor gate. For exam-
ple, each gated population in a PBMC sample could be reported as
a percent of all live cells (after exclusion of aggregates). While this is
not the same as absolute counting, it does diminish the effect of
changes in reciprocal populations. It is also easily done using the
reported percent-of-parent statistics, by simply multiplying percen-
tages of each ancestor gate up to the top level gate being reported.
Some software programs can provide such reporting automatically
by choosing the desired ancestral gate from which to report
percentages.
Finally, it is possible to report not the frequency or percentage
of cells in a gate, but rather their mean or median fluorescence
intensity (MFI). This is particularly useful when reporting data on
markers that may not have distinct positive and negative popula-
tions, but whose distribution represents a continuum (“smear”) or
for which shifts in expression level are the biologically important
readout. Examples include CD80 and CD86 expression on den-
dritic cells, whose intensity relates to cell maturation/activation.
Rather than measuring the percent of “positive” cells using a
potentially arbitrary cutoff, it makes more sense to report MFI of
the entire population. However, when comparing MFI across
100 Janet Staats et al.

experiments, day-to-day variations in staining and instrument setup

will confound such results. Calibration of MFI to measured equiv-
alent soluble fluorochrome (MESF), using beads of standardized
intensity, is recommended in such cases [40].

2.10 Special Flow cytometry experiments and therefore associated results can be
Considerations affected by artifacts in the assay, which are not considered or known
prior to performance of a particular assay. These artifacts can
include samples with too few events or difficult-to-discern rare
populations; or shifts in a cell population from one sample to
another. Visual inspection of gates on all samples is therefore critical
before finalizing the gating. In determining whether to move a gate
for a specific sample, one should attempt to maintain a fair and
unbiased approach, that is, keeping gates as similar as possible
except where clear shifts in a population warrant changing the
gate. Samples to be directly compared, such as stimulated and
unstimulated conditions from the same donor, need to have the
exact same gating applied.
When gating a large study, or when multiple operators and/or
sites are involved, it becomes necessary to create a gating template.
This is a document with default gates provided, that is the starting
point for analysis of each batch of samples. It represents not only
the hierarchy of all gates to be applied but also a “best guess” of
where those gates should lie. Movement of gates from the template
position should only be undertaken when necessary, and with the
caveats described above in mind. In general, central analysis by a
single operator (or at least verified by a single operator) is more
consistent than separate analysis by different individuals or sites
[27, 41].
Some types of artifacts can be anticipated and possibly avoided,
including many of the factors listed in Table 1. Some variables that
should be controlled to avoid excessive population shifts include:
reagent lot control, consistent washing between wells or tubes, and
avoiding degradation of problematic tandem dyes that can be idio-
syncratic between samples. It is of course always preferable to avoid
artifacts that would complicate gating in the first place, but this may
not always be perfect. An example of a “preventable” population
shift is that caused by brilliant dye interaction, as shown in Fig. 9.
Finally, a piece of idiosyncratic advice regarding gate placement
relative to X and Y axes. Even with appropriate transformation and
data display, there are almost always events that are below the axis
for negative populations. As such, gates on the negative population
need to be placed over the relevant axis to include events that are
below that axis. Depending on the plot size and visual resolution,
this difference may be insidious; two gates may look essentially
identical, but one includes events below the axis and one does
not. Experienced cytometrists therefore typically draw a gate and
then drag it well beyond the axis in order to avoid this issue.
 Flow Cytometry Gating Guidelines 101

Fig. 9 Interaction between fluorophores may affect populations. Human lysed whole blood was stained with
CD4 Brilliant Violet 421™ and CD8 Brilliant Violet 711™ in the presence (left) or absence (right) of Brilliant
Stain Buffer™. Note loss of mutually exclusive expression of CD4 and CD8 in the left plot

3 Discussion/Conclusions

In research and clinical trial settings, the application of gates to flow

cytometry data has been done largely in the absence of wide-
ranging consensus or written guidelines. This is surprising given
the pervasiveness of flow cytometry in immunology research and
immune monitoring settings. The guidelines set forth here were
initially drafted by a small group of experienced flow cytometrists
from academic, industrial, and government labs. They were further
vetted by presentation to the Federation of Clinical Immunology
Societies (FOCIS) 2017 and 2018 annual meetings. The result is an
attempt at a consensus document on the highest-level principles for
gating flow cytometry data in immunological assays.
The guidelines presented here mainly focus on types of gates
that are (or should be) common to most immunological assays.
These include acquisition thresholds (or triggers), time gates, sin-
glet gates, exclusion gates, and scatter gates. Examples of how the
definition of these gates can affect downstream data highlight their
general usefulness. We also discuss specific population gates, which
tend to be more assay-centric, for example, T cell and rare event
gates in the setting of in vitro-stimulated and/or antigen-specific T
cell assays. Finally, we discuss the quality control of gating. This can
include the use of “back-gating” to identify the effect of each gate
in a gating hierarchy, allowing for optimization of the recovery of a
desired population. It can also include the review of compensation
for artifacts that cause downstream gates to report inaccurate
results. Finally, the application of gates to multiple conditions and
donors is discussed, including the need to visually inspect all files,
102 Janet Staats et al.

creating gates that can be applied across samples while still allowing
for donor-specific corrections if needed.
Much of the advice presented here can be broadly summarized
as follows.
1. All relevant parameters in the data file should be used to most
purely and completely define populations of interest. These para-
meters include time, which can give clues to artifacts such as
clogs or bubbles in the data acquisition. They also include the
height, area, and/or width of the light scatter parameters,
which can be used in combination to exclude cell aggregates.
Expression of one or more parameters may be used to exclude
irrelevant events (exclusion gating), in addition to parameters
used to include relevant events. And backgating can be useful
to judge the adequacy of individual gates in a hierarchy, includ-
ing light scatter gates, with regard to their unbiased selection of
the population of interest. Importantly, the fact that a parame-
ter is not used in the gating scheme does not mean that it may
not affect the gated results; so visual inspection of all para-
meters in the data is recommended.
2. Once defined for a single sample, gates should be checked to ensure
fair application across all samples. This may require adjusting
gates between donors, but it should generally not entail
changes in the gate among samples or conditions to be com-
pared within the same donor. Attempts should be made to
avoid gates that slice arbitrarily through an obvious cluster of
cells, including cases where part of a cell cluster lies below the
axis of a histogram or dot plot. In some instances, an otherwise
homogeneous cell population will display a continuous distri-
bution of intensity for a given parameter. It may then be more
sensible to report the mean (or median) intensity of the popu-
lation rather than arbitrarily gating a positive and negative
subset for that parameter and reporting percent positive or
negative.
The eventual goal of these guidelines is not only accuracy in
defining populations of interest but also reproducibility between
experiments, between operators, and between donors and cohorts.
This is particularly important in longitudinal and/or multisite
studies. In these cases, it is best to have all gating verified by the
same personnel, even if initially performed by different personnel.
Further, it is advisable to continually refer to historically gated data,
to ensure reproducible treatment relative to contemporary samples,
for studies that are analyzed over a period of time.
In this document, we have tried to illustrate the most overarch-
ing principles that apply to gating flow cytometry data from immu-
nological assays. There may be exceptions to these principles, but
they are likely to be few. There are also undoubtedly situations
where one can avoid some of these recommendations without
 Flow Cytometry Gating Guidelines 103

peril. But it is also likely that following most or all of the recom-
mendations increases the confidence that artifacts and false conclu-
sions will be avoided. Furthermore, there are certainly additional
and more detailed recommendations that could be added to these
guidelines. But with increasingly detailed recommendations comes
the likelihood that they will not apply as broadly across different
situations and assays. So we have mostly avoided too much specific-
ity in these guidelines, in favor of exposing the high-level concepts
behind the gating strategies discussed. We hope that wide disper-
sion of these concepts leads to an overall higher level of accuracy
and reproducibility of flow cytometry gating.

Acknowledgments

The authors thank the EQAPOL consortium and Jennifer Enzor

for providing data and analysis examples, and the FOCIS Immu-
nophenotyping Course participants for helpful suggestions and
vetting of these guidelines.

References
1. Burel JG, Qian Y, Lindestam Arlehamn C et al 9. Fletez-Brant K, Špidlen J, Brinkman RR et al
(2017) An integrated workflow to assess tech- (2016) flowClean: automated identification
nical and biological variability of cell popula- and removal of fluorescence anomalies in flow
tion frequencies in human peripheral blood by cytometry data. Cytometry A 89:461–471
flow cytometry. J Immunol 198:1748–1758 10. Maecker HT, McCoy JP, Nussenblatt R (2012)
2. Disis ML, dela Rosa C, Goodell V et al (2006) Standardizing immunophenotyping for the
Maximizing the retention of antigen specific Human Immunology Project. Nat Rev Immu-
lymphocyte function after cryopreservation. J nol 12:191–200
Immunol Methods 308:13–18 11. Lamoreaux L, Roederer M, Koup R (2006)
3. Posevitz-Fejfár A, Posevitz V, Gross CC et al Intracellular cytokine optimization and stan-
(2014) Effects of blood transportation on dard operating procedure. Nat Protoc
human peripheral mononuclear cell yield, phe- 1:1507–1516
notype and function: implications for immune 12. Maecker HT, Frey T, Nomura LE et al (2004)
cell biobanking. PLoS One 9:e115920 Selecting fluorochrome conjugates for maxi-
4. Perfetto SP, Chattopadhyay PK, Lamoreaux L mum sensitivity. Cytometry A 62:169–173
et al (2006) Amine reactive dyes: an effective 13. Roederer M (2001) Spectral compensation for
tool to discriminate live and dead cells in poly- flow cytometry: visualization artifacts, limita-
chromatic flow cytometry. J Immunol Meth- tions, and caveats. Cytometry A 45:194–205
ods 313:199–208 14. Chevrier S, Crowell HL, Zanotelli VRT et al
5. Shankey TV, Rabinovitch PS, Bagwell B et al (2018) Compensation of signal spillover in sus-
(1993) Guidelines for implementation of clini- pension and imaging mass cytometry. Cell Syst
cal DNA cytometry. International Society for 6:612–620.e5
Analytical Cytology. Cytometry 14:472–477 15. Finak G, Perez J-M, Weng A et al (2010) Opti-
6. Perfetto SP, Roederer M (2007) Increased mizing transformations for automated, high
immunofluorescence sensitivity using 532 nm throughput analysis of flow cytometry data.
laser excitation. Cytometry A 71:73–79 BMC Bioinformatics 11:546
7. Maecker HT, Trotter J (2006) Flow cytometry 16. Novo D, Wood J (2008) Flow cytometry his-
controls, instrument setup, and the determina- tograms: transformations, resolution, and dis-
tion of positivity. Cytometry A 69:1037–1042 play. Cytometry A 73:685–692
8. Watson JV (1987) Time, a quality-control 17. (1992) Guidelines for the performance of
parameter in flow cytometry. Cytometry CD4+ T-cell determinations in persons with
8:646–649
104 Janet Staats et al.

human immunodeficiency virus infection. 28. Verschoor CP, Lelic A, Bramson JL et al (2015)
MMWR Recomm Rep 41:1–17 An introduction to automated flow cytometry
18. (1994) 1994 revised guidelines for the gating tools and their implementation. Front
performance of CD4+ T-cell determinations Immunol 6:380
in persons with human immunodeficiency 29. Mair F, Hartmann FJ, Mrdjen D et al (2016)
virus (HIV) infections. Centers for Disease The end of gating? An introduction to auto-
Control and Prevention. MMWR Recomm mated analysis of high dimensional cytometry
Rep 43:1–21 data. Eur J Immunol 46:34–43
19. (1997) 1997 revised guidelines for performing 30. Martin JC, Swartzendruber DE (1980) Time:
CD4+ T-cell determinations in persons a new parameter for kinetic measurements in
infected with human immunodeficiency virus flow cytometry. Science 207:199–201
(HIV). Centers for Disease Control and Pre- 31. Hoffman RA (2009) Pulse width for particle
vention. MMWR Recomm Rep 46:1–29 sizing. Curr Protoc Cytom Chapter 1:Unit
20. Mandy FF, Nicholson JKA, McDougal JS et al 1.23
(2003) Guidelines for performing single- 32. Wersto RP, Chrest FJ, Leary JF et al (2001)
platform absolute CD4+ T-cell determinations Doublet discrimination in DNA cell-cycle anal-
with CD45 gating for persons infected with ysis. Cytometry 46:296–306
human immunodeficiency virus. Centers for 33. Furman MI, Barnard MR, Krueger LA et al
Disease Control and Prevention. MMWR (2001) Circulating monocyte-platelet aggre-
Recomm Rep 52:1–13 gates are an early marker of acute myocardial
21. O.W. Health (2007) Laboratory guidelines for infarction. J Am Coll Cardiol 38:1002–1006
enumerating CD4 T lymphocytes in the con- 34. Nomura LE, Walker JM, Maecker HT (2000)
text of HIV/AIDS. World Health Organiza- Optimization of whole blood antigen-specific
tion Regional Office for South-East Asia, New cytokine assays for CD4(+) T cells. Cytometry
Delhi 40:60–68
22. Sutherland DR, Anderson L, Keeney M et al 35. Parks DR, Roederer M, Moore WA (2006) A
(1996) The ISHAGE guidelines for CD34+ new “Logicle” display method avoids deceptive
cell determination by flow cytometry. Interna- effects of logarithmic scaling for low signals
tional Society of Hematotherapy and Graft and compensated data. Cytometry A
Engineering. J Hematother 5:213–226 69:541–551
23. Illingworth A, Marinov I, Sutherland DR et al 36. Bagwell CB, Hill BL, Herbert DJ et al (2016)
(2018) ICCS/ESCCA consensus guidelines to Sometimes simpler is better: VLog, a general
detect GPI-deficient cells in paroxysmal noc- but easy-to-implement log-like transform for
turnal hemoglobinuria (PNH) and related dis- cytometry. Cytometry A 89:1097–1105
orders part 3—data analysis, reporting and case
studies. Cytometry B Clin Cytom 94:49–66 37. Andersen MN, Al-Karradi SNH, Kragstrup TW
et al (2016) Elimination of erroneous results in
24. McNeil LK, Price L, Britten CM et al (2013) A flow cytometry caused by antibody binding to
harmonized approach to intracellular cytokine Fc receptors on human monocytes and macro-
staining gating: results from an international phages. Cytometry A 89:1001–1009
multiconsortia proficiency panel conducted by
the Cancer Immunotherapy Consortium 38. Richards AJ, Staats J, Enzor J et al (2014)
(CIC/CRI). Cytometry A 83:728–738 Setting objective thresholds for rare event
detection in flow cytometry. J Immunol Meth-
25. Britten CM, Janetzki S, Ben-Porat L et al ods 409:54–61
(2009) Harmonization guidelines for
HLA-peptide multimer assays derived from 39. Hultin LE, Chow M, Jamieson BD et al (2010)
results of a large scale international proficiency Comparison of interlaboratory variation in
panel of the Cancer Vaccine Consortium. Can- absolute T-cell counts by single-platform and
cer Immunol Immunother 58:1701–1713 optimized dual-platform methods. Cytometry
B Clin Cytom 78:194–200
26. Nomura L, Maino VC, Maecker HT (2008)
Standardization and optimization of multipa- 40. Vogt RF, Cross GD, Henderson LO et al
rameter intracellular cytokine staining. Cyto- (1989) Model system evaluating fluorescein-
metry A 73:984–991 labeled microbeads as internal standards to cal-
ibrate fluorescence intensity on flow cyt-
27. Finak G, Langweiler M, Jaimes M et al (2016) ometers. Cytometry 10:294–302
Standardizing flow cytometry immunopheno-
typing analysis from the human immunophe- 41. Maecker HT, Rinfret A, D’Souza P et al (2005)
notyping consortium. Sci Rep 6:20686 Standardization of cytokine flow cytometry
assays. BMC Immunol 6:13
 Chapter 6

Strategies and Techniques for NK Cell Phenotyping

Chen Ziqing, Andreas Lundqvist, and Kristina Witt

Abstract
Therapies based on activating the immune system, that is, immunotherapy, are now widely implemented in
clinical praxis in patients with advanced cancer. Although cancer immunotherapy can result in long-lasting
clinical responses, the majority of patients do not respond or develop resistance. Furthermore, cancer
immunotherapy is being increasingly combined with other forms of immunotherapy or conventional cancer
therapies. It is therefore much needed to identify biomarkers that can precisely classify what patients will
benefit from the treatment without any major adverse events and to further develop the efficacy of cancer
immunotherapy. While much attention has been focused on monitoring T cell responses in cancer immu-
notherapy, recent reports have shown that NK cells also play a major role in the response to cancer
immunotherapy. The gold standard for immunoprofiling of NK cells is flow cytometry, but other technol-
ogies have emerged and include mass cytometry, multiplex immunohistochemistry, and single-cell RNA--
sequencing. In this chapter we provide a detailed protocol to profile NK cells using flow cytometry, and a
brief introduction to other techniques.

Key words NK cells, Flow cytometry, Immunoprofiling

1 Introduction

NK cells are innate lymphocytes that can kill virus infected and
malignant cells without any prior sensitization of antigens. They
play a major role in cancer immunosurveillance, and high frequency
and activity of NK cells is correlated with improved prognosis in
several different solid tumors and hematological malignancies. In
recent years, immunotherapy based on activating NK cells has been
explored in several clinical trials including bispecific antibodies or
adoptive transfer of unmodified or genetically modified NK cells
[1, 2]. Cancer immunotherapy and in particular treatment with
immunomodulatory antibodies has now established itself as one of
the pillars in cancer care. While this treatment can result in long-
lasting clinical responses in patients with advanced cancers, many
patients do not respond or develop resistance to the treatment.
Recently, reports have demonstrated that NK cells may play an impor-
tant role in the response to immunomodulatory antibodies

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_6, © Springer Science+Business Media, LLC, part of Springer Nature 2019

105
106 Chen Ziqing et al.

[3]. Thus, it is increasingly important to perform immunoprofiling of

NK cells to identify immunological predictive, pharmacodynamic, and
prognostic biomarkers. The gold standard for a comprehensive analy-
sis of immune cell subsets including NK cells in blood and tissues has
for long been flow cytometry and the growing list of fluorescently
labeled antibodies available for intracellular and cell surface markers
make flow cytometry even more applicable. In parallel, new technol-
ogies have been developed to profile NK cells. In this chapter we
describe sample processing and preparation for NK cell phenotyping
using flow cytometry. We also briefly discuss novel technologies
including mass cytometry, multiplex immunohistochemistry, RNA
sequencing, and spatial transcriptomics that have evolved in recent
years and can be used to perform immunoprofiling of NK cells.

2 Materials

2.1 Flow Cytometry 1. Ice.

and Fluorescence- 2. FACS tubes.
Activated Cell
3. 1 DPBS.
Scanning (FACS)
4. Tabletop centrifuge.
5. FACS buffer.
(a) 1 DPBS.
(b) 2% FBS.
6. FcBlock.
7. Fixation solution (see Note 18).
8. Permeabilization buffer (see Note 19).
9. Fluorochrome-labeled antibodies (see Note 6).
10. DCM (see Notes 12 and 16).

3 Methods

3.1 Flow Cytometry 1. NK cell phenotyping using flow cytometric analysis requires
single cell suspensions of viable cells. Possible starting materials
are whole blood (see Note 1), fresh polymorphonuclear cells
(PBMC) (see Note 2), frozen PBMC (see Note 3), fine needle
aspiration biopsy (see Note 4), and fresh organ or tumor mate-
rial (see Note 5).
2. The panel to study NK cells in the blood or tissue should be
designed according to the study interest and the available flow
cytometer. To simply identify human NK cells in blood or
tissue the following markers are widely used and accepted:
CD45, CD3, CD56, and CD16. Additional makers that are
 NK Cell Phenotyping 107

Fig. 1 Identification of NK cells in human PBMC. Human PBMC are stained and acquired on flow cytometry. NK
cells are identified with the following gating strategy. (1) Gate on lymphocyte population in FSC–SSC plot.
(2) Next gate on linear population in FSC-A–FSC-H plot. (3) Gate on live cells in dead cell marker–FSC plot.
(4) Within the live cells gate on CD19neg cell population in CD19–FSC plot. (5) From the CD19neg population
gate on CD56+CD3 cells in CD56–CD3 plot. In this plot the CD56dim and CD56bright cells appear as two
separate populations

studied on NK cells are described in Note 6. The gating

strategy to identify NK cells in human PBMCs is shown in
Fig. 1.
3. The panel should be designed for the flow cytometer used in
the study. Details are described in Note 7.
4. Samples can be stained directly in FACS tubes. Prepare
FACS tubes according to the number of samples and panels
(see Note 8).
5. Resuspend all samples in DPBS and count all samples.
6. Transfer 2  105 cells into each prepared FACS tubes (see
Note 9).
If nonfixable dead cell marker is used, continue with step 14.
7. Fill up to 1 ml with DPBS and centrifuge tubes at 700  g for
2 min.
8. Discard supernatant by inverting tube and blotting on tissue
(see Note 10).
108 Chen Ziqing et al.

9. Vortex each sample for 3 s to loosen cell pellet (see Note 11).
10. Repeat steps 7 and 8.
11. Vortex each sample for 3 s to loosen cell pellet and resuspend
cells in leftover DPBS.
12. Add fixable dead cell marker to samples according to the
manufacturer’s protocol (see Notes 12 and 13).
13. Incubate according to the manufacturer’s protocol.
14. Add 1 ml of FACS buffer and centrifuge tubes at 700  g for
2 min.
15. Discard supernatant by inverting tube and blotting on tissue
(see Note 10).
16. Vortex each sample for 3 s to loosen cell pellet (see Note 11).
17. Repeat steps 14–16 two times.
18. Resuspend cells to single cell suspension in leftover FACS
buffer.
19. Add FcBlock to each sample at a concentration of 2.5 μg for
1  106 cells.
20. Prepare master mix containing surface staining antibodies.
Mixes should be filled up with FACS buffer to a volume of
10–20 μl added to each sample (see Note 14).
21. Incubate samples for 20 min at 4  C in the dark (see Note 15).
22. Repeat steps 14–16 two times.
For intracellular or intranuclear staining continue with
step 26.
23. Vortex each sample for 3 s to loosen cell pellet and resuspend
cells in 200 μl DPBS.
24. If no fixable dead cell marker was used, add nonfixable dead cell
marker to all cells and incubate for 5 min at 4  C in the dark (see
Note 16).
25. Keep samples on 4  C until acquisition on flow cytometer (see
Note 17).
26. Resuspend cells in fixation solution (see Note 18).
27. Incubate samples for 30 min at 4  C in fixation solution or
according to the manufacturer’s protocol.
28. Prepare permeabilization buffer (see Note 19).
29. Add 1 ml of permeabilization buffer (Perm buffer) to each
sample centrifuge tubes at 700  g for 2 min.
30. Discard supernatant by inverting tube and blotting on tissue
(see Note 10).
31. Vortex each sample for 3 s to loosen cell pellet (see Note 11).
32. Repeat steps 29–31.
 NK Cell Phenotyping 109

33. Repeat steps 29 and 30.

34. Vortex each sample for 3 s to loosen cell pellet and resuspend
cells to single cell suspension in leftover Perm buffer.
35. Prepare master mix containing surface staining antibodies.
Mixes should be filled up with Perm buffer to a volume of
10–20 μl added to each sample (see Note 14).
36. Incubate samples for 40–60 min on ice (see Note 20).
37. Add 1 ml of permeabilization buffer (Perm buffer) to each
sample centrifuge tubes at 700  g for 2 min.
38. Discard supernatant by inverting tube and blotting on tissue
(see Note 10).
39. Vortex each sample for 3 s to loosen cell pellet (see Note 11).
40. Repeat steps 37–39.
41. Repeat steps 37 and 38.
42. Vortex each sample for 3 s to loosen cell pellet and resuspend
samples in 200 μl DPBS.
43. Keep samples on 4  C until acquisition on flow cytometer (see
Note 21).

3.2 Mass Cytometry Mass cytometry, also called cytometry by time of flight (CyTOF) is
a novel platform to perform even deeper analysis of immune cell
subsets compared with flow cytometry [4]. Immunophenotyping
by mass spectrometry provides the ability to measure >40 proteins
at a rate of 1000 cells/s, compared with 18 proteins per cell, at
>10,000 cells/s for fluorescence-based flow cytometry. Theoreti-
cally, mass cytometry is capable to measure up to 100 different
stable isotope tags, but the high-purity requirement of these tags
limits the usage down to around 40 rare earth metal tags [5]. The
sample preparation is similar to flow cytometry. In contrast to the
acquisition on a flow cytometer, the single cells are analyzed in a
mass spectrometer. The antibodies are not coupled to fluoro-
chromes but instead to heavy metals. Each heavy metal conjugated
to an antibody binding an NK cell will give a unique peak at an exact
molecular weight. The advantage is that there is no overlap
between different heavy metals; thus, no compensation needs to
be performed. More than 100 heavy metals are available for mass
cytometry, giving the opportunity to analyze in depth NK cell
phenotypes and subsets. Due to the high degree of customization,
the library of preconjugated commercially available antibodies is
limited. This requires self-conjugation of antibodies but at the same
time allows for the design of large individualized and specific
panels. Details on the design and optimization of an NK cell-
specific mass spectrometry panel have been described by Kay
et al. [6].
110 Chen Ziqing et al.

3.3 Multiplex Flow cytometry and CyTOF can identify the presence of rare
Immunohisto- immune cell subsets in single-cell suspensions. While these tech-
chemistry nologies are powerful, no information regarding spatial distribu-
tion can be obtained. The localization of immune cells and their
local environment indeed differ between different areas in the
tumor (stromal, intratumoral, and tumor margin). The under-
standing of cell localization can be of great prognostic value
[7]. Multiplex immunohistochemistry has the ability to provide
this missing information. Sequential staining of a single section
allows for the understanding of spatial cell type distribution within
a tumor and to draw conclusions on cell-specific protein expression.
The sequential stainings are individually acquired on a fluorescent
microscope and are overlapped upon image analysis. Since the
majority of antibodies are produced in mice, rats, and rabbits,
antibody stripping is required after individual rounds of staining.
Stripping of antibodies always have the risk to remove or damage
antigens. The order in which the antigens are stained requires
therefore a lot of optimization. The technology has further been
developed to allow for standardization and clinical application. For
example, the whole-slide eight-plex mIHC platform is set up to
analyze six antigens and cell nuclei in parallel in whole FFPE sec-
tions [8]. The staining procedure is standardized, and while the
secondary antibodies are not changeable, the primary antibodies
can be freely selected.
There are certainly developments to perform deeper analysis
using multiplex immunohistochemistry. Imaging mass cytometry is
one such development, combining mass cytometry with tissue
localization [9]. In this application, fresh-frozen paraffin-embed-
ded tissue or frozen sections are stained with mass spectrometry
antibodies following standard immunofluorescence protocols.
With a precisely directed laser beam, tissue pieces of 1 μm2 are
removed one after the other from the section and acquired in the
mass spectrometer. This allows to analyze the protein expression on
the tissue section to a resolution of 1 μm2. Another development is
the CO-Detection by indEXing (CODEX) technology to enable
multiplexing of antibody-tagged target epitopes. In CODEX, anti-
body binding events are rendered iteratively using DNA barcodes,
fluorescent dNTP analogs, and an in situ polymerization-based
indexing procedure [10].

3.4 Genome-Wide Omics technologies are typically considered to be based on geno-

Analysis mics and presents a panoramic view of the unbiased molecular
determinants of NK cells. To analyze NK cells using different levels
of omics approaches, different technologies can be applied, includ-
ing microarrays, RNA-sequencing, single-cell RNA sequencing,
and microRNA sequencing. In microarrays, thousands of biologic
reactions at DNA, RNA, or protein levels can be measured or even
quantified in a single experiment. With decreasing costs, RNA-seq
 NK Cell Phenotyping 111

is increasingly used for transcriptome analysis. Unlike microarray

technology which relies on fluorescent labeling, RNA-seq trans-
forms RNA into a cDNA library, which is followed by sequencing.
RNA-seq is applied to analyze the differential elements of gene
expression of the whole transcriptome in a more accurate, repro-
ducible, wider, and more reliable manner than that of other meth-
ods. The data obtained from microarray and RNA-seq technologies
represent average values of cell populations. With advances in sepa-
ration of single cells and establishment of cDNA libraries, single cell
RNA-seq technology has emerged into a technology that facilitates
the analysis of molecular profiles of a single cell from cell
populations.

4 Notes

1. Whole blood without further processing can be used for KN

cell phenotyping. Staining antibodies are directly added to
whole blood. After incubation, all samples are washed with
FACS buffer followed by red blood cell lysis. The samples are
then washed again and acquired using a flow cytometer. Start-
ing with whole blood is in particular of interest when working
with patient material. The processing is minimal and reduces
the risk of losing or damaging surface proteins. Especially when
investigating chemokines, it has been observed that expression
of some chemokine receptors (i.e., CXCR2) is lost after Ficoll
density centrifugation.
2. Prior to NK cell phenotyping in fresh blood samples, PBMC
should be isolated using Ficoll. PBMC contain between 5% and
15% NK cells in healthy donors. Having fresh PBMC as a
starting material allows not only to perform phenotyping on
the material but also allows for functional studies. However, it
has been shown that the isolation of PBMC can alter surface
proteins, in particular chemokine receptor [11].
3. Frozen PBMC can be used for NK cell phenotyping as well.
Vital freezing of PBMC is essential in order to use frozen
PBMC as starting material. It is also recommended to test
beforehand, if the staining gives a similar result on frozen
PBMC as observed on fresh PBMC. The freeze–thaw process
has been shown to alter and damage surface proteins and
should be avoided if possible [12].
4. Single cells can be obtained from fine needle aspiration biopsies
using trypsin. Following sample digestion, cells should be
passed through a 70 μm cell strainer to remove tissue clumps.
If free DNA is present in the sample, an additional treatment
with DNase is recommended.
112 Chen Ziqing et al.

5. Fresh tumor or organ material requires mechanic dissociation.

Samples should be first cut into small pieces with scalpels in
PBS, supplemented with 5% FBS, 200 U/ml penicillin, and
200 U/ml streptomycin. Tissue pieces can then be passed
through a metal sieve with 80 μm mesh size. Enzymatic diges-
tion with collagenase and dispase can be additionally per-
formed at this moment. If free DNA is present, DNase can be
added to the enzyme mix. In the final step samples should be
passed through a 40 μm cell strainer to remove remaining
tissue clumps.
6. NK cells are defined as CD45+, CD3 , CD56+ cells. NK cells
are further divided into CD56bright CD16low cytokine-produc-
ing NK cells and CD56dim CD16+ cytotoxic NK cells. The NK
cell activation status can be evaluated using the following mar-
kers resembling protein associated to activation or inhibition of
NK cells: DNAM-1, NKG2D, NKp30, NKp44, NKp46,
NKG2A. KIR receptors (inhibitory: KIR3DL2, KIR3DL1,
KIR2DL1, KIR2DL2/3; activating: KIR 3DS1,
KIR2DS1–5) binding different MHC class I molecules are
essential for the activity and function of NK cells but are not
often analyzed in NK cell phenotyping. The following costi-
mulatory proteins are expressed of NK cells and can be of
interest in the NK cell phenotyping: CD137, OX40, and
CD27). A number of immune checkpoint proteins (PD-1,
TIGIT, TIM-3, and VISTA) can be expressed on activated
NK cells and can be included in NK cell phenotyping panel.
7. The intensity of different fluorochromes varies, and
low-expression proteins should be detected with antibodies
conjugated to brighter fluorochromes. Further, spectral over-
laps and machine settings should be considered. Every flow
cytometer is unique despite having the same lasers and detec-
tors. Proteins for which gradual expression or changes are to be
evaluated should be detected with fluorochromes that have low
spectral overlap with other fluorochromes to reduce compen-
sation errors.
8. Prepare one tube for each staining panel and one unstained
tube for each sample.
9. If intracellular staining is included, it is recommended to start
with a higher cell number (300,000–500,000 cells). Fixation
and permeabilization can decrease cell number due to an
increased number of washing steps.
10. After centrifugation a small cell pellet should be visible at the
bottom of the tube. After discarding the supernatant blot tubes
directly on tissue without turning the tube back up. Cell pellet
loosens and resuspends in remaining liquid as soon as tubes are
turned back up.
 NK Cell Phenotyping 113

11. Alternatively, cells can also be resuspended using a hand

pipette. Vortexing between washes saves time.
12. Fixable dead cell markers are amine reactive dyes. They are
membrane impermeable and react with amine groups of any
protein. It is therefore important to stain cells with these dyes
in the absence of free protein to ensure optimal staining of dead
cells.
13. Add dead cell marker to unstained tube if you intend to analyze
protein expression on live NK cells.
14. Preparation of antibody master mixes for each panel is recom-
mended to ensure comparability between samples and to
reduce pipetting related errors, especially when working with
volumes in the range of 0.2–1 μl. Preparation of master mixes
further reduces the total handling time of the samples. Always
prepare the master mix for n + 1 samples.
15. Samples are best stored on ice in a closed container. If samples
are stored in the fridge, samples should be covered with tin foil
to prevent light exposure.
16. Commonly used nonfixable dead cell markers are 7-AAD and
propidium iodide (PI).
17. Nonfixed sample should be acquired within 1 h.
18. Fixation method for your samples and application should be
optimized. It is common to use PFA solutions for cell fixation
or commercially available kits for intracellular and intranuclear
staining. Resuspend samples to homogeneous single cell sus-
pension. Doublets cannot be separated any longer after
fixation.
19. Methanol and acetone are common detergents to permeabilize
fixed cells. The choice of detergent is dependent on the appli-
cation (intracellular vs intranuclear staining) and protein stabil-
ity. Commercially available kits include permeabilization
buffers and are available for both intracellular and intranuclear
staining, including the staining for phosphorylated proteins.
20. Intracellular staining requires longer incubation time than cell
surface staining.
21. Fixed samples can be stored for a few hours at 4  C prior to
acquisition. Stability of fluorochromes, especially conjugated
fluorochromes, is the time-limiting factor for storage of fixated
samples.
114 Chen Ziqing et al.

References
1. Davis ZB, Vallera DA, Miller JS, Felices M margin of colorectal cancer liver metastases
(2017) Natural killer cells unleashed: check- associated with improved survival. Oncoimmu-
point receptor blockade and BiKE/TriKE uti- nology 6(3):e1286436. https://doi.org/10.
lization in NK-mediated anti-tumor 1080/2162402X.2017.1286436
immunotherapy. Semin Immunol 31:64–75. 8. Blom S, Paavolainen L, Bychkov D, Turkki R,
https://doi.org/10.1016/j.smim.2017.07. Maki-Teeri P, Hemmes A, Valimaki K,
011 Lundin J, Kallioniemi O, Pellinen T (2017)
2. Fang F, Xiao W, Tian Z (2017) NK cell-based Systems pathology by multiplexed immunohis-
immunotherapy for cancer. Semin Immunol tochemistry and whole-slide digital image anal-
31:37–54. https://doi.org/10.1016/j.smim. ysis. Sci Rep 7(1):15580. https://doi.org/10.
2017.07.009 1038/s41598-017-15798-4
3. Hsu J, Hodgins JJ, Marathe M, Nicolai CJ, 9. Chang Q, Ornatsky O, Hedley D (2017) Stain-
Bourgeois-Daigneault MC, Trevino TN, ing of frozen and formalin-fixed, paraffin-
Azimi CS, Scheer AK, Randolph HE, Thomp- embedded tissues with metal-labeled antibo-
son TW, Zhang L, Iannello A, Mathur N, Jar- dies for imaging mass cytometry analysis. Curr
dine KE, Kirn GA, Bell JC, McBurney MW, Protoc Cytom 82:12.47.11–12.47.18.
Raulet DH, Ardolino M (2018) Contribution https://doi.org/10.1002/cpcy.29
of NK cells to immunotherapy mediated by 10. Goltsev Y, Samusik N, Kennedy-Darling J,
PD-1/PD-L1 blockade. J Clin Invest 128 Bhate S, Hale M, Vazquez G, Black S, Nolan
(10):4654–4668. https://doi.org/10.1172/ GP (2018) Deep profiling of mouse splenic
JCI99317 architecture with CODEX multiplexed imag-
4. Ornatsky O, Bandura D, Baranov V, Nitz M, ing. Cell 174(4):968–981 e915. https://doi.
Winnik MA, Tanner S (2010) Highly multi- org/10.1016/j.cell.2018.07.010
parametric analysis by mass cytometry. J Immu- 11. Naranbhai V, Bartman P, Ndlovu D,
nol Methods 361(1–2):1–20. https://doi. Ramkalawon P, Ndung’u T, Wilson D,
org/10.1016/j.jim.2010.07.002 Altfeld M, Carr WH (2011) Impact of blood
5. Atkuri KR, Stevens JC, Neubert H (2015) processing variations on natural killer cell fre-
Mass cytometry: a highly multiplexed single- quency, activation, chemokine receptor expres-
cell technology for advancing drug develop- sion and function. J Immunol Methods 366
ment. Drug Metab Dispos 43(2):227–233. (1–2):28–35. https://doi.org/10.1016/j.jim.
https://doi.org/10.1124/dmd.114.060798 2011.01.001
6. Kay AW, Strauss-Albee DM, Blish CA (2016) 12. Florek M, Schneidawind D, Pierini A, Baker J,
Application of mass cytometry (CyTOF) for Armstrong R, Pan Y, Leveson-Gower D,
functional and phenotypic analysis of natural Negrin R, Meyer E (2015) Freeze and thaw
killer cells. Methods Mol Biol 1441:13–26. of CD4+CD25+Foxp3+ regulatory T cells
https://doi.org/10.1007/978-1-4939-3684- results in loss of CD62L expression and a
7_2 reduced capacity to protect against graft-
7. Berthel A, Zoernig I, Valous NA, Kahlert C, versus-host disease. PLoS One 10(12):
Klupp F, Ulrich A, Weitz J, Jaeger D, Halama e0145763. https://doi.org/10.1371/journal.
N (2017) Detailed resolution analysis reveals pone.0145763
spatial T cell heterogeneity in the invasive
 Chapter 7

Immunophenotyping of Human B Lymphocytes in Blood

and in Adipose Tissue
Alain Diaz, Maria Romero, Daniela Frasca, and Bonnie B. Blomberg

Abstract
The human obese subcutaneous adipose tissue (SAT) contributes to systemic and B cell intrinsic inflamma-
tion, reduced B cell responses, and increased secretion of autoimmune antibodies. Immune cells are
recruited to the SAT by chemokines released by both adipocytes and infiltrating immune cells. We describe
here the characterization of B lymphocytes from the SAT and blood (control) of obese females undergoing
weight reduction surgeries (breast reduction or panniculectomy). We show how to isolate the immune cells
from the blood and SAT, how to characterize B cells and their subsets, and how to measure markers of
activation and/or transcription factors in SAT-derived B cells and B cell subsets. We also show how to
evaluate other immune cell types infiltrating the SAT, including T cells, NK cells, monocyte/macrophages,
in order to measure relative proportions of these cell types as compared to the blood.

Key words Adipose tissue (AT), Body mass index (BMI), Cardiovascular (CV), Dulbecco’s modified
Eagle’s medium (DMEM), Fetal calf serum (FCS), Free fatty acids (FFAs), Germinal center (GC),
Hanks’ balanced salt solution (HBSS), Insulin resistance (IR), Insulin sensitivity (IS), Peripheral blood
mononuclear cells (PBMC), Red blood cells (RBC), Reactive oxygen species (ROS), Room tempera-
ture (RT), Subcutaneous adipose tissue (SAT), Stromal vascular fraction (SVF), Type-2 diabetes
(T2D), Toll-like receptor (TLR)

1 Introduction

Obesity is an inflammatory condition associated with chronic acti-

vation of cells of the innate immune system and consequent local
and systemic inflammation, responsible for several chronic patho-
logic conditions including cardiovascular (CV) disease [1], type-
2 diabetes (T2D) [2–4], cancer [5], psoriasis [6], atherosclerosis
[7], and inflammatory bowel disease [8]. Increased body mass
index (BMI) has also been significantly associated with insulin
resistance (IR), which indicates the lack of appropriate response
to circulating insulin in several tissues, including pancreas, liver,
muscle, and adipose tissue (AT). Recent studies have shown
increased prevalence of obesity over the past 25 years in 68 million
individuals living in 195 countries [9], this global obesity pandemic

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_7, © Springer Science+Business Media, LLC, part of Springer Nature 2019

115
116 Alain Diaz et al.

affecting all age groups. The decreased proportion of obese patients

responding to vaccination [10–12] or achieving remission in
response to therapy [13–15] is an emerging world health problem.
The combined effect of impaired immunity and altered responses to
interventions may further affect the outcome of infection, with
obvious health and monetary consequences at the individual and
population level.
Inflammaging, the condition of increased chronic inflamma-
tion in the elderly, is an important link between obesity, IR, aging,
and age-associated diseases. Chronic low-grade (sterile) inflamma-
tion causes IR and consequent transition from metabolically nor-
mal obesity to metabolic syndrome. This occurs through both
systemic inflammation and metaflammation [16], a process
whereby excess nutrients promote chronic low-grade inflamma-
tion, whose metabolic hallmarks are high levels of lipids, free fatty
acids (FFAs), glucose, and reactive oxygen species (ROS).
The increase in AT size is characteristic of obesity. It is now well
known that the AT is not only a storage for excess nutrients but it is
an active endocrine tissue. Conversion of the AT from an insulin
sensitive (IS) to a IR state during obesity involves expansion of
adipocyte volume and remodeling of extracellular matrix compo-
nents (collagens, elastins, and the associated blood vasculature)
which parallels increases in the secretion of adipokines, proinflam-
matory cytokines and chemokines, which are involved in the
recruitment of immune cells to the AT. Failure to undergo appro-
priate remodeling in response to overnutrition is detrimental to
body metabolic homeostasis, as excess nutrients promote meta-
flammation [16, 17].
There is evidence that altered innate and adaptive immune
responses occur in the “calorie-stressed” AT [18]. Our results
recently published in humans [19] have shown that immune cells
are recruited to the obese subcutaneous AT (SAT) and increased
RNA expression of chemokines by the adipocytes (responsible for
the recruitment of immune cells to the AT) as well as RNA expres-
sion of the corresponding chemokine receptors by the immune cells
[19]. Moreover, we have shown that the SAT generates autoim-
mune antibodies. We have identified several mechanisms responsi-
ble for the release of “self” antigens in the human obese SAT,
induction of class switch, and production of autoimmune antibo-
dies. We have shown that reduced oxygen availability in the obese
SAT (leading to hypoxia and reduced mitochondrial respiration),
NK cell cytotoxicity, and DNA damage induce cell death and lead
to further release of proinflammatory cytokines, “self” protein
antigens, cell-free DNA, and lipids. All these stimulate class switch
and the production of autoimmune IgG antibodies which have
been described to be pathogenic [20]. Autoimmune antibody pro-
duction has been shown to be dependent on the expression of the
 Immunophenotyping of Human B Cells 117

transcription factor T-bet by several groups [21–23], and we have

also found T-bet expression in the human obese SAT [19].
This paper describes the characterization of B lymphocytes
from the SAT and blood (control) from obese individuals under-
going weight reduction surgeries at the University of Miami Miller
School of Medicine. We show here how to isolate the immune cells
from blood and SAT, how to characterize B cells and their subsets
and how to measure markers of activation/transcription factors in
AT-derived B cells and B cell subsets. We also show how to evaluate
other immune cell types infiltrating SAT, including T cells, NK
cells, monocyte/macrophages, in order to measure relative propor-
tions of these cell types as compared to blood.

2 Materials

2.1 Isolation 1. Fresh SAT specimen (100–200 g of tissue) and blood as control
of Lymphocytes are from females undergoing breast reduction surgery or pan-
niculectomy surgery (removal of lower abdominal fat) at the
Division of Plastic and Reconstructive Surgery at the Univer-
sity of Miami Hospital.
2. Supplemented HBSS: 1 Hanks’ balanced salt solution
supplemented with 200 nM adenosine and 1
penicillin–streptomycin.
3. Supplemented DMEM (s-DMEM): 1 High Glucose Dulbec-
co’s modified Eagle’s medium supplemented with 15 mM
HEPES, 1 mM sodium pyruvate, 1 penicillin–streptomycin,
1% BSA, and 200 nM adenosine.
4. Complete RPMI: RPMI 1640, supplemented with 10% FCS,
10 μg/mL gentamicin, 2  10 5 M 2-mercaptoethanol, and
2 mM L-glutamine.
5. Collagenase Type V (Sigma).
6. Sharp scissors and blender (for mincing adipose tissue).
7. Forceps.
8. Siliconized glass beakers.
9. Stainless steel 316 syringe needle, blunt tip point (gauge 18, L
6 in).
10. 300 μm Mesh (Spectra/Mesh Nylon, 30  30 cm square).
11. 50 mL polypropylene centrifuge tubes (BD Falcon).
12. 20 cc syringe.
13. Sigmacote (Sigma).

2.2 Staining of Blood 1. BD FACS™ lysing solution (BD).

and SVF
118 Alain Diaz et al.

2. ACK lysing solution (Dissolve 8.29 g NH4Cl, 1 g KHCO3, and

37.2 mg Na2EDTA in 1 L deionized H2O).
3. Staining buffer (Dissolve 9.8 g Hanks’ balanced salts, 0.35 g
NaHCO3, 1 g BSA, 0.2 g sodium azide in 1 L deionized H2O).
4. BD Cytofix/Cytoperm™ Fixation/Permeabilization Solution
Kit containing Fixation and Permeabilization Solution and BD
Perm/Wash™ buffer.
5. Fluorochrome-conjugated antibodies for staining of major
lymphocyte subsets:
(a) LIVE/DEAD™ Fixable Aqua Dead Cell Stain Kit
(ThermoFisher).
(b) Pacific Blue™ anti-CD45 (clone 2D1, Biolegend).
(c) APC anti-CD19 (clone HIB19, BD).
(d) FITC anti-CD3 (clone HIT3a, BD).
(e) PerCP anti-CD4 (clone L200, BD).
(f) APC-H7 anti-CD8 (clone SK1, BD).
(g) APC anti-TCR α/β (clone IP26, BD).
(h) PE/Cy7 anti-TCR γ/δ (clone B1, Biolegend).
(i) APC/Cy7 anti-CD16 (clone 3G8, Biolegend).
(j) PE anti-CD56 (clone B159, BD).
(k) APC anti-CD14 (clone M5E2, BD).
6. Fluorochrome-conjugated antibodies for staining of γ/δ T cell
subsets:
(a) FITC anti-TCR Vδ1 (clone TS8.2, InVitrogen).
(b) PE anti-TCR Vδ2 (clone B6, Biolegend).
(c) Alexa Fluor® 700 anti-CD3 (clone OKT3, Biolegend).
7. Fluorochrome-conjugated antibodies for staining of B cell sub-
sets and germinal center (GC) B cells:
(a) APC anti-CD19, PE anti-CD27 (clone M-T271, BD).
(b) FITC anti-IgD (clone IA6–2, BD).
(c) Brilliant Violet 711™-anti-CD10 (clone HI10a,
Biolegend).
(d) PE/Cy7 anti-human/mouse Bcl-6 (clone 7D1,
Biolegend).
8. Fluorochrome-conjugated antibodies for staining of GC T
cells:
(a) Anti-CD3.
(b) Anti-CD4.
(c) APC anti-CD279 (PD-1) (clone EH12.2H7, Biolegend).
(d) APC/Cy7 anti-CD185 (CXCR5) (clone J252D4,
Biolegend).
 Immunophenotyping of Human B Cells 119

9. Additional fluorochrome-conjugated antibodies for single

color control staining: Brilliant Violet 510™ anti-human/
mouse CD45R/B220 (clone RA3-6B2, Biolegend).

2.3 Detection 1. Invitrogen™ PrimeFlow™ RNA Assay Kit (ThermoFisher).

of Single Cell mRNA 2. Type 1 target probe for detection of T-bet mRNA
(Affymetrix).
3. TLR7 agonist CL097 (InVivogen).

3 Methods

3.1 Isolation of SAT- 1. Weigh the AT. The sample should be collected in HBSS or
Derived Blood and SVF similar buffer.
2. Remove connective tissue with scissors. Proceed to cut the
tissue in smaller pieces and place them in 50 mL tubes until
approximately half-filled. Remove connective tissue with
scissors.
3. Wash tissue three times by filling the tubes with supplemented
HBSS and carefully removing the buffer below the floating
tissue by aspiration with syringe and blunt needle. Use at least
2 new 50 mL tubes to collect the first wash from this step that
contains most of the blood surrounding the tissue. This will be
used for blood cell staining subsequently.
4. After the last wash, add one and a half volumes of supplemen-
ted DMEM and transfer to a food blender.
5. Blend the washed tissue in the blender, using very short strokes
to avoid damaging the cells.
6. Transfer tissue to new 50 mL tubes and centrifuge for 3 min at
1200  g.
7. Weigh 1 mg/g of tissue of collagenase type V and dissolve in
s-DMEM at 1 mg/mL.
8. After centrifugation (step 6), remove the precipitated debris
and most of the liquid with the syringe and transfer the tissue
to a siliconized beaker containing about an equal volume of
s-DMEM containing 1 mg/mL of collagenase.
9. Incubate for 1 h in a 37  C water bath, gently mixing the
reaction mix every 10 min to help the digestion of the tissue.
10. After incubation, filter the reaction mix through the 300 μm
Mesh into 50 mL tubes to eliminate undigested larger tissue
pieces. This step will require several pieces of mesh since they
get clogged rapidly, as well as squeezing the mesh by gloved
hand in order to recover most of the digested sample.
120 Alain Diaz et al.

11. Centrifuge the tubes containing the filtrate for 5 min at

300  g. Use the syringe and blunt needle to collect the pellet,
which comprises the SVF, into a new 50 mL tube.
12. Wash the SVF with s-DMEM by centrifugation for 5 min at
300  g. Discard the supernatant.
13. Lyse the RBC in the SVF by resuspending the cell pellet in
3 mL of the ACK lysing solution, incubating for 5 min at room
temperature (RT).
14. Add s-DMEM up to at least ten times the volume of ACK, and
centrifuge for 5 min at 300  g.
15. Discard the supernatant and resuspend the pellet in 10 mL of
s-DMEM. Filter through a 40 μm cell strainer to remove any
remaining clumps.
16. Proceed to do a viability count of the SVF with Trypan Blue.
17. The protocol of isolation of adipocytes and SVF from the
human SAT is adapted from [24].

3.2 Staining 1. Distribute 500,000 to 106 SVF cells in five 5 mL FACS tubes
of the SVF for Flow for antibody staining. Add 1 mL of FACS buffer and centrifuge
Cytometry Analyses for 5 min at 500  g.
2. Discard the supernatant and briefly vortex to resuspend the
pellet.
3. Add the adequate amount of the following combination of
antibodies for membrane staining in a volume of at least
50 μL of staining buffer:
Tube 1 (B cell phenotyping and GC B cells): LIVE/
DEAD™ Fixable Aqua Dead Cell Stain, Pacific Blue-CD45,
FITC-IgD, PE-CD27, APC-CD19 and Brilliant Violet
(BV) 711-CD10.
Tube 2 (T/NK-T cells phenotyping): LIVE/DEAD™ Fix-
able Aqua Dead Cell Stain, Pacific Blue-CD45, FITC-CD3,
PE-CD56, PE/Cy7-TCRγ/δ, PerCP-CD4, APC-TCRα/β,
and APC/H7-CD8.
Tube 3 (Tγδ subsets): LIVE/DEAD™ Fixable Aqua Dead
Cell Stain, Pacific Blue-CD45, Alexa Fluor 700-CD3, FITC-
TCRγδ1, PE-TCRγδ2, PerCP-CD4, APC-TCRαβ, and
APC/H7-CD8.
Tube 4 (GC T cells): LIVE/DEAD™ Fixable Aqua Dead
Cell Stain, Pacific Blue-CD45, Alexa Fluor 700-CD3, PerCP-
CD4, APC-PD1, APC/Cy7-CXCR5.
Tube 5 (Monocytes/NK cells): LIVE/DEAD™ Fixable
Aqua Dead Cell Stain, Pacific Blue-CD45, FITC-CD3,
PE-CD56, PE/Cy7-CD19, APC-CD14, APC/Cy7-CD16.
 Immunophenotyping of Human B Cells 121

4. Incubate for 20 min at RT in the dark.

5. After incubation, add 2 mL of staining buffer and centrifuge for
5 min at 500  g. Discard the supernatant and repeat the
centrifugation step.
6. Discard the supernatant and add 300 μL of BD fixation buffer
to all tubes except tubes 1 and 4. Briefly vortex to resuspend
cells.
7. Tubes 1 and 4 require additional steps for intracellular staining.
Add 250 μL of BD Fixation and Permeabilization Solution and
incubate for 20 min at 4  C in the dark.
8. Add 2 mL of BD Perm/Wash 1 buffer to the tubes and
centrifuge for 5 min at 500  g. Discard the supernatant and
vortex briefly to detach the cell pellet.
9. Add adequate amounts of PE/Cy7 anti-Bcl-6 antibody for
intracellular staining in a volume of 100 μL of BD Wash/
Perm buffer to each tube. Incubate for 20 min at RT in
the dark.
10. After incubation, repeat steps 5 and 6 for these tubes.

3.3 Staining of SAT- 1. Use the blood collected in Subheading 3.1, step 3 (“Isolation
Derived Blood for Flow of the blood and the SVF from the AT”). Centrifuge the 50 mL
Cytometry Analyses tubes containing the first wash of the AT sample for 5 min at
500  g. Carefully remove the supernatant by slowly inverting
the tube without disturbing the more viscous bloody fluid and
pellet that remain at the bottom of the tube. A volume of
approximately 500 μL of blood fluid and pellet should remain
in the tube.
2. Vortex briefly to resuspend the pellet and transfer 100 μL of
blood to four 5 mL FACS tubes for antibody staining.
3. Add the same volume and combination of antibodies for mem-
brane staining used in step 3 in Subheading 3.2.
4. Incubate for 20 min at RT in the dark.
5. Add 2 mL of BD FACS™ lysing solution to each tube, vortex
to mix and incubate for 15 min at RT in the dark.
6. Centrifuge the tubes for 5 min at 500  g. Discard the super-
natant and briefly vortex.
7. Add 2 mL of staining buffer to each tube and centrifuge for
5 min at 500  g.
8. Perform steps 6–10 from Subheading 3.2 for intracellular
staining of blood.
122 Alain Diaz et al.

SSC

CD45

CD19

CD27
20±2 6±2
8±1

BLOOD

17±1 57±5
SSC FSC Live/Dead FSC IgD

CD19
CD45

CD27
31±3 3±2
12±2

SVF

49±4 17±4
FSC Live/Dead FSC IgD

Fig. 1 Gating strategies and frequencies of B cells in the obese SAT versus blood. The blood and the SVF from
the SAT of the same individual were stained to evaluate the frequencies of B cells. Gating strategies and a
representative dot plot from one individual are shown. Means  SE from 20 individuals are shown in red. The
figure is adapted from Frasca D et al., PLOS One, 2018

3.4 Acquisition 1. Single color controls should be prepared for each experiment
and Analysis to assist with compensation. We routinely use cryopreserved
of Lymphocyte normal PBMC from healthy donors for single color staining.
Populations in SAT- 2. Follow your instrument guidelines to perform automatic com-
Derived Blood and SVF pensation using the single color controls as well as the
(see Note 1) unstained cells.
3. Acquire at least 50,000 (or as many as possible) events in the
Live/CD45+ gate.
4. After acquisition we export the files and perform the analyses at
a later time with the FlowJo 10.0.6 software. Figure 1 shows
our gating strategy and analysis of B cell subsets in blood versus
SVF from the same obese individual.

3.5 Detection 1. Dispense the cells in 12-well culture plates at a concentration of

of Single Cell T-Bet 106 cells/mL in c-RPMI (2 to 5  106 cells/well).
mRNA by Amplified 2. Two culture conditions are set up. Cells are left unstimulated or
Fluorescence In Situ stimulated with the TLR7 agonist CL097 at 5 μg/mL.
Hybridization Using 3. Place plate inside a 5% CO2 incubator at 37  C for 24 h.
PrimeFlow mRNA
4. Collect the cells with a serological pipette and count them with
Assay (see Note 2)
Trypan Blue. Up to 5  106 cells can be used per reaction.
5. Dispense the cells in FACS tubes for membrane staining. There
should be a no probe control tube per condition, that is,
unstimulated and TLR7-stimulated cells.
6. Perform membrane staining as described above with the fol-
lowing antibodies: LIVE/DEAD™ Fixable Aqua Dead Cell
Stain Kit (ThermoFisher L34966), Pacific Blue™ anti-CD45
(clone 2D1), APC Anti-CD19 (clone HIB19).
 Immunophenotyping of Human B Cells 123

Gated on Live CD45+CD19+ B cells

T-bet
Unstimulated: 3% CL097: 12%

BLOOD

Unstimulated: 11% CL097: 40%

SVF

FSC

Fig. 2 Frequencies of T-bet-positive B cells in the obese SAT versus blood as

evaluated by PrimeFlow RNA assay. SVF and PBMC from age-, gender-, and
BMI-matched obese healthy donors were stained to evaluate the frequencies of
T-bet-positive B cells. Briefly, cells were labeled with the live/dead detection kit
and then with anti-CD45 and anti-CD19 antibodies to detect B cells. For T-bet
mRNA detection, target probe hybridization was performed using type 1 (Alexa
Fluor 647) probe for T-bet. Cells were unstimulated or stimulated with the TLR7
agonist CL097. Cells were incubated for 2 h with the probe in a precisely
calibrated incubator set to 40  C. All samples were then incubated with the
PreAmplification (PreAmp) reagent for 2 h and the Amplification [9] reagent for
an additional 2 h at 40  C. After signal amplification, cells were incubated with
the labeled probe at 40  C for 1 h. Cells were washed and suspended in staining
buffer prior to acquisition. The figure is adapted from Frasca D et al., PLOS One,
2018

7. After staining, proceed to the PrimeFlow mRNA Assay proto-

col as per manufacturer’s instructions. We follow exactly the
protocol provided by the manufacturer (ThermoFisher), and it
will be briefly described in the following steps. For a better
managed workflow we perform the assay in 2 days as
recommended.
Day 1: Fixation, permeabilization and target probe
hybridization.
8. Wash once with Flow Cytometry Staining Buffer. Spin cells at
500  g for 5 min at 2–8  C.
9. Fix cells in PrimeFlow RNA Fixation Buffer 1 for 30 min at
2–8  C.
10. Wash twice with 1 PrimeFlow RNA Permeabilization Buffer
with RNase Inhibitors. Spin cells at 800  g for 5 min at
2–8  C.
124 Alain Diaz et al.

11. Fix cells in 1 PrimeFlow RNA Fixation Buffer 2 for 60 min

at RT.
12. Wash twice with PrimeFlow RNA Wash Buffer. Spin cells at
800  g for 5 min at RT (room temperature).
13. Perform Target (T-bet) Probe hybridization for 2 h at 40  C.
Invert to mix after 1 h.
14. Wash once with PrimeFlow RNA Wash Buffer. Spin cells at
800  g for 5 min at RT.
15. Wash once with PrimeFlow RNA Wash Buffer with RNase
Inhibitors. Spin cells at 800  g for 5 min at RT.
16. Store samples overnight.
Day 2: Signal amplification.
17. Perform PreAmp hybridization for 1.5 h at 40  C.
18. Wash three times with PrimeFlow RNA Wash Buffer. Spin cells
at 800  g for 5 min at RT.
19. Perform Amp hybridization for 1.5 h at 40  C.
20. Wash twice with PrimeFlow RNA Wash Buffer. Spin cells at
800  g for 5 min at RT.
21. Perform Label (fluorescent) Probe hybridization for 1 h at
40  C.
22. Wash twice with PrimeFlow RNA Wash Buffer. Spin cells at
800  g for 5 min at RT.
23. Wash once with PrimeFlow RNA Storage Buffer or Flow
Cytometry Staining buffer. Spin cells at 800  g for 5 min
at RT.
24. Analyze samples on the BD LSR Fortessa flow cytometer.
25. Single-color controls consisting of stained PBMC are used for
compensation in each experiment. We do not perform the
beads based compensation suggested in the manufacturer’s
instructions.
26. Acquire at least 50,000 (or as many as possible) events in the
Live/CD45+ gate.
27. After acquisition we export the files and perform the analyses at
a later time with the FlowJo 10.0.6 software. Figure 2 shows
the frequencies of T-bet-positive B cells in the obese SAT
versus blood from age-, gender-, and BMI-matched obese
healthy donors.

4 Notes

1. All FACS tubes containing stained blood and SVF should be

filtered to eliminate clumps and prevent clogging of the flow
cytometer. We recommend using Falcon tubes with a cell
 Immunophenotyping of Human B Cells 125

strainer cap (BD Falcon) and Pasteur pipettes to perform the

filtering. For the combination of colors that we use, the instru-
ment used for acquisition should be equipped with at least the
Violet (405 nm), Blue (488 nm), and Red (640 nm) lasers. Our
instrument is a BD LSR Fortessa HTS analyzer running the
DiVa-8 software, equipped with five lasers and capable of
19-color analysis.
2. To perform this assay we use the isolated SVF as well as PBMC
from age-, gender- and BMI-matched obese healthy donors.
The purpose is to detect the mRNA of the transcription factor
T-bet in B cells using this novel technique in combination with
flow cytometry. The transcription factor T-bet is encoded by
the tbx21 gene, and it is known to be associated with the
secretion of IgG2a/c antibodies in mice [21–23] (IgG1 in
humans), which include pathogenic (autoimmune)
antibodies [20].

Acknowledgments

This study was supported by NIH R56 AG32576 (DF/BBB), R21

AG042826 (DF/BBB), R21 AI096446 (BBB/DF), R56
AG059719 [4].

References
1. Apovian CM, Gokce N (2012) Obesity and study II. Arch Intern Med 167
cardiovascular disease. Circulation 125 (15):1670–1675. https://doi.org/10.1001/
(9):1178–1182. https://doi.org/10.1161/ archinte.167.15.1670
CIRCULATIONAHA.111.022541 7. Casas R, Sacanella E, Estruch R (2014) The
2. Hotamisligil GS (2006) Inflammation and immune protective effect of the Mediterranean
metabolic disorders. Nature 444 diet against chronic low-grade inflammatory
(7121):860–867. https://doi.org/10.1038/ diseases. Endocr Metab Immune Disord Drug
nature05485 Targets 14(4):245–254
3. Johnson AM, Olefsky JM (2013) The origins 8. Hass DJ, Brensinger CM, Lewis JD, Lichten-
and drivers of insulin resistance. Cell 152 stein GR (2006) The impact of increased body
(4):673–684. https://doi.org/10.1016/j.cell. mass index on the clinical course of Crohn’s
2013.01.041 disease. Clin Gastroenterol Hepatol 4
4. Shoelson SE, Lee J, Goldfine AB (2006) (4):482–488. https://doi.org/10.1016/j.
Inflammation and insulin resistance. J Clin cgh.2005.12.015
Invest 116(7):1793–1801. https://doi.org/ 9. Collaborators GBDO, Afshin A, Forouzanfar
10.1172/JCI29069 MH, Reitsma MB, Sur P, Estep K, Lee A,
5. Renehan AG, Tyson M, Egger M, Heller RF, Marczak L, Mokdad AH, Moradi-Lakeh M,
Zwahlen M (2008) Body-mass index and inci- Naghavi M, Salama JS, Vos T, Abate KH,
dence of cancer: a systematic review and meta- Abbafati C, Ahmed MB, Al-Aly Z, Alkerwi A,
analysis of prospective observational studies. Al-Raddadi R, Amare AT, Amberbir A, Ame-
Lancet 371(9612):569–578. https://doi.org/ gah AK, Amini E, Amrock SM, Anjana RM,
10.1016/S0140-6736(08)60269-X Arnlov J, Asayesh H, Banerjee A, Barac A,
6. Setty AR, Curhan G, Choi HK (2007) Obesity, Baye E, Bennett DA, Beyene AS, Biadgilign S,
waist circumference, weight change, and the Biryukov S, Bjertness E, Boneya DJ, Campos-
risk of psoriasis in women: nurses’ health Nonato I, Carrero JJ, Cecilio P, Cercy K, Cio-
banu LG, Cornaby L, Damtew SA,
126 Alain Diaz et al.

Dandona L, Dandona R, Dharmaratne SD, 14. Ottaviani S, Allanore Y, Tubach F, Forien M,

Duncan BB, Eshrati B, Esteghamati A, Feigin Gardette A, Pasquet B, Palazzo E, Meunier M,
VL, Fernandes JC, Furst T, Gebrehiwot TT, Hayem G, Job-Deslandre C, Kahan A,
Gold A, Gona PN, Goto A, Habtewold TD, Meyer O, Dieude P (2012) Body mass index
Hadush KT, Hafezi-Nejad N, Hay SI, influences the response to infliximab in anky-
Horino M, Islami F, Kamal R, Kasaeian A, losing spondylitis. Arthritis Res Ther 14(3):
Katikireddi SV, Kengne AP, Kesavachandran R115. https://doi.org/10.1186/ar3841
CN, Khader YS, Khang YH, Khubchandani J, 15. Sandberg ME, Bengtsson C, Kallberg H,
Kim D, Kim YJ, Kinfu Y, Kosen S, Ku T, Defo Wesley A, Klareskog L, Alfredsson L, Saevars-
BK, Kumar GA, Larson HJ, Leinsalu M, dottir S (2014) Overweight decreases the
Liang X, Lim SS, Liu P, Lopez AD, chance of achieving good response and low
Lozano R, Majeed A, Malekzadeh R, Malta disease activity in early rheumatoid arthritis.
DC, Mazidi M, McAlinden C, McGarvey ST, Ann Rheum Dis 73(11):2029–2033. https://
Mengistu DT, Mensah GA, Mensink GBM, doi.org/10.1136/annrheumdis-2013-
Mezgebe HB, Mirrakhimov EM, Mueller 205094
UO, Noubiap JJ, Obermeyer CM, Ogbo FA, 16. Hotamisligil GS (2017) Inflammation, meta-
Owolabi MO, Patton GC, Pourmalek F, flammation and immunometabolic disorders.
Qorbani M, Rafay A, Rai RK, Ranabhat CL, Nature 542(7640):177–185. https://doi.
Reinig N, Safiri S, Salomon JA, Sanabria JR, org/10.1038/nature21363
Santos IS, Sartorius B, Sawhney M,
Schmidhuber J, Schutte AE, Schmidt MI, 17. Pearce EL, Pearce EJ (2013) Metabolic path-
Sepanlou SG, Shamsizadeh M, ways in immune cell activation and quiescence.
Sheikhbahaei S, Shin MJ, Shiri R, Shiue I, Immunity 38(4):633–643. https://doi.org/
Roba HS, Silva DAS, Silverberg JI, Singh JA, 10.1016/j.immuni.2013.04.005
Stranges S, Swaminathan S, Tabares-Seisdedos- 18. Grant RW, Dixit VD (2015) Adipose tissue as
R, Tadese F, Tedla BA, Tegegne BS, Terkawi an immunological organ. Obesity 23
AS, Thakur JS, Tonelli M, Topor-Madry R, (3):512–518. https://doi.org/10.1002/oby.
Tyrovolas S, Ukwaja KN, Uthman OA, 21003
Vaezghasemi M, Vasankari T, Vlassov VV, Voll- 19. Frasca D, Diaz A, Romero M, Thaller S,
set SE, Weiderpass E, Werdecker A, Wesana J, Blomberg BB (2018) Secretion of autoim-
Westerman R, Yano Y, Yonemoto N, Yonga G, mune antibodies in the human subcutaneous
Zaidi Z, Zenebe ZM, Zipkin B, Murray CJL adipose tissue. PLoS One 13(5):e0197472.
(2017) Health effects of overweight and obe- https://doi.org/10.1371/journal.pone.
sity in 195 countries over 25 years. N Engl J 0197472
Med 377(1):13–27. https://doi.org/10. 20. Winer DA, Winer S, Shen L, Wadia PP,
1056/NEJMoa1614362 Yantha J, Paltser G, Tsui H, Wu P, Davidson
10. Frasca D, Ferracci F, Diaz A, Romero M, MG, Alonso MN, Leong HX, Glassford A,
Lechner S, Blomberg BB (2016) Obesity Caimol M, Kenkel JA, Tedder TF,
decreases B cell responses in young and elderly McLaughlin T, Miklos DB, Dosch HM, Engle-
individuals. Obesity 24(3):615–625. https:// man EG (2011) B cells promote insulin resis-
doi.org/10.1002/oby.21383 tance through modulation of T cells and
11. Ovsyannikova IG, White SJ, Larrabee BR, Grill production of pathogenic IgG antibodies. Nat
DE, Jacobson RM, Poland GA (2014) Leptin Med 17(5):610–617. https://doi.org/10.
and leptin-related gene polymorphisms, obe- 1038/nm.2353
sity, and influenza A/H1N1 vaccine-induced 21. Gerth AJ, Lin L, Peng SL (2003) T-bet regu-
immune responses in older individuals. Vaccine lates T-independent IgG2a class switching. Int
32(7):881–887. https://doi.org/10.1016/j. Immunol 15(8):937–944
vaccine.2013.12.009 22. Naradikian MS, Myles A, Beiting DP, Roberts
12. Sheridan PA, Paich HA, Handy J, Karlsson EA, KJ, Dawson L, Herati RS, Bengsch B, Linder-
Hudgens MG, Sammon AB, Holland LA, man SL, Stelekati E, Spolski R, Wherry EJ,
Weir S, Noah TL, Beck MA (2012) Obesity is Hunter C, Hensley SE, Leonard WJ, Cancro
associated with impaired immune response to MP (2016) Cutting edge: IL-4, IL-21, and
influenza vaccination in humans. Int J Obes 36 IFN-gamma interact to govern T-bet and
(8):1072–1077. https://doi.org/10.1038/ijo. CD11c expression in TLR-activated B cells. J
2011.208 Immunol 197(4):1023–1028. https://doi.
13. George MD, Baker JF (2016) The obesity epi- org/10.4049/jimmunol.1600522
demic and consequences for rheumatoid arthri- 23. Peng SL, Szabo SJ, Glimcher LH (2002) T-bet
tis care. Curr Rheumatol Rep 18(1):6. https:// regulates IgG class switching and pathogenic
doi.org/10.1007/s11926-015-0550-z autoantibody production. Proc Natl Acad Sci
 Immunophenotyping of Human B Cells 127

U S A 99(8):5545–5550. https://doi.org/10. interactions between adipocytes and tumor

1073/pnas.082114899 cells recruit and modify macrophages to the
24. Santander AM, Lopez-Ocejo O, Casas O, mammary tumor microenvironment: the role
Agostini T, Sanchez L, Lamas-Basulto E, of obesity and inflammation in breast adipose
Carrio R, Cleary MP, Gonzalez-Perez RR, tissue. Cancers (Basel) 7(1):143–178. https://
Torroella-Kouri M (2015) Paracrine doi.org/10.3390/cancers7010143
 Chapter 8

Multiparametric Flow Cytometry Analysis of Naı̈ve, Memory,

and Effector T Cells
Ankit Saxena, Pradeep K. Dagur, and Angélique Biancotto

Abstract
Polychromatic flow cytometry enables the detection and characterization of markers which are helpful in
defining phenotype of various cell subsets. Here we describe flow cytometry-based method to characterize
phenotype of naı̈ve, memory, and effector T cells. Being able to differentiate these cells is crucial in
understanding immune response, and immune profiling. Naı̈ve T cells enable the body to fight off new,
unrecognized infections and diseases, and memory T cells are enriched for response to recall antigens.
Furthermore, the antigen-experienced T cell populations can be broadly divided into effector and memory
cell compartments, both of which are needed for sustaining a responsive immune system. Simplistically, the
effector T cells require active antigenic stimulation to eliminate pathogens. On the other hand, memory T
cells are described as cells which remain present in the absence of antigenic stimulation and have the capacity
to expand rapidly upon secondary challenges. Recently, with the identification of central and effector
memory T cell subsets, tremendous efforts have been devoted to characterize markers on the surfaces of
these cells. Though, various markers have been used to identify the subsets, no single marker that segregates
one subset from the other has been described. Thus, multiple markers are needed to subset the cells in order
to characterize them. Here we report the verification of a nine-color panel (CD3, CD4, CD8, CD45RO,
CD28, CD95, CCR7, Live/Dead Aqua, dump channel-CD19, CD14, CD56, CD16) that can successfully
identify six distinct CD4 and CD8 T cell populations within the naı̈ve and effector cell subsets from human
donors.

Key words Immunophenotyping, Multicolor flow cytometry, Whole blood, T cells, Memory T cells,
Naı̈ve T cells

1 Introduction

T cell phenotype has long been used as means of functionally

classifying T cell subsets [1, 2]. The development of techniques
for proteins detection at the single-cell level allowed detailed cor-
relations between the functional properties of T-cells and their
phenotype [3]. Following positive and negative selection in thymus
CD4+ and CD8+ T cells are released in the periphery as mature
naı̈ve T cells. Mature CD4 and CD8 T cells in the extrathymic
environment are long-lived cells which can remain in interphase

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_8, © Springer Science+Business Media, LLC, part of Springer Nature 2019

129
130 Ankit Saxena et al.

for many weeks or months [4, 5]. In response to antigenic chal-

lenge naı̈ve T (TN) cells with specificity for a certain antigenic
epitope proliferate and differentiate and later die >95%, after anti-
genic clearance leaving behind a small pool of long-lived memory
cells [6, 7]. The process of conversion of a naı̈ve T cell to memory is
a sequential process where less differentiated cells give rise to more
differentiated progenies in response to antigenic stimulation or
homeostatic signaling. These subsets are often associated with
signature molecules and display phenotypic and gene specific prop-
erties which are useful to define them as cluster of cells. Though the
progression of the naı̈ve and effector memory cell subsets can be
broadly staged as different subsets such as naı̈ve T (TN), T-stem cell
memory (TSCM), T-central memory (TCM), T-terminal memory
(TTM), T-effector memory (TEM), and T-terminal effector (TTE)
[7]. While naı̈ve T cells are fairly homogeneous, diversity becomes
extreme in the antigen-experienced memory compartment. The
precise identification of these subsets is pivotal for both basic as
well as clinical immune monitoring of patients undergoing immu-
notherapy for T cell-based therapeutic approaches [8].
Since multiparameter flow cytometry approaches are becoming
more popular, if not routine, in immunology laboratories, it would
be logical that human T cell differentiation should be delineated
using a minimum set of canonical markers, that is, CD45RO
(or CD45RA), CCR7, CD28, and CD95. The differential expres-
sion of these markers allows the identification of six subsets in the
peripheral blood of healthy humans: TN, TSCM, TCM, TTM, TEM,
and TTE cells (Fig. 1).
Multicolor flow cytometric analysis of T cells has revealed that a
single marker or a combination of two markers does not allow the
identification of pure TN cells [2]. Ideally, the more markers are
included to define TN cells in a polychromatic panel, the better. We
describe a nine-color panel to identify the abovementioned subsets
(Table 1).

2 Materials

Prepare all solutions using distilled water (dH2O) and analytical

grade reagents. Prepare and store all reagents at room temperature
(RT) (unless indicated otherwise). Diligently follow all waste dis-
posal regulations when disposing of waste materials.
1. Sample material: human peripheral blood, bone marrow
(in anticoagulant—heparin, EDTA etc.), lymphoid tissue.
2. RBC lysis buffer—ammonium chloride lyse (10 concentra-
tion) NH4Cl (ammonium chloride) 8.02 gm; NaHCO3
(sodium bicarbonate) 0.84 gm; EDTA (disodium) 0.37 gm;
dissolve in 100 mL with dH2O. Store at 4  C for 6 months.
 Immunophenotyping Naı̈ve, Memory, and Effector T Cells 131

Fig. 1 T cell Heterogeneity among naı̈ve and memory subsets. CD45R0, CCR7, CD28, and CD95 helps in
identifying six major subsets of T cells. While differentiating (through activation) from TSCM to TCM, TTM, TEM
and culminating in TTE cells, memory T cells progressively lose or acquire specific surface protein markers.
Following encounter with antigen, these naive T cells develop into effectors, for which phenotype is highly
dynamic and unpredictable. When the antigen is cleared, surviving effector T cells that survive return to
memory state. Bottom-Multiparametric flow cytometry identification of multiple T cell subsets among naı̈ve
and memory compartment in the peripheral blood of healthy individuals according to expression of CD45RO,
CCR7, CD28, and CD95

Working solution dilute 10 mL 10 concentrate with 90 mL

dH2O. Refrigerate until use.
3. Staining buffer—1 phosphate buffered saline (PBS): dissolve
8 gm NaCl, 0.2 gm KCl, 1.15 g Na2HPO4, and 0.2 gm
KH2PO4 in 800 mL distilled water (dH2O), 10% FCS or 1%
bovine serum albumin (BSA). Adjust the pH to 7.4 and the
final volume to 1 liter (see Note 1).
4. Fluorochrome labeled antibody diluted to the appropriate con-
centration as determined by titration and viability dye (Table 2)
(see Note 2).
5. 12  5-mm round-bottom polystyrene test tubes or 96-well
round-bottom, V-bottom or flat bottom microtiter plates for
high-throughput screening (HTS) (see Note 3).
6. 40–100-μm nylon mesh for isolating cells from lymphoid
tissues.
Table 1
Instrument configuration. The panel was optimized for an LSR II with the listed optical elements

Optical filters

Laser Laser Band

wavelength power Laser Spectral range for Dichroic Dichroic pass
(nm) (mW) type detector (nm) #1 (nm) #2 (nm) (nm) Fluorochrome
633 200 DPSS 750–810 – 740 780/60 APC-Cy7
650–670 685 – 660/20 APC
532 150 DPSS 760–800 – 740 780/40 PE-Cy7
640–680 690 640 660/40 PE-Cy5
562–587 600 – 575/25 PE
488 100 DPSS 505–525 685 505 515/20 FITC
405 100 DPSS
590–620 630 595 605/40 BV605
564–606 595 630 660/40 BV650
540–580 557 535 560/40 AquaBlue
425–475 535 – 450/50 BV421
In our set-up the dichroic filters act as long pass filters (Dichroic #2) for one detector, sending longer wavelengths
through the respective band pass filters, and reflecting shorter wavelengths down the optical path to the next dichroic
filter (Dichroic #1). DPSS Diode pumped solid state

Table 2
Antibody panel details: antigen specific mAb used for various cell surface antigens

Specificity Fluorochrome Ab clone Vendor Catalog number Volume

CD3 BV450 UCHT1 BD Biosciences 560365 3.5 μL
CD4 APC-Cy7 RPA-T4 BD Biosciences 557871 3.5 μL
CD8 BV605 SK19 BD Biosciences 56416 3.5 μL
CD28 PE CD28.2 BD Biosciences 555730 15.0 μL
Dead cells AquaBlue – Life Technologies L34957 100 μLa
CCR7 FITC 150503 R&D FAB197F 10 μL
CD45RO PE-Cy5 UCHL-1 BD Biosciences 555494 15 μL
CD95 APC DX2 BD Biosciences 558814 15 μL
CD20 PE-Cy7 L27 BD Biosciences 335793 15 μL
CD56 PE-Cy7 NCAM16.2 BD Biosciences 335809 15 μL
CD14 PE-Cy7 M5E2 BD Biosciences 557742 3.5 μL
CD16 PE-Cy7 3G8 BD Biosciences 557742 3.5 μL
APC allophycocyanin, APC-Cy7 allophycocyanin–cyanine H7 tandem, FITC fluorescein isothiocyanate, PE phycoery-
thrin, PE-Cy7 phycoerythrin–cyanine 7 tandem, Pe-Cy5 peridinin–cyanine 5 tandem, BV450 brilliant violet 450, BV605
brilliant violet 605 violet
aAquaBlue was prediluted 3.5 μL of Aqua blue stock in 1 mL of FACS buffer
 Immunophenotyping Naı̈ve, Memory, and Effector T Cells 133

7. Centrifuge with adaptors for 5 mL, 12  5-mm round-bottom

test tubes, 50- and 15 mL falcon tubes or 96-well round-
bottom microtiter plates.
8. Fixation buffer: 2% paraformaldehyde in 1 PBS.
9. Compensation beads (BD Biosciences).
10. Flow cytometer.
11. FlowJo analysis software (Tree Star Inc.).

3 Methods

3.1 Blood Sample 1. Whole Blood Lysis: collect blood in sodium heparin, sodium
Collection and Cell citrate, or EDTA tube and keep it at RT on a nutator. Transfer
Preparation 5–10 mL blood with a disposable Pasteur pipette into a 50 mL
falcon tube and add 40–45 mL of RBC lysis buffer. Place the
50 mL falcon tube on a nutator for 2 min. Centrifuge tube at
300  g for 5 min at 25  C. Carefully remove the lysis buffer
and break the cell pellet by slow vortex or gentle tapping.
Repeat addition of 45–50 mL of RBC lysis buffer and keep
cells on nutator for 1 min. Centrifuge tube at 300  g for 5 min
at 25  C followed by lysis buffer removal and pellet breaking.
Now add 10 mL of staining buffer (RT) to the cells and give a
final wash to replace remnant RBC lysis buffer with the staining
buffer. Pellet the cells (5 min, 300 g) and aspirate the superna-
tant. Centrifuge tube at 300  g for 5 min at 25  C to make a
final cell concentration of 1–2  106cells/mL (see Note 4).
2. Add fresh staining buffer and adjust cell suspension to a con-
centration of 1–2  106 cells/mL in ice-cold PBS, 10% FCS. If
large clumps seen the sample may be passed through the
40–100 μm mesh to get rid of clumps and obtain single cell
suspension. Cells are usually stained in polystyrene round-
bottom 12  75 mm2 Falcon tubes. However, they can be
stained in any container for which you have an appropriate
centrifuge (e.g., test tubes, microcentrifuge tubes, and
96-well round-bottomed microtiter plates) (see Note 5).
3. Add titrated amount of fluorochrome-conjugated antibodies
and live/dead aqua stain along with FcR blocker or in the
presence of 2 μL of normal mouse serum to reduce background
staining.
4. Incubate cells at 4  C for 15 min (see Note 6).
5. Add 2 mL of staining buffer and wash the cells (5 min, 300 g).
Resuspend the pellet in 0.4–0.5 mL of staining buffer.
6. If sample acquisition is to be done after 24 h, then fix cells using
fixation buffer. Add 0.5 mL of fixation buffer to the cells and
incubate at RT for 15–20 min in dark (see Note 7).
134 Ankit Saxena et al.

7. Pellet the cells (5 min, 500 g) and decant the fixation buffer-
containing supernatant (paraformaldehyde is toxic: handle with
care; dispose of according to material safety regulations) (see
Note 8).
8. Resuspend cells with 0.5 mL of staining buffer keep it in dark
and cold till samples are acquired on cytometer. (It is advised to
run processed samples as soon as possible or within 72 h of
staining.)

3.2 Data Collection Specific methods depend on the available make and model of flow
on Flow Cytometer cytometer. It is important to appropriately establish forward and
side scatter gates to exclude debris and cellular aggregates from
analysis. If cells are stained with multiple fluorescent labels, atten-
tion must be paid to balance the signal amplification among photo-
multiplier tubes (PMTs) in a way that particular fluorophore is
brighter in its respective PMT in order to minimize potential
spillover between fluorochromes and detectors.
1. Single-color controls for each fluorochrome is prepared and
unstained cells are kept as negative control. Single-color con-
trols can be prepared both by using cells or compensation
beads as long as bright signal is observed in each respective
detector PMT (see Note 9).
2. Here a BD LSR-II was used to acquire stained samples
(Table 3) using BD FACS DIVA v8.0 (BD) acquisition and
analysis software.

Table 3
Rationale for panel development. The table highlights important points to be considered in order to
successfully design and execute a multicolor panel. Points to be considered are ordered under
priority rating followed by rationale and reagents concerned

Priority rating Category Rationale for priority level Reagents concerned

1 Dump channel Detector with least number of Monoclonal antibodies against
good conjugates available CD16, CD14, CD56, and CD20
human antigen
2 Live dead To gate out dead cells AquaBlue live/dead dye
3 T cell subset Identify CD4+ and CD8+ T Monoclonal antibodies against CD3,
markers cell compartments CD4, CD8
4 Differentiation Reason for development of CCR7 and CD45RO human antigen
markers panel
5 Differentiation Identification of TTE and Monoclonal antibodies against
markers: TSCM subsets within naı̈ve- CD95 and CD28 human antigen
additional like cells
 Immunophenotyping Naı̈ve, Memory, and Effector T Cells 135

3. Open DIVA software and connect to the instrument. Look for

the green light appearing on the right bottom corner of the
software to ensure proper connection. Once the instrument is
connected, establish a new experiment in browser window and
add appropriate number of tubes to acquire.
4. When the workstation is connected to the cytometer and an
experiment is open, a gray pointer icon is displayed next to
tubes in the browser. To activate acquisition controls, click the
icon next to the tube you want to acquire to set as current tube
pointer. The icon turns green and the tube becomes the active
tube in the acquisition dashboard.
5. Under tab view select, Browser, Inspector, Cytometer, Acqui-
sition, dashboard, Worksheet panels.
6. Click the “Parameters” tab in the “Cytometer” window and
delete all the parameters that you do not need. To delete
parameters, click the selection button next to each parameter
that you want to delete. Hold down the Ctrl key to select more
than one. After you are finished selecting, click the “delete”
button. if needed, parameters can be added or modified by
clicking the “Add” button.
7. Use the experiment layout to create labels, enter values for
keywords, and enter acquisition criteria for each tube in an
experiment.
8. Add or change labels. Select the field(s) listing the fluoro-
chromes to be labeled, and type to enter a label.
9. Once you have created your experiment, you are ready to create
compensation controls. Select experiment > compensation
setup > create compensation controls. Alternatively, right-
click the experiment icon and select compensation setup >
create compensation controls. A dialog appears where you can
add or delete controls, define label-specific controls, or change
the order of the compensation controls.
10. Leave the “Include separate unstained control tube/well”
checkbox selected when you are running unstained sample as
one of your compensation controls.
11. Record data for the unstained and single-color controls
approximately 5000 events for beads or cells, using appropriate
signal amplification by PMTs to ensure that bright signal is
observed in each respective detector PMT (see Note 10; com-
pensation controls can only be acquired on normal worksheet,
not on global worksheet).
12. At the end, select Experiment > Compensation setup > Cal-
culate compensation.
136 Ankit Saxena et al.

13. Enter a compensation setup name, click Link and Save, and
click OK. To apply compensation values to that experiment’s
cytometer settings, click “Apply Only.”
14. Toggle back to global work sheet (in case gating strategy
changes in different tube then use “sheet” instead of “Global
work sheet”).
15. Use menu commands or plot buttons to create dot plots,
contour plots on worksheet.
16. To create a plot using a plot button, click the appropriate
button on the worksheet toolbar, and click once on the work-
sheet to draw a plot of default size.
17. Acquire cells at an event rate of 3000–5000 events per second
and record data.
18. Use a series of plots and establish a gating hierarchy to ensure
recording of debris and doublet free live T cells (at least
250,000 single live T cells to ensure sufficient number of rare
subsets to analyze in recorded data file, if needed).
19. Postacquisition, the fcs files should be transferred as follows:
Right-click on the Experiment and select Export fcs Files. Save
files in the desired folders. This will transfer all your compensa-
tion files as well as sample fcs files.

3.3 Data Analysis 1. On Flow-jo software open a new workspace and add compen-
sation fcs files and samples fcs files in separate group folders.
2. If required, compensation can be recalculated using compensa-
tion fcs files and newly calculated compensation can be applied
to the sample fcs files for further data analysis (see Note 11).
3. Open the fcs files and start creating gates as follows.
4. Define the gated populations as shown in (Fig. 2a, b).
5. Lymphocyte gate: Create bivariant polychromatic plot and gate
click FSC-A on X-axis and SSC-A on Y axis. Make a polygonal
gate around lymphocyte population as shown. Please ensure in
active gate check box gate events inside is checked on the flow-
jo plot.
6. Singlet gating: select gate and right-click, select drill down
bivariant polychromatic plot with FSC-A on X axis and
FSC-H on Y axis for preselected lymphocytes. Make a gate
around singlets. If the area scaling for FSC detector and Blue
LASER is properly adjusted while data recording, then singlets
will be visible on a diagonal line when FSC-H and FSC-A
parameters are plotted as X and Y axis of bivariate plot. Doub-
lets will fall toward FSC-A but not FSC-H and can be gated out
to identify single cells.
7. Live cell gating: Gate Aqua LD negative subsets as live cells.
 Immunophenotyping Naı̈ve, Memory, and Effector T Cells 137

8. Gate dump negative cells excluding out B cells, NK cells, and

Monocytes.
9. Draw a histogram and select CD3 cells.
10. On gated CD3 cells Select CD4 and CD8 populations.
11. Select CCR7 on Y axis and CD45RO on X axis for both CD4
and CD8 subsets and analyze as follows.
12. Divide this population into four subsets based on the presence
or absence of CD45RO and CCR7 as CD45RO+CCR7+,
CD45RO+CCR7–, CD45RO–CCR7+, CD45RO–CCR7–
on CD4 and CD8 T-cell subsets.
13. Right-click on the CD45RO–CCR7+ gate and select option
“Drill down” from drop down menu, make subset gate on
bivariate plot for CD95 and CD28. Define CD28 bright
CD95+ cells as TSCM and CD95–CD28+ as TN.

Fig. 2 Gating strategy for identifying naı̈ve and memory T cell subsets. (a)The sequential gating strategy
defines gated total lymphocytes from whole blood lysed sample. (b) Lymphocytes were gated for singlets, live
and dump channel which excluded B, T, NK, and Monocytes. CD3 gated cells were further gated for CD4, CD8
followed by CD45RO and CCR7. On the basis of CD45RO and CCR7 the memory and naı̈ve T cells were defined
as follows. From CD45RO–CCR7+ subset gate on CD95 and CD28. TSCM– CD45RO–CCR7+ CD28 bright CD95+,
TN– CD45RO–CCR7+ CD95–CD28+, TCM– CD45RO+CCR7– CD95+CD28+, TTM– CD45RO+CCR7– CD95
+CD28+, TEM– CD45RO+CCR7– CD95+CD28–, TTE– CD45RO–CCR7–CD28– and CD95–
138 Ankit Saxena et al.

Fig. 2 (continued)

14. Drill down CD45RO+CCR7+ subsets as defined above and

identify CD95+CD28+ expressing cells as TCM.
15. Drill down CD45RO+CCR7– subset and identify CD95
+CD28+ expressing cells as TTM whereas CD95+CD28– as
TEM subset.
16. Drill down CD45RO–CCR7 subsets and identify CD28–
and CD95– as TTE cells.
17. Drag the gated population from work space to layout editor to
prepare gating strategy schemes and to display data.

4 Notes

1. Use of azide in FACS buffer should be avoided especially if cells

are to be sorted and used later for functional studies. Azide
inhibits various metabolic functioning of cells, which could
hamper the downstream functional studies of sorted cells.
 Immunophenotyping Naı̈ve, Memory, and Effector T Cells 139

Also, unfixed cells kept long in azide buffer may show more
death than normal.
2. To avoid nonspecific binding of antibodies/reagents each
fluorochrome-conjugated antibody/reagent should be opti-
mally titrated before use.
3. High-throughput screen (HTS) should be set according to the
manufacturer’s instructions; please refer to the manual. It is
Important to note that there is a dead volume of 30 μL for
plates of all kind. The samples should preferably be run at a low
flow rate between 1 and 2 μL/s for up to 106 cells.
4. Avoid prolonged (>5 min) incubation with ACK lysis buffer as
it will significantly affect the lymphocyte viability.
5. Supplementing the medium with 20 μg/mL DNase helps to
dissociate cell aggregates that result from DNA released by
dead cells. This is especially helpful if working with cryopre-
served cells as freeze–thaw may cause cell lysis.
6. Cell staining at 4  C in the dark is a preferred method as it
avoids antibody capping as well as photobleaching.
7. If there are any questions about the pathogenicity of the cells
involved, then fixation of cells after staining and before running
them through a flow cytometer is highly recommended for
immunophenotyping purposes, not for live cell sorting.
8. Fixative solution should be removed after incubation by wash-
ing cells with FACS buffer. Prolonged fixation may alter fluo-
rochrome properties and may lead to spillover in other
channels.
9. Compensation beads should not be washed—as it may lead to
loss of beads.
10. Optimize PMT voltages using unstained and single colors in
order to obtain minimal spectral overlap.
11. Compensation matrix may still require to be adjusted during
analysis. After adjusting the compensation electronically
update biexponential transformation. On FlowJo—Select Plat-
form<Custom transformation<Compensation matrix. Select
appropriate compensation matrix. And appropriately adjust
Biexponential scales in order to visualize the population clus-
ters on scales.

References
1. Maecker HT, McCoy JP, Nussenblatt R (2012) 2. De Rosa SC, Herzenberg LA, Herzenberg LA
Standardizing immunophenotyping for the et al (2001) 11-color, 13-parameter flow cyto-
human immunology project. Nat Rev Immunol metry: identification of human naive T cells by
12(3):191–200. https://doi.org/10.1038/ phenotype, function, and T-cell receptor diver-
nri3158 sity. Nat Med 7(2):245–248. https://doi.org/
10.1038/84701
140 Ankit Saxena et al.

3. Picker LJ, Singh MK, Zdraveski Z et al (1995) 6. Zhang N, Bevan MJ (2011) CD8(+) T cells: foot
Direct demonstration of cytokine synthesis het- soldiers of the immune system. Immunity 35
erogeneity among human memory/effector T (2):161–168. https://doi.org/10.1016/j.
cells by flow cytometry. Blood 86 immuni.2011.07.010
(4):1408–1419 7. Mahnke YD, Brodie TM, Sallusto F et al (2013)
4. Sprent J, Cho JH, Boyman O et al (2008) T cell The who’s who of T-cell differentiation: human
homeostasis. Immunol Cell Biol 86 memory T-cell subsets. Eur J Immunol 43
(4):312–319. https://doi.org/10.1038/icb. (11):2797–2809. https://doi.org/10.1002/
2008.12 eji.201343751
5. Sprent J, Tough DF (1994) Lymphocyte life- 8. Golubovskaya V, Wu L (2016) Different subsets
span and memory. Science 265 of T cells, memory, effector functions, and
(5177):1395–1400 CAR-T immunotherapy. Cancers (Basel) 8(3).
https://doi.org/10.3390/cancers8030036
 Chapter 9

Immunophenotyping of Human Regulatory T Cells

Janet Staats

Abstract
Regulatory T cells, also known as Tregs, play a pivotal role in maintaining homeostasis of the immune
system and self-tolerance. Tregs express CD3, CD4, CD25, and FOXP3 but lack CD127. CD4 and CD3
identify helper T lymphocytes, of which Tregs are a subset. CD25 is IL-2Rα, an essential activation marker
that is expressed in high levels on Tregs. FOXP3 is the canonical transcription factor, important in the
development, maintenance, and identification of Tregs. CD127 is IL-7 receptor, expressed inversely with
suppression, and is therefore downregulated on Tregs. Flow cytometry is a powerful tool that is capable of
simultaneously measuring Tregs along with several markers associated with subpopulations of Tregs,
activation, maturation, proliferation, and surrogates of functional suppression. This chapter describes a
multicolor flow cytometry-based approach to measure human Tregs, including details for surface staining,
fixation/permeabilization, intracellular/intranuclear staining, acquisition of samples on a flow cytometer,
plus analysis and interpretation of resulting FCS files.

Key words Tregs, nTregs, iTregs, Effector tregs, Immunophenotyping, Flow cytometry, FoxP3,
CD127, Suppressor T cells

1 Introduction

Regulatory T cells (Tregs) orchestrate immune responses, playing

an essential role in immune tolerance, against self and nonself
antigens, and inflammation [1–3]. Their ability to suppress inflam-
mation is directly related to the production of cytokines IL-10 and
TGF-β [1, 3]. Given their critical role in regulating the immune
system, Tregs have been implicated in allergy, transplantation,
infectious diseases, neoplastic disease, pregnancy, diabetes, autoim-
mune disease, colitis, atherosclerosis, and even sleep disorders [4–
28]. A deeper understanding of Tregs, as well as their various
subsets and states of being, may improve the existing standard of
care for these diseases and disorders.
Because tregs are continually sensing and responding to their
microenvironment, they are composed of diverse and dynamic
subpopulations and exist in many different states of being
[29]. Treg subpopulations include natural Tregs (nTregs), induced

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_9, © Springer Science+Business Media, LLC, part of Springer Nature 2019

141
142 Janet Staats

regulatory T cells (iTregs), Tr1, Th3, and Tfh regulatory T cells

[30–32]. In the face of inflammation, these subpopulations
respond to their environment by changing their phenotype. There-
fore, Tregs exist in various states of being, including naı̈ve (nvTreg),
resting (rTreg), memory (mTreg), effector (eTreg), activated
(aTreg), and proliferating Tregs [1, 3, 7, 33–39]. These phenotyp-
ically distinct subpopulations, and their states of being, are identi-
fied using cellular markers, where a marker is a molecule, usually a
protein, that is located on or within a cell. Many different markers
have been used to identify Tregs. The most commonly used Treg
markers are CD3, CD4, CD25, FOXP3, CD127, CD45RA,
CD45RO, CD194, HLA-DR, and Ki67 [7, 35]. CD3 is the
T-cell receptor, used in conjunction with CD4 to identify helper
T cells, of which Tregs are a subset. CD25 is interleukin-2 receptor
alpha or IL-2Rα, a crucial activation marker expressed in high levels
on Tregs, including nvTreg, mTreg, and eTregs [1, 3, 36, 40,
41]. It is important to note, however, that nonregulatory subsets
of T cells, including effector T cells, also express high levels of
CD25. FOXP3 and CD127 are the two quintessential markers of
suppressor function, essential for identifying Tregs. FOXP3 is the
canonical Treg transcription factor, required for Treg development,
maintenance, and function [1–3]. Similar to CD25, FOXP3 is also
expressed in subsets of activated T cells that lack suppressive func-
tion [3, 42, 43]. In humans, unlike mice, FOXP3 exists in different
isoforms, resulting from posttranscriptional modifications [19, 44,
45]. FOXP3 is intranuclear, requiring fixation and permeabilization
for flow cytometry assays. Fixed cells are not suitable for viable cell
sorting; therefore, CD127 was identified as a surrogate for FOXP3
[41, 46]. Expression of CD127, interleukin-7 receptor alpha or
IL-7Rα, is inversely correlated to suppressor function, meaning the
lack of CD127 expression is a useful means of identifying Tregs
[3, 7, 35, 40, 41]. One key intracellular marker for Tregs is Ki67, a
proliferation marker [7]. rTregs proliferate in response to stimula-
tion with specific antigen, becoming aTregs. CD45RA and
CD45RO are CD45 isoforms that are mutually expressed on lym-
phocytes based on their maturational status. They may be used
interchangeably or together to identify Treg subsets that exist in
various maturational states. rTregs are CD45RA+ and CD45RO ,
while mTreg, eTreg, and aTreg are CD45RO+ and CD45RA
[1, 3, 7, 35]. Chemokine receptors are involved in cell trafficking
and play an important role in the migration of Tregs to inflamma-
tory tissue and interactions with antigen presenting cells. Chemo-
kine receptor expression is dependent upon the maturational state
of Tregs. CD194, chemokine receptor 4 (CCR4), is expressed on
terminally differentiated eTregs, the most suppressive phenotype
[33, 35, 37, 38]. Finally, MHC-Class II or HLA-DR is upregulated
on activated lymphocytes, including aTregs [1, 3, 34, 35]. How-
ever, nvTregs lack HLA-DR. While there is no single marker that
 Immunophenotyping of Human Regulatory T Cells 143

can be used to identify Tregs, a combination of markers, called a

panel, must be employed to identify, qualify, and quantify Tregs.
Flow cytometry is the ideal tool for simultaneously measuring
the expression of multiple markers on or within a single cell, and
can be used to measure Tregs, their subsets, and their states of
being [47–50]. This chapter describes how to measure human
Tregs using flow cytometry for the purposes of immunophenotyp-
ing. Flow cytometry requires samples to be in a single cell suspen-
sion in order to establish a laminar sample flow through the point of
acquisition. Human peripheral blood leukocytes are derived from a
liquid sample and therefore require no manipulation to achieve a
single cell suspension. Samples are prepared or stained, using a
procedure that tags each of the markers of interest with a fluoro-
phore that can be measured by the cytometer. When several fluo-
rescently labeled reagents are used together in one assay, they are
referred to as a cocktail. The cocktail of reagents and their
corresponding combination of markers or panel are identified a
priori in advance of staining cells [51–54]. Optimized, standar-
dized, and harmonized Treg panels have been previously estab-
lished [7, 35, 55–59]. The method for staining cells is
determined by the location of the markers. Markers on the surface
of the cell are readily accessible; however, intracellular or intra-
nuclear markers often require fixation, to preserve the integrity of
the markers, followed by permeabilization, to gain access to the
intracellular or intranuclear compartments.
A flow cytometer has three main components: fluidics, optics,
and electronics [47, 60]. The fluidic system injects the sample into a
stream and introduces the sample, one cell at a time, for interroga-
tion by the LASER(s). The optical system includes one or more
LASERs, to excite the fluorophores from the various reagents
on/in the cells, and collection optics. The collection optics capture
the light emitted from the fluorophores plus light scattered by the
cells based on their size and complexity. The electronic system
converts the photons of light collected by the optics into a pulse,
a digital signal that is interpreted by the software [61]. The digi-
tized signals, collected from all of the various cells in a sample, are
saved into a single data file using a standard format called flow
cytometry standard or FCS file [62–64]. Flow cytometry software
reads the FCS files and enables analysis of the various markers that
are expressed in specific patterns by the cell types present in the
sample.
The optical systems of a cytometer are not foolproof; as a result
some fluorophores are excited by more than one laser, albeit not to
the same degree, and the optical filters specific for one fluorophore
will collect light from neighboring fluorophore(s). This is called
spectral overlap [65, 66]. To deconvolute data collected from the
various fluorophores, a compensation matrix is created by the flow
cytometry software and used identify the fluorescence for one
144 Janet Staats

specific marker from all of the other fluorescent signals present on

the same cell [67, 68]. An operator then verifies the compensation
matrix and analyzes the data by creating a series of sequential
two-dimensional plots, where each plot displays fluorescence inten-
sity of two different markers, X and Y. Some cells express a high
amount of fluorescence, and are said to be positive, and some cells
express low or no fluorescence and are said to be dim or negative,
respectively. The combination of positive and negative populations
creates patterns that are unique for a given population of interest.
Electronic gate(s) or analysis region(s) are placed on the population
(s) of interest based on these expression patterns. The act of placing
gates and regions is called “gating” or analysis. The process of
gating is highly subjective and; therefore, introduces a significant
source of variability for flow cytometry-based assays (see Chapter 5,
[7, 35, 69]. Regions are placed on plots for the purpose of gen-
erating statistical measures from a given population, such as fre-
quency of parent (%) or median fluorescence intensity (MFI). The
statistical data may be exported for subsequent bioinformatics or
summary statistics.
Today, most cytometers are capable of detecting at least four
colors simultaneously. Therefore, the premise of this Treg assay is to
focus on a small set of four essential markers—the base panel. The
required base panel includes CD3, CD4, CD25, and CD127 or
FOXP3, where CD3 and CD4 are anchor markers, to gate on
helper T cells, and CD25 is the key activation marker that is used
with one marker of suppressor function, either CD127 or FOXP3,
to identify Tregs. CD127 is preferred over FOXP3 when a surface
staining assay is more desirable, either to conserve money or time.
While the use of FOXP3 to identify Tregs is well established,
there is precedence in the literature that some clones of anti-
FOXP3 antibodies may not work as well as others, causing Treg
measurements to vary by choice of clone, and the choice of per-
meabilizing agent also introduces variance in the measured results
[28, 70–72]. It is important to note that many of the clonal studies
predated the discovery that humans have different isoforms of
FOXP3; thus, the variation observed between clones may in part
be related to different epitopes expressed by the FOXP3 isoforms.
When using FOXP3 to enumerate Tregs by flow cytometry a
nuclear permeabilizing reagent is required, along with a fixative to
preserve the surface membranes. There are essentially two types of
permeabilizing agents. A gentler solution, used to access intracellu-
lar antigens, and a harsher solution, required to access intranuclear
antigens. Different permeabilizing agents use different mechanisms
to create holes in the cellular membranes. Some reagents are deter-
gent based, permeabilizing cells only when the detergent is present,
and others create permanent holes in the membranes. Both the
clone and permeabilizing agent should be considered potential
sources of variance in Treg assays.
 Immunophenotyping of Human Regulatory T Cells 145

If the cytometer is not limited to four colors and a modestly

expanded Treg panel is more desirable, then the base panel may be
expanded by including a handful of optional Treg consensus mar-
kers; these include a fixable vital dye, CD45RA, CD45RO, CD194,
HLA-DR, and Ki67. Expanding the Treg panel will exponentially
increase the complexity of the assay and may not be necessary. To
help inform the decision to expand the panel, the next few para-
graphs provide guidance for when an expanded Treg panel may
strengthen the overall results obtained from this Treg assay. The
Subheading 3 will include provisions for all of the commonly used
Treg markers, including the essential base panel and optional con-
sensus markers.
There are two approaches to expanding the Treg panel that may
be used together or separately. The first approach is to focus on the
overall accuracy of the Treg measurement, and the second focuses
on the Treg state of being. Accuracy of the Treg measurement can
be improved by combining two different suppressor markers
together in the same panel. The combination of FOXP3 and
CD127 utilizes one marker that correlates positively with suppres-
sion and a second marker that is inversely correlated with suppres-
sive activity. These two markers appear to identify slightly different
Treg populations; albeit, with a high degree of overlap between the
populations identified using CD127 and FOXP3 [73]. Two differ-
ent clones of FOXP3 addresses the issue that not all clones of
FOXP3 perform the same and FOXP3 exists in different isoforms
[74]. The use of two Treg markers is recommended only when it is
critical to hone the accuracy of the true FOXP3+ Treg
measurement.
If you have access to an instrument that is capable of measuring
up to 8-colors simultaneously and the scientific goal is to character-
ize the Treg state of being, then including the optional Treg
consensus markers (a vital dye, CD45RA, CD45RO, CD194,
HLA-DR, and Ki67) will add value to your results. The vital dye
is useful when working with less pristine samples, such as overnight-
shipped blood, cryopreserved/thawed PBMC, or samples collected
from patients with conditions or medications that severely affect
lymphocytes (some chemotherapies and immunotherapies). The
remaining optional markers represent the combined markers from
two unique and highly standardized panels; select one of these two
panels based on your specific needs.
Both of the highly standardized panels were meticulously
developed, optimized, and standardized by teams of expert immu-
nologists who gathered specifically to create a standardized Treg
assay for the purposes of immune monitoring and comparing data
across laboratories and across human clinical trials. The Human
Immunology Project Consortium (HIPC) standardized not only
the panel of markers used to identify Tregs, but also standardized
the assay and approach to data analysis [35, 75]. The HIPC Treg
146 Janet Staats

assay uses a surface staining protocol with lyophilized plates con-

taining a pretitered antibody cocktail that consists of the following
eight markers: a vital dye, CD3, CD4, CD25, CD127, CD45RO,
CD194, and HLA-DR. The HIPC assay is used to measure Tregs in
general, as well as eTregs, aTregs, and mTregs. Of note, the HIPC
assay was designed to measure multiple leukocyte subsets, includ-
ing Tregs; this chapter focuses only on the Treg portion of the
HIPC panels. In a separate and independent harmonization effort,
the CIMT Immunoguiding Program, CIP, systematically com-
posed a ranking of Treg markers, tested the markers to identify
differences and overlap across the most commonly used Treg defi-
nitions, then evaluated the resulting panels for their ability to
measure Tregs in various human tissues, primarily in the context
of neoplastic disease [7]. Like HIPC, CIP also provided a robust
gating strategy to guide the data analysis. The CIP harmonized
panel includes the following seven markers: CD3, CD4, CD25,
CD127, FOXP3, Ki67, and CD45RA. It is used to measure Tregs
in general, as well as eTregs, nvTregs, aTregs, and mTregs. Both
standardized panels include an essential base panel to identify
Tregs. However, the two panels differ in the staining method,
surface versus intracellular/intranuclear, and markers used to iden-
tify the Treg states of being. Note, different markers are used by the
two standardized panels to identify similar states of being. For this
reason, it is critical to use the same definition of a population, both
in terms of markers and gating strategy, throughout the course of a
given study. The essential marker list included in this chapter
encompasses all of the markers in both the HIPC and CIP standar-
dized Treg panels, enabling the use of either panel in the Treg assay
described below.
This chapter describes how to measure human Tregs using
analytical flow cytometry. The Treg assay preparation method
includes a surface stain plus fixation and permeabilization. This
permeabilization method resembles the one used by CIP and has
been reported to be the most reliable intranuclear procedure
[7]. To complete the Treg assay, the preparation method must be
used in conjunction with a panel of markers to identify Tregs. Given
the complexity involved in developing flow cytometry panels, a list
of the most common Treg markers is provided. These include CD3,
CD4, CD25, FOXP3, CD127, a vital dye, CD45RA, CD45RO,
CD194, HLA-DR, and Ki67. The common markers may be used
to in three different combinations or panels. To complete the assay,
select the panel best suited to address the scientific question. The
base panel is required to identify Tregs, and can be configured as a
four-color assay. If the based panel is used with CD127, and not
FOXP3, it is a surface only procedure. If used with FOXP3, and not
CD127, fixation and permeabilization is required. To measure Treg
states of being such as aTreg, mTreg, and the highly suppressive
eTreg, there are two highly standardized panel options, HIPC and
 Immunophenotyping of Human Regulatory T Cells 147

CIP panels. Both standardized panels include the required base

panel to identify Tregs. The HIPC Treg panel utilizes a simple
surface staining procedure, while the CIP panel requires permeabi-
lization and is more challenging technically. For all three panels,
there is the option to incorporate a vital dye, second clone of
FOXP3, and/or both suppressive markers, CD127 and FOXP3.
Note, incorporating these additional options into one of the seven
to eight-color standardized panel options may require a higher
dimensional flow cytometer, with up to 12-color capabilities. The
use of nonhuman or solid tissues, cell sorting, and more advanced
Treg assays, designed measure the plethora of subpopulations of
Tregs, their respective states of being, or function, are beyond the
scope of this chapter.

2 Materials

1. Personal protective equipment, based on institutional

guidelines.
2. Cytometer, capable of measuring at least four colors simulta-
neously, or up to 12 colors for the HIPC or CIP Treg panels.
3. Flow cytometry analysis software.
4. Centrifuge, refrigerated, with buckets to hold staining tubes or
plates, depending on which is used in the assay.
5. Microfuge.
6. FACSWash buffer: PBS containing 0.5% hiFBS. Use a serolog-
ical pipette to remove 2.5 mL PBS from a newly opened bottle
and add 2.5 mL hiFBS to the PBS. Cap the bottle and invert
5 times to mix. Store at 4  C for up to 1 month after prepara-
tion (see Note 1).
7. Blocking agent: commercially available, store according to
manufacturer recommendations until expiration (see Note 2).
8. Red blood cell lysing agent: commercially available, usually as a
concentrate, prepare and store according to the manufacturer’s
instructions (see Note 3).
9. Fixable vital dye stain: commercially available in several colors,
prepare and store according to the manufacturer’s instructions
(see Note 4).
10. Fluorescent conjugated monoclonal Abs (mAb) for surface and
intracellular/intranuclear staining: Essential mAb reagents
include CD3, CD4, and CD25, plus CD127 and/or FOXP3.
Additional antibody reagents include HLA-DR, CD194, Ki67,
plus CD45RA and/or CD45RO. They are commercially avail-
able; store at 4  C until expired (see Note 5).
148 Janet Staats

(a) Basic panel with CD127 includes CD3, CD4, CD25, and
CD127. Requires surface stain only.
(b) Basic panel with FOXP3 includes CD3, CD4, CD25, and
FOXP3. Requires surface and intranuclear stain. CD4,
CD25, CD127, FOXP3, Ki67, and CD45RA.
(c) HIPC panel CD3, CD4, CD25, CD127, CD194,
CD45RO, and HLA-DR. Requires surface stain only.
(d) CIP panel includes CD3, CD4, CD25, CD127, FOXP3,
CD45RA, and Ki67. Requires surface stain and intracel-
lular/intranuclear stain.
11. Nuclear fixation/permeabilizing solution (Fix/Perm): com-
mercially available, usually as a concentrate, prepare and store
according to the manufacturer’s instructions (see Note 6).
12. Permeabilization buffer (Perm buffer): commercially available,
usually comes with the Fix/Perm as a concentrate, prepare and
store according to the manufacturer’s instructions.
13. Compensation beads (see Note 7).
14. 1% Fixative: PBS with 1% formalin. Add 5 mL 10% formalin to
a 45 mL PBS in a 50 mL conical tube, cap and mix by inverting
several times, store at ambient temperature for up to 1 month
after preparation (see Note 8).
15. Vacuum manifold, if staining in plates (see Note 9).

3 Methods

1. Read all steps and Notes prior to performing the assay, as much
advanced preparation is required to perform this assay.
2. Perform all steps on ice or at 4  C unless otherwise noted,
including centrifugation.
3. Store buffers on ice until use unless otherwise noted.
4. Vital dye stain.
(a) Add 100 μL whole blood or viable PBMC viable cells to
the staining tubes or wells (see Note 10).
Note, for the remainder of the assay, through sample
acquisition, the samples must be protected from light and
kept on ice, unless otherwise noted (see Note 11).
(b) Wash cells by adding FACSWash buffer, 2 mL for tubes or
100 μL for wells.
(c) Centrifuge tubes for 5 min or plates for 3 min at 350  g
and 4  C (see Note 12).
(d) Decant tubes or aspirate wells to remove supernatant (see
Note 13).
 Immunophenotyping of Human Regulatory T Cells 149

(e) Add 100 μL vital dye stain at a final concentration based

on pretitration (see Notes 14 and 15).
(f) Mix cells with the vital dye by vortexing tubes for 3 s at a
moderate setting of 3 or pipetting the entire volume in
wells up and down 10 times (see Note 16).
(g) Incubate cells with the vital dye for 30 min at ambient
temperature.
(h) During the vital dye incubation prepare the surface mAb
stain mix, a cocktail of all conjugated mAbs at the optimal
concentrations in FACSWash buffer (see Note 17).
(i) Wash cells by adding FACSWash buffer, 2 mL for tubes or
200 μL for wells.
(j) Centrifuge tubes for 5 min or plates for 3 min at 350  g
and 4  C.
(k) Decant tubes or aspirate wells to remove supernatant.
5. Surface stain.
(a) If using liquid mAbs, microfuge mAb reagent vials for
5 min immediately prior to use (see Note 18).
(b) Add fluorescently labeled monoclonal antibodies (mAb)
to cells at pretitered concentrations in a final volume of
100 μL (see Notes 15 and 19).
(c) Mix cells with surface mAb mix by vortexing tubes for 3 s
at a moderate setting of 3 or pipetting the entire volume in
wells up and down 10 times.
(d) Incubate cells with surface mAb mix for 30 min.
(e) Wash with FACSWash buffer by adding 2 mL in tubes or
100 μL in wells.
(f) Centrifuge tubes for 5 min or plates for 3 min at 350  g
and 4  C.
(g) Decant tubes or aspirate wells to remove supernatant.
6. RBC lyse.
(a) Add ambient temperature 1 lysing solution to cells,
1 mL in tubes or 100 μL in wells (see Note 20).
(b) Mix cells with lyse by vortexing tubes for 3 s at a moderate
setting of 3 or pipetting the entire volume in wells up and
down 10 times.
(c) Incubate cells with lyse for 10 min at ambient tempera-
ture, or according to the manufacturer’s
recommendations.
(d) Wash with FACSWash buffer by adding 2 mL to tubes or
100 μL to wells.
(e) Centrifuge tubes for 5 min or plates for 3 min at 350  g
and 4  C.
150 Janet Staats

(f) Decant tubes or aspirate wells to remove supernatant.

(g) If performing surface stain only assay skip to Fixation step;
otherwise, continue with permeabilization and intracellu-
lar/intranuclear staining.
7. Permeabilization.
(a) Permeabilize cells by adding fixation/permeabilization
reagent, 1 mL for tubes or 100 μL for wells (see Notes 6
and 21).
(b) Mix cells with Fix/Perm reagent by vortexing tubes for 3 s
at a moderate setting of 3 or pipetting the entire volume in
wells up and down 10 times.
(c) Incubate cells with Fix/Perm for 45 min at ambient tem-
perature or according to the manufacturer’s
recommendations.
(d) Wash cells with Perm buffer by adding 2 mL to tubes or
100 μL to wells.
(e) Centrifuge tubes for 8 min or plates for 5 min at 500  g
and 4  C (see Note 22).
(f) Decant tubes or aspirate wells to remove supernatant.
8. Block.
(a) Add 5–15 μL blocking agent, depending upon the man-
ufacturer’s recommendations.
(b) Mix cells with block by vortexing tubes for 3 s at a mod-
erate setting of 3 or pipetting the entire volume in wells up
and down 10 times.
(c) Incubate cells with block for 15 min.
(d) During the block incubation prepare the intracellular/
intranuclear mAb stain mix (see Notes 15, 17, and 23).
9. Intracellular/Intranuclear stain.
(a) Microfuge mAbs.
(b) Add fluorescently labeled monoclonal antibodies (mAb)
to cells in a final volume of 100 μL.
(c) Mix cells with intracellular/intranuclear mAb mix by vor-
texing tubes for 3 s at a moderate setting of 3 or pipetting
the entire volume in wells up and down 10 times.
(d) Incubate cells with intracellular/intranuclear mAb mix for
30 min.
(e) Wash with Perm buffer by adding 2 mL to tubes or
100 μL to wells.
(f) Centrifuge tubes for 8 min or plates for 5 min at 500  g
and 4  C.
 Immunophenotyping of Human Regulatory T Cells 151

(g) Decant tubes or aspirate wells to remove supernatant.

(h) Repeat wash two more times for a total of three washes,
using 2 mL for tubes and 200 μL for wells.
10. Fix.
(a) After removing supernatant add 200 μL 1% Fixative to
cells.
(b) Mix cells with Fixative by vortexing tubes for 3 s at a
moderate setting of 3 or pipetting volume in wells up
and down 10 times (see Notes 8 and 24).
(c) Store at 4  C in the dark until acquisition.
11. Stain compensation bead controls.
(a) Mix compensation beads by vortexing at a setting of 8 for
5 s.
(b) Add 50 μL or one drop of Negative compensation beads
to a staining tube or well (see Note 25).
(c) Add 50 μL or one drop of Positive compensation beads to
separate staining tubes or wells, one for each fluorescent
marker used in the panel (see Note 25).
(d) Add conjugated mAb to each respective compensation
bead tube, matching the Ab concentration used to stain
cells (see Note 26).
(e) Mix beads with mAbs by vortexing tubes for 3 s at a
moderate setting of 3 or pipetting the entire volume in
wells up and down 10 times.
(f) Incubate beads with mAbs for 20 min.
(g) Add 2 mL FACSWash buffer to each tube or well.
(h) Centrifuge tubes or plates at 350  g for 5 min and 4  C.
(i) Decant tubes or aspirate wells to remove supernatant.
(j) Add 200 μL 1% Fixative to each tube or well (see Notes 7,
8, and 24).
(k) Mix beads with 1% Fixative by vortexing tubes for 3 s at a
moderate setting of 3 or pipetting the entire volume in
wells up and down 10 times.
12. Instrument setup and acquisition.
(a) Verify the optical configuration of the cytometer will
accommodate the fluorophores used for Treg assay (see
Note 27).
(b) Empty waste tank (see Note 28).
(c) Fill sheath tank (see Note 28).
(d) Verify cytometer performance is acceptable for use (see
Note 29).
152 Janet Staats

(e) Enter annotation for samples and compensation beads in

cytometer, including samples, conditions, markers, and
fluorophores.
(f) Calibrate cytometer to set fluorescence voltage, scatter
gains, and detection threshold (see Note 30).
(g) Use unstained cells to set collection and storage gates (see
Note 31).
(h) Acquire compensation tubes.
(i) Calculate compensation matrix and apply compensation
(see Note 32).
(j) Acquire cells (see Note 33).
(k) Clean the cytometer (see Note 34).
(l) Transfer data from cytometer to network or analysis work-
station immediately following acquisition (see Note 35).
13. Compensation.
(a) Create a compensation matrix using the compensation
beads (see Note 36).
(b) Apply the compensation matrix to all of the files (see
Note 37).
(c) Using a file that contains all of the markers, verify com-
pensation values are correct by visualizing each X axis
versus all of the respective Y axes (see Note 38 and Fig. 1).
(d) Manually adjust compensation as necessary (see Note 39).
(e) Save the adjusted compensation matrix.
14. Gating.
(a) Follow the gating scheme provided in Fig. 2 (see
Note 40).
(b) Use sequential two-dimensional gating, to create gates for
Time, Singlets (FSC and SSC), and Lymphocytes (see
Chapter 5).
(c) From the Lymphocyte gate, display CD3 versus CD4
and create a CD3+CD4+ gate to identify Helper T
lymphocytes.
(d) From the Helper T cell gate, display CD25 versus either
CD127 or FOXP3, depending upon the panel used.
Create the Treg gate as CD25+CD127low of CD25
+FOXP3+ (see Figs. 2 and 3).
(e) To gate on markers included in the HIPC panel, follow
the gating scheme provided in Fig. 4. Gate on Tregs,
using either CD25 versus CD127 or FOXP3, then gate
on CD194+ to identify effector or eTregs.
 Immunophenotyping of Human Regulatory T Cells 153

Fig. 1 N  N compensation plot. Dot plots in Fig. 1 represent an N  N plot for using to visualize compensation
errors. Overcompensation and undercompensation errors are noted. The over compensation of Zombie from
the HLA-DR creates a subnegative HLA-DR population that is visible in the HLA-DR versus CD45RO dot plot, as
well as other Y axes markers paired with HLA-DR. Subnegative values are usually caused by
overcompensation

(f) From the CD194+ gate, create CD45RO+ and

CD45RO regions to identify memory or mTregs and
naı̈ve or nTregs, respectively.
(g) Also from the CD194+ gate, create HLA-DR+CD45RO
+ regions to identify activated memory Tregs.
(h) To gate on markers included in the CIP panel, follow the
gating scheme provided by Santegoets et al. [7].
15. Controls.
(a) Many types of controls are used in flow cytometry
[76]. The essential controls for multicolor flow cytometry
are compensation controls, single stained controls used to
154 Janet Staats

Fig. 2 Treg gating scheme. Cells obtained from a normal donor were surface stained with the HIPC panel using
lyoplates and acquired on a BD LSRFortessa. CD25 versus CD127 is used to identify Tregs. The gates are
drawn as described for Time, Zombie-, Singlets (FSC and SSC), Lymphocytes, CD3+CD4+ Helper T cells, and
CD25+CD127low Tregs. The proportion of Tregs is 8.23% of Helper T cells, in the normal range of
approximately 5–9%. The gates are listed above each respective dot plot and green arrows indicate the
gating hierarchy from one gate to the next. The number of events are shown for the whole FCS file (925428)
and CD3+CD4+ (labeled CD4+) (177801); 8.23% of 177801 is 14626 Treg events collected from a total of
2  106 viable PBMCs added to the staining well. This is a sufficient number of events to continue with
analyzing Treg states of being, if desired

create compensation matrices to distinguish one fluores-

cence from another.
(b) Hard-dyed microbeads are commonly used to assess cyt-
ometer performance [77, 78]. Hard-dyed beads contain
fluorescence molecules inside the bead itself, shielding it
from many harmful elements found in the microenviron-
ment. The choice for what bead to use and how to use it is
generally up to flow core managers. Instrument manufac-
turers design quality control measures into software
packages to facilitate routine instrument performance
assessments. These are also based on hard-dyed beads.
(c) Controls used to calibrate cytometers include a number of
options ranging from hard-dyed beads to stained beads to
stained cells. These assay-specific settings are up to flow
core managers and sometimes left to the end user’s dis-
cretion. The most commonly used control to calibrate
cytometers is hard dyed beads.
(d) Fluorescence minus one (FMO) controls contain all of the
fluorescence markers in a panel except for one (see Figs. 3
and 5) [79]. For example, a FOXP3 control for a basic
Treg panel would contain a vital dye stain, CD3, CD4,
 Immunophenotyping of Human Regulatory T Cells 155

Fig. 3 CD25 versus FOXP3 Treg analysis region. Cells obtained from an HIV+
patient were stained for surface CD25 and intranuclear FOXP3 (PCH101 clone)
using the optimized panel in OMIP 006 and acquired on an LSRII [56]. The
sample in the left plot is stained with FOXP3 (Y axis). The frequency of Tregs in
the HIV+ sample are low, but close to the lower limit of the normal range. The
control on the right plot is a gating control. It contains all markers except FOXP3,
like an FMO control. However, unlike an FMO control, the gating control includes
an isotype control for PE, used at the same concentration of the FOXP3 mAb. In
this instance the gating control was not helpful due to a shift in the PE negative.
The shift in the PE negative population is an artifact with two possible causes.
Negative drag-up created by improper washing that leaves unbound conjugated
mAbs in solution that is excited by the LASER. Staining fixed/permeabilized cells
with intracellular and intranuclear markers. This example is likely a combination
of the two artifacts. In this example, the PE gate was placed between the visually
discrete populations of FOXP3 and FOXP3+ populations in the sample, called
an internal negative control. If a dimmer fluorophore or different clone for FOXP3
had been used, the resulting FOXP3 staining might not have been so clearly
separated, making the analysis much more challenging. Proper washing and
blocking the intracellular or intranuclear mAb helps to reduce these artifacts

and CD25 but not FOXP3. FMO controls are used to

identify the boundary between positive fluorescence and
negative fluorescence, taking into account any increase in
background from neighboring detectors. They are helpful
controls to use when deciding where to place boundaries
of fluorescence positive and negatives when creating gates
and analysis regions. The downside is that they are costly
to incorporate in terms of both cells and reagent, and my
not always work, as when there is negative drag-up or
sometimes with intracellular/intranuclear markers. At an
absolute minimum, FMO controls should be used during
panel development to inform decisions regarding degree
of interference from spillover.
(e) Internal negative controls are populations of cells that do
not express the marker of interest but are present in a tube
156 Janet Staats

Fig. 4 HIPC panel gating scheme. Cells obtained from a normal donor were surface stained with the HIPC panel
using lyoplates and acquired on a BD LSRFortessa. CD25 versus CD127 is used to identify Tregs. The gates
are drawn as described for Time, Zombie-, Singlets (FSC and SSC), Lymphocytes, CD3+CD4+ Helper T cells,
and CD25+CD127low Tregs. In addition to the Treg markers, the HIPC panel includes markers to measure the
Treg states of being, namely maturation and activation. From the Treg gate, eTregs are identified as CD194+
and from the eTregs memory (CD45RO+) and activated memory (HLA-DR+CD45RO+) Tregs are also
identified. CD25, CD127, CD194, CD45RO, and HLA-DR gates and regions were set using FMO controls.
Figures 2 and 4 are created from the same data file

that is stained with all of the fluorescence markers in the

panel [80]. In Fig. 3, an internal control was used to set
the FOXP3 analysis region when the FMO control failed.
In this example, we are using the biological knowledge
about the sample, that not all Helper T cells express
FOXP3. Internal negative controls only work when
using a mixed population of cells, where some cells
express the antigen of interest and some do not express
the antigen of interest.
(f) Biological controls consist of normal donor cells that are
either fresh, individual samples or in bulk form from
leukopaks or buffy coats, or prepared as a bulk commercial
product. Figures 2 and 4 were obtained using a commer-
cial product derived from stabilized buffy coat leukocytes.
Biological controls provide helpful information regarding
 Immunophenotyping of Human Regulatory T Cells 157

Fig. 5 FMO control for HLA-DR. Cells obtained from a normal donor were surface
stained with the HIPC panel using lyoplates and acquired on a BD LSRFortessa.
CD25 versus CD127 is used to identify Tregs. Markers for maturation and
activation, namely, CD194, CD45RO, and HLA-DR, were included in the panel.
The plot on the left is a sample stained with all markers in the panel (a vital dye,
CD3, CD4, CD25, CD127, CD194, CD45RO, and HLA-DR). The plot on the left is
an FMO control for HLA-DR, stained with all markers in the panel except for
HLA-DR. The X quadrant was placed using the HLA-DR FMO control

the overall quality of the assay performed and can be used

to assess staining, permeabilization, and analysis, depend-
ing upon how the control is designed. When the same
biological control is used across multiple assays, the results
can be used as a measure for overall variability between
assays.
16. Reporting flow cytometry results.
(a) Flow cytometry results are reported using three units:
frequency, absolute counts, and fluorescence intensity.
(b) Frequency is based on the number of cells that express a
particular marker, or combination of markers relative to
gate. Most commonly, frequency is reported as the per-
centage of the parent, meaning a proportion of the parent
gate in the gating hierarchy. For example, Tregs are
reported as a frequency of Helper T cells, CD4+CD3+
lymphocytes. It is possible to report Tregs as a proportion
of lymphocytes; this would be frequency of grandparent.
Most software packages enable users to report frequencies
for a given population as a proportion of any gate in the
gating hierarchy. Frequency is a qualitative measure.
(c) Absolute counts represent a means for converting a fre-
quency to a quantitative measure. Specific counting con-
trols are required to report values in absolute counts or
cells/μL [81–84]. Counting controls are added to the
158 Janet Staats

original sample during the staining process. This is called a

single-platform method, meaning both the counting and
the frequency measures are obtained using one instru-
ment, the flow cytometer. Reporting Tregs as cells/μL
requires both the use of whole blood samples and count-
ing controls. A second, more variable dual-platform
method, may be used generate absolute counts using the
flow cytometry-derived frequencies and the hematology-
derived absolute blood count and differential. This
method requires a means of relating the flow cytometry
analysis to the whole blood used in the hematology assay,
usually the percentage of lymphocytes.
(d) Fluorescence intensity is directly proportional to the den-
sity of the antigen expressed on or inside a cell. The units
of fluorescence intensity, median or mean (median/mean
fluorescence intensity or MFI), are qualitative. With the
appropriate controls, MFI values can be converted to
quantitative units of antibody binding, such as antibody
binding capacity (ABC) or molecules of equivalent fluo-
rescence intensity (MESF) [85–88].

4 Notes

1. Flow cytometry wash buffers: at a minimum include must be

buffered to neutrality and isotonic, such as PBS (calcium and
magnesium free). Common ingredients included in flow cyto-
metry wash buffers are FBS, EDTA, sodium azide, and
DNAse/RNase. Protein, usually in the form of FBS, is added
to protect cells from apoptosis, prevent heterophilic antibody
interference, and block nonspecific binding of antibodies
[89]. FBS must be heat-inactivated prior to use. Heat inactiva-
tion removes complement proteins that might otherwise cause
unwanted cell lysis. Heat-inactivated FBS is commercially avail-
able or may be prepared in the lab by placing FBS in a 56  C
water bath for 1 h. Heat-inactivated FBS is aliquoted and
frozen for up to 1 year or until expiration. Each aliquot can
be subjected to a total of three freeze–thaw cycles, then discard.
Use FBS at a final concentration from 0.5% to 10%, depending
upon how much blocking is needed. EDTA prevents cell
clumping and is most useful when specimens are clumpy, such
as derived from solid tissue or subjected to cryopreservation
and thawing. If EDTA is needed, use at a final concentration
from 0.5 mM to 5 mM. Sodium azide is a preservative. It also
inhibits capping, antigen internalization and shedding of sur-
face antigens. If sodium azide is needed, use at a final concen-
tration of 0.1–1%. If the sample contains dead cells, usually
 Immunophenotyping of Human Regulatory T Cells 159

seen in disaggregated tissue or thawed cells, the sample will also

contain a fair amount of extracellular DNA and RNA derived
from the dead cells. DNA and RNA are sticky and lead to cell
clumping. To remove extracellular DNA and RNA, add DNAse
and RNase to the wash buffer at a final concentration of
25–50 μg/mL.
2. Blocking agents contain immunoglobulin and are used to bind
Fc receptors present on cells. They may be comprised of simply
serum or a mixture of antibodies specific to Fc receptors.
During flow cytometry assays, that use antibodies as a tool to
stain cells, the Fc portion of the staining antibody (Ab) may
also bind to Fc receptors on the surface of cells expressing FcR,
causing increased background or nonspecific staining. When
fluorescently tagged Abs bind via the FcR and not binding their
specific antigen on the cell, the resulting expression pattern is a
diagonal streak across the dot plot, running from roughly the
double negative region of the plot toward the double positive
region of the plot. The nonspecific binding or higher back-
ground interferes with the ability to accurately classify cells as
positive or negative during analysis. For this reason, Fc recep-
tors are often blocked during the staining process. Some cell
types express more FcR than others and some patient cells are
more prone to nonspecific binding. Blocking is especially
important for intracellular/intranuclear staining, such as
FOXP3.
3. Red blood cells or RBCs, when present in samples, obfuscate
white cells during analysis and increase the overall variability of
the measurements. Obvious red cell contamination, and even
observing a small number of red cells in the cell pellet, warrants
consideration to remove RBCs. There are two approaches to
removing RBCs. The first method is to lyse RBCs and the
second is to use CD45 gating, either with or without
CD235a. There are many commercially available lysing solu-
tions and most work by increasing osmotic pressure inside the
cell through a net influx of NH4Cl, causing the red cells to
burst. Most commercially available lysing agents are sold as
concentrates and may or may not include a fixative. Because
lysing agents work by changing osmolarity, it is important to
read the manufacturer’s instructions carefully and use the
appropriate diluent, usually either water or PBS. Using the
inappropriate diluent or not diluting properly, may cause an
overlysis or underlysis. Overlysis may also lyse some leukocytes
and under lysing results in red cells remaining after treatment
with the lysing agent. Both over and under lysing can change
the observed scatter properties on a forward versus side scatter
plot. Lysing agents do not work well with nucleated RBCs.
Some samples contain nucleated RBCs, such as those obtained
160 Janet Staats

from pediatric patients, patients receiving chemotherapeutic

agents or radiation, or patients who might be reticulating for
other reasons. Nucleated RBCs do not lyse well. When nucle-
ated red cells are present or when there is a need to avoid lyse in
general, one can use specific markers to identify white cells and
red cells through gating. CD45 is a pan-leukocyte marker and
can be used to effectively to identify lymphocytes when com-
bined with either SSC. CD235a is a red blood cell marker, can
be used to remove red cells making it easier to identify the
CD45+ white blood cells. The addition of CD45 is preferred
when using samples that are derived from tissue that contains
scarce leukocytes or when accuracy of the resulting measure-
ment is paramount.
4. Use a fixable vital dye to exclude dead cells [90, 91]. Small-
molecule dyes such as PI, AO, and 7AAD, will not work with
the fixative used in Fix/Perms or the 1% Fix. Small molecule
dyes are also notorious for binding the tubing inside of the
cytometer tubing, causing a significant increase in background
and decrease in signal-to-noise ratio.
5. A list of the most commonly used Treg markers is provided,
including consensus markers used to identify Treg states of
being. The reagent used to identify each respective marker is
a fluorescent conjugated mAb. For a given study, optimal
results are obtained when using the same clones, fluorophores,
and lots for each respective reagent.
(a) Most of the commercially available mAb clones are accept-
able for use; however, the choice of FOXP3 clone matters
[28, 69–71]. To know which anti-FOXP3 clone works
best for your particular application, it is highly recom-
mended that you compare anti-FOXP3 clones using
your own samples.
(b) Not all fluorophores are equal and it is important to
understand how the fluorophores work and how to obtain
the best performance for each dye [92–96]. The process
of determining which fluorophore is conjugated to each
respective mAb is called panel design and requires knowl-
edge about the fluorophores, expression pattern of each
marker (how it relates to the other markers of interest,
whether it is coexpressed on the same cell or on a different
lineage), plus knowledge about the optical configuration
of the cytometer to be used [51, 53, 97]. The general rule
for panel development is that markers with a high antigen
density, such as CD3 and CD4, are conjugated to fluor-
ophores with dim or moderate fluorescence intensity,
while markers with low antigen density, such as FOXP3
and CD127 low on Tregs, are conjugated to fluorophores
 Immunophenotyping of Human Regulatory T Cells 161

with bright intensity. Additionally, markers that are coex-

pressed on a single cell, such as CD4 and CD3, should be
conjugated with fluorophores that are on opposing ends
of the fluorescence spectrum and on different LASERs.
Overlapping fluorescence emission from markers
expressed on the same subset leads to spreading error
and masks the low or dim fluorescence positive cells.
Because there are many fluorophores with emission in
the 600 nm range, spreading error is most likely to happen
when using fluorophores within the 600 nm range. A
proper panel design minimizes the effect spreading error
has on the ability to classify cells as positive or negative
appropriately [51, 56, 97]. Investigators must work with
their flow core manager to identify fluorescent conjugates
that work on the cytometers available and optimize panel
design.
(c) For the HIPC panel, the combination of fluorophores has
been carefully designed to work well for most cytometers
[35, 75]. The HIPC panel is available commercially in a
lyoplate format that contains prealiquoted cocktail mix of
all fluorescently conjugated mAbs in a lyophilized pellet.
The prealiquoted kit eliminates the need for titering mAbs
and reduces overall pipetting time and error.
(d) Should there be a need or desire to develop a panel that is
not listed in this chapter, some pointers for selecting
markers to include in a panel are provided: Begin by
stratifying markers into four levels of priority. The first
priority is the end points of the assay; these markers repre-
sent the values that are reported in the final results. When
simply measuring the presence of Tregs, the end-point
markers are CD25 with either FOXP3 or CD127
(or both). However, there are instances in the literature
that suggest that it is not the presence of Tregs but their
state of being or functional activity that is clinically rele-
vant. In this instance, markers for Tregs and markers for
their respective state of being would be considered the
first tier of markers, including CD194, HLA-DR, Ki67,
and CD45RO and/or CD45RA. The second priority tier
is gating. Since Tregs are a subset of helper T cells, CD4
together with the T cell receptor, CD3, is used as anchor
gates to identify the subset of lymphocytes that contain
Tregs. Arguably, if the Treg state of being is the scientific
end point for a given study, then one might consider the
Treg markers CD25, FOXP3, and/or CD127 as the sec-
ond priority tier. However, since they are essential to
scientific end point for Treg state of being, there are
included in the first tier. The first and second priority
162 Janet Staats

tiers always make it into the final panel; therefore, our

minimal recommended panel includes CD3, CD4,
CD25, and FOXP3 and/or CD127. The third priority
tier includes cleanup markers—dead cell markers to
exclude nonviable cells that are present in less pristine
samples, such as whole blood that has been shipped over-
night, cryopreserved cells, disaggregated tissue or samples
from patients with conditions or medications that severely
affect lymphocytes, such as chemotherapy or immu-
notherapies. The third priority tier is sometimes included
in the panel but not always; this is determined by instru-
ment configuration, available reagents/conjugates, time
to develop the panel, and overall costs. The final priority
tier contains subcategories that are based on findings in
the literature, preliminary data, and finally hunches. The
fourth priority tier is rarely included in the panel; this is
determined by instrument configuration, available
reagents/conjugates, time to develop the panel, and over-
all costs. No fourth priority Treg markers are included in
this chapter.
(e) Knowing what how to treat each type of fluorophore is
critical to a successful flow cytometry assay [90, 92–
96]. In general, fluorescence conjugated reagents should
be kept in the dark, on ice, and in neutral pH solutions,
irrespective of whether the fluorescence is inside the mAb
vial or in a staining tube. Tandem dyes, such as PE-Cy7,
degrade easily and must be treated with great care, includ-
ing during transport. A degraded tandem dye might work
well when stained by itself. However, the only way to
know if a tandem dye has degraded is to stain cells or
beads with the dye and observe how much, if any, fluores-
cence is present in the fluorescence donor; this is PE in the
PE-Cy tandem. Also, many fluorophores are sensitive to
fixatives. When using fixatives, test the fluorophores with
the fixative to know how it might affect the fluorescent
results.
6. The choice of staining protocol may lead to very different
results; for this reason, it is highly recommended that you
compare nuclear permeabilization reagents using your own
samples and your FOXP3 clone of choice to select the reagents
that work best for your assay. Permeabilizing agents are
designed for different purposes. There are permeabilization
reagents for accessing intracellular antigens and intranuclear
antigens. Permeabilization reagents for intracellular antigens
will not work well for intranuclear markers, while permeabiliza-
tion reagents for intra-nuclear markers work well for intracellu-
lar markers. Both types of permeabilization reagents may affect
 Immunophenotyping of Human Regulatory T Cells 163

expression of surface antigens. FOXP3 is a transcription factor

and an intranuclear permeabilization reagent is required to
measure FOXP3 expression. Ki67 is present in the cytoplasm
and an intracellular permeabilization reagent is required to
measure Ki67 expression. Intranuclear permeabilization
reagents also work for measuring Ki67. Nuclear permeabiliza-
tion buffers have been compared in the literature; the
eBioscience FOXP3 permeabilization buffer is currently the
most widely used for Treg assays [28, 70, 72]. Fixation is
used together with the permeabilization reagent to maintain
integrity of surface membrane and markers. Read the manu-
facturer’s recommended instructions carefully when using per-
meabilization reagents and buffers. If the wrong diluent is used
to create the Fix/Perm or Perm Wash, or if the wrong buffer is
used to wash, then the permeabilization result may be subopti-
mal and yield poor FOXP3 or Ki67 staining. Prepare the Fix/-
Perm and Perm buffer fresh and discard unused portion after
using.
7. To conserve cells and for optimal signal, compensation beads
are used in place of cells as single-stained controls for compen-
sation [66–68, 79, 92, 98]. Compensation beads bind the same
fluorescent conjugated mAbs used for staining cells via the Fc
portion of the mAb. There are positive and negative compen-
sation beads. Positive compensation beads bind mAb and are
positive for the respective Ab used to stain the beads. Negative
compensation beads do not bind mAb and therefore have no
fluorescent tag; they are used as a negative control. Positive and
negative compensation beads can be included in the same
staining tube or separately. When positive and negative com-
pensation beads are stained in the same tube, then a positive
and negative gate must be created for the given fluorescence in
each tube. When positive and negative compensation beads are
stained in separate tubes, then the negative can be used as a
universal negative during compensation. A universal negative is
the negative for all fluorophores used in the given panel and
saves gating time during analysis. Binding mAb to beads is not
the same as binding mAb to cells; include some protein in the
compensation bead wash buffer, and a fixative, such as 2–5%
FBS, to help stabilize the mAb binding to beads. Otherwise,
the fluorescence from the positive compensation beads might
appear as a smear instead of a single bright peak. Compensation
beads are smaller and less complex than cells; for this reason,
the FSC and SSC detector settings will likely be different
between compensation beads and cells. However, the proper
use of the compensation beads requires that the fluorescent
detector settings for the beads be identical to the settings used
for cells acquired in the same experiment. The fixable vital dye
164 Janet Staats

is not an mAb but rather binds to free amines inside or on the

cell. There are special compensation beads available to use with
the fixable vital dyes, called arc beads. Arc beads are a separate
product from standard compensation beads. They are also
smaller than the regular compensation beads, but can usually
be seen using the same scatter channel settings as the regular
compensation beads when the settings are adjusted properly.
8. Fixation is not a necessary part of the assay, but can be very
helpful when used as the buffer cells are stored in after staining
and prior to acquisition. Fixatives kill cell-associated infectious
agents, when cells are stored in fix for at least 24 h, and are
required by some facilities based on the type of sample being
used and the relative biological safety risk. Fixatives also stabi-
lize binding of fluorescent conjugated mAbs to cells. If acqui-
sition is delayed beyond 6 h, fixation will preserve the
fluorescence intensity and integrity of staining if present during
the delay. Fixatives can increase the level of autofluorescence
observed in otherwise unstained cells and some fluorescent
dyes are sensitive to some fixatives. For this reason, testing
the combination of fluorescent dyes and fixatives is recom-
mended prior to using the combination with test samples.
9. A vacuum manifold is used to aspirate supernatant from wells.
Flicking plates to remove supernatant leads contamination
between wells, higher backgrounds, and overall increased
assay variability from inconsistent washing. To ameliorate pro-
blems caused by flicking plates, use a vacuum manifold. Vac-
uum manifolds standardize supernatant aspiration when using
plates and have been shown to reduce background noise and
improve overall assay variability [99, 100]. Using vacuum
manifolds properly requires practice. Setting the vacuum pres-
sure inappropriately is the most common problem; pressure set
too high aspirates cells along with the supernatant and pressure
set too low does not remove all of the supernatant which leads
to increased backgrounds or drag-up of the negative. Check
the total number of events in each resulting FCS file to know
whether there is any unexpected cell loss. If the assay has been
performed properly, and the manifold used correctly, the files
will contain a consistent number of events for a given sample. If
the pressure is too low, then supernatant remains after aspira-
tion, leading to high backgrounds and increased assay variabil-
ity. Not holding the manifold perpendicular to the plate creates
uneven pressure when aspirating the supernatant, such that
some wells do not receive enough vacuum pressure; this also
leads to increased variability. To set the vacuum properly, add
water to a tray, place the manifold tip just under the surface of
the water, not resting on the bottom of the tray. While holding
the manifold in position, very slowly turn on the vacuum
 Immunophenotyping of Human Regulatory T Cells 165

pressure. Apply as little pressure as possible, then very slowly

increase pressure until the water just begins to aspirate. The
pressure should be set so that there is just enough pressure to
pull the water gently from the tray. Avoid pressures high
enough to aspirate the most of the liquid at once. After the
vacuum pressure is set, and before using the manifold to aspi-
rate actual samples, practice using the manifold with a plate. To
practice with a plate, fill the wells in a plate with water and use
the manifold to aspirate water from the wells in the plate.
Observe or measure the residual volume remaining in wells
after aspirating the supernatant. If the volumes are too high
the pressure is too low, if the volumes are too low the pressure
is too high, if the volumes are inconsistent, then the manifold is
not properly positioned relative to the plate during aspiration.
Continue practicing with the manifold until the residual
volumes for all wells are as expected. Each manifold has a
known residual volume to expect; consult the manufacturer
for more details. Remove water from the manifold tip before
using to aspirate supernatant from cells by wiping the tips with
clean gauze. Residual water may lyse cells. If not used properly,
vacuum manifolds may pose a safety risk. To minimize the
safety hazard, do not apply more vacuum pressure than what
is needed to aspirate the supernatant and implement a holding
chamber, to house the aspirated supernatant waste between the
vacuum line and source of sample. Consulting with institu-
tional safety is recommended before installing or using a vac-
uum manifold.
10. Cell staining may be performed in either tubes or plates. The
number of cells to add per tube or per well ranges from
0.5  106 to 4  106 viable leukocytes, depending upon the
number, type, and location of the markers. The number of cells
recommended is based on the fact that the normal range for
Tregs is approximately 5–9% of CD4 lymphocytes, meaning
the gating hierarchy or the denominator influences the number
of cells to be added, smaller denominators require more cells.
For an assay that measures Tregs (CD3, CD4, CD25, and
CD127 or FOXP3), 0.5–1.0  106 viable cells is the minimum
number of cells to add per tube or well. In normal blood the
number of white cells range from approximately 1–2  106/
mL, meaning that 250–500 μL of whole blood will be needed
to achieve 0.5  106 cells. For Treg assays that measure surface
expression of CD127 and do not require permeabilization,
0.5  106 cells is sufficient. Permeabilization causes the cells
to be more buoyant and increases cell loss; therefore, 1  106
cells is the minimum for an assay that utilizes the intranuclear
marker FOXP3 to identify Tregs. To measure Treg states of
being, the end value to report is a proportion of Tregs, a
166 Janet Staats

smaller subset of CD4 lymphocytes. To accurately report Treg

states of being a minimum of 2–4  106 viable leukocytes is
needed. Use two million cells for the HIPC surface assay and
four million for the CIP intracellular/intranuclear assay. To
recover a sufficient number of cells for analysis, add at least
the recommended minimum number of cells (more when
available), wash when necessary but avoid unnecessary wash-
ing, and permeabilize only when necessary.
11. Keep fluorescent reagents and samples stained with fluorescent
reagents protected from light at all times when not directly
using them. Fluorophores are sensitive to light and if exposed
will result in diminished fluorescence intensity.
12. Do not leave the centrifuge unattended while cells are in the
centrifuge. Do not leave cells in the centrifuge after spinning.
Cells left in a pellet become hypoxic rapidly, inducing cell
death. Leaving cells on a pellet longer than necessary is a
leading cause of poor-quality flow cytometry assay results.
Dead and dying cells may still stain fluorescent markers; how-
ever, their expression may be variable in intensity, and/or
falsely negative or positive. Avoid unnecessary cell death after
centrifugation by working as quickly as possible to remove cells
from the centrifuge, remove the supernatant, and resuspend
cells. Use the centrifugation and incubation times to prepare
for the next step in advance so that cells do not have to sit on a
pellet waiting for the next step. Advanced preparation enables
one to resuspend cells in the next solution quickly and effi-
ciently, improving viability and flow cytometry results. To help
with troubleshooting, position tube labels opposite from the
rotor when putting tubes into the bucket and the bucket into
the centrifuge (see Note 13). After centrifugation, the cell
pellet will be positioned toward the side of the tube that is
furthest away from the rotor. If tubes are positioned as noted,
the pellet will be located on the side of the tube with the label.
Follow these recommendations for all centrifugation steps.
13. Do not disturb the cell pellet when removing the tubes or plate
from the centrifuge or when removing the supernatant. After
centrifugation, visually inspect each tube or well for a visible
pellet. If no pellet is observed in any tube or well, do not decant
the supernatant and consider repeating the centrifugation step.
Check the centrifuge settings. When repeating a centrifugation
step, first resuspend cells by vortexing or pipetting; this will
minimize the time cells are left on a pellet. If using tubes, note
the position of the cell pellet prior to decanting the superna-
tant. If there are too few cells to see a pellet, the pellet can
still be identified by marking the tube relative to the rotor
(see Note 12). To minimize cell loss, decant the supernatant
 Immunophenotyping of Human Regulatory T Cells 167

from the tubes with the pellet positioned on the top of the
tube. Follow these recommendations for all decanting steps.
14. Practice using pipettes prior to using them to pipette small
volumes of vital dye or antibodies. When pipetting smaller
volumes check the pipette tips to verify that the volume is
accurate. Quality tips are graduated and have lines for verify
the volumes. After dispensing the liquid verify that the liquid is
no longer in the tip. If pipettes are not used properly the
appropriate amount of liquid is not added stain the cells,
resulting in variable results, usually missing or lower staining
than expected. Refer to manufacturer instructions for more
details. Calibrate pipettes every 6 months to ensure accuracy.
Follow these recommendations for all steps that involve pipet-
ting small volumes.
15. Fluorescent reagents, including vital dyes and mAbs, should be
pretitered, under the conditions of the assay, to determine the
optimal concentration for each reagent [101, 102]. Optimal
staining of vital dyes and fluorescent conjugated mAbs is
achieved by using saturating amounts. Saturation is deter-
mined through titrations based on mass, usually performed
using serial twofold dilutions. Reagents to be used for surface
stain must be titered as a surface stain using the same tempera-
tures, wash buffer and fixative as will be used in the final assay.
Reagents to be used for intracellular or intranuclear stain must
be titered as intracellular or intranuclear stain using the same
temperatures, Fix/Perm, Perm buffer, and fixative as will be
used in the final assay.
16. Mix cells thoroughly after adding each reagent added. When
mixing cells avoid creating air bubbles as these increase surface
tension, leading to cell death. After mixing cells, visually
inspect the tubes or wells to see if a pellet remains. If a pellet
is observed, cells were not mixed properly and should be
remixed until no pellet remains. If cells are not mixed properly,
the stain results will be poor quality. In addition, cells that are
not resuspended properly are left on a pellet, causing increased
cell death and assay variability. Avoid overmixing cells, they
only need to be mixed enough to resuspend the cell pellet in
the added solution. Overmixing might also result in poor
quality results.
17. Create mAb mixes immediately prior to use. Do not allow
them to set longer than 30 min prior to use. Use the appropri-
ate buffer to create mAb mixes, FACSWash buffer for surface
stain mix and Perm buffer for intracellular/intranuclear stain
mix. If more than one brilliant dye is used to stain cells, then a
commercially available stabilization buffer that is specific for
brilliant dyes should be used to prepare the mAb mix. The mAb
168 Janet Staats

mix should be prepared such that a final volume of 100 μL will

include all of the mAbs at their optimal concentration when
added to the cells in their residual volume. Residual volume is
usually around 35 μL but should be measured for each opera-
tor to ensure accuracy. If using preconfigured lyoplates, such as
for the HIPC panel, follow the manufacturer’s instructions for
staining.
18. Microfuge the mAbs to pellet aggregates of the antibody
and/or fluorophore. When pipetting mAb from the reagent
vial, avoid disturbing the pellet and pipette liquid from the top
of the vial, avoid touching the pipette tip to the bottom of the
vial. Fluorophore aggregates appear in a freckling pattern
higher than the cell staining and increase assay variability.
19. Surface staining mAb reagents include CD3, CD4, CD25,
CD127, CD45RO, CD45RA, HLA-DR, and CD194,
depending upon the preferred Treg panel. CD194 is CXCR4.
Chemokines recirculate in the membrane when incubated at
37  C; for this reason, optimal staining for chemokines is
observed when prestaining at 37  C for 10 min prior to
continuing with the remainder of the surface stain [101]. If
using prealiquoted lyoplates, follow the manufacturer’s
instructions for staining.
20. Lysis is required when using whole blood but optional for
PBMCs. Lysis, when used properly, improves results by reduc-
ing noise (see Note 3).
21. Intranuclear permeabilization is required for FOXP3 detec-
tion. Intracellular permeabilization is required for Ki-67.
22. Permeabilized cells are more buoyant than cells that have not
been permeabilized. Thus, increased centrifugal force and time
are needed to pellet permeabilized cells. More harsh Fix/Perm
reagents may necessitate additional centrifugal force and time.
If you recover overall fewer cells than expected in the final data
files, first verify that the operator is acquiring all of the sample,
then verify that the postperm centrifugation is set to a higher
force and longer time.
23. Intracellular/intranuclear mAb reagents include FOXP3 and
Ki76. CD3 and CD4 may also be added during the intracellu-
lar/intranuclear staining step, if using clones that are made
against fixed antigens, such as the SK7 and SK3, respectively,
and the fluorescent conjugates are titered for intracellular/
intranuclear usage. The buffer used to create the intracellu-
lar/intranuclear mAb mix must be the Perm buffer and not
FACSWash buffer. If more than one brilliant dye is used, then
the commercially available brilliant dye stabilization buffer
must be added, comprising 10–20% of the final mix volume.
 Immunophenotyping of Human Regulatory T Cells 169

24. Resuspending cells in a fixative protects samples in case longer

storage time is needed and inactivates infectious agents. If not
resuspending cells in fixative, reconstitute cells in FACSWash
buffer for surface stained cells and Perm buffer for intranuc-
lear/intracellular stained cells.
25. Compensation beads by be stained and acquired in either tubes
or plates. If using lyoplates, reconstitute beads according to the
manufacturer recommendations.
26. Optimal bead staining may require more volume of mAb than
what is used to stain cells. The amount of mAb needed to stain
compensation beads should be predetermined during panel
development. The fluorescence intensity of beads should be
the same as or higher than the fluorescence intensity of cells
when stained with the same mAb.
27. Compare LASER (excitation) and filter (emission) optics
against the suggested excitation and emission spectra for each
fluorescence dye used in the panel. Flow core operators and
manufacturers can provide you with the optical configuration
of their equipment. In addition, most companies that sell
fluorescent conjugated mAbs offer technical support to facili-
tate panel design as well as on-line charts and tools to help
identify which optics are compatible with their specific fluores-
cent products. Ideally, the process of comparing collection
optics on the cytometer with fluorescent reagents occurs both
prior to panel develop to guide the process and during the
assay, to confirm that each respective fluorophore is being
excited by the appropriate LASER and collected in the appro-
priate detector.
28. Begin each acquisition with full sheath tank and empty waste
tank. Depending upon the instrument used, when a tank is half
full or less the sheath pressure changes, causing fluorescence to
shift and the event rate to change. These changes affect the
resulting data by increasing variability.
29. Verify instrument performance by confirming that the perfor-
mance criteria are within acceptable limits based on established
ranges. Instrument performance criteria and their respective
acceptable ranges may vary from lab to lab. If there are no
established performance criteria for a given instrument, Per-
fetto et al. describes a very useful method for qualifying instru-
ment performance [77].
30. Set fluorescence PMT voltages by running a fluorescent stan-
dard control and setting the median fluorescence to a prede-
termined value within the acceptable range. Instrument
calibration is derived from the instrument verification used by
manufacturers or as described by Perfetto et al. [77, 78]. Detec-
tor settings for scatter (FSC, SSC), as well as the detection
170 Janet Staats

threshold, is set by running unstained cells and placing the

lymphocytes on scale. The entire lymphocyte population
should be visible, separate and distinct from debris, monocytes,
and dead cells. Do not exclude lymphocytes by setting the
detection threshold too high. Calibrating the instrument
before each experiment (and prequalifying your reagents)
ensures positive fluorescence and scatter settings are standar-
dized. Standardized data enable centralized gating, applying
the same gates to all files across all experiments, and improve
assay reproducibility.
31. Set collection gate to acquire 25,000–50,000 CD3+ lympho-
cytes for Treg assays or 50,000–100,000 CD3+ lymphocytes
for Treg state of being assays. This may mean that the entire
volume of the tube is aspirated. The increased number of cells
acquired is necessary for HIPC and CIP panels to evaluate
states of being with some degree of statistical confidence. Do
not hard-gate data, meaning collect all of the events and not
just lymphocytes, as it may be necessary to adjust the lympho-
cyte gate during analysis.
32. Compensation can be set either during or after acquisition. It is
not necessary to set compensation during acquisition; how-
ever, compensating data during acquisition facilitates collecting
a predetermined number of cells, such as CD3+ lymphocytes.
33. Clogs and air bubbles are the most common problem encoun-
tered when acquiring samples on a cytometer. They can render
the data in a file uninterpretable. While they are not entirely
preventable, their frequency and effect on resulting data can be
minimized. To avoid clogs, inspect all samples after staining
and before acquisition. All samples with any visible clumps
should be filtered prior to acquisition. To avoid bubbles, fill
the sheath tank and empty the waste tank prior to each acqui-
sition. Also, be vigilant at keeping track of the event rate during
the acquisition of each tube or well. If event rate changes
dramatically (higher or lower), or declines steadily over three
consecutive refreshes, then there is a clog or air bubble. When
this happens, stop the acquisition and troubleshoot to remove
clog or air bubble before continuing. Observing the event rate
is the only opportunity to identify and correct acquisition
problems before they are permanently recorded into data files.
34. After each acquisition, stringently follow either the manufac-
turer or lab-specific cleaning protocol. Flow cytometers are
fluidic instruments containing salt solutions. For the fluidics
to work properly they must be cleaned regularly. After cleaning,
place a tube filled half-full with deionized water on the sample
port. Do not allow the sample port to dry. When they are not
cleaned regularly or properly, the number of clogs from cells,
 Immunophenotyping of Human Regulatory T Cells 171

parts of cells, or salt solutions increase significantly and are

more severe when they occur, often requiring service engineers
to correct.
35. The biggest risk of data loss occurs while the data are located
only on the cytometer workstation. To minimize the risk of
data loss, transfer data files from the cytometer either during
acquisition or while cleaning the cytometer. Verify all files are
successfully transferred immediately after transferring them
from the cytometer workstation. All data files are removed
from the cytometer workstation during regular maintenance
procedures, usually each month.
36. Ensure that there is a stained compensation bead for each
fluorophore used in the panel [67, 68, 79, 92, 98]. In order
for the compensation matrix to be calculated properly, confirm
that the parameter selected in the file matches the appropriate
channel for the marker that was stained. For example, when PE
is collected in the Green E detector on the cytometer, then the
data file for PE compensation bead should use the Green E
parameter. If Green D parameter is selected for the PE com-
pensation bead, then there is mismatch and the resulting com-
pensation will be inaccurate. Some software have
autocompensation calculations that can make these types of
mismatch errors when the compensation bead is brighter in
the secondary detector (Green D in this example) than it is in
the primary detector (Green E in this example). Staining that is
brighter in a secondary channel than in the primary channel
happens when the instrument is not calibrated properly and
necessitates significantly more compensation, if the primary
signal is to be resolved (see Notes 29 and 30)
37. Check that the appropriate compensation matrix is applied to
all files. Some software packages are notorious for changing
compensation matrix, especially reverting back to an acquisi-
tion matrix. It is helpful to show the compensation in the
workspace window throughout the entire analysis process, to
ensure that the proper compensation matrix has been applied
to each file.
38. To verify compensation is set properly, use N  N plots,
meaning to view all X parameters versus all Y parameters in a
given file simultaneously. Some software packages generate
N  N plots automatically, if not then the plots need to be
created manually. When compensation is set properly the
populations appear symmetrical in a two-dimensional dot
plot. Compensation errors are visible as a “hook” or parabola,
where positive events are turning either toward or away from
the respective axis. If the hook is away from the axis (toward the
double positive quadrant), then the compensation error is said
172 Janet Staats

to be undercompensation. If the hook is toward the axis, then

the compensation error is said to be overcompensation. If a
sample appears to be both over and undercompensated simul-
taneously, and the error remains even after adjusting compen-
sation, this is spreading error. Spreading error cannot be
corrected by adjusting compensation; it can only be prevented
by using quality panel design [51–53, 97].
39. Correct undercompensation by increasing the compensation
value. Correct overcompensation by decreasing the compensa-
tion value. Adjust compensation manually by using the match-
ing means method. In a two-dimensional dot plot, with X
fluorescence versus Y fluorescence, view the cell population
along the X axis first. There are cells that are positive for X
and negative for X. The Y mean of the cells that are positive for
X should match the Y mean for the cells that are negative for
X. For mutually exclusive populations, like CD4 or CD8 T
cells, the populations should appear symmetrical when com-
pensation is set accurately. Repeat this process for each combi-
nation of X versus Y for all parameters in the file. Because
compensation from one parameter affects other parameters in
the file, it may be necessary to repeat the entire process of
checking all combinations of X versus Y for two to three passes.
Biexponential transformation helps to visualize the negative
[103–105]. When biexponential transformation is calculated
in the presence of a severe overcompensation error, then the
resulting biexponential display can create a valley artifact,
where the negative population becomes merged with the posi-
tive population [106]. Correct the compensation error then
recalculate the biexponential transformation to correct the
valley artifact. The more markers used in a panel, the more
complex and time-consuming compensation will be. A well-
designed panel [together with appropriate instrument settings
(see Note 29 and 30)] requires much less manual tweaking of
the compensation matrix. If help is required for compensation,
core operators or technical applications specialists can provide
assistance.
40. Data analysis, including compensation, as well as the placement
of gates and analysis regions, is the single largest source of
variability in flow-based assays [69, 99, 107–109]. To amelio-
rate overall assay variability, great care has been taken to stan-
dardize panels, such as the HIPC and CIP panels, reagents
such as the HIPC lyoplates, and instruments through calibra-
tion. After standardizing panels, reagents, and instruments, the
cell populations should fall in the same relative positions for
each dot plot displayed during analysis. The conundrum is that
the actual placement of gates and regions, to separate negative
and positive populations, is highly subjective and relies upon
 Immunophenotyping of Human Regulatory T Cells 173

the expertise of the operator performing the analysis. To

reduce the variability associated with the placement of gates, a
gating template or uniform gates may be employed. Gating
templates are template file formats, created in the analysis
software and applied across multiple files and/or multiple
experiments without moving the gates. Uniform gating is
achieved by applying the same gates across all samples from a
given donor. In either case, there are exceptions to the rule.
Time gates are changed with each file to include all of the
events that are acceptable in the file. Fluorescence from fixable
vital dyes seems to shift more than conjugated mAbs, causing
the live cell gate to be moved even within a single donor. The
same is true for a scatter lymphocyte gate. Light scatter proper-
ties are less stable than fluorescence, owing to changes within
the microenvironment of each tube during the staining pro-
cess. As a result, scatter gates sometimes need to be moved,
even within a single donor. When following the methods
described in this chapter, it is rare that gates or regions drawn
on markers from fluorescent conjugated mAbs need to be
moved within a donor. It is feasible that fluorescent conjugated
mAb gates and regions might need to be adjusted slightly
between donors, to accommodate different backgrounds;
however, shifts in the positive populations are usually not
observed. When shifts in the positive population occur, they
are usually due to staining, acquisition, or instrument errors.
Careful note taking during the staining and acquisition steps,
documenting any errors with each tube, can be helpful in
tracking these types of errors and deciding which data may be
excluded based on errors.

References

1. Sakaguchi S et al (2010) FOXP3+ regulatory 6. Borrow P, Moody MA (2017) Immunologic

T cells in the human immune system. Nat Rev characteristics of HIV-infected individuals
Immunol 10(7):490–500 who make broadly neutralizing antibodies.
2. Li X, Zheng Y (2015) Regulatory T cell iden- Immunol Rev 275(1):62–78
tity: formation and maintenance. Trends 7. Santegoets SJ et al (2015) Monitoring regu-
Immunol 36(6):344–353 latory T cells in clinical samples: consensus on
3. Vignali DA, Collison LW, Workman CJ an essential marker set and gating strategy for
(2008) How regulatory T cells work. Nat regulatory T cell analysis by flow cytometry.
Rev Immunol 8(7):523–532 Cancer Immunol Immunother 64
4. Hall BM (2016) CD4+CD25+ T regulatory (10):1271–1286
cells in transplantation tolerance: 25 years 8. Tosolini M et al (2011) Clinical impact of
on. Transplantation 100(12):2533–2547 different classes of infiltrating T cytotoxic
5. Falivene J et al (2015) Th17 and Th17/Treg and helper cells (Th1, th2, Treg, th17) in
ratio at early HIV infection associate with patients with colorectal cancer. Cancer Res
protective HIV-specific CD8(+) T-cell 71(4):1263–1271
responses and disease progression. Sci Rep 9. Nishikawa H, Sakaguchi S (2014) Regulatory
5:11511 T cells in cancer immunotherapy. Curr Opin
Immunol 27:1–7
174 Janet Staats

10. Saito T et al (2016) Two FOXP3(+)CD4(+) T autoimmune and inflammatory diseases. Nat
cell subpopulations distinctly control the Rev Immunol 15(5):283–294
prognosis of colorectal cancers. Nat Med 22 25. Reissig S et al (2017) Elevated levels of Bcl-3
(6):679–684 inhibits Treg development and function
11. Figueiredo AS, Schumacher A (2016) The T resulting in spontaneous colitis. Nat Com-
helper type 17/regulatory T cell paradigm in mun 8:15069
pregnancy. Immunology 148(1):13–21 26. Tan HL et al (2013) T regulatory lympho-
12. Engler JB et al (2017) Glucocorticoid recep- cytes and endothelial function in pediatric
tor in T cells mediates protection from auto- obstructive sleep apnea. PLoS One 8(7):
immunity in pregnancy. Proc Natl Acad Sci U e69710
S A 114(2):E181–e190 27. Ni K et al (2015) Th17/Treg balance in chil-
13. Feyaerts D et al (2017) Human uterine lym- dren with obstructive sleep apnea syndrome
phocytes acquire a more experienced and tol- and the relationship with allergic rhinitis. Int J
erogenic phenotype during pregnancy. Sci Pediatr Otorhinolaryngol 79(9):1448–1454
Rep 7(1):2884 28. Grant J et al (2009) Validated protocol for
14. Guzman-Genuino RM, Diener KR, Regu- FoxP3 reveals increased expression in type
latory B (2017) Cells in pregnancy: lessons 1 diabetes patients. Cytometry B Clin
from autoimmunity, graft tolerance, and can- Cytom 76(2):69–78
cer. Front Immunol 8:172 29. Hori S (2014) Lineage stability and pheno-
15. Cipolletta D et al (2012) PPAR-gamma is a typic plasticity of Foxp3(+) regulatory T cells.
major driver of the accumulation and pheno- Immunol Rev 259(1):159–172
type of adipose tissue Treg cells. Nature 486 30. Wing JB et al (2017) A distinct subpopulation
(7404):549–553 of CD25( ) T-follicular regulatory cells loca-
16. Qiao YC et al (2016) Changes of regulatory T lizes in the germinal centers. Proc Natl Acad
cells and of proinflammatory and immuno- Sci U S A 114(31):E6400–e6409
suppressive cytokines in patients with type 31. Weinstein JS et al (2016) TFH cells progres-
2 diabetes mellitus: a systematic review and sively differentiate to regulate the germinal
meta-analysis. J Diabetes Res 2016:3694957 center response. Nat Immunol 17
17. Serr I et al (2016) Type 1 diabetes vaccine (10):1197–1205
candidates promote human Foxp3(+)Treg 32. Sage PT, Sharpe AH (2016) T follicular regu-
induction in humanized mice. Nat Commun latory cells. Immunol Rev 271(1):246–259
7:10991 33. Sugiyama D et al (2013) Anti-CCR4 mAb
18. Visperas A, Vignali DA (2016) Are regulatory selectively depletes effector-type FoxP3
T cells defective in type 1 diabetes and can we +CD4+ regulatory T cells, evoking antitumor
fix them? J Immunol 197(10):3762–3770 immune responses in humans. Proc Natl Acad
19. Joly AL et al (2018) Alternative splicing of Sci U S A 110(44):17945–17950
FOXP3 controls regulatory T cell effector 34. Steinborn A et al (2012) Pregnancy-
functions and is associated with human ath- associated diseases are characterized by the
erosclerotic plaque stability. Circ Res 122 composition of the systemic regulatory T cell
(10):1385–1394 (Treg) pool with distinct subsets of Tregs.
20. Noval Rivas M et al (2015) Regulatory T cell Clin Exp Immunol 167(1):84–98
reprogramming toward a Th2-cell-like line- 35. Maecker HT, McCoy JP, Nussenblatt R
age impairs oral tolerance and promotes food (2012) Standardizing immunophenotyping
allergy. Immunity 42(3):512–523 for the human immunology project. Nat Rev
21. Bacher P et al (2016) Regulatory T cell speci- Immunol 12(3):191–200
ficity directs tolerance versus allergy against 36. Li Z et al (2015) FOXP3+ regulatory T cells
aeroantigens in humans. Cell 167 and their functional regulation. Cell Mol
(4):1067–1078.e16 Immunol 12(5):558–565
22. Palomares O et al (2017) Mechanisms of 37. Li N et al (2015) The abnormal expression of
immune regulation in allergic diseases: the CCR4 and CCR6 on Tregs in rheumatoid
role of regulatory T and B cells. Immunol arthritis. Int J Clin Exp Med 8
Rev 278(1):219–236 (9):15043–15053
23. Miyara M, Ito Y, Sakaguchi S (2014) TREG- 38. Iellem A et al (2001) Unique chemotactic
cell therapies for autoimmune rheumatic dis- response profile and specific expression of che-
eases. Nat Rev Rheumatol 10(9):543–551 mokine receptors CCR4 and CCR8 by CD4
24. Klatzmann D, Abbas AK (2015) The promise (+)CD25(+) regulatory T cells. J Exp Med
of low-dose interleukin-2 therapy for 194(6):847–853
 Immunophenotyping of Human Regulatory T Cells 175

39. Donnelly C et al (2018) Optimizing human 54. McKinnon KM (2018) Multiparameter con-
Treg immunotherapy by Treg subset selection ventional flow cytometry. Methods Mol Biol
and E-selectin ligand expression. Sci Rep 8 1678:139–150
(1):420 55. Wingender G, Kronenberg M (2015) OMIP-
40. Liu W et al (2006) CD127 expression 030: characterization of human T cell subsets
inversely correlates with FoxP3 and suppres- via surface markers. Cytometry A 87
sive function of human CD4+ T reg cells. J (12):1067–1069
Exp Med 203(7):1701–1711 56. Murdoch DM, Staats JS, Weinhold KJ (2012)
41. Seddiki N et al (2006) Expression of interleu- OMIP-006: phenotypic subset analysis of
kin (IL)-2 and IL-7 receptors discriminates human T regulatory cells via polychromatic
between human regulatory and activated T flow cytometry. Cytometry A 81(4):281–283
cells. J Exp Med 203(7):1693–1700 57. Moncunill G et al (2014) OMIP-024: pan--
42. Allan SE et al (2007) Activation-induced leukocyte immunophenotypic characteriza-
FOXP3 in human T effector cells does not tion of PBMC subsets in human samples.
suppress proliferation or cytokine production. Cytometry A 85(12):995–998
Int Immunol 19(4):345–354 58. Mahnke YD, Beddall MH, Roederer M
43. Tran DQ, Ramsey H, Shevach EM (2007) (2013) OMIP-015: human regulatory and
Induction of FOXP3 expression in naive activated T-cells without intracellular staining.
human CD4+FOXP3 T cells by T-cell recep- Cytometry A 83(2):179–181
tor stimulation is transforming growth factor- 59. Biancotto A et al (2012) OMIP-004:
beta dependent but does not confer a regu- in-depth characterization of human T regu-
latory phenotype. Blood 110(8):2983–2990 latory cells. Cytometry A 81(1):15–16
44. Mailer RKW (2018) Alternative splicing of 60. Givan AL (2011) Flow cytometry: an intro-
FOXP3-virtue and vice. Front Immunol duction. Methods Mol Biol 699:1–29
9:530 61. Snow C (2004) Flow cytometer electronics.
45. Sambucci M et al (2018) FoxP3 isoforms and Cytometry A 57(2):63–69
PD-1 expression by T regulatory cells in mul- 62. Data file standard for flow cytometry (1990)
tiple sclerosis. Sci Rep 8(1):3674 Data file standards committee of the society
46. Hartigan-O’Connor DJ et al (2007) Human for analytical cytology. Cytometry 11
CD4+ regulatory T cells express lower levels (3):323–332
of the IL-7 receptor alpha chain (CD127), 63. Seamer L (2001) Data file standard for flow
allowing consistent identification and sorting cytometry, FCS 3.0. Curr Protoc Cytom.
of live cells. J Immunol Methods 319 Chapter 10: p. Unit 10.2
(1–2):41–52
64. Spidlen J et al (2011) Flow cytometry data
47. McKinnon KM (2018) Flow cytometry: an standards. BMC Res Notes 4:50
overview. Curr Protoc Immunol
120:5.1.1–5.1.11 65. Feher K et al (2016) Multispectral flow cyto-
metry: the consequences of increased light
48. Shapiro HM (2018) Flow cytometry: the collection. Cytometry A 89(7):681–689
glass is half full. Methods Mol Biol 1678:1–10
66. Bagwell CB, Adams EG (1993) Fluorescence
49. Perfetto SP, Chattopadhyay PK, Roederer M spectral overlap compensation for any number
(2004) Seventeen-colour flow cytometry: of flow cytometry parameters. Ann N Y Acad
unravelling the immune system. Nat Rev Sci 677:167–184
Immunol 4(8):648–655
67. Roederer M (2002) Compensation in flow
50. Ashhurst TM, Smith AL, King NJC (2017) cytometry. Curr Protoc Cytom. Chapter 1:
High-dimensional fluorescence cytometry. p. Unit 1.14
Curr Protoc Immunol 119:5.8.1–5.8.38
68. Szaloki G, Goda K (2015) Compensation in
51. McLaughlin BE et al (2008) Nine-color flow multicolor flow cytometry. Cytometry A 87
cytometry for accurate measurement of T cell (11):982–985
subsets and cytokine responses. Part I: panel
design by an empiric approach. Cytometry A 69. McNeil LK et al (2013) A harmonized
73(5):400–410 approach to intracellular cytokine staining
gating: results from an international multi-
52. Mahnke Y, Chattopadhyay P, Roederer M consortia proficiency panel conducted by the
(2010) Publication of optimized multicolor cancer immunotherapy consortium (CIC/-
immunofluorescence panels. Cytometry A 77 CRI). Cytometry A 83(8):728–738
(9):814–818
70. Sun YL et al (2013) Application and effects of
53. Mahnke YD, Roederer M (2007) Optimizing mouse Foxp3 antibody and fixation/permea-
a multicolor immunophenotyping assay. Clin bilization buffer on the detection of CD4+
Lab Med 27(3):469–85, v
176 Janet Staats

regulatory T cells in various mammal species. 84. Hensley-McBain T et al (2014) Optimization

Genet Mol Res 12(4):6535–6545 of a whole blood phenotyping assay for enu-
71. Pillai V, Karandikar NJ (2008) Attack on the meration of peripheral blood leukocyte popu-
clones? Human FOXP3 detection by lations in multicenter clinical trials. J
PCH101, 236A/E7, 206D, and 259D Immunol Methods 411:23–36
reveals 259D as the outlier with lower sensi- 85. Lenkei R et al (1998) Performance of calibra-
tivity. Blood 111(1):463–464. author reply tion standards for antigen quantitation with
464-6 flow cytometry. Cytometry 33(2):188–196
72. Law JP et al (2009) The importance of Foxp3 86. Rossmann ED et al (2007) Performance of
antibody and fixation/permeabilization calibration standards for antigen quantitation
buffer combinations in identifying CD4 with flow cytometry in chronic lymphocytic
+CD25+Foxp3+ regulatory T cells. Cytome- leukemia. Cytometry B Clin Cytom 72
try A 75(12):1040–1050 (6):450–457
73. Klein S et al (2010) CD127(low/ ) and 87. Gratama JW et al (1998) Flow cytometric
FoxP3(+) expression levels characterize differ- quantitation of immunofluorescence inten-
ent regulatory T-cell populations in human sity: problems and perspectives. European
peripheral blood. J Invest Dermatol 130 Working Group on Clinical Cell Analysis.
(2):492–499 Cytometry 33(2):166–178
74. Presicce P et al (2010) Association of two 88. Wang L et al (2002) Quantitating fluores-
clones allows for optimal detection of human cence intensity from fluorophores: practical
FOXP3. Cytometry A 77(6):571–579 use of MESF values. J Res Natl Inst Stand
75. Finak G et al (2016) Standardizing flow cyto- Technol 107(4):339–353
metry immunophenotyping analysis from the 89. Bolstad N, Warren DJ, Nustad K (2013) Het-
human immunophenotyping consortium. Sci erophilic antibody interference in immuno-
Rep 6:20686 metric assays. Best Pract Res Clin Endocrinol
76. Maecker HT, Trotter J (2006) Flow cytome- Metab 27(5):647–661
try controls, instrument setup, and the deter- 90. Perfetto SP et al (2006) Amine reactive dyes:
mination of positivity. Cytometry A 69 an effective tool to discriminate live and dead
(9):1037–1042 cells in polychromatic flow cytometry. J
77. Perfetto SP et al (2012) Quality assurance for Immunol Methods 313(1–2):199–208
polychromatic flow cytometry using a suite of 91. Perfetto SP et al (2010) Amine-reactive dyes
calibration beads. Nat Protoc 7 for dead cell discrimination in fixed samples.
(12):2067–2079 Curr Protoc Cytom. Chapter 9: p. Unit 9 34
78. Hoffman RA (2005) Standardization, calibra- 92. Maciorowski Z, Chattopadhyay PK, Jain P
tion, and control in flow cytometry. Curr Pro- (2017) Basic multicolor flow cytometry.
toc Cytom. Chapter 1: p. Unit 1.3 Curr Protoc Immunol 117:5.4.1–5.4.38
79. Roederer M (2001) Spectral compensation 93. Chattopadhyay PK et al (2012) Brilliant vio-
for flow cytometry: visualization artifacts, lim- let fluorophores: a new class of ultrabright
itations, and caveats. Cytometry 45 fluorescent compounds for immunofluores-
(3):194–205 cence experiments. Cytometry A 81(6):
80. Hulspas R et al (2009) Considerations for the 456–466
control of background fluorescence in clinical 94. Hulspas R et al (2009) Flow cytometry and
flow cytometry. Cytometry B Clin Cytom 76 the stability of phycoerythrin-tandem dye
(6):355–364 conjugates. Cytometry A 75(11):966–972
81. Scott LE et al (2017) Choosing a new CD4 95. Johansson U, Macey M (2014) Tandem dyes:
technology: can statistical method compari- stability in cocktails and compensation con-
son tools influence the decision? Cytometry siderations. Cytometry B Clin Cytom 86
B Clin Cytom 92(6):465–475 (3):164–174
82. Langenskiold C et al (2018) Determination of 96. Baumgarth N, Roederer M (2000) A practical
blood cell subtype concentrations from frozen approach to multicolor flow cytometry for
whole blood samples using TruCount beads. immunophenotyping. J Immunol Methods
Cytometry B Clin Cytom 94(4):660–666 243(1–2):77–97
83. Nicholson JK et al (1997) Evaluation of a 97. Nguyen R et al (2013) Quantifying spillover
method for counting absolute numbers of spreading for comparing instrument perfor-
cells with a flow cytometer. Clin Diagn Lab mance and aiding in multicolor panel design.
Immunol 4(3):309–313 Cytometry A 83(3):306–315
 Immunophenotyping of Human Regulatory T Cells 177

98. Byrd T et al (2015) Polystyrene microspheres 103. Bagwell CB (2005) Hyperlog-a flexible
enable 10-color compensation for immuno- log-like transform for negative, zero, and pos-
phenotyping of primary human leukocytes. itive valued data. Cytometry A 64(1):34–42
Cytometry A 87(11):1038–1046 104. Herzenberg LA et al (2006) Interpreting flow
99. Staats JS et al (2014) Toward development of cytometry data: a guide for the perplexed. Nat
a comprehensive external quality assurance Immunol 7(7):681–685
program for polyfunctional intracellular cyto- 105. Tung JW et al (2004) New approaches to fluo-
kine staining assays. J Immunol Methods rescence compensation and visualization of
409:44–53 FACS data. Clin Immunol 110(3):277–283
100. Jaimes MC et al (2011) Quality assurance of 106. Novo D, Wood J (2008) Flow cytometry his-
intracellular cytokine staining assays: analysis tograms: transformations, resolution, and dis-
of multiple rounds of proficiency testing. J play. Cytometry A 73(8):685–692
Immunol Methods 363(2):143–157 107. Richards AJ et al (2014) Setting objective
101. Lugli E et al (2013) Identification, isolation thresholds for rare event detection in flow
and in vitro expansion of human and nonhu- cytometry. J Immunol Methods 409:54–61
man primate T stem cell memory cells. Nat 108. Chan C et al (2008) Statistical mixture mod-
Protoc 8(1):33–42 eling for cell subtype identification in flow
102. Lamoreaux L, Roederer M, Koup R (2006) cytometry. Cytometry A 73(8):693–701
Intracellular cytokine optimization and stan- 109. Maecker HT et al (2005) Standardization of
dard operating procedure. Nat Protoc 1 cytokine flow cytometry assays. BMC Immu-
(3):1507–1516 nol 6:13
 Chapter 10

Immunophenotyping of Human Innate Lymphoid Cells

Sara Trabanelli, Alejandra Gomez-Cadena, and Camilla Jandus

Abstract
In the last years, the family of innate lymphocytes has been growing following the discovery of innate
lymphoid cells (ILCs). ILCs are lymphocytes able to rapidly produce a wide range of soluble mediators in an
antigen-independent fashion. So far, three main subsets of ILCs have been discovered, ILC1, ILC2, and
ILC3, expressing respectively the transcription factors T-bet, GATA3, and Rorγt and secreting distinct
types of cytokines. After their discovery, several studies showed that different pathologies, such as allergic
airway diseases and inflammatory disorders, are sustained by dysfunctional ILCs before adaptive immune
sets in. In this regard, considerable efforts are currently performed to harmonize the identification and
monitoring of ILCs in healthy and pathologic conditions to streamline a uniform immunophenotyping.
Standardized ILC monitoring techniques will accelerate our understanding of these effector innate immune
cells and ultimately facilitate their targeting in the context of infection, cancer, autoimmune disease, and
transplantation.

Key words Human, ILCs, Flow cytometry, Staining, Isolation, Cytokines, Transcription factors,
Expansion

1 Introduction

Innate lymphoid cells (ILCs) are the most recently identified family
of lymphocytes that belong to the innate immune system [1].
Indeed, they lack rearranged antigen-specific receptors, such as
the T-cell receptor (TCR, specifically expressed by T lymphocytes)
or the B-cell receptor (BCR, specifically expressed by B lympho-
cytes), on their cell surface. As a consequence, ILCs do not recog-
nize, bind, and respond to specific antigens but to soluble factors or
to molecules expressed on other immune and nonimmune cell
types. Therefore, they rapidly respond to signals coming from the
extracellular environment and constitute one of the first-line
defense of our body. For this reason, ILCs are considered “early
sentinels” enriched at mucosal and barrier surfaces. They are also
present in primary and secondary lymphoid organs, and they can be
found circulating in the peripheral blood.

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_10, © Springer Science+Business Media, LLC, part of Springer Nature 2019

179
180 Sara Trabanelli et al.

ILCs differentiate from the common lymphoid precursor in the

fetal liver or in the bone marrow [2]. Their development is similar
to the one of natural killer (NK) cells, since it requires the expres-
sion of the common γ chain of the interleukin-2 (IL-2) receptor
and the transcriptional repressor Id2. Different from NK cells,
though, ILC differentiation is dependent on GATA3 and IL-7Rα
expression. Further sequential engagement of transcription factors,
cytokines and microbial signals is critical for the development of
three distinct groups of mature ILCs. ILC1 express T-bet, are
responsive to IL-12, and produce IFN-γ. ILC2 highly express
GATA3, are responsive to the alarmins IL-25, IL-33 and TSLP,
and produce IL-4, IL-5, IL-9, IL-13, and amphiregulin (Areg).
ILC3 express RORγt, are responsive to IL-1β and IL-23, and
produce IL-17 and/or IL-22. Notably, because of their transcrip-
tional and cytokine profiles, mature ILCs are considered the innate
counterpart of CD4 T helper cells. Indeed, ILC1 mirror the Th1,
ILC2 the Th2, and ILC3 the Th17/22 T helper subsets. However,
similar to CD4 T helper cells, this classification is not totally fixed,
since heterogeneity and plasticity are prominent features of ILC
behavior across all subsets and even beyond ILC themselves.
Indeed, interconversion between NK cells and ILCs has also been
reported [3].
The different ILC subsets actively participate to the initial
phases of immune responses against different pathogens or during
physiological and nonphysiological situations that alter homeosta-
sis. In particular, ILC1 have been described as important players in
fighting intracellular bacteria and protozoa, as well as in sustaining
chronic inflammation. ILC2 are involved in the resolution of hel-
minth infections and in allergic diseases and asthma. ILC3 are
engaged during immune responses against extracellular bacteria
and fungi; they are involved in intestinal homeostasis and partici-
pate in lymphoid tissue development.
To date, only few protocols are available for the identification
and functional characterization of human ILCs. Since this recently
described population of innate lymphocytes has become the focus
of extensive investigations that aim at understanding their physio-
logical and pathological roles in humans, uniform and robust
methodologies for their careful identification and functional evalu-
ation are essential [4].
This chapter represents a detailed description of materials,
methods and notes of laboratory techniques, with the aim to pro-
mote common procedures leading to standardization of assays,
ultimately allowing for wide application and improved understand-
ing of the role of the different ILC subsets in physiological and
nonphysiological processes [5].
 Immunophenotyping of Human Innate Lymphoid Cells 181

2 Materials

2.1 Isolation 1. Heparinized human blood.

of Peripheral Blood 2. 50 mL tubes.
Mononuclear Cells
3. Lymphoprep™ (Axis-Shield PoC AS), store at 4  C in the dark.
(PBMCs) from Human
Peripheral Blood 4. PBS, store at room temperature.
5. RPMI 1640, GlutaMAX™-I medium (Gibco). Store at 4  C.
6. 10 mL and 25 mL sterile pipettes.

2.2 Isolation 1. Pair of sterile scissors.

of Human 2. Petri dishes.
Mononuclear Cells
3. 70 μm and 100 μm cell strainers.
(MNCs) from Tonsils
4. 50 mL tubes.
5. Plunger of a plastic syringe.
6. 10 mL Sterile pipette.
7. Lymphoprep™ (Axis-Shield PoC AS), store at 4  C in the dark.
8. PBS, store at room temperature.
9. RPMI 1640, GlutaMAX™-I medium (Gibco). Store at 4  C.

2.3 Extracellular 1. FACS buffer: PBS supplemented with 5 mM EDTA, 0.2% BSA,
Staining of Ex Vivo and 0.2% NaN3. Store at 4  C.
Human ILCs 2. LIVE/DEAD™ fixable green dead cell stain kit for 488 nm
excitation (Invitrogen) (see Note 1).
3. Monoclonal antibodies:
(a) Lineage (Lin): CD4 (clone RPA-T4); CD8 (clone
MEM-31) (ImmunoTools). CD3 (clone UCTH1);
CD14 (clone HCD14); CD15 (clone HI98); CD19
(clone H1B19); CD20 (clone 2H7); CD33 (clone
HIM3-4); CD34 (clone 561); CD203c (clone NP406);
FcεRI (clone AER-37) (Biolegend). CD16 (clone
RMO52) (Beckman Coulter), all FITC conjugated.
(b) CD127 BV421 (clone AO19D5); CD56 Alexa
700 (clone HCD56); CD117 (c-Kit) APC-Fire (clone
104D2); CRTH2 PE (clone BM16); CD161 (clone
HP-3G10) PE-Cy7; PerCPCy5.5 NKp44 (clone P44-8)
or NKp46 (clone 9E2) (Biolegend) (see Notes 2 and 3).
4. 5 mL V-bottom tubes or 96-well plates.
5. 5 mL U-bottom tubes.
182 Sara Trabanelli et al.

2.4 Fluorescence- 1. Anti-human CD3, anti-human CD14 MicroBeads (Miltenyi

Activated Cell Sorting Biotec). Store at 4  C.
(FACS) to Isolate 2. LD Columns (Miltenyi Biotec).
Human ILC Subsets
3. MACS® Manual Separators—MidiMACS™ (Miltenyi Biotec).
4. Sorting buffer: PBS supplemented with 5 mM EDTA and 0.2%
BSA. Store at 4  C.
5. Same antibodies described for the extracellular staining (see
Subheading 2.3) (see Note 3).
6. 1.5 mL Eppendorf tubes.
7. Sorting collection buffer: 100% fetal calf serum (FSC) (Gibco).

2.5 Transcription 1. Monoclonal antibodies: T-bet PE-CF 594 (clone O4-46);

Factor (TF) Analysis RORγt PE (clone Q21-559) (BD).
on Human ILCs (Fig. 2) GATA-3 APC (clone 16E10A) (eBioscience™) (see Notes
2–4).
2. Foxp3/Transcription Factor Staining Buffer Set
(eBioscience™).
3. FACS buffer: PBS supplemented with 5 mM EDTA, 0.2%. BSA
and 0.2% NaN3. Store at 4  C.
4. 5 mL V-bottom tubes or 96-well plates.
5. 5 mL U-bottom tubes.

2.6 Functional 1. 96-well plate.

Analysis 2. ILC culture medium: RPMI 1640, GlutaMAX™-I supplemen-
of Human ILCs ted with 8% human serum (HS) (recommended is the use
2.6.1 Intracellular
of pooled A+ healthy donor’s serum), 1% PSG Mix (penicillin
Staining for Evaluation and streptomycin) (Gibco), 1.5 mM L-glutamine (Gibco),
of Cytokine Production by 0.24 mM L-asparagine (Sigma), 0.55 mM L-arginine (Sigma),
Human ILCs
1% HEPES buffer (Animed), 1% from the 100 mM sodium
pyruvate (Gibco), kanamycin 100 (Gibco), 0.1% 2-
β-mercaptoethanol (Gibco). Store at 4  C.
3. Recombinant human (rh) cytokines: 100 IU/mL IL-2
(Roche); 5 ng/mL rhIL-7; 50 ng/mL rhIL-12; rhIL-15;
rhIL-25; rhIL-1β; rhIL-23; rhTSLP; rhIL-33 (PeproTech).
rhIL-18 (R&D). Store stock dilutions according to the manu-
facturer’s recommendations or at 80  C, respectively.
4. GolgiPlug and GolgiStop (BD Biosciences) or Brefeldin A
(Sigma). Store stock dilutions according to the manufacturer’s
recommendations or at 20  C, respectively.
5. 50 ng/mL PMA (Phorbol 12-myristate 13-acetate) (Sigma)
and 500 μg/mL ionomycin (Sigma). Store stock dilutions at
80  C.
 Immunophenotyping of Human Innate Lymphoid Cells 183

6. Fixation buffer: PBS supplemented with 1% formaldehyde, 2%


D-glucose and 5 mM NaN3. Store at 4 C.

7. 0.1% saponin (Sigma). Store stock dilutions and aliquots at

20  C.
8. Monoclonal antibodies: IFN-γ (clone 4S.B3); TNF-α (clone
Mab11); IL-4 (clone MP4-25D2); IL-5 (clone TRFK5); IL-9
(clone MH9A4); IL-13 (clone JES10-5A2); IL-17A (clone
eBio64DEC17); IL-22 (clone 2G12A41); LT-α (clone
359-81-11).

2.6.2 Quantification 1. 96-well plate.

of Cytokine Secretion 2. ILC culture medium (see Subheading 2.6.1).
in Human ILC Culture
3. Recombinant human (rh) cytokines: 100 UI/mL IL-2
Supernatants
(Roche); 5 ng/mL rhIL-7; 50 ng/mL rhIL-12; rhIL-15;
rhIL-25; rhIL-1β; rhIL-23; rhTSLP; rhIL-33 (PeproTech).
rhIL-18 (R&D). Store stocks dilutions according to the man-
ufacturer’s recommendations or at 80  C, respectively.
4. Multiplex immunoassay for cytokine detection kit (i.e., MSD
Th1/Th2 cytokines, LEGENDplex™ Human Th Cytokine
Panel, BD CBA Human Inflammatory Cytokines Kit, Lumi-
nex® Multiplex Kit R&D or single staining cytokine array).

2.7 In Vitro 1. 96-well plates.

Expansion 2. ILC culture medium (see Subheading 2.6.1).
and Cryopreservation
3. Recombinant human (rh) cytokines: 100 U/mL rhIL-
of Human ILCs
2 (Roche); 5 ng/mL rhIL-7 (PeproTech).
4. FCS sterile (Gibco).
5. Dimethyl sulfoxide (DMSO), sterile.
6. RPMI 1640, GlutaMAX™-I medium (Gibco). Store at 4  C.
7. Freezing solution: pure FCS supplemented with 10% DMSO.
Make fresh as required.
8. Round-bottom 1.8 mL cryogenic vials.

3 Methods

3.1 Isolation 1. Transfer freshly collected heparinized human blood in 50 mL

of Peripheral Blood tubes (see Note 5).
Mononuclear Cells 2. Dilute the blood at 1:1 ratio with sterile PBS.
(PBMCs) from Human
3. Prepare 50 mL tubes with the Lymphoprep™ solution. The
Peripheral Blood ratio of Lymphoprep™ to diluted blood is 1:1.5.
4. Carefully and slowly, overlay the diluted blood on the Lym-
phoprep™ layer, creating a sharp Lymphoprep™–blood
184 Sara Trabanelli et al.

interface (see Note 6). Cap the tube, avoiding accidental

contamination.
5. Centrifuge the tube at 528  g for 20 min with slow accelera-
tion and without brake, in a swinging bucket rotor, at room
temperature (see Note 7). This step allows for the formation of
a density gradient that, by centrifugation, stratifies on different
layers peripheral blood mononuclear cells and other cells such
as erythrocytes and polymorphonuclear leukocytes. It creates
an opalescent ring at the Lymphoprep™–medium surface
where you will find the PBMCs.
6. Carefully harvest the PBMC layer, using a 5 or 10 mL sterile
pipette. Transfer the cells into a new 50 mL tube. Fill up the
tube with PBS or RPMI 1640, GlutaMAX™-I, mix well, and
centrifuge for 5 min, at 528  g with break, at room
temperature.
7. Remove the supernatant, add PBS or RPMI 1640,
GlutaMAX™-I, then centrifuge the PBMCs at 234  g for
10 min without brake at room temperature (see Note 8).
8. Wash the cells in PBS or RPMI 1640, GlutaMAX™-I (5 min at
528 g).
9. Resuspend the pellet in ILC culture medium and count the
cells before using the cell suspension (see Note 9).

3.2 Isolation of 1. Put 3 mL of RPMI 1640, GlutaMAX™-I in a petri dish with

Human Mononuclear tonsils (see Note 5).
Cells (MNCs) from 2. Cut tonsils into small pieces in the petri dish using sterile
Tonsils scissors (see Notes 10 and 11).
3. Take the pieces and ground through a 100 μm sterile strainer
using a plunger of a plastic syringe above a 50 mL tube. Rinse
the filter two or three times with additional 5 mL RPMI 1640,
GlutaMAX™-I.
4. Filter again the obtained cell suspension, this time through a
70 μm sterile strainer and rinse the strainer at least 2 times with
5 mL RPMI 1640, GlutaMAX™-I. Complete up to 25 mL
with the same medium (see Note 12).
5. Carefully overlay this cell suspension on 15 mL of
Lymphoprep™.
6. Centrifuge at 760  g for 20 min with slow acceleration and
without brake at room temperature.
7. Collect the mononuclear cell layer into a 15 mL tube, wash
with PBS or RPMI 1640, GlutaMAX™-I, 10% FCS and cen-
trifuge for 5 min at 528  g.
8. Remove supernatant, resuspend very well the cell pellet with
ILC culture medium and count the cells.
 Immunophenotyping of Human Innate Lymphoid Cells 185

3.3 Extracellular 1. Transfer the cells into V-bottom tubes or V- or U-bottom

Staining of Ex Vivo 96 well plates.
Human ILCs 2. Wash with PBS and centrifuge for 5 min at 528  g. At the
same time, prepare the LIVE/DEAD solution referred before
according to the manufacturer’s recommendations in plain
PBS (see Subheading 2.3) (see Note 13).
3. Discard the supernatant.
4. Add 100 μL of LIVE/DEAD solution on the pelleted cells.
Incubate for at least 20 min and up to 1 h at 4  C in the dark.
5. Prepare extracellular antibody mix using FACS buffer (see
Note 14).
6. Wash with 250 μL (V bottom tubes) or 150 μL (96 well plates)
of FACS buffer and centrifuge for 5 min at 528  g.
7. Resuspend the pellet with 50 μL of extracellular antibody mix.
Incubate for 30 min at 4  C (see Note 15).
8. Wash as in step 6.
9. Resuspend in 200 μL of FACS buffer if there is no further
intracellular staining. Transfer the stained cells in U-bottom
tubes compatible with your flow cytometry analysis
instrument.

3.4 Fluorescence- 1. Wash cell suspension with sterile sorting buffer.

Activated Cell Sorting 2. Deplete CD3+ and CD14+ cells using magnetic cell sorting
(FACS) to Isolate according to the manufacturer’s recommendations.
Human ILC Subsets
3. Centrifuge and resuspend the pellet with 50 μL of the antibody
cocktail (see Note 14).
4. Incubate for 30 min at 4  C (see Note 15).
5. Wash with sorting buffer: add 2 mL of sorting buffer and
centrifuge for 5 min at 528  g.
6. Resuspend the cells in sorting buffer at 10  106 cells/mL (see
Note 16).
7. To isolate ILC1, ILC2, and ILC3 classical subsets the gating
strategy presented in Fig. 1 can be used (see Note 17).
8. Recover your cells either in sterile 1.5 mL Eppendorf tubes or
5 mL tubes with 200 μL of sorting collection buffer.
9. Centrifuge the collected cells 5 min at 528  g and then
aspirate carefully the supernatant.
10. Resuspend in ILC culture medium according to your needs.

3.5 Transcription 1. After the extracellular staining described in Subheading 3.3,

Factor Analysis samples are washed and resuspended in 100 μL of fixation/
on Human ILCs permeabilization buffer, prepared following the manufacturer’s
recommendations. Incubate for 20 min at room temperature.
186 Sara Trabanelli et al.

Fig. 1 Representative example of human ILC subset identification by multiparametric flow cytometry. ILCs
were identified within the peripheral blood lymphocyte region on the basis of their forward (FSC) and side
scatter (SSC) profiles (FSC low and SSC low) and by excluding from the analysis doublets (FSC H/FSC W dot
plot, followed by SSC A/SSC W dot plot). Total ILCs were gated as Lin CD127+ cells. Then, the different ILC
subsets were determined according to the expression of CRTH2 vs cKit: ILC1 CRTH2 cKit , ILC2
CRTH2+cKit+/ , and ILC3 CRTH2 cKit+. The dot plots shown are the result of a minimum of 106 MNCs
acquired on a flow cytometer. Data were analyzed using FlowJo™ software (TreeStar)

Fig. 2 Representative example of specific transcription factor expression by human ILCs. Total ILCs and ILC
subsets were stained as in Fig. 1, then the cells were fixed and permeabilized and stained for T-bet, GATA3,
and RORγt expression

2. Wash with permeabilization buffer 1 solution: add 100 μL if

using 96-well plates or 200 μL if using V-bottom tubes and
centrifuge for 5 min at 528  g.
3. Prepare the transcription factor (TF) antibody mix using the
permeabilization buffer.
4. Resuspend the pelleted cells in 50 μL of TF antibody mix.
Incubate for 30 min minimum at room temperature (see
Note 18).
5. Add 100 or 200 μL of Permeabilization buffer solution and
centrifuge for 5 min at 528  g.
6. Resuspend the cells in 200 μL of FACS buffer and transfer the
cell suspension into U-bottom tubes compatible with your flow
cytometry equipment.
 Immunophenotyping of Human Innate Lymphoid Cells 187

3.6 Functional 1. After obtaining the total ILC suspension or isolating ILC sub-
Analysis sets, ILCs can be stimulated to assess their intracellular cyto-
of Human ILCs kine production. To this purpose, different ILC subsets must
be stimulated with their specific cytokine mix. ILC1 mix: hrIL-
3.6.1 Intracellular 2 (100 U/mL); hrIL-7 (5 ng/mL); hrIL-12; hrIL-18; hrIL-15
Cytokine Staining (50 ng/mL). ILC2 mix: hrIL-2 (100 U/mL); hrIL-7 (5 ng/
for Evaluation of Cytokine mL); hrIL-33; hrIL-25 and TSLP (50 ng/mL). ILC3 mix:
Production by Human ILCs hrIL-2 (100 U/mL); hrIL-7 (5 ng/mL); hrIL-1β; hrIL-23
(50 ng/mL) (see Note 19).
2. Incubate at 37  C with 5% of CO2 overnight with 2 μg/mL of
Brefeldin A (see Note 20).
3. 1 h before starting the staining, stimulate with 50 ng/mL PMA
and 500 ng/mL ionomycin.
4. At the end of the stimulation, harvest the cells and transfer
them to V-bottom tubes or 96-well plates and proceed with the
extracellular staining (see Subheading 3.3).
5. Wash with FACS buffer and centrifuge for 5 min at 528  g.
6. Resuspend the pellet in 100 μL of fixation buffer (see Subhead-
ing 2.6.1). Incubate for 30 min at room temperature.
7. Add 100 μL, for the 96-well plates, or 200 μL, for the
V-bottom tubes, of a 0.1% saponin solution (see Subheading
2.6.1) and centrifuge for 5 min at 528  g.
8. Prepare your cytokine antibody mix in the 0.1% saponin solu-
tion (see Note 21).
9. Incubate for 30 min at 4  C.
10. Finally wash the cells with the 0.1% saponin solution, centri-
fuge for 5 min at 528  g and resuspend your cells in 200 μL of
FACS buffer. Transfer the stained cells in U-bottom tubes
compatible with your flow cytometer (Fig. 3).

3.6.2 Quantification 1. Collect the supernatant after 24 and/or 48 h stimulation of

of Cytokine Secretion isolated ILC subsets (see Subheading 3.6.1) (see Note 22).
in Human ILC Culture 2. Supernatants must be stored at 20  C for short periods of
Supernatants time or at 80  C for longer periods (see Note 23).
3. Proceed with any multiplex quantitative protocol to assess
soluble cytokine, according to the manufacturer’s
recommendations.

3.7 In Vitro 1. Culture the sorted ILC subsets in 96-well plates, with ILC
Expansion culture medium supplemented with rhIL-2 (100 U/mL) and
and Cryopreservation rhIL-7 (5 ng/mL) for all the subsets.
of Human ILCs 2. Place the plates in an incubator at 37  C, with 5% CO2.
3. Refresh the medium every 48 or 72 h, and split the cells if
necessary (see Note 24).
188 Sara Trabanelli et al.

Fig. 3 Representative example of cytokine production evaluated by intracellular staining of human ILCs. ILCs
were cultured with different cytokine cocktails overnight in the presence of brefeldin A. As negative control,
PBMCs were cultured in absence of stimulating cytokines for the same period of time. After the culture, total
ILCs and ILC subsets were stained as in Fig. 1, then the cells were fixed and permeabilized. Intracellular
expression of IFN-γ, IL-5, and IL-22 was evaluated in each subset, respectively
 Immunophenotyping of Human Innate Lymphoid Cells 189

4. For cryopreservation of expanded ILCs, prepare the freezing

solution and put it on ice.
5. Label the cryogenic vials and put them on ice.
6. Count cells, wash them with RPMI 1640, GlutaMAX™-I, and
resuspend in 1 mL of ice-cold freezing medium. Transfer to the
cold cryogenic vials (see Note 25).
7. Close the tubes and mix gently. Immediately transfer the vials
at 80 , or into dedicated freezing containers.
8. Cryogenic vials can be transferred to liquid nitrogen 24 h after
freezing at earliest.

4 Notes

1. According to the configuration of your flow cytometer, other

LIVE/DEAD Fixable dyes can be used to exclude dead cells.
2. This fluorochrome combination is our standard format for
successful ex vivo monitoring of human ILCs. However, this
format can be modified according to the configuration of your
flow cytometer.
3. Antibodies should be titrated prior use in flow cytometry to
determine the optimal dilution to be used for staining. In
addition, always keep the antibody mix on ice or at 4  C, in
the dark, to avoid fluorochrome dissociation during the prepa-
ration of the mix, and rapidly store antibodies back after use.
Also, staining should be performed avoiding the exposure to
direct light.
4. In order to combine transcription factors and ILC surface
marker expression for discrimination of subsets find the best
fluorochrome combination. One option is to remove the
NKp46 marker and include CRTH2 conjugated with the dye
PerCPCy5.5.
5. Always wear gloves when handling human blood or tissue and
perform PBMCs isolation in sterile conditions, under a clean
laminar flow hood.
6. High-quality separation of PBMCs after density gradient cen-
trifugation depends on a sharp interface between lymphocytes
and the separation solution after the layering step. This step can
be achieved by a slow overlay of the diluted peripheral blood
onto the Lymphoprep™ solution using a 10 mL pipette. The
second method involves inclining the two tubes (one contain-
ing the Lymphoprep solution and the other the diluted blood)
in an upside down “V” position with the rims firmly touching
each other, such that the blood can be carefully and continu-
ously poured onto the solution along the side of the tube.
190 Sara Trabanelli et al.

Avoid any shaking of the tubes to avoid mixing of the two

solutions and progressively straighten the receiving tube as it
becomes filled.
7. If the blood is stored for more than 2 h, increase the centrifu-
gation time to 30 min.
8. This step allows to eliminate platelets by lower speed centrifu-
gation and without brake. For small amounts of blood, like
patient samples, this step can be removed.
9. After this step, a red blood cell lysis step can be performed if
erythrocytes are contaminating your isolated PBMC. If you
progress to fluorescence-activated cell sorting, this step is
highly recommended.
10. All the instruments used for tissue dissociation should be ster-
ilized either by autoclaving them or by using 70% ethanol.
11. Keep always tissues on ice for all the steps.
12. For tonsils, an enzymatic digestion is not necessary; however, if
you are working with tissues that are harder and compact (e.g.,
lungs, intestine), perform an additional digestion step. For one
piece of sample, 5 mL of digestion medium is needed (385 μL
Liberase TL (Roche); 1.6 μL DNAse I (Invitrogen) in
4613.4 μL of 10% FCS (Gibco) RPMI 1640 medium). After
preparing the medium, add 5 mL of the solution and incubate
for 20 min at 37  C, mixing regularly by inverting the tube.
Depending on the samples, the digestion might take different
incubation times. Obtained dissociated sample should be small
enough to be aspirated using a 10 mL pipette and as close as
possible to a single cell suspension.
13. Viability dyes like LIVE/DEAD™ must be prepared in plain
PBS and not in staining buffer, in order to avoid quenching of
the staining by the BSA contained in the buffer.
14. CD3 T cell contamination is really common when sorting
human ILCs. Therefore, we suggest to prepare a Lineage
cocktail excluding the anti-CD3 antibody, that will be added
in a different fluorochrome. One option is to use the same
staining panel for extracellular markers as the one presented in
Subheading 2.3, but adding an anti-CD3 antibody conjugated
to APCeF780 and use an anti-CD117 antibody labeled with
APC instead of APC-Fire.
15. For some extracellular markers such as chemokine receptors it
is recommended to incubate cells at room temperature instead
at 4  C.
16. To avoid very long sorting sessions, and to ensure high purity
of sorted ILCs, a predepletion of CD3+, CD14+, and CD19+
cells using magnetic cell sorting is highly recommended.
 Immunophenotyping of Human Innate Lymphoid Cells 191

17. In our gating strategy we define ILC3 as lineage negative

CD127+CRTH2 cKit+ cells, as defined by Spits et al. [6, 7]
However, in a recent publication, it has been shown that in
human peripheral blood, lineage negative
CD127+CRTH2 cKit+ cells represent ILC precursors [8].
18. In order to obtain better transcription factor staining, 1 h
room temperature incubation is recommended.
19. The stimulation of purified ILC subsets can be performed in
96-well plates according to the cell numbers obtained after
sorting. If you are using cell suspensions from total human
PBMCs or tissues, 48- or 24-well plates can be used, plating
1 or 2  106 cells/mL.
20. Stimulation periods can vary from an overnight incubation to
up to 1 week, if the effect of compounds of interest is assessed
on cytokine-stimulated human ILCs. For longer incubation
times, add Brefeldin A only during the last overnight incuba-
tion period.
21. If you plan to concomitantly stain for cytokine production and
transcription factor analysis, use the protocol described in Sub-
heading 3.5 and the fixation/permeabilization kit
(eBioscience™).
22. This supernatant will contain detectable amounts of the cyto-
kines used to stimulate the different ILC subsets. Only cyto-
kines specifically produced by ILCs should be assessed in the
supernatant. For example, for ILC3, the stimulation media
contains hrIL-2; hrIL-7; hrIL-1β and hrIL-23, so cytokines
like IL-22; IFN-γ; TNF-α; IL17-A or IL-17F should be pref-
erentially assessed.
23. Avoid repetitive thawing and freezing of the supernatants. If
you plan measurements of cytokine in supernatants at different
times, freeze several independent small aliquots of superna-
tants, for single thawing and use.
24. ILCs have a slow proliferation rate. If a fast proliferation rate is
observed, this might indicate that the culture is contaminated
with Lineage positive cells, most probably CD3+ lymphocytes.
Regularly check the phenotype of your cultured cells by flow
cytometry (see Subheading 3.3). If necessary, perform an addi-
tional sorting to eliminate any Lineage+ contaminants.
25. DMSO is a toxic compound. Perform all the manipulation on
ice, in order to slow the DMSO entry into the cells.

References
1. Eberl G, Colonna M, Di Santo JP et al (2015) 2. Diefenbach A, Colonna M, Koyasu S et al
Innate lymphoid cells: a new paradigm in immu- (2014) Development, differentiation, and
nology. Science 348(6237):aaa6566
192 Sara Trabanelli et al.

diversity of innate lymphoid cells. Immunity 41 immunopathology and immunotherapy. Nat

(3):354–365 Immunol 17(7):755–757
3. Vivier E, Artis D, Colonna M et al (2018) Innate 6. Spits H, Artis D, Colonna M et al (2013) Innate
lymphoid cells: 10 years on. Cell 174 lymphoid cells--a proposal for uniform nomen-
(5):1054–1066 clature. Nat Rev Immunol 13(2):145–149
4. Trabanelli S, Gomez-Cadena A, Salomé B et al 7. Sonnenberg GF, Mjösberg J, Spits H et al
(2017) Human innate lymphoid cells (ILCs): (2013) SnapShot: innate lymphoid cells. Immu-
towards a uniform immunophenotyping. Cyto- nity 39(3):622–622.e1
metry B Clin Cytom 8. Lim A, Li Y, Lopez-Lastra S et al (2017) Systemic
5. Cording S, Medvedovic J, Aychek T et al (2016) human ILC precursors provide a substrate for tissue
Innate lymphoid cells in defense, ILC differentiation. Cell 168(6):1086–1100.e10
 Chapter 11

Enumeration of Plasmacytoid Dendritic Cells in Peripheral

Blood and Bone Marrow by Flow-Cytometric Analysis
Abdullah Alsuwaidan, Franklin Fuda, Weina Chen, and Mingyi Chen

Abstract
Dendritic cells (DCs) are powerful antigen presenting cells that are involved in regulating immune
response. Plasmacytoid dendritic cells (pDCs) are subtype of DCs that present in small quantity in the
bone marrow, peripheral blood, and lymph nodes. They are important component of our immune system in
normal condition and diseases. They activate T cells and play a critical role in immune tolerance. In this
chapter we review the immunophenotypic features of pDCs and provide a practical protocol for pDCs
enumeration in the peripheral blood and bone marrow samples.

Key words Flow cytometry, Dendritic cells, Plasmacytoid dendritic cells, Immunophenotyping

1 Introduction

Dendritic cells (DCs) are highly specialized cells that link innate and
adaptive immune response [1]. DCs are a heterogeneous popula-
tion of cells that regulate the immune response through their
function as potent antigen presenting cells. They have a special
capacity to present the antigen to the naı̈ve T cells which lead to
T cell activation, and maturation. Additionally, depending on the
inducing stimulus, they play a critical role in immune tolerance [2].
Two subtypes of DCs have been well described, conventional
dendritic cells (cDCs) and plasmacytoid dendritic cells (pDCs)
both derived from a common DC precursor arising from the
CD34+ hematopoietic stem cell (HSC). Precursor pDC
(pre-pDC) differentiate into pDCs in the bone marrow then
released into peripheral blood where they reside in lymphoid tissues
such as lymph nodes, tonsils, spleen, thymus, bone marrow, and
Peyer’s patches.
Mature pDCs stimulate distinct types of CD4 T helper (Th-1)
cells in response to endogenous antigens. Likewise, pDCs can be
stimulated by exogenous antigens like viruses [3]. Activated pDCs
secret type I interferon that activate and enhance T cells and natural

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_11, © Springer Science+Business Media, LLC, part of Springer Nature 2019

193
194 Abdullah Alsuwaidan et al.

killer (NK) cells [4,–6]. Thus, pDCs play a critical role in variety of
immune-mediated pathophysiological process such as graft-versus-
host disease (GvHD), allograft rejection, and autoimmune disor-
ders [2, 7].
Following hematopoietic stem cell transplantation (HSCT),
pDC facilitate HSC engraftment and induce tolerance to prevent
GvHD [5]. Several studies reveal the roles of pDC in immune
reconstitution after HSCT for hematolymphoid malignancies and
disease relapse, patient survival and risk of GvHD [1, 8]. The
higher percentages of pDCs, after allogeneic HSCT for leukemia,
are associated with successful bone marrow engraftment. On the
other hand, lower pDCs count predicts relapse, death, and acute
graft-versus-host disease [1]. In fact, it has been reported that the
percentage of pDCs following HSCT is an independent indicator
for adverse clinical outcomes [1, 8].
pDCs population are present in low quantity in bone marrow,
peripheral blood and lymph nodes. They show plasmacytoid mor-
phology and can be defined by their immunophenotype. In general,
normal mature pDCs in the bone marrow express CD4, CD22dim,
CD33, CD36, CD38, CD45, CD123, CD303, and HLA-DR,
and they are negative for CD5, CD11b, CD13, CD15, CD34,
CD56, and CD64 [8, 9]. Precursors of pDC often express CD34.
Of note, subtle immunophenotypic variations have been described
in pDCs in lymph nodes compared to bone marrow [9]. Further-
more, neoplastic pDCs may show immunophenotypic aberrancy as
in cases of blastic plasmacytoid dendritic cell neoplasm (e.g., CD56
expression [10]).
Given the critical role of pDCs in the immune system and
particularly in allogeneic HSCT, this chapter provides a practical
approach to identify and quantify pDCs using multiparametric flow
cytometry. Our protocol is designed to distinguish pDCs from
other hematopoietic cells present in bone marrow.

2 Materials

2.1 Preparation Anticoagulated bone marrow or peripheral blood sample usually

of Bone Marrow collected in heparinized or ethylenediaminetetraacetic acid (EDTA)
Mononuclear Cells tubes. It is ideal to analysis the sample within 48 h of collection.

2.2 Cell Staining 500 mL Dulbecco’s phosphate-buffered saline (DPBS) (10);

Buffer (PBS/Azide/BSA 22.75 g sodium azide; 50.0 g bovine serum albumin (BSA) (1%);
(1% PAB)) Adjust the pH to 7.2–7.4 with HCl and final volume to 5 liters with
additional deionized H2O; Stored at 2–8  C, brought to room
temperature before use.
 Enumeration of Plasmacytoid Dendritic Cells in Peripheral Blood and Bone. . . 195

2.3 1% PFA Cell 100 mL of paraformaldehyde (PFA) at 10% concentration; into

Fixation Solution 900 mL of phosphate-buffered saline (PBS, 10); Stored at
2–8  C, brought to room temperature before use.

2.4 RBC Lysis Buffer 80 g of NH4Cl (ammonium chloride); 10 g of KHCO3 (potassium

bicarbonate); 3.7 g of EDTA (ethylenediaminetetraacetic acid);
Adjust the pH to 7.2–7.4 with HCl and final volume to 1 liter
with additional deionized H2O; Use sterile filter, 0.22 μm poly-
ethersulfone (PES) filter; Stored at 2–8  C, brought to room
temperature before use.

2.5 Monoclonal The listed antibody panels (Tables 1, 2, and 3) are suggested
Antibodies monoclonal antibodies combinations [11] to allow the enumera-
tion of the pDCs among other myeloid and lymphoid cells using
four- or ten-color flow cytometry, (see Notes 1–3).

Table 1
Suggested monoclonal antibodies for 10-color flow cytometry panels

Surface marker Fluorochrome Clone

CD2 FITC S5.2
CD3 BV421 SK7 (Leu4)
CD4 PerCP-Cy5.5 SK3 (Leu3a)
CD5 PerCP-Cy5.5 L17F12
CD8 APC-H7 SK1
CD11b APC D12
CD13 PE L138
CD15 FITC MMA/HI98
CD22 APC S-HCL-1
CD34 APC-R700 8G12
CD36 FITC FA6.152
CD38 PE-Cy7 HIT2
CD45 V500-C 2D1
CD56 BV605 NCAM16.2
CD64 APC-R700 MD22
CD123 BV421 9F5/7G3
CD303 PE-Cy7 BDCA-2
HLA-DR APC-H7 L243
196 Abdullah Alsuwaidan et al.

Table 2
Antibodies panel for four-color flow cytometry

Tube FITC PE PerCP APC

1 CD36 CD123 HLA-DR CD22
2 CD4 CD123 HLA-DR CD56
3 CD34 CD123 CD45 CD303

Table 3
Antibodies panel for 10-color flow cytometry

APC-
Tube FITC PE PerCP-Cy5.5 PE-Cy7 APC R700 APC-H7 BV421 V500-C BV605
1 CD15 CD13 CD4 CD38 CD11b CD64 HLA-DR CD123 CD45 –
2 CD36 CD34 CD4 CD303 CD22 – HLA-DR CD123 CD45 CD56

3 Methods

1. In general, it is advisable to stain 500,000 cells in each tube

with appropriate volume of monoclonal antibodies and acquire
at least 100,000 cellular events for analysis.
2. Several pretest factors that are critical and can affect the final
result such as bone marrow collection method, anticoagula-
tion, and transportation.
3. Testing the sample as early as possible is encouraged, ideally
within 24 h or in less than 48 h of collection.
4. All testing should be carried out in sterile condition (see Notes
4 and 5).

3.1 Specimen 1. Resuspend the anticoagulated specimen by inverting the sam-

Processing ple several times.(see Note 6).
2. Decant the specimen into 50 mL conical tube.
3. Dissolve clots and remove particles.
4. Use filter to remove undissolved particles (70 μm nylon mesh)
if needed.
5. Add 3 mL NH4Cl (Ammonium chloride) lysis buffer into each
tube and vortex.
6. Incubate for 10 min at room temperature (RT).
7. Centrifuge for 7 min 800  g at RT.
 Enumeration of Plasmacytoid Dendritic Cells in Peripheral Blood and Bone. . . 197

8. Discard the supernatant.

9. Resuspend the cell pellet in 50 mL 1% PAB (PBS/Azide/BSA)
and mix gently.
10. Centrifuge for 7 min 800  g at RT.
11. Discard the supernatant.
12. Resuspend the cell pellet in 40 mL PAB (1%) and mix gently.
13. Centrifuge for 7 min 800  g at RT.
14. Discard the supernatant.
15. Resuspend the cell pellet in 40 mL PAB (1%) and mix gently.
16. Count cells of the filtered bone marrow using the Coulter AcT
diff 2 Analyzer.

3.2 Monoclonal 1. Pipet appropriate quantity of patient sample to achieve

Antibody Staining 500,000 cells in each 12  75 mm tube.
2. Pipet 2–3 mL of PAB (1%) into each tube.
3. Add the precise amount of each monoclonal antibodies or
antibody cocktail to the appropriate tube.
4. Vortex each tube gently.

5. Incubate for 15–20 min in the refrigerator (2–8 C) in
the dark.
6. Pipet 2–3 mL of PAB (1%) into each tube.
7. Vortex each tube gently.
8. Add 250 μL of 1% PFA fixative and vortex.
9. Run the tubes or store them in the refrigerator (2–8  C) until
they can be acquired on the flow cytometer, preferably within
1 h of the staining.

3.3 Data Acquisition 1. Sequential gating strategies are recommended to identify and
quantify the pDCs (see Notes 1, 7 and 8), (Fig. 1).
(a) Gate out all doublets and nonviable cells (see Notes 9
and 10).
(b) Gate all HLA-DR+/CD45+ and all CD4+ cells
population.
(c) Exclude T lymphocytes (CD3+).
(d) Exclude monocytes (CD14+/CD64+/CD36+).
(e) Exclude all events that are CD123– and/or HLA-DR–.
(f) Report the percentage of pDCs as pDC% of total viable
events.

3.4 Data 1. Nonneoplastic pDCs in the bone marrow and peripheral blood
Interpretation frequently express CD4, CD22dim, CD33, CD36, CD38,
CD45, CD123, CD303, and HLA-DR [12, 13], and lack
198 Abdullah Alsuwaidan et al.

Fig. 1 Multiparametric flow cytometry analysis of a normal bone marrow specimen. The multicolor flow
cytometry panel includes several monoclonal antibodies such as CD2, CD4, CD5, CD11b, CD11c, CD13, CD15,
CD22, CD33, CD38, CD45, CD56, CD64, CD123, CD303, HLA-DR. pDCs (red population) are found just below
monocytes (green population), that is, same size to monocytes but less SSC, in the FSC/SSC plot. Additionally,
pDCs are slightly dimmer for CD45 with less SSC compared to monocytes in the CD45/SSC plot. In CD123/
HLA-DR plot, pDCs are double positive for CD123/HLA-DR compared to basophils (blue population) that are
CD123+/HLA-DR–. In this case, the nonneoplastic pDCs are characteristically positive for CD4, CD22, CD38,
CD45, CD123, CD303, and HLA-DR, and largely negative for CD2, CD5, CD11c, CD13, CD15, CD33, CD56, and
CD64

other myelomonocytic or lymphoid markers (see Note 2 and 7

and Notes 11–13).
2. In the forward/side scatter (FSC vs. SSC) plot, pDCs are
found just below monocytes in the two-dimensional scatter
plot (i.e., same size to monocytes, but less SSC) [9].
3. In the CD45 vs. SSC plot, pDCs had slightly dimmer for CD45
and less SSC compared to monocytes [9].
4. In CD3 vs. CD4 plots, the cluster of nonmonocytic CD4+/
CD3 are mainly pDCs providing no aberrant T cells are
present.
5. In the CD45 vs. CD34 plot, pDCs may show maturation
spectrum where early pre-pDCs (Stage I) are CD34+ and
 Enumeration of Plasmacytoid Dendritic Cells in Peripheral Blood and Bone. . . 199

CD45 dimmer than mature pDCs which are CD34 [14], (see
Note 14).

4 Notes

1. A panel of monoclonal antibodies are used to better character-

ize the pDCs and discriminate these cells from other hemato-
lymphoid cells that share some immunophenotypic findings.
2. It has been reported that a small subset of pDCs may express
CD7 or CD2 particularly in the bone marrow samples. How-
ever, the majority of pDCs should be negative for these markers
[7]. Also, CD33 can be negative in minority of cases.
3. The antibodies must be validated and titered for optimal signal-
to-noise ratio under the settings to be used, particularly if the
antibodies are pre-cocktailed.
4. Reagents should be used according to the manufacturer’s
instructions.
5. Daily QC, calibration controls, compensation procedure and
preventive maintenance must be performed according to the
manufacturer’s recommendations.
6. It is crucial to evaluate the specimen for hemodilution which
may affect the accurate pDCs count.
7. Clinical and morphologic correlation:
(a) Knowing the clinical presentation and the indications for
testing is essential in order to address the clinician
concerns.
(b) In cases with suspected increased pDCs, an examination of
cytospin morphology of the corresponding sample is
recommended. Additionally, correlation with the bone
marrow biopsy and clot section is required to provide
accurate assessment.
8. In normal condition, pDCs show reproducible staining pat-
tern. They usually present within or close to blasts gate. While
they can be detected using pregating strategy, we recommend
the use of cluster analysis method that involves identifying
multiple populations in ungated data
9. It is important to gate-out doublets “doublet discrimination”
during the pDC analysis by examining forward scatter area over
height (FSC-A and FSC-H) then exclude all events that do not
represent single cells.
10. Excluding the nonviable cells or debris can be performed
gating-out events with very low forward scatter in
200 Abdullah Alsuwaidan et al.

FSC-A vs. SSC-A plot. The other option to remove the nonvi-
able cells is to use a DNA-binding dye, such as 7-amino-acti-
nomycin D (i.e., 7-AAD) which can provide a good assessment
of sample integrity. However, it is essential to ensure that
excluding nonviable cells do not interfere with the analysis of
cell of interest.
11. Our protocol can be incorporated into routine evaluation of all
bone marrow and peripheral blood samples for pDCs, particu-
larly if there is increase in nonmonocytic CD4+/CD3
population.
12. The use of isotype controls is essential to exclude false-positive
events (nonspecific staining).
13. It is advised to include Fluorescence Minus One (FMO) con-
trols in multiparametric flow cytometry which help in identify-
ing the proper gating boundaries.
14. Based on our experience, the early immature pDC may down-
regulate CD303 expression. This is mainly seen in bone mar-
row sample.

References

1. Reddy V, Iturraspe JA, Tzolas AC (2004) Low 7. Elze MC, Ciocarlie O, Heinze A (2015) Den-
dendritic cell count after allogeneic hemato- dritic cell reconstitution is associated with
poietic stem cell transplantation predicts relapse-free survival and acute GVHD severity
relapse, death, and acute graft-versus-host dis- in children after allogeneic stem cell transplan-
ease. Blood 103(11):4330–4335. Epub 2004 tation. Bone Marrow Transplant 50
Feb 12 (2):266–273
2. Abe M, Wang Z, de Creus A (2005) Plasmacy- 8. Su RJ, Green R, Chen M (2018) Enumeration
toid dendritic cell precursors induce allogeneic of bone marrow plasmacytoid dendritic cells by
T-cell hyporesponsiveness and prolong heart multiparameter flow cytometry as a prognostic
graft survival. Am J Transplant 5 marker following allogeneic hematopoietic
(8):1808–1819 stem cell transplantation. Blood Cells Mol Dis
3. McKenna K, Beignon AS, Bhardwaj N (2005) 69:107–112
Plasmacytoid dendritic cells: linking innate and 9. Pilley J, Chen W, Fuda F et al (2017) The
adaptive immunity. J Virol 79(1):17–27. immunophenotypic appearance of plasmacy-
Review toid dendritic cells in various specimens. Mod
4. Tomasello E, Naciri K, Chelbi R (2018) Pathol 30(Suppl):370A
Molecular dissection of plasmacytoid dendritic 10. Swerdlow SH, Campo E, Jaffe ES et al (2017)
cell activation in vivo during a viral infection. WHO classification of tumours of haemato-
EMBO J 37(19):e98836 poietic and lymphoid tissues, 4th edn. IARC
5. Auletta JJ, Devine SM, Waller EK (2016) Plas- Press, Lyon, pp 174–177
macytoid dendritic cells in allogeneic hemato- 11. Chen W, Luu HS (2017) Immunophenotyping
poietic cell transplantation: benefit or burden? by multiparameter flow cytometry. Methods
Bone Marrow Transplant 51(3):333–343 Mol Biol 1633:51–73
6. Swiecki M, Colonna M (2015) The multiface- 12. Alculumbre S, Pattarini L (2016) Purification
ted biology of plasmacytoid dendritic cells. Nat of human dendritic cell subsets from peripheral
Rev Immunol 15(8):471–485 blood. Methods Mol Biol 1423:153–167
 Enumeration of Plasmacytoid Dendritic Cells in Peripheral Blood and Bone. . . 201

13. Setoodeh S, Fuda F, Chen W et al (2017) Flow 14. Martı́n-Martı́n L, Almeida J, Hernández-
cytometric immunophenotypic features of Campo PM (2009) Immunophenotypical,
blastic plasmacytoid dendritic cell neoplasm: a morphologic, and functional characterization
single institution experience. Mod Pathol 30 of maturation-associated plasmacytoid den-
(Suppl):376A–377A dritic cell subsets in normal adult human bone
marrow. Transfusion 49(8):1692–1708
 Chapter 12

Immunophenotyping of Circulating Endothelial Cells

and Endothelial Microparticles
Nicholas Wanner and Kewal Asosingh

Abstract
Flow-cytometric detection of circulating endothelial cells and endothelial microparticles is an essential tool
in studies of vascular diseases. Here we describe the principles and detailed methods for human blood
sample processing, storage, labeling, and gating of circulating endothelial elements.

Key words Endothelial, Flow cytometry, Microparticle, Immunophenotyping

1 Introduction

Studies of circulating endothelial cells and endothelium-derived

microparticles are gaining interest in vascular biology. Enumeration
and immunophenotyping of these circulating elements are becom-
ing a standard flow cytometry application. Circulating microparti-
cles and endothelial cells have been reported as markers of
angiogenesis, endothelium damage, or vascular remodeling in sev-
eral diseases [1–5].
Endothelial cells in the blood circulation are infrequent. Thus,
all best practices for rare event flow cytometry are highly recom-
mended to identify these populations successfully [6–10]. The
signal-to-noise ratio in rare event detection requires the acquisition
of many events and stringent elimination of artifacts and debris to
gate for true endothelial cells. Nonspecific binding of antibodies
must be reduced by the use of Fc-block or blocking serum before
and during the incubation with test antibodies. This can also be
facilitated by using a dump channel to exclude nonspecific and
nonendothelial subsets. Another benefit of a dump channel is an
enrichment of the rare event in the remaining population. A nuclear
stain is included in the panel to discriminate true nucleated cells and
cell fragments. DNA stain will also allow for the identification of
hypodiploidy in cells representing dead and dying cells, which can

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_12, © Springer Science+Business Media, LLC, part of Springer Nature 2019

203
204 Nicholas Wanner and Kewal Asosingh

display nonspecific antibody binding and should be excluded from

analysis. In addition to the above steps all other general flow
cytometry gating best practices apply. Time gating should be per-
formed to select for events collected only during periods of laminar
flow in the flow cell. Clusters of cells or cell doublets can be
eliminated by analyzing the pulse height and width and excluding
events that do not have a width in proportion to the height. Any
remaining cell debris can be removed by discriminating events with
light scatters too low to be cells.
Flow cytometry of circulating microparticles is an evolving
technology. Guidelines for consensus best practices are still
emerging [11–15]. As detailed below, special precautions should
be considered to reduce background noise originating from parti-
cles in buffers and antibody aggregates. Antibodies titers should be
determined using the microparticles of interest in titration experi-
ments, and the presence of correct vesicles should be validated by
treating samples with a detergent. Aggregation and swarming of
microparticles should be prevented.

2 Materials

2.1 Blood Sample 1. 10% Formaldehyde: 16% formaldehyde, 0.9% saline. A new
Processing 10 mL ampule of 16% electron microscopy-grade formalde-
hyde is diluted with 6 mL of saline to create a 10% solution.
Store at room temperature for up to 2 weeks.
2. 10% Triton X-100: Triton X-100 detergent, 0.9% saline. Add
100 mL of 0.9% saline to a glass jar. Add 20 mL of Triton
X-100 to the saline. Measure out an additional 80 mL of saline
and with the pipet used to add the Triton X-100, add the saline.
Add a magnetic stirring flea and allow all particles to dissolve
while stirring on a magnetic stirrer. Aliquot 932 μL into cryo-
vials and store in liquid nitrogen.
3. Lyse/perm Buffer: A vial of 10% Triton X-100 is thawed then
added to 39 mL of 0.9% saline. The buffer is prewarmed to
37
C before using.
4. Phosphate Buffered Saline (PBS): 137 mM sodium chloride,
2.7 mM potassium chloride, 9.8 mM phosphate buffer. Does
not contain calcium or potassium.
5. Wash Buffer: 4% fetal bovine serum (FBS), PBS. 20 mL of FBS
is added to 480 mL of PBS. The solution is filtered through a
0.22 μm filter. Store at 4
C.
6. Freezing Medium: RPMI 1640, 20% FBS, 10% glycerol. Com-
bine 35 mL of RPMI 1640 with 10 mL of FBS and 5 mL of
glycerol. Store at 20
C.
 Endothelial Flow Cytometry 205

2.2 Cell Surface 1. Antibody diluting Buffer: 1% bovine serum albumin (BSA),
Staining for Flow PBS. Weigh and add 5 g of BSA to 500 mL of PBS with a
Cytometry stirring flea and stir until dissolved. Filter the solution through
a 0.22 μm filter. Store at 4
C.
2. Blocking Buffer: antibody diluting buffer, 10% normal goat
serum (NGS) Abcam, Human Trustain Fc-block, Biolegend.
Amount made is based on number of samples being stained.
250 μL made per sample; 25 μL of NGS and 6.25 μL of
Fc-block added to 218.75 μL of diluting buffer.
3. 5 mL Polystyrene Round-Bottom Tubes.
4. Antibodies: CD3 PE-Cy7, Biolegend, use at 1/1600. CD19
PE-Cy5, Biolegend, use at 1/27. CD45 Alexa Fluor 700, Bio-
legend, use at 1/800. CD34 FITC, BD Bioscience, use at 1/5.
All antibodies are diluted in antibody diluting buffer.
5. 40 ,6-Diamidino-2-phenylindole (DAPI), Life Technologies,
use at 1/400.
6. AbC Total Antibody Compensation Bead Kit, Life
Technologies.
7. Ultra Rainbow Calibration Kit, Spherotech.
8. CS&T Research Bead, Becton Dickinson.

2.3 Microparticle 1. Prostaglandin: 10 mM prostaglandin stock is diluted 1/1000

Isolation and Staining when added to plasma (1 μL of 10 mM prostaglandin to 1 mL
for Flow Cytometry of plasma).
2. Annexin-V Binding Buffer: 10 mM HEPES, 140 mM NaCl,
2.5 mM CaCl2. Add 69 mg of CaCl2, 2.5 mL of 1 M HEPES,
and 7 mL of 5 M NaCl to 150 mL of diH2O. Stir until the
CaCl2 has dissolved then use a graduated cylinder to bring the
volume to 250 mL with diH2O. Store at 4
C.
3. r-hirudin: DSM Nutritional Products, 2000 ATU/vial, use at
1/1000.
4. Apogee Mix for Flow Cytometer Performance Assessment,
Apogee.
5. SPHERO Rainbow Calibration Particles, 2.08 μm,
Spherotech.

3 Methods

3.1 Blood Processing 1. Pipette 4 mL of whole blood per 50 mL tube. If less than 4 mL
for White Blood Cells of blood is available, add saline to bring the volume to 4 mL.
2. Add 2.6 mL of 10% formaldehyde to the blood and incubate
for 10 min at room temperature (RT).
206 Nicholas Wanner and Kewal Asosingh

3. After 10 min add 40 mL of lyse/perm buffer prewarmed to

37
C and thoroughly vortex. Incubate for 20 min at RT (see
Note 1).
4. Centrifuge for 10 min at 1000  g. Aspirate the supernatant
and wash the cells by adding 5 mL of wash buffer. Incubate
cells in wash buffer for 15 min at RT.
5. Centrifuge again for 10 min at 1000  g and repeat the aspira-
tion, washing and incubation step.
6. During the second incubation, count cells to determine the
number of vials to freeze in. There should be between
3.5  106 and 7  106 cells per vial.
7. Suspend cells in freezing medium and freeze in cryovials at
80
C for 1–2 days before moving to liquid nitrogen. Alter-
natively samples can proceed with the staining process.

3.2 Cell Surface 1. Thaw aliquots of cells in cryovials at room temperature for
Staining for Flow 15 min.
Cytometry 2. Wash cells of freezing medium by diluting suspended cells with
4 mL of PBS in 5 mL polystyrene round bottom tubes.
3. Centrifuge samples at 900  g for 5 min (all subsequent spins
will be at 900  g for 5 min), then aspirate the supernatant.
4. Resuspend each sample in 1 mL of blocking buffer by pipetting
up and down. Incubate cells at RT in the dark for 20 min.
5. Remove an aliquot for the unstained sample.
6. Centrifuge all samples and aspirate.
7. The cells that are stained with DAPI are resuspended in 75 μL
of diluted DAPI (1/400 in PBS). Vortex the cells and incubate
for 15 min at RT in the dark.
8. Wash with 1 mL of diluting buffer then centrifuge.
9. Resuspend the DAPI stained cells in 300 μL of diluting buffer
and divide into the DAPI comp, FMOs, and sample tubes
(75 μL per tube) (see Note 2).
10. Add 200 μL of diluting buffer to all tubes except the DAPI
comp tube. Put the DAPI comp tube aside and keep at 4
C in
the dark.
11. Centrifuge remaining samples and aspirate.
12. Add 100 μL of appropriate antibody dilutions to respective
tubes and vortex (see Note 3). Incubate at room temperature
on the platform lab shaker for 30 min.
13. Wash samples with 500 μL of diluting buffer and centrifuge.
14. Aspirate the supernatant and resuspend in 100 μL of BD
FACSFlow, Becton Dickinson and keep at 4
C until acquisi-
tion on the flow cytometer (see Note 4). See Fig. 1 for gating
strategy of endothelial cells.
 Endothelial Flow Cytometry 207

Fig. 1 Circulating endothelial cell gating strategy. Time gating (a), and aggregate exclusion (b) corrected for
fluidic disturbances and cell doublets. DAPI was used to select cells in G0/G1 (c). CD19+ (d) and CD3+ (e) cells
were excluded, then CD45 cells were selected (f) using the CD45 FMO. Finally CECs were gated using CD34
(g). The CD34 FMO was used to set gate boundaries

3.3 Isolation 1. Blood is collected into cell preparation tubes with sodium
of Microparticles heparin to prevent coagulation and platelet activation. Samples
for Flow Cytometry are processed as soon as possible.
2. Platelet-rich plasma is isolated by centrifuging whole blood at
150  g for 20 min with the centrifuge brake off (see Note 5).
3. The supernatant is platelet-rich plasma (PRP) and it is collected
from the tube and placed in a 15 mL tube. For each 1 mL of
PRP, 1 μL of prostaglandin is added to prevent platelet
activation.
4. Centrifuge PRP at 150  g for 10 min. The same brake
instructions from step 2 are followed in this step.
5. The supernatant is again collected and placed in a separate
15 mL tube, then centrifuged at 1500  g for 15 min with
the acceleration and brake of the centrifuge at max.
6. Collect the supernatant and centrifuge at 2500  g for 15 min
(see Note 6).
7. Collect the supernatant again and centrifuge a second time at
2500  g for 15 min.
8. The supernatant is collected then aliquoted in 250 μL portions
before being frozen at 80
C.

3.4 Staining 1. Platelet free plasma (PFP) is thawed at room temperature and
of Microparticles 18 μL of PFP per sample is transferred to an Eppendorf tube.
for Flow Cytometry 2. A reference sample containing 25 μL of PFP is also thawed and
24 μL of PFP is transferred to an Eppendorf tube.
3. AV binding buffer is prepared by adding r-hirudin to the
buffer at a 1 to 1000 dilution then filtering the buffer twice
through a 0.1 μm filter. R-hirudin will prevent aggregation of
microparticles.
4. 400 μL of AV binding buffer is added to all tubes.
5. The tubes are centrifuged at 15,000  g for 5 min (all
subsequent spins are at this speed and length of time) in
order to pellet microparticles (MPs).
208 Nicholas Wanner and Kewal Asosingh

6. A p1000 pipet is used to remove 403 μL of supernatant from

the sample tubes which leaves a remainder of 15 μL of super-
natant in the tubes. 404 μL of supernatant from the reference
tube is removed which leaves a remainder of 20 μL of superna-
tant in the tube (see Note 7).
7. The residual volumes are mixed with a pipet to suspend the
MPs, then 5 μL is aliquoted into separate tubes.
8. For the samples the tubes are unstained, AV only, and
AV + CD144.
9. For the reference sample the tubes are unstained, AV only, CD
144 only, and AV + CD144.
10. To the CD144 only and AV + CD 144 tube, 20 μL of CD144
antibody (see Note 8) is added to the tube, pipetting up and
down to thoroughly mix (see Note 9).
11. Samples incubate for 30 min at room temperature in the dark.
12. After the incubation, samples that were stained with CD144
are washed with 400 μL of binding buffer then centrifuged at
15,000  g for 5 min.
13. For the samples that were washed, the supernatant is aspirated
down to the 100 μL mark (see Note 10).
14. 95 μL of AV binding buffer is added to the unstained and AV
only tubes to bring the volume to 100 μL.
15. 5 μL of AV is added to 100 μL of supernatant in the AV only
and AV + CD144 tubes then pipetted up and down to mix (see
Notes 11 and 12).
16. Samples incubate for 30 min at room temperature.
17. After the incubation, samples are washed with 400 μL of
binding buffer then centrifuged at 15,000  g for 5 min.
18. The supernatant is aspirated down to the 100 μL mark then the
wash/centrifuge step is repeated.
19. The supernatant is aspirated to the 100 μL mark then 50 μL of
AV binding buffer is added for a final volume of 150 μL.
20. Samples can be stored at 4
C protected from light until they
are run on the cytometer (see Note 13). See Fig. 2 for gating
strategy of endothelial-derived microvesicles.

4 Notes

1. This protocol utilizes a one-step RBC lysis and fixation and

permeabilization step [16, 17]. This procedure allows intracel-
lular immunophenotying and preserves epitopes for phosflow.
In addition, fixation before cryopreservation provides better
light scatter resolution than unfixed cryopreserved cells
 Endothelial Flow Cytometry 209

Fig. 2 Gating of endothelial-derived microvesicles in plasma. Small Angle Light Scatter (SALS) and Large
Angle Light Scatter (LALS) were used to gate the microvesicle population based on size (a). Annexin-V+ events
were selected on a LALS/Green fluorescence channel plot (b). Annexin-V+ region was gated based on the
unstained sample (d). CD144 (VE-cadherin)+ events in the annexin-V+ gate were selcted on an LALS/orange
fluorescence channel plot (c) based on an Annexin-V only control (e) Samples were acquired at a flow rate of
6.01 μL/min for 2 min. At least 5000 annexin-V+ CD144-PE+ microparticles were acquired

[16, 17]. Cryopreservation also allows for batch staining and

analysis. Users can stain fresh samples, if desired.
2. The DAPI stained cells are aliquoted into four tubes. The tubes
are DAPI comp, CD45 FMO, CD34 FMO, and complete
panel.
3. 100 μL of antibody dilutions are made for each tube The CD45
FMO dilution will be have all antibodies (CD3, CD19, and
CD34) except CD45. The CD34 FMO will be stained with all
antibodies (CD3, CD19, and CD45) except CD34. The sam-
ple tube will be stained with all antibodies (CD3, CD19,
CD45, and CD34). In this protocol CD34 is used to identify
endothelial cells. Other antibodies have been cited in the liter-
ature as markers of endothelial cells and can be substituted for
CD34 at the discretion of the user. In this case, it should be
validated that the fixation and permeabilization did not com-
promise antigenicity.
210 Nicholas Wanner and Kewal Asosingh

4. Samples were run on a BD Fortessa. CS&T beads were run

daily for cytometer set up. Six peak rainbow beads were run
before each experiment to calibrate fluorescence channels.
5. If your centrifuge is able, the brake can be set to the minimum
setting when the centrifuge is spinning at 50  g or less.
6. Centrifuging at 2500  g will pellet the platelets while keeping
MPs in the supernatant.
7. It is important to not aspirate too much supernatant leaving
less than 15 μL. 5 μL must be aliquoted 3 times so at least
15 μL is needed. More than 15 μL can remain, but it is
important to note exactly how many μLs remain.
8. All antibody vials should be centrifuged at 15,000  g for
10 min before used in the experiment in order to pellet
aggregates.
9. Additional markers for endothelial derived MPs were tested
including CD31 and CD34. CD144 was found to be the
most highly expressed on vesicles.
10. Eppendorf tubes with 100 μL gradations should be used in
order to aspirate to a specific volume of 100 μL.
11. AV is used as a general MP marker, but may not be binding to
all MPs. Other MP markers can be used if the user determines
it to be advantageous [18].
12. Presence of true MPs should be validated by dissolving vesicles
by treating samples with detergent, such as Triton-X 100 [19].
13. Samples were run on the Apogee A50 Micro-cytometer. Size
beads and rainbow calibration beads were run in order to
standardize size resolution and fluorescence respectively.

References
1. Goon PK, Lip GY, Boos CJ, Stonelake PS, 5. Curtis AM, Edelberg J, Jonas R, Rogers WT,
Blann AD (2006) Circulating endothelial Moore JS, Syed W, Mohler ER 3rd. (2013)
cells, endothelial progenitor cells, and endo- Endothelial microparticles: sophisticated vesi-
thelial microparticles in cancer. Neoplasia cles modulating vascular function. Vasc Med
8:79–88 18:204–214
2. Del Papa N, Colombo G, Fracchiolla N, Mor- 6. Zimmerlin L, Donnenberg VS, Donnenberg
onetti LM, Ingegnoli F, Maglione W, Comina AD (2011) Rare event detection and analysis
DP, Vitali C, Fantini F, Cortelezzi A (2004) in flow cytometry: bone marrow mesenchymal
Circulating endothelial cells as a marker of stem cells, breast cancer stem/progenitor cells
ongoing vascular disease in systemic sclerosis. in malignant effusions, and pericytes in disag-
Arthritis Rheum 50:1296–1304 gregated adipose tissue. Methods Mol Biol
3. Bertolini F, Shaked Y, Mancuso P, Kerbel RS 699:251–273
(2006) The multifaceted circulating endothe- 7. Donnenberg AD, Donnenberg VS (2007)
lial cell in cancer: towards marker and target Rare-event analysis in flow cytometry. Clin
identification. Nat Rev Cancer 6:835–845 Lab Med 27:627–652. viii
4. Dignat-George F, Boulanger CM (2011) The 8. Roederer M (2008) How many events is
many faces of endothelial microparticles. Arter- enough? Are you positive? Cytometry A
ioscler Thromb Vasc Biol 31:27–33 73:384–385
 Endothelial Flow Cytometry 211

9. Tibbe AG, Miller MC, Terstappen LW (2007) 16. Chow S, Hedley D, Grom P, Magari R, Jacob-
Statistical considerations for enumeration of berger JW, Shankey TV (2005) Whole blood
circulating tumor cells. Cytometry A fixation and permeabilization protocol with red
71:154–162 blood cell lysis for flow cytometry of intracellu-
10. Hedley BD, Keeney M (2013) Technical issues: lar phosphorylated epitopes in leukocyte sub-
flow cytometry and rare event analysis. Int J populations. Cytometry A 67:4–17
Lab Hematol 35:344–350 17. Chow S, Hedley D, Shankey TV (2008) Whole
11. Nolan JP (2015) Flow cytometry of extracellu- blood processing for measurement of signaling
lar vesicles: potential, pitfalls, and prospects. proteins by flow cytometry. Curr Protoc
Curr Protoc Cytom 73(13.14):11–16 Cytom. Chapter 9: Unit 9 27
12. Erdbrugger U, Lannigan J (2016) Analytical 18. de Rond L, van der Pol E, Hau CM, Varga Z,
challenges of extracellular vesicle detection: a Sturk A, van Leeuwen TG, Nieuwland R, Cou-
comparison of different techniques. Cytometry mans FAW (2018) Comparison of generic fluo-
A 89:123–134 rescent markers for detection of extracellular
13. Stoner SA, Duggan E, Condello D, vesicles by flow cytometry. Clin Chem
Guerrero A, Turk JR, Narayanan PK, Nolan 64:680–689
JP (2016) High sensitivity flow cytometry of 19. Rose JA, Wanner N, Cheong HI, Queisser K,
membrane vesicles. Cytometry A 89:196–206 Barrett P, Park M, Hite C, Naga Prasad SV,
14. Nolan JP, Moore J (2016) Extracellular vesi- Erzurum S, Asosingh K (2016) Flow cyto-
cles: great potential, many challenges. Cytome- metric quantification of peripheral blood cell
try B Clin Cytom 90:324–325 beta-adrenergic receptor density and urinary
endothelial cell-derived microparticles in pul-
15. Nolan JP, Duggan E (2018) Analysis of indi- monary arterial hypertension. PLoS One 11:
vidual extracellular vesicles by flow cytometry. e0156940
Methods Mol Biol 1678:79–92
 Chapter 13

Rare Event Phenotyping and Molecular Characterization:

Circulating Tumor Cells
Moen Sen, Ling Wang, Liping Yu, and Erica L. Carpenter

Abstract
Noninvasive isolation of circulating tumor cells (CTCs) from patient blood samples allows for interrogation
of valuable molecular and phenotypic information useful for disease diagnosis and monitoring response to
therapy. However, CTCs are extremely rare relative to red and white blood cells (R/WBC), thus making
CTC isolation from unmanipulated whole blood very time-consuming. Moreover, single CTC analysis
often requires hand-picking, a step that can result in more CTC loss and compromised cell integrity. Here
we describe an automated flow cytometry-based approach for isolation and analysis of single, viable CTCs
that combines gentle RBC lysis and magnetic, no-wash negative-depletion of WBCs, followed by a highly
adaptable sorting protocol for rare cells of interest. Multiparametric flow-cytometric panels allow probing
of numerous extracellular markers for immunophenotyping, while whole transcriptome analysis contributes
to molecular characterization of individual CTCs. Index sorting links single CTC proteogenomics
information.

Key words Circulating tumor cells, Acoustic cell enrichment, Magnetic depletion, Flow-cytometric
analysis, Molecular indexing, Phenotyping

1 Introduction

Circulating tumor cells (CTCs) are rare cells shed from the primary
tumor into blood. The enumeration and characterization of CTCs
isolated from the blood, also known as liquid biopsy, can serve as a
valuable alternative to repeated invasive tumor biopsies for diagnos-
ing and monitoring disease. However, isolation and characterization
of CTCs has been challenging due to the fact that CTCs are extremely
rare, present at 1–10 cells per ten billion blood cells, and also hetero-
geneous [1]. CellSearch®, the only US Food and Drug Administra-
tion (FDA) approved CTC characterization device has been used for
the enumeration of CTCs in patients with metastatic breast, colorec-
tal or prostate cancer. However, CellSearch-based phenotyping is
restricted to one open channel, and additional steps are necessary
for single cell isolation and molecular analysis [2–6].

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_13, © Springer Science+Business Media, LLC, part of Springer Nature 2019

213
214 Moen Sen et al.

Fig. 1 Schematic overview of sample preparation, flow-cytometric phenotyping, and whole-transcriptome

molecular analysis of circulating tumor cells

In this chapter, we describe an integrated workflow for the

isolation, immunophenotyping and downstream molecular charac-
terization of rare cells using Yellow Fluorescent Protein (YFP)
positive CTCs isolated from the blood of tumor-bearing KPCY
(Kras;p53;Cre;YFP alleles) mice as a model [7]. The YFP lineage
is expressed in all cells of pancreatic origin, including CTCs, thus
facilitating rare cell selection among YFP–/CD45+ WBCs. The
workflow is simple and integrates isolation, phenotyping, and
molecular characterization of single CTCs. The protocol is com-
posed of several key steps, as illustrated in Fig. 1, including: (1) sam-
ple preparation involving staining cells for extracellular markers
followed by magnetic labeling of WBCs, (2) gentle, no-wash RBC
lysis, (3) magnetic depletion of WBCs, (4) acoustic focusing result-
ing in reduction of blood cells and debris and consequent enrich-
ment of rare CTCs, (5) analysis and sorting of cells on a flow
cytometer, and (6) whole transcriptome analysis of sorted single
CTCs. While broad variations of the protocol have been described
by other groups and us, the elimination of washes and a separate
isolation step preceding molecular analysis greatly increases the
workflow efficiency and minimizes cell loss [2, 8–16]. In addition,
the combination of immunophenotyping with transcriptional
 CTC Phenotypic and Molecular Characterization 215

analysis at the single cell level affords truly high dimensional char-
acterization of CTCs [16, 17]. This protocol increases sensitivity
and provides better rare cell characterization by combining enrich-
ment and index sorting of rare cells with downstream single cell
whole transcriptome analysis [18–20].

2 Materials

2.1 Mouse Blood 1. 70% ethanol.

Sample Preparation 2. 27 G 16 mm 1 mL insulin syringe.
3. Heparin.
4. 15 mL conical tube.
5. Aluminum foil.
6. Shaker.
7. Staining antibodies against WBCs (CD45) and known CTC
markers such as EPCAM and ECAD.
8. IMag™ Magnetic beads.
9. G Biosciences RBC Lysis Buffer.
10. 5 mL polypropylene tube.

2.2 Setting Up Pre- 1. Ficoll® Paque.

enrichment Platform 2. BD FACSFlow™ sheath solution or equivalent sheath solution
(BD Focus) for flow cytometry.
3. 50 mL Falcon conical tube.
4. Labview GUI.
5. Fluorescence-activated cell sorter (FACS) flow cytometer
sorter.
6. BD Focus (prototype being developed by BD Biosciences, San
Jose, California). The platform includes an analog pressure
control system, a magnetic separator with plastic tubing for
conducting magnetic separation, an acoustic-driven microflui-
dic device (chip) and fluidic tubing that passes sample from
sample station through magnetic tubing, acoustic chip and
sample inlet on FACS.

2.3 Single Cell 1. BD™ Precise WTA Single Cell Encoding Plate, 96 Well
Sorting (BD Biosciences).
2. Flow-cytometric cell sorter (also used in Subheading 2.2,
item 5).
216 Moen Sen et al.

2.4 Whole 1. Pipettes (1 μL–1000 μL volume capability).

Transcriptome 2. RNase-free filter pipette tips.
Analysis
3. Microcentrifuge tubes, 1.5 and 2.0 mL tubes.
4. 0.2 mL PCR 8-strip tubes.
5. UVP UV3 HEPA PCR Workstation (UVP) or equivalent.
6. PCR plate spinner (VWR) or equivalent.
7. Thermal cycler with heated lid.
8. Tube Magnet.
9. Magnetic separation stand for 0.2 mL tubes.
10. 2100 analyzer (Agilent Technologies) or equivalent.
11. Qubit Assay Tubes.
12. Qubit 3.0 Fluorometer (Thermo Fisher Scientific) or
equivalent.
13. Illumina sequencer (Example: Illumina NextSeq 500 Sequenc-
ing System).
14. AMPure XP Reagent (Beckman Coulter).
15. Pure ethyl alcohol (ethanol), Molecular Biology Grade.
16. Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific).
17. Agilent DNA High Sensitivity Kit (Agilent Technologies).
18. 80% (v/v) ethanol (ethyl alcohol): In a new 50 mL conical
tube, add 4.0 mL of nuclease-free water and 16 mL of ethanol.
Vortex and mix well. Prepare fresh and use
1 day.
19. BD™ Precise WTA Single Cell Kit (elution buffer included)
(BD Biosciences).

3 Methods

3.1 Mouse Blood All steps involving preparation of blood sample should be per-
Sample Preparation formed at room temperature
1. Coat the syringe and 15 mL conical tube with an anticoagulant
like heparin.
2. Euthanize the mouse according to appropriate ethical
guidelines.
3. Spray 70% ethanol to disinfect the mouse.
4. Insert the 1 mL syringe right below the rib cage at a 45 angle
and slowly draw blood out by cardiac puncture (see Note 1).
Transfer the blood to a heparin coated 15 mL conical tube.
5. Set out lysis buffer (4 the volume of blood collected) at room
temperature for use in step 8.
 CTC Phenotypic and Molecular Characterization 217

6. Add staining antibodies against markers of interest at appro-

priate concentrations and incubate for 20 min at room temper-
ature covered in aluminum foil on a shaker (see Notes 2 and 3).
This includes, for example, antibodies against known CTC
markers like EPCAM and ECAD as well as the CD45-PE
antibody that serves to both label and, together with
the beads described in the next step, deplete unwanted
CD45+ WBCs.
7. Add magnetic beads to deplete unwanted cells at a volume of
1:10. Thus, for 1 mL of whole blood, add 100 μL of magnetic
beads. For example, beads conjugated to anti-PE are used to
deplete unwanted WBCs labeled with CD45-PE (see Note 4).
Incubate for 20 min at room temperature (RT) covered in
aluminum foil on a shaker.
8. Add 4 RBC lysis buffer (4 times the volume of whole blood)
and incubate for 20 min at room temperature covered in alu-
minum foil on a shaker (see Note 5).
9. Transfer the diluted, stained blood sample into a 5 mL poly-
propylene tube through a cell strainer cap (see Note 6).

3.2 Set Up BD Focus The BD Focus should be set up while the blood sample is being
processed. In our lab, the BD Focus set up and mouse blood sample
preparation is conducted simultaneously by two people.
1. Prepare wash buffer for the BD Focus washing step by making
20% Ficoll® Paque in BD FACSFlow™ sheath solution.
2. Turn on the BD Focus.
3. Start the LabVIEW GUI (BD Biosciences, California) that
monitors flow rates read from flow meters (Sensirion, Switzer-
land) on the sorter.
4. Push the magnet tubing into the magnetic field to engage
magnetic separation.
5. Place a 50 mL Falcon conical tube containing 50 mL fresh wash
buffer in the “wash port,” and empty 50 mL Falcon conical
tube on the “waste” port.
6. Connect the sample line of BD Focus to the sample line of the
FACS system. Turn on to backflush BD Focus sample line with
sheath fluid. The fluidic lines are primed when the wash buffer
is pushed to the waste and sample tube.
7. Monitor air in the chip via MicroViewer, to make sure all air is
pushed from the chip by the flow of saline from the sheath tank.
8. Place a 50 mL Falcon conical tube containing 50 mL fresh wash
buffer in the “wash port,” an empty 50 mL Falcon conical tube
on the “waste port,” and a 5 mL polypropylene test tube with
FACS buffer on BD Focus (Fig. 1).
218 Moen Sen et al.

9. Adjust pressure regulators on BD Focus to achieve positive

flow rates on sample line to FACS, wash buffer to chip, and
waste line from chip to waste tube. Target values are sample line
to FACS 60–70 μL/min, wash line 90–110 μL/min, and waste
line 130–140 μL/min (see Note 7).
10. Check and set a frequency of 2.17 MHz and 15 volts on the pzt
driver board to set the acoustic focusing chip at its optimal
performance for cell focusing. The pzt frequency is the fre-
quency at which the chip vibrates. Chip performance is sensi-
tive to the frequency and amplitude. The frequency does not
need to be changed until the chip is replaced. The manufac-
turer provides a suggested frequency and amplitude for each
chip they fabricate.
11. Stop the pressure regulators when target flow rates are
achieved. Replace the sample tube with the 5 mL polypropyl-
ene tube containing the experimental sample.

3.3 Run Sample on 1. Start pressure regulators and let fluidics get to equilibrium
BD Focus state, which typically takes a minute. Target rare cells are
enriched as the sample undergoes magnetic depletion of
unwanted cells and acoustic focusing that washes RBC lysate/
debris away.
2. Monitor event rate on flow cytometer and adjust pressure
regulators to control sample flow rate.

3.4 Flow-Cytometric 1. Draw dot plots in FACS software and employ appropriate gating
Analysis strategies to identify and characterize rare events. Gate out
debris based on size (FSC v SSC). This is followed by gating
on “live cells” that are DAPI negative. Following live/dead
selection, live YFP+ CD45 rare CTCs are gated on for sorting.
Gene expression profiling by whole transcriptome analysis is
routinely performed by us to verify cancer cell identity in sorted
CTCs [16]. As shown in the representative plots in Fig. 2, in one
experiment 16 (0.11% of all events, 0.16% of live cells) YFP+
CTCs were detected in the blood of a tumor-bearing KPCY
mouse (We have observed CTCs in the range of 1–254 cells/
mL in KPCY mice with high metastatic burden) [16]. To better
visualize the rare YFP+ events, the properties of the YFP gate
were edited on FACS analysis software to show up as larger dots
on the FACS plots (Fig. 2). Panels of multiple markers can be
optimized to minimize spectral overlap to phenotype CTCs.

3.5 Single Cell 1. Set up the flow sorter to sort per the manufacturer’s
Sorting instructions.
2. Chill the 96 well plate holder of the sorter at 4  C for 10 min.
3. Thaw the sealed 96 well Precise WTA plate at room tempera-
ture and spin at 377  g for 10 s to collect the reagents in each
 CTC Phenotypic and Molecular Characterization 219

Fig. 2 Gating strategy for the detection of YFP+ CD45 circulating tumor cells

well. Place the plate on ice, slowly peel the seal from the plate
and place it in the chilled plate holder of the sorter.
4. Sort single cells into the 96 well encoding plate and leave 3–6
wells empty for nontemplate controls. A few pools of 10 or
more cells per well can also be included as sample input con-
trols (see Note 8).
5. Immediately place the plate on ice, seal the plate with a plate
cover, vortex and spin down at 4  C, 377  g for 10 s and
freeze at 80  C (see Note 9).

3.6 Whole Before conducting the procedure for WTA library prep, dedicate
Transcriptome two isolated workspaces in the laboratory as preamplification and
Analysis postamplification workplaces. In our laboratory, we used a PCR
workstation as the preamplification workspace, and a separate clean
3.6.1 Perform Reverse space on the bench as the postamplification workspace.
Transcription, Pool and In the preamplification workspace:
Purify the Product
1. Briefly spin the WTA plate for 10 s at RT on the PCR plate
spinner and incubate the plate at 65  C for 3 min (Program 1 in
the preamplification thermal cycler). Cool on ice.
2. Prepare the RT enzyme master mix sufficient for a 96-well
plate: 324 μL of Nuclease Free Water, 240 μL of Precise RT
Buffer, 12 μL of RNase Inhibitor, and 24 μL Reverse Tran-
scriptase, for a total volume of 600 μL. Use a multichannel
pipette to dispense 5 μL of the mix into each of the 96 wells.
3. Seal the plate followed by a brief vortex and 10 s spin at
377  g at room temperature (RT). Run Program 2 in the
preamplification thermal cycler: 42  C for 30 min ! 80  C for
5 min ! hold at 4  C.
4. Spin the plate for 10 s at 671  g at RT and pool all reactions into
one single 2 mL tube by using a multichannel pipette and then a
single channel pipette. The final volume is approximately 900 μL.
5. Purify the pooled samples using 900 μL of AMPure XP beads.
6. Vortex to mix well and incubate at room temperature for
5 min. Place tube on magnet and wait 5 min until the liquid
is clear and separating from the beads. Keep the tube on the
220 Moen Sen et al.

magnet and carefully remove the liquid and discard. Gently

rinse the beads using 2 mL 80% (v/v) ethanol (prepare fresh
and use
1 day). Immediately remove all residual ethanol
solution with a pipette. Repeat this wash step once.
7. Remove the residual ethanol as much as possible and leave the
tube open on the magnet to air dry the beads for 3 min. To
elute the sample from the beads, remove the tube from the
magnet and add Elution Buffer in an amount as instructed in
the protocol. Wet the beads with the elution buffer and vortex
well to mix. Wait 1 min and return the tube to the magnet. Wait
5 min until the solution is clear. Recover the purified product
by carefully pipetting the solution into a new tube.
8. Leave the tube open to dry the beads at RT for 3 min before
eluting with 72 μL of Elution Buffer in a new 0.2 mL
PCR tube.

3.6.2 Perform Second In the preamplification workspace

Strand Synthesis and Purify
1. Dilute the Precise SS Enzyme Mix in a 1.5 mL tube by adding
the Product
38 μL of the Precise SS Synthesis Buffer and 2 μL of the Precise
SS Enzyme Mix. Gently mix and place on ice.
2. Prepare the second strand synthesis mix by adding 8 μL of the
diluted Precise SS Enzyme Mix to 72 μL of the pooled samples
in the 0.2 mL PCR tube from Subheading 2.4, bringing the
volume up to 80 μL. Gently mix and place on ice.
3. Perform ss synthesis and run Program 3 in the preamplification
thermal cycler: 16  C for 5 min ! hold at 4  C.
4. Immediately put the second strand synthesis products on ice
and add 5 μL of Precise SS Stop Solution. The total volume is
85 μL.
5. Purify the ss synthesis products using 100 μL of AMPure XP
beads, and wash twice using 200 μL of 80% ethanol (v/v) on
the magnet per Subheading 3.6.1, step 6.
6. Leave the tube open to dry the beads at RT for 3 min before
eluting in a new 0.2 mL PCR tube with 55 μL of Elution Buffer
per Subheading 3.6.1, step 7. Proceed immediately to next step.

3.6.3 Ligate the cDNA to In the preamplification workspace:

the Precise WTA Adapter
1. Prepare the cDNA ligation master mix by slowly pipetting
and Purify the Product
15 μL of Precise Ligase to 55 μL of the ss synthesis products
in the 0.2 mL PCR tube from Subheading 3.6.2 for a total
volume of 70 μL. Briefly vortex and spin the tube at 671  g
for 10 s and keep it at room temperature.
2. Prepare the cDNA ligation mix by adding 3 μL of Precise WTA
Adapter and 5 μL of Precise Ligase into the 70 μL of the cDNA
ligation master mix bringing it to a total of 78 μL.
 CTC Phenotypic and Molecular Characterization 221

3. Perform ligation by running Program 4 in the preamplification

thermal cycler: 23  C for 30 min ! hold at 4  C.
4. Briefly centrifuge the products at 671  g for 10 s, and then
place them at room temperature.
5. Purify the 78 μL of ligation products using 75 μL of AMPure
XP beads and wash twice using 200 μL of 80% ethanol (v/v) on
the magnet per Subheading 3.6.1, step 6.
6. Leave the tube open to dry the beads at RT for 3 min before
eluting with 55 μL of Elution Buffer in a new 0.2 mL PCR tube
per Subheading 3.6.1, step 7. Proceed immediately to the
next step.

3.6.4 Amplify the In the preamplification workspace:

Ligation Products and
1. Prepare the PCR master mix by adding the following compo-
Purify the Product
nents to the 0.2 mL tube containing the 55 μL of purified
ligation products from Subheading 3.6.3: 25 μL of nuclease-
free water, 20 μL of Precise WTA PCR Primer, and 100 μL of
the Precise WTA PCR Mastermix, for a total volume of 200 μL.
2. Gently mix by pipetting (vortex may not be gentle enough) and
place on ice. Pipette 50 μL of the PCR master mix into each of
the three new 0.2 mL PCR tubes on ice for a total of four tubes
with 50 μL each.
In the postamplification workspace:
3. Amplify the ligation products by running Program 5 in the
postamplification thermal cycler: 98  C for 30 s ! 20 cycles of
(98  C for 10 s ! 58  C for 15 s ! 72  C for 3 min) ! 72  C for
5 min ! hold at 4  C.
4. Combine the four PCR reactions into a new 1.5 mL tube. This
is the WTA product.
5. Purify the 200 μL of the WTA product using 200 μL of
AMPure XP beads, and wash twice using 1 mL of 80% ethanol
(v/v) on the magnet per Subheading 3.6.1, step 6.
6. Leave the tube open to dry the beads at RT for 3 min before
eluting with 20 μL of Elution Buffer in a new 1.5 mL PCR tube
per Subheading 3.6.1, step 7.

3.6.5 Perform Quality Store the purified WTA product at 20  C for
6 months or
Control on the WTA Product perform QC on the WTA product before proceeding to Subhead-
ing 3.6.6.
In the postamplification workspace:
1. Use 2 μL of the WTA product to measure the concentration
with the Qubit dsDNA HS Assay according to the
222 Moen Sen et al.

manufacturer’s instructions. WTA product concentrations are

typically 5 ng/μL.
2. Dilute the WTA product to about 1–3 ng/μL and use 1 μL to
measure the WTA product size and purity by running a High
Sensitivity DNA Chip on the Agilent Bioanalyzer per the man-
ufacturer’s instructions. The WTA product sizes are typically
>300 bp and peak around 2000 bp.

3.6.6 Perform Random In the postamplification workspace:

Primer Extension and Purify
1. Prepare the priming master mix: Calculate the volume (X) for
the Product
50 ng of WTA product from the estimated concentration, and
then add 34-X μL of nuclease-free water, 10 μL of Precise WTA
Library Random Primer Mix, and X μL of 50 ng WTA product
in a new 0.2 mL PCR tube, for a total volume of 44 μL.
2. Denature and prime the WTA product by running Program 6
in the postamplification thermal cycler: 95  C for 2 min ! 4  C
for 5 min ! hold at 4  C. Place the denatured WTA product
on ice.
In the preamplification workspace:
1. Prepare the primer extension reaction mix in a new 1.5 mL
tube by adding 7.5 μL of Precise Reaction Buffer and 1.5 μL of
Precise Extension Enzyme. Gently mix by pipetting (not vor-
tex) and place on ice. Add 6 μL of the mix to the 0.2 mL PCR
tube containing the 44 μL of the WTA product for a total of
50 μL primer extension reaction mix. Gently mix (not vortex)
by pipetting and place on ice.
2. Run Program 7 in the postamplification thermal cycler: 37  C
for 30 min ! 80  C for 20 min ! hold at 4  C for Primer
extension of the WTA product.
3. Purify the 50 μL of primer extension product using 35 μL of
AMPure XP beads, and wash twice using 200 μL of 80%
ethanol (v/v) on the magnet per Subheading 3.6.1, step 6.
4. Leave the tube open to dry the beads at RT for 3 min before
eluting with 20 μL of Elution Buffer in a new 0.2 mL PCR tube
per Subheading 3.6.1, step 7.

3.6.7 Amplify the Primer In the preamplification workspace:

Extension Product and
1. Prepare the library amplification master mix in a new 0.2 mL
Purify the Amplified Library
PCR tube by adding 25 μL of Precise WTA PCR Mastermix,
2.5 μL of Precise WTA Library Forward Primer, and 2.5 μL of
Precise WTA Library Index Primer 1 or Primer 2 (see Note 10),
for a total volume of 30 μL.
 CTC Phenotypic and Molecular Characterization 223

In the postamplification workspace:

2. Prepare the amplification mix by adding 20 μL of the primer
extension product from Subheading 3.6.6 to the 30 μL of the
library amplification master mix for a total of 50 μL. Gently mix
and briefly spin the tube for 10 s at 671  g at RT.
3. Amplify the amplification mix by running thermal cycler Pro-
gram 8 in the postamplification thermal cycler: 98  C for
30 s ! 12 cycles of (98  C for 10 s ! 65  C for 15 s ! 72  C
for 20 s) ! 72  C for 2 min ! hold at 4  C.
4. Purify 50 μL of amplified library using 35 μL of AMPure XP
beads, and wash twice using 200 μL of 80% ethanol (v/v) on
the magnet per Subheading 3.6.1, step 6.
5. Leave the tube open to dry the beads at RT for 3 min before
eluting with 20 μL of Elution Buffer in a new 0.2 mL PCR tube
per Subheading 3.6.1, step 7. The purified and amplified
library can be stored at 20  C for
6 months or proceed to
Subheading 3.6.8.

3.6.8 Perform Quality In the postamplification workspace:

Control on the Amplified
1. Measure the concentration of the amplified library using 2 μL
Library
of the amplified library with the Qubit dsDNA HS Assay Kit
according to the manufacturer’s instructions. Library concen-
trations range from 10 ng/μL to 30 ng/μL.
2. Verify the amplified library size and purity by diluting the
amplified library to 1–3 ng/μL and use 1 μL of the diluted
amplified library to run a High Sensitivity DNA Chip on the
Agilent Bioanalyzer according to the manufacturer’s instruc-
tions. The amplified library consists of fragments ranging from
300 to 1000 bp.

3.6.9 Sequence the Use an Illumina platform to sequence the libraries and generate
Amplified Libraries and 75  75 paired-end reads for a targeted range of 250,000–350,000
Analyze the Data paired reads per cell (see Note 10). Use the BD™ Precise Whole
Transcriptome Assay Analysis Pipeline on the Seven Bridges Geno-
mics platform to demultiplex and map reads from the sequencing
files, and calculate unique Molecular Indexes for each target with
built-in correction algorithms. Use the BD Genomics Data View
software (see Note 11) or open-source R statistical software for
secondary analysis such as dimensionality reduction, differentiated
gene expression, pathway analysis, and correlation of the Molecular
Indexes measured by the BD Precise™ assay with the fluorescence
intensity of the protein measured by flow cytometry at the single
cell level.
224 Moen Sen et al.

4 Notes

1. Move the needle up and down or change the angle of the

needle slowly if blood is not filling into the syringe. You should
expect 0.6–1 mL of blood from a single cardiac puncture.
2. Titrations to determine appropriate antibody concentrations
should be done using positive- and negative-control cell lines.
In our laboratory, we use extracellular markers conjugated to
fluorochromes to avoid an extra labeling step with a secondary
antibody.
3. Fluorescence-minus-one (FMO) controls can be used to set
gates [21]. If using FMOs, blood collected should be divided
accordingly between the sample and FMO controls. FMO
controls are then treated with magnetic beads and RBC lysis
similar to the sample.
4. We titrated IMag™ magnetic beads to determine a concentra-
tion that maximizes depletion without causing rare cell loss on
the magnet. Titrations may need to be done to optimize the
best conditions for each laboratory. In our laboratory, we use
CD45-PE (BD Pharmingen; Catalog 561087) at a concentra-
tion of 1:100 and BD IMag™ Anti-R-PE Magnetic Particles at
a concentration of 1:10.
5. For our workflow, we use G Biosciences RBC lysis buffer and
have demonstrated that our sample processing has minimal
effects on gene expression [16].
6. Passing the blood through the cell strainer cap prior to running
the sample on the BD Focus is important to prevent any
clogging in the fluidic path, acoustic chip or nozzle.
7. Monitor the flow rates recorded by the flow meter. Stable
laminar flow in the main flow channel of the acoustic chip is
needed to achieve high cell recovery. If sudden change of any
flow rate is observed the system needs to be stopped. Cleaning
of fluidics (priming with DI followed by 1 PBS) needs to be
performed before the analysis and sorting is resumed.
8. We suggest an event rate at or lower than ¼ of the pzt fre-
quency (frequency at which the chip vibrates) of the sorter to
make the sorting efficiency over 80%. Event rate can be opti-
mized by adjusting sample concentration and sample flow rate.
For example, with the 200 μm nozzle on the sorter and a pzt
frequency of 2017 kHz (2.17 MHz), 400–500 per second
event rate results in 90% efficiency. The sample flow rate on
BD Focus can be tuned by adjusting the pressure regulator.
9. Use a cell sorter per the manufacturer’s instructions to sort
cells directly into the Precise WTA Single Cell Encoding Plate,
preferably on a prechilled 96-well aluminum cooler block. The
 CTC Phenotypic and Molecular Characterization 225

sealed WTA Single Cell Encoding Plate with sorted cells can be
stored at 80  C for
6 weeks.
10. The Precise WTA Library Index Primers 1 or 2 are used to
multiplex multiple libraries for sequencing. The plate indexes
for the Precise WTA assay are equivalent to Illumina indexed
adapters D701 (ATTACTCG) and D702 (TCCGGAGA). If
running multiple plates, normalize all libraries to an equivalent
concentration and pool them to prepare one sequencing run
per Illumina guidelines.
11. External release of the BD Genomics Data View software is
available at http://bitbucket.org/CRSwDev/dataview. Instal-
lation of the version 9.3 (R2017b) of the MATLAB Runtime is
required for running the Data View v1.2.2.

References
1. Ting DT, Wittner BS, Ligorio M et al (2014) 7. Rhim AD, Mirek ET, Aiello NM et al (2012)
Single-cell RNA sequencing identifies extracel- EMT and dissemination precede pancreatic
lular matrix gene expression by pancreatic cir- tumor formation. Cell 148(1-2):349–361.
culating tumor cells. Cell Rep 8 https://doi.org/10.1016/j.cell.2011.11.025
(6):1905–1918. https://doi.org/10.1016/j. 8. Gorges TM, Kuske A, Röck K et al (2016)
celrep.2014.08.029 Accession of tumor heterogeneity by multiplex
2. Yee SS, Carpenter EL (2017) Enumeration, transcriptome profiling of single circulating
dielectrophoretic capture, and molecular anal- tumor cells. Clin Chem 62(11):1504–1515.
ysis of circulating tumor cells BT - circulating https://doi.org/10.1373/clinchem.2016.
tumor cells: methods and protocols. In: 260299
Magbanua M, Jesus M, Park JW (eds) Circulat- 9. Carpenter EL, Rader J, Ruden J et al (2014)
ing tumor cells. Springer, New York, NY, pp Dielectrophoretic capture and genetic analysis
193–202 of single neuroblastoma tumor cells. Front
3. Cohen SJ, Punt CJA, Iannotti N et al (2008) Oncol 4:201. https://doi.org/10.3389/fonc.
Relationship of circulating tumor cells to 2014.00201
tumor response, progression-free survival, and 10. Peeters DJE, De Laere B, Van Den Eynden GG
overall survival in patients with metastatic colo- et al (2013) Semiautomated isolation and
rectal cancer. J Clin Oncol 26(19):3213–3221. molecular characterisation of single or highly
https://doi.org/10.1200/JCO.2007.15. purified tumour cells from CellSearch enriched
8923 blood samples using dielectrophoretic cell sort-
4. De Bono JS, Scher HI, Montgomery RB et al ing. Br J Cancer 108(6):1358–1367. https://
(2008) Circulating tumor cells predict survival doi.org/10.1038/bjc.2013.92
benefit from treatment in metastatic castration- 11. Karabacak NM, Spuhler PS, Fachin F et al
resistant prostate cancer. Clin Cancer Res 14 (2014) Microfluidic, marker-free isolation of
(19):6302–6309. https://doi.org/10.1158/ circulating tumor cells from blood samples.
1078-0432.CCR-08-0872 Nat Protoc 9:694–710. https://doi.org/10.
5. Cristofanilli M, Budd GT, Ellis MJ et al (2004) 1038/nprot.2014.044
Circulating tumor cells, disease progression, 12. Issadore D, Chung J, Shao H et al (2012)
and survival in metastatic breast cancer. N Ultrasensitive clinical enumeration of rare cells
Engl J Med 351:781–791. https://doi.org/ ex vivo using a micro-hall detector. Sci Transl
10.1056/NEJMoa040766 Med 4(141):141ra92. https://doi.org/10.
6. Allard WJ, Matera J, Miller MC et al (2004) 1126/scitranslmed.3003747
Tumor cells circulate in the peripheral blood of 13. Harb W, Fan A, Tran T et al (2013) Mutational
all major carcinomas but not in healthy subjects analysis of circulating tumor cells using a novel
or patients with nonmalignant diseases. Clin microfluidic collection device and qPCR assay.
Cancer Res 10(20):6897–6904. https://doi. Transl Oncol 6(5):528–538. https://doi.org/
org/10.1158/1078-0432.CCR-04-0378 10.1593/tlo.13367
226 Moen Sen et al.

14. Rhim AD, Thege FI, Santana SM et al (2014) precise and absolute gene expression measure-
Detection of circulating pancreas epithelial cells ment by single-molecule counting. Anal Chem
in patients with pancreatic cystic lesions. Gas- 86(6):2867–2870. https://doi.org/10.1021/
troenterology 146(3):647–651. https://doi. ac500459p
org/10.1053/j.gastro.2013.12.007 19. Fu GK, Xu W, Wilhelmy J et al (2014) Molec-
15. Stott SL, Hsu C-H, Tsukrov DI et al (2010) ular indexing enables quantitative targeted
Isolation of circulating tumor cells using a RNA sequencing and reveals poor efficiencies
microvortex-generating herringbone-chip. in standard library preparations. Proc Natl
Proc Natl Acad Sci 107(43):18392–18397. Acad Sci 111(5):1891–1896. https://doi.
https://doi.org/10.1073/pnas.1012539107 org/10.1073/pnas.1323732111
16. Bhagwat N, Dulmage K, Pletcher CH et al 20. Yu L, Sa S, Wang L, Dulmage K, Bhagwat N,
(2018) An integrated flow cytometry-based Yee SS, Sen M, Pletcher CH, Moore JS,
platform for isolation and molecular character- Saksena S, Dixon EP, Carpenter EL (2018)
ization of circulating tumor single cells and An integrated enrichment system to facilitate
clusters. Sci Rep 8. https://doi.org/10. isolation and molecular characterization of sin-
1038/s41598-018-23217-5 gle cancer cells from whole blood. Cytometry
17. Lang JE, Scott JH, Wolf DM et al (2015) A 93(12):1226–1233. https://doi.org/10.
Expression profiling of circulating tumor cells 1002/cyto.a.23599
in metastatic breast cancer. Breast Cancer Res 21. Mahnke YD, Roederer M (2007) Optimizing a
Treat 149:121–131. https://doi.org/10. multi-colour immunophenotyping assay. Clin
1007/s10549-014-3215-0 Lab Med 27:469–46v. https://doi.org/10.
18. Fu GK, Wilhelmy J, Stern D et al (2014) Digi- 1016/j.cll.2007.05.002
tal encoding of cellular mRNAs enabling
 Chapter 14

Quality Control of Immunophenotyping

Bruce Greig

Abstract
Applications of immunophenotyping using flow cytometry offer precise and accurate means for providing
information used to both diagnose and monitor disease; they serve as a standard platform for many research
endeavors that study discrete populations of biological entities. The proper use of this highly sophisticated
technology requires daily and ongoing monitoring of both the instrument and the methodology. Best
practices for this begin with quality control (QC) procedures designed to set up and monitor the
instrument performance, the reagents, and the results to ensure that they are working properly both on
the day of use and over time. If the results of those QC procedures are outside of acceptable then recording
the corrective action taken must also be included in the quality control records. Quality assurance (QA) is a
way to know that the three phases of testing, namely, preanalytic, analytic, and postanalytic procedures, are
being followed. This chapter describes the procedures used to assess quality control as it pertains to flow
cytometry and immunophenotyping in all three phases of testing.

Key words Quality control, Quality assurance, Flow cytometry, Levey–Jennings, Immunophenotyp-
ing, Monoclonal antibodies, Immunofluorescence, Color compensation, FMO controls, Proficiency
testing, Method validation, Accuracy, Precision

1 Introduction

Quality control defined: a set of procedures performed each day for

the continuous and immediate monitoring of laboratory work to
confirm whether the results are reliable and can be used to diag-
nose, monitor, or ascertain the presence of disease in clinical appli-
cations or to produce results that are accurate, precise, and
reproducible in a research setting.
Why perform quality control in immunophenotyping? The
reasons are as follows:
1. It insures that all measurands meet a standard.
2. It monitors day-to-day consistency and reliability.

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_14, © Springer Science+Business Media, LLC, part of Springer Nature 2019

227
228 Bruce Greig

3. It provides confidence in results.

4. It insures reproducibility.
5. It addresses regulatory requirements.
Quality control (QC) and quality assurance (QA) as it pertains
to immunophenotyping is a broad topic that covers many different
components each of which is used to ensure that the results of a
flow cytometry assay are reliable. QC is what you do to ensure your
experiment or assay will do what it is supposed to do at each
step from setup to acquisition to analysis to interpretation to the
final answer. QA is the ongoing monitoring of all quality control
procedures that reflect the overall acceptability of the results. QC
procedures should demonstrate that everything being used to pro-
duce results is in proper working order prior to running patient
samples in the clinical lab or biological specimens used in a
research lab.
The results of the QC procedures on both the instrument and
the reagents used by the laboratory should be readily available for
inspection purposes [13, 14] as well as provide a means for long-
term monitoring to detect trends and exceptions. Flow cytometry
laboratories that perform clinical testing must meet and maintain
several regulatory requirements found on the College of American
Pathologists (CAP) Flow Cytometry Inspection Checklist. Figure 1:
Quality control requirements of the CAP flow cytometry inspection
checklist is a partial list of these requirements.
Flow cytometers used in immunophenotyping perform several
different measurements and assurance of how well it is working is
tantamount to being confident in the results it produces. Once the
samples have been acquired additional analytic steps are necessary

Fig. 1 Quality control requirements of the CAP flow cytometry inspection

checklist [1]
 Quality Control of Immunophenotyping 229

to ensure that the interpretation of the results is accurate. All these

considerations can be put into three categories:
1. Preanalytic quality control: quality control as it pertains to the
steps taken to prepare an assay including the specimen require-
ments and reagents used to set up the assay [4].
2. Analytic quality control: quality control as it pertains to the
instrument used to perform the immunophenotyping assay
that ensures it is operating properly.
3. Postanalytic quality control: quality control as it pertains to the
results generated by an immunophenotyping assay that ensures
that they are accurate and precise.
Each of these steps is described in this chapter along with
diagrams and references for further explanations.

2 Materials

1. QC beads, generally sourced from the instrument manufac-

turer or other commercial vendors.
(a) Alignment/performance check beads.
(b) Compensation beads.
2. PBS buffer with or without albumin
3. Monoclonal antibodies (various vendors).
4. 7-AAD viability dye
5. Ammonium chloride RBC lysing reagent
6. Quantitative commercial controls used in immunophenotyp-
ing procedures.
(a) Normal control
(b) Abnormal control
7. Qualitative controls for tests that use descriptive results
i.e. “positive”, “negative”, “present”, “absent” and/ or “not
detected”.
8. Miscellaneous laboratory equipment
(a) Refrigerators and freezers
(b) Centrifuges
(c) Water baths and incubators
(d) Biological hoods
230 Bruce Greig

3 Methods

3.1 Preanalytic Proper specimen collection and timely processing are critical pre-
Quality Control analytical quality control steps required for performing immuno-
Procedures phenotyping procedures. Specimens used in flow cytometry testing
can be obtained from several sources and in the clinical laboratory
3.1.1 Specimen these may include peripheral blood, bone marrow aspirates, solid
Requirements tissue biopsies, fine needle aspirates, and several types of body fluids
such as CSF, pleural fluid, synovial fluid, and vitreous fluid. In a
research setting this list may also include frozen tissue, animal
samples, plant samples, sea water, and a variety of other sources.
Each specimen has specific handling requirements that should be
fully described in the procedures of the laboratory.
The following list includes the most widely used specimen types
and their specific collection requirements for immunophenotyping.

3.1.2 Peripheral Blood All blood or bone marrow samples must be in an anticoagulant if
and Bone Marrow Aspirates the assay is for testing either white blood cells (WBCs) and/or red
blood cells (RBCs) [25]. The most commonly used anticoagulants
for immunophenotyping include EDTA, (sodium) heparin, and
ACD (acid citrate dextrose). All are designed to prevent clotting
while preserving the WBCs however, there are advantages and
disadvantages to each and is discussed below. Another important
consideration is the hold time between specimen collection and
testing. This can be critical depending on the anticoagulant used.
For the times listed in this section the laboratory should do their
own validation of each anticoagulant to determine what is
acceptable.
1. EDTA (ethylenediaminetetraacetic acid) preserves lymphocyte
viability up to 72 h if refrigerated, 48 h at room temperature.
This is the most common anticoagulant used in the clinical
laboratory for testing lymphocyte subsets since it can also be
used for performing white blood counts (WBCs) and differen-
tials necessary for calculating the absolute lymphocyte counts.
Granulocytes, however, may begin to deteriorate in less than
8 h due to the chelation of the calcium in the plasma by the
EDTA which makes this anticoagulant less desirable for flow
assays that focus on myeloid lineage populations. These tubes
usually have a purple stopper [2].
2. Sodium heparin prevents the formation of fibrin from fibrino-
gen and thrombin activation thus preventing clotting of the
sample. It may cause some RBCs to clump and, in some cases,
impart a blue background on Wright-Giemsa stains so it is not
suitable for hematology assays [3]; however, it works quite well
on bone marrow aspirates and is the anticoagulant of choice for
these samples. Samples in heparin are generally good for up to
 Quality Control of Immunophenotyping 231

48 h if refrigerated. These tubes usually have a light green

stopper.
3. ACD (acid citrate dextrose) is the same anticoagulant used to
collect whole blood donations for transfusion. It has many
qualities that make it highly suitable for longer and more
stabilized specimens and is more conducive to maintaining
cell integrity and viability than either EDTA or heparin. How-
ever, it is a liquid additive and thus will make absolute counts
inaccurate since it changes the volume of whatever is added to
it, so it should not be used for either flow cytometry single-
platform absolute counts or Hematology quantitative results.
Samples in ACD are good for up to 48 h at room temperature
and up to 4 days if refrigerated making it much preferred in
laboratories that are not open all day every week. These tubes
usually have a pale-yellow stopper.

3.1.3 Solid Tissue Tissue samples submitted for flow cytometry testing may be from
Biopsies and Fine Needle many different sources. In a Hematopathology setting this would
Aspirates likely include a lymph node biopsy, soft tissue from an organ such as
the gut, and even skin punches submitted by a dermatologist. If the
sample is submitted from the same place as the lab doing the testing
in most cases it will be reasonably fresh and intact. Care should be
taken to insure good viability by using a transport media with the
tissue such as RPMI with or without albumin that has nutrients and
buffers that greatly improve the chances of maintaining viability.
Using other media like plain PBS is discouraged because the sam-
ples will deteriorate much more quickly compared to a transport
media like RPMI. The use of any fixative in the media is not
acceptable since this will be toxic to the tissue and prohibit any
possibility of doing a flow cytometry assay which requires viable
cells. Some laboratories use frozen tissue banks that allow the lab to
save rare or study samples in a way that they are not damaged by the
subfreezing temperature. Previously frozen samples will need to be
thawed and washed before processed to a single cell suspension.
Follow published procedures that describe the steps necessary to
gently bring the sample back to room temperature prior to making
single cell suspensions. Making a single cell suspension of the solid
tissue can be achieved several different ways including:
1. Grinding the sample with a mortar and pestle and rinsing
with RPMI.
2. Using a tuberculin-sized needle that is filled with RPMI and
repeatedly injecting it into the tissue in a petri dish to extract
the cells of interest and leave behind the sample connective
tissue. The cell suspension can then be washed with RPMI and
concentrated prior to staining.
232 Bruce Greig

3. Adding an enzyme solution such as collagenase that digests the

connective tissue leaving behind the cells of interest. Though
effective in practice there may be higher cell damage using this
procedure [6].
4. Using mechanical devices that disaggregate the connective
tissue such as the Becton Dickinson Medimachine™. Small
pieces of the tissue are put into a special cup called a Medicon™
which is then inserted into the Medimachine. RPMI is added to
the Medicon, the cup capped, and the machine turned on. A
spindle turns a raised metal lip that gradually forces the tissue
through a bladed mesh as the spindle turns. This process then
strips away the connective tissue leaving a single cell suspension
in the RPMI at the bottom of the Medicon. The RPMI which
now includes the separated cells is then withdrawn from the
Medicon cup using a syringe, filtered, and then processed for
staining [7]. Some laboratories perform cell counts after these
steps to insure adequate staining based on concentrations that
are appropriate to use with previously made antibody cock-
tails (a cocktail is two or more antibodies in the same tube).

3.1.4 Fine Needle FNAs are frequently submitted for flow cytometry testing in the
Aspirates (FNAs) context of hematopathology procedures. Tissue samples collected
with this technique are a much less invasive procedure compared to
obtaining a tissue biopsy. The samples are collected by using a very
thin needle (22 gauge or smaller) that is inserted into an area of
questionable tissue like a lymph node or other soft tissue. The cells
are then aspirated into the syringe and added to a tube containing
transport media such as RPMI. Since the cells are already in sus-
pension very little manipulation of the sample is necessary prior to
using it in an immunophenotyping procedure. It can then be
washed, concentrated, and resuspended in RPMI prior to staining.
These samples may have some peripheral blood contamination that
happens during collection and thus will need to include a RBC
lysing step as part of the staining procedure.

3.1.5 Body Fluids Body fluids submitted for flow cytometry immunophenotyping can
and Cerebrospinal Fluid be collected from several different sources. These may include
(CSF) peritoneal, bronchial, pleural, synovial, pericardial, vitreous, and
cerebrospinal fluids. In most situations the question for testing is
to find and identify an underlying malignancy such as lymphoma or
leukemia. Usually the body fluid itself if collected the same day will
maintain the cell’s viability if transported quickly and refrigerated.
CSF however, is collected using a laparoscopic procedure and even
though these specimens are usually sent to the lab with no addi-
tional media in the tube it is highly recommended that the sample
be resuspended in fresh RPMI as soon as possible to prevent the
rapid deterioration of any cells present. Studies have shown that
CSF with RPMI added may have a much higher rate of recovery
compared to samples that do not include a transport media [10].
 Quality Control of Immunophenotyping 233

3.2 Preanalytic All reagents used in an immunophenotyping experiment must be

Quality Control tested for efficacy prior to being used in a procedure [4]. For
Procedures example, PBS (phosphate buffered saline) and RBC lysing agents
such as ammonium chloride (NH4Cl), can be tested quite easily
3.2.1 Reagent Quality using a sample type that is commonly submitted for testing such as
Control peripheral blood. Set up a sample using the new reagents, then
acquire on the instrument and review the results for acceptability. Is
Buffers and Lysing the light scatter okay? Are the RBCs lysed as expected? Record the
Solutions results of testing in a Reagent QC log to track performance.
Include the lot number if applicable, the testing specimen used
and whether the results were as expected, usually by either “pass”
or “fail” depending on criteria included in the procedure. As with
all QC records, the supervisor or designee should review the results
monthly.

3.2.2 Antibody Quality Primary reagents like monoclonal antibodies require a more thor-
Control ough control process than other test reagents [9]. Quality control
of antibodies should include the following procedures:
1. Antibody specificity: does the antibody react with the intended
antigen [12]? This aspect of antibody QC requires specific
specimens to prove efficacy of the reagent. For example, if
CD33 is being tested, the control material should include
myeloid lineage cells such as granulocytes and monocytes,
since that antigen is fully expressed on those cells. Lympho-
cytes present in those samples can serve as a negative control.
Other antibody control materials include the following:
(a) Commercial quality controls: These are samples that
include a list of expected results for several antibodies in
the package insert. These controls are generally used for
daily testing of an antibody cocktail panel (a cocktail is two
or more antibodies in the same tube), which is required
for clinical applications like peripheral blood lymphocyte
subsets or stem cell quantitation assays. These commercial
controls will generally have known qualitative and quanti-
tative values for most of the antibodies used for immuno-
phenotyping in a clinical laboratory. See Fig. 2:
Commercial control package insert taken from a commer-
cial whole blood control listing.
(b) In-house controls: Other control samples that can be used
to check antibody specificity may be sourced from previ-
ously tested material that has been confirmed to be posi-
tive or negative for the antibody depending on the
population of interest. If alternate testing is performed
in the laboratory, such as immunohistochemistry (IHC),
the flow cytometry antibody results can be compared to
the results of the alternate assay such as a tissue section
stained for CD20. However, this choice may only offer
234 Bruce Greig

Phenotype3,a WBC Type4 %Positive Cell1,b Expected Ranges6 Absolute Number7,c Expected Range4
+
CD2 Lymphocytes 85.0 73.0 - 97.0 – –
CD3+a Lymphocytes 78.4 68.4 - 88.4 1625 1275 - 1975
CD3+/CD4+a Lymphocytes 51.1 45.1 - 57.1 1059 879 1239
CD5+ Lymphocytes 74.6 64.6 - 84.6 - -
CD3+/CD8+a Lymphocytes 25.3 19.3 - 31.3 524 354 - 694
CD3-/CD16+ Lymphocytes 9.2 4.2 - 14.2 – –
CD3-/CD16+.CD56+ Lymphocytes 10.5 5.5 - 15.5 218 73 - 363
CD19+ Lymphocytes 11.2 6.2 - 16.2 232 82 - 382
CD20+ Lymphocytes 11.1 6.1 - 16.1 – –
CD3-/CD56+ Lymphocytes 8.6 3.6 - 13.6 – –
HLA . DR+ Lymphocytes 15.8 9.8 - 21.8 – –
CD14+ Monocytes 85.0 70.0 - 100.0 – –
CD45+ Leukocyte Gate ................. CD45 is to be used for gating purposes only.8 ............

CD34+ Progenitor Cels 0.04 0.01 - 0.07 2.8 0.6 - 5.0

Lymphocytec 31.4 21.4 - 41.4 2074 1674 - 2474
Monocyte 8.4 3.4 - 13.4 – –
Granulocyte 57.1 47.1 - 67.1 – –
Absolute9WBCc(x103/µL) - - 6.6 6.1 - 7.1

Fig. 2 Example of a package insert from a whole blood commercial control listing reactivity and expected
quantitative values for antibodies used in Immunophenotyping [28]

limited comparisons, since in many cases the number of

antibodies used in Flow Cytometry applications may be
greater than those found on the Histology test menu. The
College of American Pathologists (CAP) also offers cell
controls that can be used for a limited number of other
rare antibodies such as CD103 and CD1a. Some labora-
tories with subzero freezers may cryopreserve samples that
have previously been tested and later used for quality
controlling new shipments of antibodies.
(c) Proficiency testing samples: If the laboratory participates
in an external proficiency testing program such as the CAP
Survey, there may be leftover samples that are still ade-
quate for quality control testing purposes that can be
stained with a variety of antibodies. Returned results of
these surveys can then be compared to the antibodies
being tested for quality in the lab. This may not be practi-
cal if there is a considerable time delay before the pheno-
type of the testing material is known, unless the lab has
some way of preserving the cells.

3.2.3 Antibody Most antibody manufacturers provide directions that specify the
Concentration amount of antibody to use for samples typically submitted for
and the Stain Index Immunophenotyping. Knowing the concentration of the sample
is also necessary to insure saturation staining [11, 25]. For example,
if a sample has a high concentration of cells, then the recommended
amount of antibody per test may not be enough to insure adequate
staining. In other cases, the amount of antibody listed may be
 Quality Control of Immunophenotyping 235

Fig. 3 The stain index (SI) calculation [22, 33]. The SI is the ratio of the difference
between the positive mean intensity (MFI) and negative mean intensity (MFI)
populations divided by two times the standard deviation (SD) of the negative
population. The SI is used to compare fluorescence intensity of one antibody
versus another or when comparing the concentration of one antibody versus a
different dilution of the same antibody

several times more than is necessary and the excess antibody may
negatively impact the staining by increasing the background signal.
The best antibody concentration to use can be determined by
calculating the stain index (SI). The SI is the ratio of the difference
between the positive mean channel result and the negative mean
channel result divided by two times the standard deviation (SD) of
the negative population. See Fig. 3: Stain index calculation.
The stain index can be used to compare different antibodies
and it can also be used to compare different concentrations of the
same antibody. In some cases, the SI will improve using a lower
concentration than what is recommended since the background
signal width may be smaller than it is using a higher concentration.
The higher the SI, the better the resolution between the positive
and negative populations. Different SIs can often be seen with
different concentrations. The difference can be due to one of two
reasons: a lower intensity of the positive peak or a wider SD of the
negative signal due to excess unbound antibody. Titrating the
antibody is advantageous for determining the optimal
236 Bruce Greig

Fig. 4 Antibody concentration versus CD19 PEcy7 expression. Over the course of
different titrations or dilutions the stain index (SI) improves using lower
concentrations of the antibody. There is no noticeable loss in CD19
intensity (Y-axis); however, the background signal is much greater at higher
concentrations of the antibody and thus has a lower stain index. Ideally, the
standard deviation of the negative signal should be somewhere near that of
unstained autofluorescence

concentration to use in the experiment. Different concentrations

may have little effect on the overall intensity of the positive peak;
however, in some cases the background or negative signal width
decreases because of less unbound antibody and thus a better SI. An
example of how the optimal antibody concentration can be calcu-
lated is seen in Fig. 4: Antibody concentration versus CD19 PEcy7
expression.
The titration procedure compares serial dilutions of the anti-
body after calculating the stain index of each dilution. The results
are then graphed using the SI versus the concentration of the
antibody. Determining the best titer of the antibody should be
included in the quality control procedure for any new lot number
or new shipment prior to using it in an assay. This is a requirement
in many clinical applications especially for lab developed
tests (LDT). It is okay to use a different amount than what is
recommended by the manufacturer if you can document that it
still performs as intended.

3.2.4 Antibody Titration 1. Make serial dilutions of the antibody beginning with the
Procedure recommended volume followed by dilutions of 1:2, 1:4, 1:8,
etc. The diluent used can be simple PBS buffer with or without
calf serum which is reliable for antibodies that have a high
expression such as CD45 and CD8. In some instances, espe-
cially very weak expressing antibodies should be diluted with
unlabeled or unconjugated antibody to insure adequate anti-
body saturation staining.
 Quality Control of Immunophenotyping 237

Fig. 5 Antibody concentration versus stain index. As the antibody concentration

increases the stain index (SI) increases until it levels off. The optimal concentra-
tion should be somewhere close to the value of the stain index that begins to
cease changing (in the example shown, approximately 0.5 μl/test)

2. Include a tube with no antibody (unstained) for comparative

purposes of the negative peak of the population vs an
unstained/ background signal.
3. Use the sample type that is typically tested such as peripheral
blood, bone marrow aspirate, or tissues. It is important to have
both positive and negative populations in each tube, so cell
lines might not be the best choice if no negative populations are
typically found in those samples.
4. Process all samples following the standard staining procedures
used in your laboratory.
5. Run the samples on the flow cytometer with the same instru-
ment settings used for actual testing purposes.
6. After acquiring the samples, graph the Stain Index versus the
antibody concentration to determine the best overall concen-
tration to use. See Fig. 5: Antibody concentration versus stain
index as an example.

3.2.5 Multiantibody An antibody cocktail is whenever two or more antibodies are com-
Cocktails bined to make a multiuse reagent. As part of panel design the
antibody cocktail is constructed using combinations that are used
for immunophenotyping applications. There are several advantages
to using cocktails. They can save time when setting up multiple
experiments and are likely to lower or prevent pipetting errors that
could happen if the single cocktail was made up each time the
experiment was performed. Another advantage to using cocktails
is that you can standardize delivery volumes. This makes test setups
238 Bruce Greig

easier since many single reagents use different volumes and in a

multicolor cocktail this can be very labor intensive and prone to
error. The cocktail can be made in advance and for enough tests to
last a few days or weeks.
1. Initial QC testing: Any cocktail must be proven to be accurate
before using it for testing or reporting purposes. Initial quality
testing of the cocktail should include setting it up on a fresh
sample along with a second tube of the in-use cocktail. Results
should agree for both reactivity and quantitation if applicable.
Shelf life of the cocktail must be validated as part of implemen-
tation of the cocktail. There are recommended practice guide-
lines [4, 8] for use however each lab should determine its own
pass or fail requirements.
2. Ongoing testing: If the cocktail is kept over several days, it
should be tested each week for acceptability using positive and
negative controls. Previously tested samples are good sources
of QC material. Loss of any antibody or a definitive change in
intensity means that the cocktail should be discarded. Criteria
for acceptability should be included in the QC procedure.
Document results of the quality control and keep records of
the testing.

3.2.6 Antibody Cocktail The full cocktail procedure can be found in Notes section. The
Procedure (Example) example shown in the recipe is for a 10 test, 4-color cocktail for
T-lymphocyte subset testing using titrated antibodies and buffer
that is aliquoted using 100 μl cocktail/test. Use a previously deter-
mined titer amount for each antibody in the cocktail and add
enough buffer to create a standardized aliquot for testing such as
100 μl/test. The additional volume of the buffer will also allow
better mixing.

3.2.7 Conclusion, Every experiment used in Immunophenotyping uses specific sam-

Preanalytic Quality Control ples from a wide range of sources. Each laboratory should develop
their own specific sample requirements that include the instructions
for collection, the media used to transport the sample, and the
storage and conditions required for testing. In addition, the use
of any reagents to perform the test must also be checked for
suitability using controls that confirm each of the preparation
steps of the assay. Records of all quality controls must be kept in a
concise way that ensures easy retrieval for inspection purposes as
well as to monitor performance over time. Upon successful com-
pletion of these preanalytical steps the Immunophenotyping testing
can now proceed to the analytic phase.
 Quality Control of Immunophenotyping 239

3.3 Analytic Quality Introduction: Quality control of the immunophenotyping assay

Control Procedures continues with consideration of the testing that includes the flow
cytometer itself. These very sensitive and highly adaptable plat-
3.3.1 Instrument Quality
forms are rather simple in terms of how they work. Basically, a
Control
fluidics system consisting of either buffered saline or deionized
water provides a means for the sample to be introduced into a
flow cell chamber. The flow cell has a very specific shape that
through the principle of hydrodynamic focusing takes the cells or
other objects in a random mixture and puts them into a single file
by forcing them through a small opening as they exit the flow cell.
Upon exiting the flow chamber the cell stream is bisected by one or
more light sources, typically lasers with monochromatic emissions.
The intersection of the cells with the light produces several specific
light signals that are then directed by fiber optic cables or mirrors
and filters into electronic detectors commonly known as photo-
multiplier tubes (PMTs). The PMTs then convert those signals into
a digital format for data analysis using software programs.
Despite the simplicity of its operation the quality of the results
can be greatly affected by the performance of any or all these
components of acquisition: fluidics, optics, and electronics. Daily
instrument maintenance and quality control is paramount to ensur-
ing reliable results. This section of Immunophenotyping Quality
Control will consider and describe the steps to perform on the
instrument as preparation for running test samples. See the Notes
section for a step-by-step guide for setting up a flow cytometer for
immunophenotyping assays. Details of the procedures and expla-
nation of each step in the analytic QC of the instrument is found in
the following section.

3.3.2 Instrument Quality 1. Fluidics system startup: The flow cytometer fluidics system is
Control: Flow Cytometer truly the lifeblood of the instrument. Proper care and cleaning
Startup and Maintenance of the fluidics system is synonymous with good quality control
and helps ensure high quality results. Prior to turning on the
instrument check the levels of the fluidics containers. They are
usually found either on a fluidics cart or other storage system
near the instrument. This includes the sheath fluid, cleaning
solutions, or DI water, etc. that may be a part of the whole
fluidics system. If the instrument has a waste container dispose
of anything from the last time it was used. Be sure to add bleach
to the container to minimize the danger of any hazardous waste
that may result from using biological samples. Most flow cyt-
ometers have built-in programs that provide a means for the
sheath fluid and other solutions to be primed and distributed
throughout the system including the flow cell and ultimately
into the waste container. As part of the instrument startup a
series of priming and flushing cycles takes place to be sure that
everything is working and that all the fluidics lines are free of air
bubbles. Sometimes when the instrument is shut down for
240 Bruce Greig

several hours bubbles can come out of the sheath fluid solution
still in the instrument much like what is observed on the side of
a glass of water left standing overnight. If those bubbles are not
adequately removed as part of the fluidics startup they may
cling to the inside of the flow cell and create focusing issues
much like any other obstruction in the fluidics system.
2. Fluidics clean cycle: Samples routinely run on the instrument
such as tissue cell suspensions and bone marrow specimens may
have high amounts of debris or fibrin or other nonintact cell
residue that can build up on the inside of the sample insertion
rod, fluidics tubing, or flow cell. After starting up the instru-
ment the fluidics system should be put through a short cleaning
cycle that includes running dilute bleach (10%) for 5 min fol-
lowed by a DI water sample for 5 min. Running a series of
cleaning samples helps to remove the buildup and prevent clogs
or other obstructions that might be present in the insertion rod
or fluidics tubing. Some labs prefer to do this cleaning
sequence one or more times a day. A good rule of thumb is
that for every 1 min you run dilute bleach it should be followed
for the same amount of time by running a tube of DI water.
This is to adequately remove any residual bleach from the
fluidics system. A full system clean of 60 min (30/30) or
more should be performed at least once a month to more
thoroughly clean the fluidics lines and flow cell chamber.
3. Fluidics function checks as part of bead QC.
As mentioned before, issues with fluidics are probably the
most common cause of instrument-related problems. Most
every instrument has a quality control program that uses latex
beads to measure and standardize each parameter of the instru-
ment operation. Usually performance issues are first indicated
by out of range bead QC results including diminished resolu-
tion of either or both light scatter parameters, forward scatter
(FSC) and side-scatter (SSC). A well-maintained instrument
that has been set up and optimized should demonstrate that
the FSC/SSC dot plot of a lysed blood sample has at least three
distinct WBC populations corresponding to the lymphocytes,
monocytes, and granulocytes. On the other hand, if the instru-
ment has population resolution issues seen by light scatter due
to fluidics problems the populations tend to drift downward
toward the threshold on the FSC/SSC dot plot (see Fig. 6:
Good light scatter resolution versus poor light scatter
resolution).
If the flow cytometer has debris buildup over time in the
fluidics lines or the sample insertion rod or if the stream is
obstructed in some way the Forward Scatter Channel (FSC)
signal is usually the first to show this problem. There will either
be a gradual loss of the signal resulting in a lower location on
forward scatter of a population or an increase in the detector
 Quality Control of Immunophenotyping 241

[Ungated] FS INT / SS INT [Ungated]

1000 Debris : 8.03% 1000 Debris : 0.00%

800 800
SS INT

SS INT
600 600

400 400

200 200

0 0
0 200 400 600 800 1000 0 500 1000
FS INT FS INT

Fig. 6 Good light scatter resolution versus poor light scatter resolution: Left histogram: RBC-lysed peripheral
blood sample with optimized and finely tuned FSC/SSC signals shows granulocyte (blue) populations. Right
Histogram: FSC/SSC dot plot with poor optical alignment, possibly due to fluidics issues. Notice the loss of FSC
output versus SSC resolution

gain if you are using settings that are autoadjusted by the bead
QC software to meet a target range for each peak channel.
Eventually these changes to the detector settings will exceed
the defined limits of the bead QC program and the results will
have one or more parameter failures.
4. Laser performance: The solid-state lasers found on most flow
cytometers are considerably more dependable and less prone to
failure than the high-power water-cooled lasers that were used
on many of the research-type instruments years ago. However,
they still need time to warm up and equilibrate. Usually the
time required to run the fluidics start-up cycle and cleaning
steps is typically long enough for the lasers to warm up. In most
cases the instrument will have an indication of when it is ready
to be used following the startup time for the laser.
5. Flow cell, optics, and electronics QC using commercial beads.
(a) Introduction to the Bead QC procedure: Following
instrument and fluidics startup, initial cleaning, and laser
warm-up the flow cytometer must then be checked for
proper operation. This is typically accomplished using
latex beads or other materials, usually sourced from an
instrument vendor, to ascertain the overall function of
the fluidics, electronics, and optical alignment. These
quality controls checks are where the “quality” of Quality
Assurance begins in flow cytometry. The CAP checklist
requires that performance of function checks on all instru-
ments follow a defined schedule [1]. For results from a
flow cytometer to be reliable and accurate, the setup and
monitoring of the flow cytometer performance should
always be checked when the instrument is started up
each day. Some labs may require this as often as once
each shift if the instrument is used continuously.
242 Bruce Greig

Fig. 7 QC bead signal. QC Beads run each day are placed in a specific channel of
each detector by the software automatically adjusting the detector gain or
voltage to match the baseline target value

To help ensure that the flow cytometer is fully func-

tional and each event signal is meeting a set standard, most
instrument manufacturers use bead QC programs that
provide information on all the major components of
the instrument. After establishing a baseline, the perfor-
mance is then monitored on an ongoing basis to be sure
that it is still operating properly. These latex beads are
polychromatic and thus can be detected across all the
PMTs and light scatter detectors. The beads have different
sizes and signal intensities so that linearity and precision
can be calculated along with PMT performance. If for
instance the PMT voltage required to place the bead signal
in the previously established baseline channel is outside
the software limits, then that particular parameter will be
flagged as failing (see Fig. 7: QC bead signal).
6. Focal point/alignment check: The alignment of the laser to the
flow cell stream is critical and has a direct impact on the quality
of the results. Most flow cytometers have one or more laser
light sources that are directed by a series of lenses and prisms to
a fine focal point at the exit point of the flow cell chamber. If
the laser or the flow cell is not properly aligned or focused, then
the resolution of one or more signals may be suboptimal. An
instrument with good alignment means that the laser light
intersecting the sample stream is precise and finely tuned. A
bright, sharp signal provides better resolution and sensitivity
 Quality Control of Immunophenotyping 243

Fig. 8 Optical alignment and fine focusing. Good optical alignment equates to a
fine focus which can provide a sharper, clearer signal that gives better resolution
and higher sensitivity (right image) compared to an instrument with poor optical
alignment (left image) that is blurred or out of focus

compared to a signal that is not quite as precise (see Fig. 8:

Optical alignment and fine focusing).
Running QC beads each day does not actually change the
alignment of the light with the flow cell but provides a way of
knowing just how the performance compares to the baseline
standard established earlier. Performing an actual alignment of
the instrument laser or flow cell is not something done rou-
tinely, especially in a clinical setting. That procedure is typically
performed by a service engineer as part of an initial setup of the
instrument or as part of regularly scheduled preventative main-
tenance or whenever a major component like a laser or a PMT
is replaced. Some research flow cytometers may allow the user
to change or adjust the optical alignment, but this is the excep-
tion, not the rule since in most instruments the flow cell itself
may be semipermanently attached to the optical bench.
7. PMT performance: Besides checking the alignment of the laser
and the flow cell the results of running QC beads are indicators
of PMT performance. After creating a baseline setup for each
signal, the daily performance of the QC beads is measured
against the results of the baseline measurements. Each of
these performance checks provides information about the
detector efficiency, sensitivity, resolution, precision, linearity,
background (noise) level, and laser delay. By definition:
(a) Detector efficiency is the maximum signal output with
minimum PMT voltage. A failing PMT will have a loss
244 Bruce Greig

of efficiency over time and subsequently require more

voltage to put the emitted signal within the program’s
baseline range of the scale. An increase in PMT voltage
may then result in a higher background signal and
subsequent loss of sensitivity.
(b) Sensitivity is the ability to distinguish true positive signals
from true negative signals. One way to determine PMT
sensitivity is to run polychromatic beads and calculate the
stain index (SI). The SI equation uses the fluorescence
intensity of the positive peak versus the negative peak and
can be calculated using this equation:
SI ¼ MFI of positive bead  MFI of the unlabeled bead=2
 SD of unlabeled bead
(SI ¼ stain index, MFI ¼ mean fluorescence intensity, SD ¼ stan-
dard deviation)
The standard deviation (SD) of the negative peak is a
reflection of the width of the background signal. Some
fluorochromes have a broad negative signal, while others
may have one that is much narrower and thus have a lower
SD and in many cases a higher SI.
The SI calculation on some instruments is performed
by the bead QC software, but it can also be put into and
calculated on a spreadsheet if you know the result values
to use in the equation. Knowing the relative stain index of
each detector and monitoring it on at least a weekly basis
is highly recommended. Each laboratory should establish
an acceptable range, usually 2 SD of the mean SI for
each PMT. If over time, the bead sensitivity results show a
downward trend it may be due to instrument issues such
as faulty fluidics, dirty optics, a failing laser or even elec-
tronic noise. If this trend continues for a certain amount
of time the service engineer should be notified if minor
troubleshooting does not resolve the issue.
(c) Signal resolution is the ability to detect discrete signals in a
mixture where differences are very small. Good resolution
can also lead to higher sensitivity. One way of determining
how well the instrument can resolve multiple populations
with small differences is to run beads that have varying
intensities. There should be good separation between
each of the populations throughout the range of the
detector including very weak signals versus negative sig-
nals and very bright signals being on scale and not crushed
against the right axis. If the signals are not clearly differ-
entiated it may indicate an alignment/focal point problem
often seen when there are fluidics issues especially if it
happens in more than one channel. Cleaning the flow
 Quality Control of Immunophenotyping 245

150

Count
100

50

Fig. 9 Multiple fluorescent intensities demonstrating high resolution. A finly

tuned flow cytometer with proper alignment and PMT settings can resolve
multiple populations which demonstrate discrete differences in intensity

cell using a manufacturer recommended cleaning solution

can oftentimes improve the quality of the results (see
Fig. 9: Multiple fluorescent intensities demonstrating
high resolution).
(d) Precision is the ability to detect the same signal repeatedly
within a defined range of the baseline. The measurement
of precision is the coefficient of variation (CV) of multiple
results. High CVs mean poor precision which negatively
impacts the reproducibility of your results.
(e) Linearity is the property of bead signals to be detected in a
straight line over time. If the detector has good linearity, a
signal that has twice the output of the first signal should
then be 2 brighter than the first signal and so on for
subsequent results with increasing positive peaks
[21]. This is very important for experiments that measure
DNA ploidy where the G2 + M population should be
2 the G1/G0 population [5, 23, 24]. If it is not, then the
results might be misinterpreted as being aneuploid. Good
linearity means that the reliability of multiple signals is
expectedly proportional.
(f) Background level is sometime referred to as the noise
level. It is the extraneous signal detected when no other
signal is expected. It is present on all PMT outputs, but it
may be higher in some compared to others. Knowing the
background signal is important because it may impact the
sensitivity if it is higher than the otherwise negative signal
produced by the cells being measured. High background
levels that increase over time are often seen in PMTs or
other instrument electronics that are beginning to fail or
need recalibration. If routine troubleshooting does not
improve the background signal it may require attention by
a service engineer.
246 Bruce Greig

(g) Laser delay is the time difference between the bead/cell

being excited by the first laser followed by signal output
from the second or subsequent lasers. It is used to unite
the signals produced by multiple lasers exciting different
fluorochromes on the same cell. If the delay is too long it
can result in missing or abbreviating detection of the
second and following signals. Generally, laser delay issues
can be traced back to fluidics problems. When some-
thing slows or impedes the fluidics stream such as a par-
tially obstructed flow cell, it may sometimes cause laser
delay errors. Performing regular instrument maintenance
like cleaning will usually prevent this problem.
8. Levey–Jennings charts.
Besides the bead QC indicating the overall operation of the
instrument on a specific day it also provides a way to monitor
the performance of several parameters over the course of days
or weeks to demonstrate reliability and consistency. Most flow
cytometers are rather stable from day to day however, the
alignment or precise focal point can drift or deteriorate over
time due to any number of reasons including a fluidics issue,
faulty electronics, laser failure, or some other unknown cause.
There may be a gradual decrease in performance in any of the
electronics or lasers that may predict a future failure. The bead
QC program used should include graphical displays of all
measured results plotted over time. These are known as
Levey–Jennings (L-J) charts and they allow the user to observe
or detect changes, trends, or outliers that can in some cases
predict future problems that can be addressed before they
totally fail. See Fig. 10: Levey Jennings Charts.

Fig. 10 Levey–Jennings graph of daily PMT QC settings over time. The voltage of the FL8 detector required to
meet the baseline specifications is gradually rising over time which may indicate an impending failure of that
PMT
 Quality Control of Immunophenotyping 247

If only one detector shows a trend, then the problem may

be a faulty PMT. If more than one detector from the same laser
shows this problem, it may indicate that the laser performance
is the potential problem. In some cases, the issue may a dam-
aged fiber-optic cable that transmits the signals from the flow
cell to the PMTs. Generally, a problem like this must be
addressed by a service engineer. If more than one parameter is
out over multiple lasers, then the problem is more likely to be
the fluidics or an impending failure of the laser or even the
instrument electronics or circuit boards or interface of the
instrument and workstation. Comments on L-J charts about
outliers or trends should be included in the results as well as any
action taken to troubleshoot the issue observed.
9. QC Bead results summary.
Performing routine maintenance and running QC beads to
check fluidics, alignment, and sensitivity is the first step in
preparing the instrument to run experiment samples. Each
lab should determine how and when to respond to a QC
failure. In some cases, performing a system clean of the fluidics
and flow cell will often fix the problem and in other cases more
detailed troubleshooting is required. If the problem continues
then a service call to the instrument vendor may be necessary to
find and fix the issue. Documentation of the QC results and
monitoring performance over time helps ensure that testing
will be successful. After successfully completing the QC beads
procedure the instrument is almost ready for test samples;
however, an important next step is to optimize the instrument
settings using actual samples. This step is described in the next
section.
10. Optimizing the instrument settings.
Optimizing the instrument settings means getting the
maximum performance or the best signal–noise ratio for each
detector [22]. The signal–noise ratio is the difference between
the median positive signal and the median negative signal.
Think of optimizing as fine-tuning the instrument settings as
compared to a course adjustment that happens when running
only the QC beads. Do not confuse optimizing the settings
with setting up compensation. Color compensation setup fol-
lows optimizing the instrument. In the clinical lab think of this
as moving the instrument from being “bead-ready” to one that
is “sample-ready.” Why? Because in some cases, bead signal
intensity will be different from cell signal intensity and using
stained cell samples to optimize the bead settings may improve
the performance of the instrument. This is especially important
for experiments designed to identify discrete populations
which are sometime present with both normal and abnormal
cells. If the default setting is too low there may be a loss of
248 Bruce Greig

Fig. 11 Optimizing PMT settings. In Histogram a, WBCs gated on lymphocytes and monocytes stained with
CD4, the PMT setting is too low and as a result the intermediate population (monocytes) is difficult to resolve.
In Histogram b, the default PMT setting is too high so CD3+ T cells gated on lymphocytes appears to be
off-scale when in fact are still on scale; however, the brightest events are piled up against the right axis
and may be visually underestimated. In Fig. c, a lysed whole blood sample, the PMT setting is optimal. CD33
expression shows clear resolution between the bright (monocytes), intermediate (granulocytes), and negative
(lymphocytes) WBC populations

sensitivity in picking up the weak signal or there may be poor

separation between the positive signal and the negative signal
which might be hiding an intermediate signal. On the other
hand, if the bead PMT setting is too high the positive signal
might be only partially observed if it is pushed up against the
far axis. This makes the upper end of the signal look like it is off
scale when instead it is simply piled up in the last channel and
may be erroneously underestimated. See Fig. 11: Optimizing
PMT Settings, for examples of PMT settings that are too low,
too high, and then optimal using a stained cell sample with
bright, intermediate, and negative populations present.
How often is it necessary to optimize? Most flow cyt-
ometers are stable and give reproducible results day after day
so the frequency of performing optimization really depends on
a few specific situations:
(a) Whenever an instrument is initially installed.
(b) After any major maintenance such as replacing a laser,
realigning the flow cell, changing out a PMT, or changing
the configuration of the PMTs.
(c) Change the QC Bead Baseline configuration or after
setting up a baseline using a new lot number of QC beads.
A good rule of thumb is to check the settings using a
fully stained sample and observing each PMT for optimal
performance at least once a month.
11. Optimization methods.
Forward Scatter (FSC)/ Side Scatter Optimization: Start
by running a cell sample such as lysed whole blood (LWB)
 Quality Control of Immunophenotyping 249

[Ungated] FS INT / SS INT

1000 Debris : 8.03%

800

600

SS INT
400

200

0
0 200 400 600 800 1000
FS INT

Fig. 12 FSC versus SSC with good separation and resolution. Forward scatter
versus side scatter of a lysed whole blood sample demonstrating good
separation between all 3 WBC populations with the noise threshold set just
below the lymphocytes to allow visualization of intact cells versus debris

using the default settings from the bead QC. The first differ-
ence you will notice is that the light scatter settings and possi-
bly the threshold will need to be adjusted since beads are often
much smaller than cells. Adjust the FSC/SSC light scatter
settings until you can see good separation between the lym-
phocyte, monocyte, and granulocyte populations. Be sure to
set the threshold to just below the lymphocyte population to
minimize the noise and debris events but still show the
entire population of lymphocytes present. Keep the granulo-
cyte population on scale however, take advantage of the entire
FSC/SSC area to see each population clearly like that seen in
Fig. 12: FSC versus SSC with good separation and resolution.
PMT Optimization: Optimizing the fluorescence PMTs
can be performed several different ways and each has advan-
tages and disadvantages, but they should all allow the user to
see an improvement in signal separation between the bright
positive, intermediate, and negative populations present. Two
of the most popular methods used to optimize the PMT set-
tings are the Unstained Sample Method and the Single-stained
Sample method. Note: optimization should be performed with
the compensation settings turned off. That step follows
optimizing the instrument settings.
(a) Unstained sample method: This method has been used
for several years and is relatively easy to perform and not
very expensive since no antibodies are required. Set up the
instrument starting with the default PMT settings from
250 Bruce Greig

Fig. 13 Unstained sample method for optimizing PMT settings. Unstained sample optimization method gated
on lymphocytes from a lysed whole blood specimen. The PMT settings are adjusted so that the leading edge of
the negative signal is just beyond the background signal in each detector

the bead QC. Run an unstained lysed whole blood sample

and draw a gate on the lymphocyte population on forward
versus side light scatter. One at a time, adjust the PMT
settings for each parameter in your sample testing panel
until the leading edge of the negative signal goes just
beyond the background noise signal. Continue until
each PMT has been set. See Fig. 13: Unstained sample
method for optimizing PMT settings.
One advantage to this method is that it does not
require any antibodies, so it is rather easy to set up. A
disadvantage to using this method, however, is that you
will not know until you run a positive-stained sample
whether the positive peak will be on scale if the setting is
too high, or whether a weak positive population can be
differentiated from a negative or bright positive popula-
tion if the setting is too low. Use the unstained results as a
starting point but also include a fully stained sample that
shows both positive and negative populations to be sure
that the all of the populations are visible on the scale of
each PMT.
(b) Single-stained sample method: This method uses a panel
of lysed whole blood samples that are single-stained with
each conjugate to be used in the experiment. The advan-
tage of this method is that you can visualize the full range
of signal intensities of multiple populations: bright, inter-
mediate, and negative for each PMT. For the single-
stained samples, choose an antibody that has multiple
 Quality Control of Immunophenotyping 251

levels of expression depending on the populations pres-

ent. Three antibodies that are useful for this method are:
l CD4: gated on lymphocytes and monocytes, stains
positive for T-helper lymphocytes (bright), monocytes
(intermediate), and is negative for CD8+ T cells and
granulocytes.
l CD33: Another good marker for visualizing three or
more populations. Monocytes are strongly positive,
granulocytes stain with an intermediate intensity, and
lymphocytes are negative.
l CD8: gated on lymphocytes, stains T-cytotoxic sup-
pressor cells very bright and natural killer cell lympho-
cytes with a heterogeneous mid-intensity, and CD4+ T
cells are negative. CD8 is one of the brightest antibo-
dies available and gated on lymphocytes it will usu-
ally express intensity at the highest decade possible for
that particular conjugate; however, CD8 may not be
the best antibody to use for optimization simply
because the intermediate population
(NK lymphocytes) is oftentimes not very distinct com-
pared to the intermediate populations observed using
CD4 or CD33. See Fig. 14: Stained cell method for
optimizing PMTs
Adjust the PMT setting so that the bright peaks are
on scale, the intermediate peak can be clearly seen, and
the negative population is also clearly defined, just
beyond the background signal. This includes seeing a

Fig. 14 The Stained cells method for optimizing PMTs. In Histogram a, WBCs gated on lymphocytes and
monocytes stained with CD4, three separate populations can be clearly visualized: T-Helper cells stain bright,
monocytes have an intermediate intensity, and the T-cytotoxic suppressor lymphocytes are negative. In
Histogram b, CD33 is used to optimize the PMT. Ungated WBCs stained with CD33 show monocytes are bright,
granulocytes intermediate, and lymphocytes are negative. In Histogram c, WBCs gated on lymphocytes
express bright CD8, intermediate positive for NK cells, and are negative for T-helper lymphocytes
252 Bruce Greig

difference between discrete, intermediate populations

versus bright or negative populations. The panel you
choose to optimize with should include as many of the
same single color-conjugates as you have available
using the antibodies previously suggested or use others
that will give you the three separate populations. CD45
is not a good choice for this method simply because in
normal peripheral blood there is no negative popula-
tion of intact WBC’s. Defining the difference between
negative and positive is an important aspect of the
method. If an antibody is not found in every color
used in the experiment, then use a combination panel
that includes one or more of the other choices in the
missing colors of the primary antibody conjugate. Of
course, you can also use antibodies that you will be
testing specimens with to see if they look okay; how-
ever, be sure that the results are on scale and discrete
differences can be detected between negative and posi-
tive results. See the Notes section for complete instruction
of this procedure: Single-stained Sample Optimization
Procedure, which describes this method of optimizing
the instrument.
12. Color compensation.
Once the instrument PMT settings are established the next
step in Analytic Quality Control is to set up the color compen-
sation. Compensation is the mathematical correction for spec-
tral overlap between different fluorophores. If two or more
color conjugates are used in an experiment, especially if they
are excited by the same laser, there is a high probability that
some of each signal will overlap into the adjacent signal. This
overlap must be accounted for and adjusted to insure signal
fidelity, each color independent of the adjacent or overlapping
signal. Compensation correction is especially important in dis-
tinguishing weakly expressing antigens from negative popula-
tions since the bright signal of one fluorochrome might spill
over and be detected as a dimly positive population of an
adjacent fluorochrome. Even though band pass filters allow
specific wavelengths of light to pass through to the detector,
while blocking others, there may still be a signal from one
fluorophore overlapping with the signal of another fluoro-
phore. This can incorrectly make a single positive or even
negative signal appear to now have a weak positive signal. See
Fig. 15: PE versus PC7, uncompensated versus compensated.
For the sample that is positive for one antibody to not be
incorrectly counted as positive for the other antibody a correc-
tion calculation, also known as compensation, must be applied
to the signal spillover. Now, when the compensation is applied
 Quality Control of Immunophenotyping 253

Weak PE positive population

5 P1
10

8C COMP QC 10-15-2018-Lymphoma
4
10

150
CD4 PE PE-A

Count
3

100
10
2

50
10
0

0
-76

2 2 3 4 5 2 3 4 5
-10 0 10 10 10 10 0 10 10 10 10
-327 -185
CD2 PC7 PE-Cy7-A CD4 PE PE-A

Fig. 15 PE versus PC7 undercompensated and showing a weak PE positive population. PE and PC7-stained
cells appear to express both PC7 and two populations of PE + PC7 dual-positive populations incorrectly due to
undercompensation of PE-PC7 in PE
104
104

A B
103
103
HLA-DR APC-H7

HLA-DR APC-H7
102
102

101
101

100
100

100 101 102 103 104 100 101 102 103 104
CD34 APC CD34 APC

Fig. 16 Different compensation settings of APC-H7 versus APC. Histogram a: Overcompensation between
APC-H7 and APC appears to show two separate populations and Histogram b: APC-H7 and APC, correctly
compensated, shows a single population that is positive for both CD34 and HLA-DR

there should be no false-positive populations present. Con-

versely, if the compensation correction is too much it could
potentially cause a weak dual-positive population to disappear
and instead be only positive for one marker by being drawn
into the axis of the adjacent conjugate. For example, see Fig. 16,
CD 34 versus HLA-DR (see Fig. 16: Different compensation
settings of APC-H7 versus APC).
This dot plot is from a clinical flow cytometry leukemia
panel. The CD34 versus HLA-DR appears to have both a dual-
positive and a CD34 exclusive population implying that there
are two different blast populations; however, the signals are
254 Bruce Greig

actually overcompensated. Why does this matter? Because

immature cells that are CD34 positive and HLA-DR negative
may be mistaken for promyelocytes. Promyelocytic leukemia
(PML) is a very dangerous disease that, if not identified early,
can, in many cases, cause disseminated intravascular coagula-
tion (DIC), an internal bleeding disorder, so identifying it is
imperative. Correctly compensated, the population in question
in this example is actually dual positive for both CD34 and
HLA-DR and turns out to be AML not APL.
13. Setting up compensation on a flow cytometer.
In some cases, the same beads used for checking instru-
ment operation and alignment might also include a setup for
color compensation so that both procedures can be done
simultaneously. There are advantages and disadvantages to
this approach depending on the application. If the same anti-
bodies and conjugates are used all the time, then this method
may be preferred. However, if there are multiple panels with a
wide variety of antibodies and conjugates in the experiment,
separately running QC beads from compensation beads may be
a better option to insure more specific compensation settings
for each panel.
Begin by setting up and running a panel that includes both
unstained and single-color stained beads or cells with one tube
for each conjugate that will be used in the experiment
[15]. After acquiring that panel compensation is then calcu-
lated by a compensation software program. A portion of the
first signal is subtracted from the total of the overlapping signal
and vice-versa for each of the colors where there is spillover.
The portion of signal subtracted from the total signal is known
as the compensation value and the total compensation calcu-
lated for the experiment is called a compensation matrix. See
Fig. 17: Compensation matrix for an 8-color experiment
In early versions of flow cytometers that could detect 4 or
fewer colors setting up the compensation could be done just by
looking at one signal versus another and making the necessary
adjustments to correct for any under- or overcompensation.
More recent flow cytometers can detect upward of 10 or more
colors; some research platforms can detect 20 or more colors.
This means calculating the compensation now requires soft-
ware programs to take into consideration every possible com-
bination of fluorochromes. For instance, to set up
compensation for a four-color experiment could require up to
12 calculated values versus the number for a 10-color experi-
ment which could be up to 90 different calculated values.
(a) Label-specific compensation: Most compensation soft-
ware programs allow the user to include both a generic
as well as label- or lot-specific antibodies in experiments
 Quality Control of Immunophenotyping 255

Compensation Matrix for an 8-Color Experiment

Signal into FITC into PE into PerCP-cy5.5 into PE-cy7 into APC into APC-H7 into BV421 into BV 510

FITC 100.00% 14.50% 3.00% 0.20% 0.00% 0.02% 0.00% 3.12%

PE 1.00% 100.00% 22.00% 0.70% 0.02% 0.02% 0.02% 0.03%

PerCP-cy5.5 0.01% 0.00% 100.00% 12.00% 1.61% 7.03% 0.03% 0.05%

PE-cy7 1.00% 5.50% 7.50% 100.00% 0.03% 10.00% 0.03% 1.00%

APC 0.00% 0.11% 1.00% 0.09% 100.00% 8.50% 0.02% 0.03%

APC-H7 0.01% 0.02% 0.06% 0.53% 1.06% 100.00% 0.04% 0.07%

BV421 0.01% 0.02% 0.06% 0.00% 0.00% 0.02% 100.00% 16.50%

BV 510 0.17% 0.16% 0.11% 0.01% 0.00% 0.01% 10.00% 100.00%

Fig. 17 Compensation matrix for an eight-color experiment. Each fluorochrome in the experiment is shown
with its respective compensation setting versus the other fluorochromes being used. The diagonal values are
all 100% because a color cannot be compensated against itself. The higher the value the more overlap each
signal has with the opposing signal. When designing an experiment, the use of color-conjugates with very high
spectral overlap should be avoided for those markers that are used to find very discrete populations since they
might be lost due to the high compensation correction value

where there is more than one color being detected in the

same PMT in a multitube panel. For instance, PE-cy5 and
PerCP cy5.5 are both acquired using the same PMT;
however, the spillover of PE-cy5 into the adjacent PE
channel will be different from the spillover of PerCP 5.5
into PE and will therefore need a separate compensation
value. Another reason to consider using label-specific set-
tings is when tandem dyes are used in the experiment. In
some cases, the tandem dye will have a different back-
ground signal compared to the same antibody/ conjugate
from a different lot number. This changes the signal–noise
ratio along with the compensation needed to account for
that difference. These settings are now lot-specific as
opposed to label specific. Label- and lot-specific settings
allow both antibodies to be used in the same experiment
since the compensation matrix now includes calculations
for both antibodies.
The compensation panel can be set up a few different
ways. Some instrument software compensation programs
use multicolor beads along with a nonlabeled (unstained)
bead. Compensation software programs require a nega-
tive/ unstained population of some kind. This establishes
the background signal from which everything beyond this
signal is counted as positive. The unstaineded bead/ cells
are first acquired using the PMT voltages previously deter-
mined during optimziation. Single colored/stained beads
256 Bruce Greig

are subsequently acquired, one for each conjugate/lot

number in the experiment. After all the tubes are acquired
the software then calculates and applies mathematical
compensation to each signal until there is no spillover
from another conjugate. The file format used for flow
cytometry results is designated with “.fcs” as part of each
file name.
All results acquired as fcs files version 3.0 and above
are acquired uncompensated; however, the compensation
calculation is included as a keyword in the file which can
then be applied during analysis or even turned off. This
gives the user the option of using the calculated compen-
sation values that are included with the file or even adjust-
ing it with the analytical software to a more suitable setting
if desired. FCS versions earlier than 3.0 have fixed com-
pensation values that cannot be adjusted post-acquisition.
Most instrument vendors supply compensation beads
that are either surface coated with the colors of the con-
jugates being used or beads that are coated with anti-IgG,
also known as “capture" or "comp" beads,”. Capture or
comp beads include both an unstained bead and another
bead which is coated with a surface anti-mouse IgG-kappa
that makes it possible to use any antibody/ conjugate to
setup compensation in your experiment as long as it is an
IgG subclass. The beads are setup and stained just like cells
however they do not require a lysing step, just washing 1x
after the primary incubation (see Fig. 18: Compensation
setup using capture beads).
The advantage of using capture/comp beads is that
unlike cells they are nonspecific and will react with almost
any antibody. The disadvantage of using beads is that they
do not always demonstrate the same intensity as a stained
cell. If the comp bead signal is very bright or is at the end
of the histogram scale turning down the PMT voltage to
lower the signal is not an option since it changes the
setting you worked hard to optimize. Another disadvan-
tage of using comp beads is that they will not react with
some panel reagents such as 7-AAD or Syto 6 which are
dyes, not antibodies. Reactive cells on the other hand, may
be difficult to come by especially for rare antigens. They
should, however, more closely resemble the signal inten-
sity of cells used in your experiment so a compensation
setup using stained cells may be more accurate than using
beads.
An additional consideration when setting up compen-
sation using stained cells is deciding what sample types to
use. The stained control samples need to be at least as
bright as any that will be used in the experiment since it is
 Quality Control of Immunophenotyping 257

Fig. 18 Compensation setup using capture/ comp beads which are beads coated
with anti-mouse IgG can be used to set up a compensation panel. The beads
react with any antibody as long as it is IgG-class. The advantages are that they
will work almost universally, be prepared in advance, and can be used with rare
antibodies. However, the bead staining may show a different intensity compared
to stained cells, and autofluorescence may be an issue with certain lasers.
Beads cannot be used for nonantibody stains like 7-AAD or Syto-6

the brightest color signals that can spill over into the
adjacent channels. One of the best sample types and anti-
bodies to use in the compensation panel is CD8-stained
lymphocytes. CD8 is a high-density antigen on lympho-
cytes and thus whatever conjugate is used CD8 will be as
bright or even more bright than other common lympho-
cyte antigens. It is also negative on CD4+ T-lymphocytes
so now there is both a positive and negative population
present for calculating the compensation. If CD8 is not
available in every conjugate of your panel, choose a differ-
ent one that will have a positive signal intensity of at least
the third decade on a log scale. Other antibodies to con-
sider using in a stained cell compensation panel would be
either/or both CD33 and CD14 gated on monocytes,
and CD16 gated on granulocytes. CD33 gated on all
WBCs in LWB should demonstrate a bright positive
monocyte population, intermediate positive granulocytes,
and be negative on lymphocytes. CD14 will be very bright
on monocytes and negative on lymphocytes. CD16 will be
very bright on neutrophils, moderately positive on NK
lymphocytes, and negative on monocytes. Avoid using
cell cultures in the comp panel unless your experiment
uses these cell sources, due to the increased background
or noise or autofluorescent signal they often have com-
pared to normal lymphocytes.
258 Bruce Greig

In most cases the calculated compensation from the

controls will be acceptable for most if not all the para-
meters in your experiment. An example Compensation
setup procedure is found in the Notes section of this
chapter.
14. Compensation controls: Before using the optimized and now
compensated instrument settings to test actual samples from
your experiment additional controls should be run using the
optimized and now-compensated PMT settings to be sure that
the compensation is correct. This requires running an addi-
tional panel for demonstrating and adjusting the compensation
if necessary. The compensation controls setup procedure is
found in the Notes Section of this chapter (Section 4.5).

3.3.3 Compensation Once compensation is set up it is important to test the settings

Controls using single-stained cells. Cells are preferred over beads since the
samples in your experiment will in most cases be cell-based, and it is
okay to use the same cells you used to create the compensation to
check the new compensation settings. There are different
approaches to how best to do this using stained samples [16,
17]. The next section describes a procedure that many flow labs
use.
1. Fluorescence-Minus-One (FMO) controls
FMO controls are one of the most widely used ways for
checking compensation settings. For this procedure stained cell
samples are set up that include every conjugate found in the
experiment minus 1 color. For instance, for a 10-color experi-
ment using FMO-controls see the example below. "FL" stands
for flourescence color; the number stands for the PMT:
(a) Tube 1: (blank)/FL2/FL3/FL4/FL5/FL6/FL7/FL8/
FL9/FL10.
(b) Tube 2: FL1/(blank)/FL3/FL4/FL5/FL6/FL7/FL8/
FL9/FL10.
(c) Tube 3, 4, etc. until each color has been omitted once.
Continue building the panel for all 10 tubes omitting a
different color in each tube.
(d) Run the FMO panel on the flow cytometer using the PMT
and compensation settings previously determined.
(e) After acquiring each tube in the FMO panel analyze and
create two-parameter dot plots that include the offset or
biexponential display of each negative/blank channel ver-
sus each of the other positive channels. See Fig. 19: Log
versus offset/biexponential dot plots which shows the
difference between a dot plot with and without an offset
display.
 Quality Control of Immunophenotyping 259

104

104
103

103
CD4 PE

CD4 PE
102

102
1
10

1
10
0
0
10

0
10 101 102 103 104 0 101 102 103 104
CD3 APC CD3 APC

Fig. 19 Log versus offset/biexponential dot plots. The dot plot on the left shows the origin of each axis at the
bottom. The dot plot on the right is the same dot plot with the axis offset/biexponential setting turned on
allowing you to see events both above and below the origin (blue arrow) for demonstrating compensation

Fig. 20 Compensation examples using FMO controls. The positive CD4 result is shown versus an unlabeled
detector as a way of demonstrating compensation between the two PMTs. The ideal compensation setting (dot
plot a) shows the mean fluorescence intensity (MFI) of the positive peak to be very close to the MFI of the
negative peak. Undercompensation in the middle figure (b) shows the MFI of the positive peak to be
significantly greater than the negative peak MFI. The MFI of the right dot plot (c) shows the positive MFI to
be significantly less than the negative peak MFI and thus is overcompensated

The next example (Fig. 20: Compensation examples

using FMO controls) illustrates what (a) correct,
(b) undercompensation, and (c) overcompensation look
260 Bruce Greig

like using three different settings. Adjust the compensa-

tion settings to get an even distribution of positive and
negative spread on the baseline otherwise known as the
"fishtail" pattern.
Once you have reviewed and made any final adjust-
ments to the compensation the settings can now be used
in your experiment.
2. Alternate method for checking compensation settings: FMX
controls
The FMO panel is a good method for checking compensa-
tion settings but it has some drawbacks. It can be an expensive
(up to 90 antibodies in a 10 tube FMO panel) and time-
consuming approach to checking compensation. However,
another similar but less expensive compensation control proce-
dure is to set up the same panel of 10 tubes but instead of
leaving out one conjugate in each tube, leave out all but one so
that you have a positive result in one conjugate that can be
compared to the all of the other negative parameters. Include
an additional tube that has reactive antibodies for each param-
eter in your experiment. It will be used to make any final
adjustments to the previously calculated compensation and
should now demonstrate how well the compensation settings
work. Again, run the FMX control tubes using the previously
determined instrument settings, create offset dot plots of every
possible combination and compare the positive MFI result in
each file to the other unstained/ negative parameters. An
example of the FMX-panel would look something like this:
(a) Tube 1: CD8 FITC/FL2 blank/FL3-blank, FL4-blank,
FL5-blank, etc.
(b) Tube 2: FL1 blank/CD8 PE/FL3-blank, FL4-blank,
FL5-blank, etc.
(c) Tube 3: FL1 blank/FL2 blank/CD8 PerCP/FL4-blank,
FL5-blank, etc.
(d) Tubes 4 or more, until each color has been used once.
(e) Include a fully stained all-color control tube as the last
tube in the panel. See the example below.
CD20 FITC/CD2PE/CD5 PerCPcy5.5/CD8 PE-cy7/
CD3 APC/CD7 or CD14∗ APC-H7/CD4 BV421/CD19
BV510.
(f) Acquire the FMX panel on the instrument using the pre-
viously calculated compensation settings. Analyze each file
one at a time and look at each positive result versus the
other fluorescence parameters used in the panel and be
sure that the biexponential or offset setting is turned
on. Now you should be able to recognize compensation
 Quality Control of Immunophenotyping 261

Fig. 21 FMX CD8 FITC control versus Each PMT. CD8 FITC versus all other colors. In this display, CD8 FITC
versus each other color shows good compensation settings except for the CD8 FITC versus BV421 which is
slightly overcompensated and the same versus BV510 which requires additional compensation

that is acceptable between two colors versus compensa-

tion settings that need adjustment (see Fig. 21: FMX
CD8 FITC Control vs each PMT). This display demon-
strates that the CD8 FITC versus PerCP and the same
versus BV510 require additional compensation.
Follow this same example for each FMX control and
change the X-axis to whatever parameter is positive versus
the negative channel on the Y-axis for all subsequent
tubes. Hint #1: If you set this up once and save the analysis
that includes each possible color combination it can be
used again for checking the compensation of a fully
stained sample in your experiment. That compensation
control is described in the next section.
3. Fully stained compensation control.
The last compensation control tube should, if possible,
include all antibody conjugates used in the experiment. The
eight-color cocktail used in the example includes several lym-
phocyte and monocyte antibodies found in normal peripheral
blood. Some will coexpress and some will be exclusive. Use the
results of this tube to do one last check of each combination to
insure no under- or overcompensation exists. Hint #2: This
cocktail can also be used to monitor compensation each day or
periodically without setting up and staining multiple tubes.
Keep a copy of the analysis worksheets showing all the fluores-
cent combinations for future reference so that if questions arise
about the compensation in an experiment, analyze the results in
question and use the previously saved dot-plot combination
worksheet to quickly identify any compensation issues. By
observing each conjugate versus the others, you should be
262 Bruce Greig

Fig. 22 Example of eight-color compensation. Multiple dot plots of every combination in an eight-color
experiment using a cocktail that includes CD8/4/5/2/3/7/56 or 16/19 + 8. CD19 and CD8 are paired with the
same conjugate to observe both coexpression and single staining. This example only shows two of the
possible combinations, FITC versus X and PE versus X. Instrument compensation settings can now be
monitored daily using a normal whole blood control with the multicolor cocktail

able to identify any compensation issues and make minor

adjustments if necessary. See Fig. 22 Example of 8-Color com-
pensation shows 2-parameter dot plots of multiple detectors in
an experiment.
This example only shows half of the possible combinations,
FITC versus X and PE versus X. Instrument compensation
settings can now be monitored daily using a normal whole
blood control stained with the multicolor compensation
control cocktail.
4. Compensation settings summary
Multicolor flow cytometry experiments are only as good as
the compensation settings used to correct for spectral overlap
of all the color conjugates found in the panel. Once the com-
pensation is set up initially, daily/ weekly monitoring is recom-
mended to ensure that those settings are still valid. Minor
adjustments can then be made to the compensation to correct
for any drift. If on the other hand the PMT settings change by
more than 10%, or there is major maintenance performed, the
compensation setup should be repeated since either of those
situations will likely change the spectral overlap of some or all
 Quality Control of Immunophenotyping 263

of the conjugates being used. An experiment that has been set

up on a flow cytometer that is properly QCed, optimized, and
color-compensated helps insure that results are accurate, pre-
cise, and dependable.

3.4 Miscellaneous Immunophenotyping laboratories have other instruments besides

Laboratory the flow cytometer that are required for setting up an assay. Exam-
Equipment QC ples include refrigerators and freezers, centrifuges, biological flow
hoods, and manual or automatic pipettes to name a few. These
devices must be maintained to insure proper operation and suitabil-
ity for use. If the refrigerator or freezer that you store antibodies in
has wide fluctuations in temperature over time it could damage the
reagents. A centrifuge has rotors, safety latches, and timers that
should be checked regularly to insure safe and proper operation.
Biological flow hoods are safety devices that should have regular
scheduled maintenance to ensure that potential hazardous material
is contained. Most laboratories use private companies who special-
ize in maintaining this sort of equipment. Records of regular main-
tenance of miscellaneous equipment and daily monitoring of the
temperatures is considered a good laboratory practice and is
required by some laboratory regulation agencies. [29]

3.4.1 Analytic Quality Daily monitoring of the flow cytometer performance including
Control Conclusion alignment checks, background and sensitivity limits, fluidics, PMT
performance, and laser output, is a standard of best practice in flow
cytometry laboratories and may also be required for regulatory
compliance. It helps insure consistency and reliability on a day-to-
day basis. Tracking the results of QC over time may reveal trends
that could predict future instrument problems. Understanding
how each component of the instrument is performing: fluidics,
optics, alignment, and electronics helps insure that results are reli-
able, accurate, and precise. Besides the flow cytometer, keeping
maintenance and operation records of other miscellaneous equip-
ment in the laboratory such as refrigerators, freezers, and centri-
fuges is also important for meeting quality assurance best practice
standards [30].

3.5 Postanalytic Once the instrument has successfully passed the analytic quality
Quality Control control procedures the final step of verification is to run sample
Procedures controls. This includes both quantitative and qualitative controls
[18]. For example, clinical laboratories that report quantitative
3.5.1 Quantitative Assay results such as lymphocyte subset percent and absolute counts, or
Quality Controls stem cell percent and absolute counts, are required to run at least
two levels of commercial or verified controls each day the test is
performed. The controls should include both a normal control as
well as an abnormal control so that the full range of reportable
results is represented. Quantitative commercial controls come with
package insert values that must also be verified by the laboratory
264 Bruce Greig

prior to being put into use. Oftentimes these control values have a
wide range of acceptable limits and therefore the laboratory should
not only verify those limits but calculate ranges of their own based
on multiple runs. Using the mean 2 standard deviations after
10 times will usually be enough to generate statistical ranges that
satisfy the requirement. The laboratory calculated control ranges
should now reflect what the reports on both normal and abnormal
results will be and should not be below the lower limit of detection
(LLOD).
For those assays in which a commercial control is unavailable
the use of a known positive sample verified by other means or by
another laboratory that performs similar testing can be an accept-
able substitute. Records of the controls and how the results were
calculated are required for verification and acceptability for use.

3.5.2 Qualitative Assay Qualitative controls, also known as method controls, test not only
Quality Controls the instrument but the staining and processing procedures from
setup to acquisition to analysis. Every antibody in an experiment
should be tested for accuracy prior to being put into use whether as
part of a cocktail or as a standalone reagent. This quality testing was
mentioned previously in the preanalytic quality control section.
Some commercial controls used for quantitative testing also include
a list of additional antibodies that will react with the control for
qualitative purposes. Method controls on the other hand are those
controls used in an experiment that prove that the setup and stain-
ing procedures are working as indicated. As was also mentioned in
the preanalytic QC section, material tested using alternate methods
such as histology and hematology can be used as method controls.
If proficiency testing material is available from a recent survey, it too
can be used as a method control.

3.5.3 Isotype Controls Isotype controls are basically antibodies of the same immunoglob-
ulin class as the primary antibody including both the heavy and
light chain molecules but lacking the specific component of that
antibody to react with the target antigen or are directed against an
irrelevant target. Isotype controls can serve as experiment controls
and many labs use these to demonstrate the background staining or
nonspecific staining present when using the test antibody. Setting a
statistical gate on the positive population was based on where the
negative signal ended for the isotype control. But there are some
inherent limitations to using isotype controls for this purpose
[19]. In some cases, the isotype control may not exactly match
the target antibody and may react with an unexpected antigen,
what is known as nonspecific binding. This may introduce uncer-
tainty about what is positive versus what could be nonspecific
staining. Another difference in using an isotype control is that it
may have a different fluorescence to protein (F/P) ratio versus the
target antibody. If the F/P ratios are different then the background
 Quality Control of Immunophenotyping 265

staining may be different and now the negative vs positive bound-

ary becomes subjective.
There are situations, however, where the use of an isotype
control can be helpful. Immunophenotyping panels designed to
study intracellular antigens can be challenging due to the presence
of more Fc- receptors inside the cell than on the cell surface which
may result in more nonspecific binding of the antibodies and thus a
higher background signal. An isotype control can be used to know
what that nonspecific staining looks like to more accurately know
where to gate the positive population. Another useful application of
isotype controls is when a fluorochrome is being used for the first
time and you want to see what typical background staining looks
like. This provides a starting point for optimizing the signal as part
of setting up the instrument.
Rare event detection is another situation in which the isotype
control may serve as a useful reference for determining positive
versus negative especially in those samples that may have weak
expression or if a low density antigen is being investigated. How-
ever, a better strategy is to create a gating scheme that with succes-
sive structure, such as counting stem cells using the ISHAGE
protocol [31], can preclude the need for an isotype control. As
the population of interest is refined, the sequential gates used in the
analysis may provide a more accurate assessment of determining
positive versus negative results rather than running a separate iso-
type control.

3.5.4 Postanalytic Quality Internal/intraassay Controls are method controls that occur natu-
Controls Internal/Intraassay rally in a sample and demonstrate what positive versus negative
Controls looks like for each same cell types being studied. For instance,
lysed whole blood WBC’s include lymphocyte, monocyte, and
granulocyte populations. An antibody like CD33 reacts with
monocytes and granulocytes but is negative on lymphocytes.
CD19 is positive on B cells but is negative on T and NK cells as
well as other myeloid populations present. In each of these exam-
ples you have one or more positive populations and one or more
negative populations. In most bone marrow specimens other exam-
ples can be used within the tube to confirm different antibody
reactivities, both positive and negative, using the same approach.
See Fig. 23: Using intraassay controls.
An advantage of using internal/intraassay controls is that now
the test antibody’s reactivity can be described because the condi-
tions, F/P ratio, and isotype are all in one place compared to a using
a separate isotype control tube.
Each lab should establish their own criteria for describing the
use of internal controls as a method and cocktail quality control
check. Abnormal populations typically found in malignancy may
aberrantly express antigens or be missing antigens that are expected
266 Bruce Greig

Fig. 23 Using intraassay controls. Antibody reactivity can be assessed using

intraassay controls. Histogram A shows lymphocytes (L) in green, monocytes
(M) in red, and granulocytes (G) in gold on forward/side light scatter. In
Histogram B, lymphocytes can be used to assess antibody performance since
in the example shown CD19 is positive for B cells and negative for T cells,
monocytes, and granulocytes. CD33 is correctly positive on monocytes (bright)
and granulocytes (weak) and is negative on lymphocytes

to be present on normal cells. In these cases, clear descriptions of

normal reactivity should be included with each antibody combina-
tion used in the experiment as part of the procedure Interpretation
of Results section.

3.5.5 Postanalytic Quality Flow cytometer list mode data is stored in a specific format that
Control Methods Using includes not only a record of every signal or event but also the time
Time as a Parameter point at which that event occurred. This additional measurement is
known as the Time parameter. The Time parameter can be a useful
quality control indicator to alert you to problems with the sample
acquisition, especially the fluidics [20]. For example, when an
acquisition is either interrupted by something briefly blocking the
fluidics or if the sample runs dry, these incidents can result in a loss
of signal or the inclusion of bubble events in the list mode
file results. Bubbles will look like collected events especially on
light scatter. Looking at a histogram of time versus another param-
eter should normally show a steady sequence of events, but if there
is an interruption during the acquisition or bubbles present, the
histogram will demonstrate a noticeable change in the pattern or
shift of a parameter. An interruption in the acquisition can be
identified using the Time dot plot. Results before the interruption
should look continuous; however, the results after an interruption
may produce a false signal that could change the appearance of your
data. The good news is that you can still salvage an otherwise faulty
 Quality Control of Immunophenotyping 267

Fig. 24 Using Time as a gating parameter. The FSC/SSC histogram on the left shows what happens when a
sample is run until empty. This introduces air bubbles, which appear as the multiple black dots in the upper
area of the histogram. The middle histogram, Time versus CD45 shows a clear break where the bubbles
started being acquired. A gate is then drawn around the “clean” events to omit the bubble events. The bubble-
free-gated FSC/SSC histogram on the right no longer shows any bubble events

acquisition without having to set up or acquire it again by gating

away from the problem. For example using the Time versus FL4 or
other fluorescence parameter, draw a gate that now includes only
the events before the interruption or when air was aspirated. Now
the newly gated dot plots should be clear of the previously acquired
interruption or air bubbles. Subsequent dot plots should now be
based on the Time gate (see Fig. 24: Using time as a
gating parameter).

3.5.6 Postanalytic Quality Another issue that can impact the quality of your results is when the
Control Methods Using sample contains doublets or cellular aggregates. Doublets are when
Singlet Gates two or more cells for whatever reason are bound together as they
pass through the flow cell chamber and are counted as one event.
Using a singlet gate is an important consideration especially in
experiments that measure DNA ploidy. If two cells with a 2N
DNA content (resting, G1/G0 phase) are measured as one cell
they could be mistaken as one cell having twice as much (prolifer-
ating, G2 + M) content. Samples with numerous doublets would
then falsely skew the number of proliferating cells versus resting
cells. Doublets can also be seen when specimens undergo several
centrifugation steps such as preparing bulk lysed samples. Bulk
lysing is often used in experiments that look for rare events, espe-
cially minimal residual disease (MRD) leukemia cases. To capture
these rare events more specimen is required than that used under
routine circumstances. In a bulk lysing procedure, using 1 ml of
sample instead of 100 μl increases the number of cells tenfold. Now
you have significantly more WBCs than before, and here is where
the problem of doublet aggregate combinations is likely to occur
268 Bruce Greig

(x 1,000)
200 250
5
10
CD19 PerCP-Cy5-5-A

Singlets
4

150
10

FSC-H
100
3
10

50
2
10

2 3 4 5
50 100 150 200 250
10 10 10 10 FSC-A (x 1,000)
CD7 FITC-A

Doublets

Fig. 25 Using a singlet gate to omit doublet events. In this example, the blue population before singlet gating
appear to be positive for both CD19, a pan-B cell marker and CD7, a T cell marker. Normal B cells should be
negative for CD7 and normal T cells should be negative for CD19. Using a Singlet gate that excludes doublets
shows the red populations to be as expected for both CD7 and CD19. Singlet gating is a good practice
regardless of the sample type since this phenomenon may happen quite frequently

especially if there are multiple centrifugation steps during the sam-

ple lysing and washing. If doublets are included in the analysis, the
results may be misidentified as something abnormal such as a cell
appears to coexpresses both a T cell antibody and a B cell antibody
when in fact it is two separate cells that are stuck together also
known as a doublet. To eliminate doublets from the analysis, draw a
dot plot of forward scatter (FSC) area versus forward scatter height.
Please note that some instruments measure width instead of height
and if so, use that signal instead versus the FSC-Area. Doublets
have a larger surface area than single cells and can be seen falling
toward the FSC-Area axis. Draw a gate around the singlet popula-
tion which is found along the diagonal axis of the dot plot to now
exclude the doublet events. This should be the first gate used in the
analysis so that subsequent events are only those found within the
singlet gate (see Fig. 25: Using a singlet gate to omit doublet
events).

3.5.7 Postanalytic Quality Samples received for immunophenotyping if collected locally or

Control Methods Using within a day’s distance will usually have high levels of viability
a Viability Gate especially if the recommended temperature and transport media
are used to deliver the samples. However, in some circumstances
even though all efforts are made to ensure that the sample is viable
the sample by its very nature of collection may contain nonviable
cells. Apheresis specimens submitted for stem cell counts is a prime
example. The mechanical means by which the product is collected
as well as the length of time to perform the apheresis procedure can
result in some portion of the cells dying. Since CD34+ stem cell
counts are typically less than 1% the need for accurate gating to
 Quality Control of Immunophenotyping 269

Fig. 26 Using 7-AAD to identify and gate out nonviable cells. Events in dot plot A have a high number of
nonviable cells which take up the 7-AAD dye, demonstrated by the very bright expression of the 7-AAD on the
Y-axis. The dead cells are seen in dot plot B, FSC/SSC as the secondary population on the SSC axis above the
lymphocytes. Dot plot C, SSC/CD19, includes the dead cells which dilute the actual percentage of B cells
present in the sample. Gating on just the cells that are 7-AAD negative in dot plot D cleans up the analysis by
now omitting the dead cells. The SSC/CD19 dot plot F () shows only the viable cells with a more accurate CD19
result

omit the nonviable cells is crucial to the experiment. Other sample

types with the potential for increased numbers of dead cells would
include CSF and tissue biopsies especially if they lack adequate or
appropriate transport media like RPMI. Using a viability dye such
as 7-AAD in the immunophenotyping panel can be very beneficial
for identifying the dead cells which take up the 7-AAD dye since the
cell membrane of dead cells is usually permeable[26]. Including a
viability gate in your gating scheme is highly recommended and
even required if you perform clinical testing on samples over 24 h
old (see Fig. 26: Using 7-AAD to identify and gate out of the
nonviable cells for an example of how to remove dead cell events
from the analysis).

3.6 External Quality Another regulatory requirement for clinical labs performing immu-
Assessment (EQA) nophenotyping is to participate in an external quality assessment
program (EQA) also known as proficiency surveys. Laboratory
regulatory agencies such as the College of American Pathologists
require that a laboratory reporting clinical results of any assay
participate in an EQA that is designed to test each aspect of the
testing and reporting procedure. The laboratory’s performance is
then compared and graded against other laboratories performing
the same testing. Proficiency surveys typically occur 2–3 times per
270 Bruce Greig

Plot of the Relative Distance of your Results from

Target as Percentages of allowed Deviation
Survey -100------------Mean-----------+100
B Lymph Diff Ag CD19+
FL-A 2017
% FL-C 2016
FL-B 2016
-100 -80 -60 -40 -20 0 20 40 60 80 100

NK Lymph CD16+/56+/CD3
% FL-A 2017
FL-C 2016
FL-B 2016
-100 -80 -60 -40 -20 0 20 40 60 80 100

Fig. 27 Proficiency results over time. The CD19% over three different surveys shows results that fall both
above and below the mean of the acceptable results. Conversely, the results of the NK Lymphocytes
consistently run above the mean during the same period. This may be indicative of an analysis gating issue
or antibody that might need to be further investigated for suitability

year and will usually include two or more samples with each mail-
ing. Proficiency testing must follow the laboratory’s written proce-
dures for all the steps used to test and report patient samples. This
not only provides information about the validated procedures but
also measures the performance of the instruments used, acceptabil-
ity of the reagents, and should over time even demonstrate poten-
tial issues that may not be as obvious on a day to day basis. For
instance, if the returned results for CD19 in an immunodeficiency
panel are acceptable but consistently trend higher than the mean of
the other participants it may indicate either a reagent, gating issue,
or instrument problem that might otherwise have gone unnoticed
(see Fig. 27: EQA results over time).

3.7 Quality Control All testing in Immunophenotyping requires that records of quality
Records control performance be maintained and retained for a certain
period of time[27]. There is a saying that goes, “If you did not
document what happened then it did not happen.” Quality assur-
ance and quality control records are the proof of that purpose. Each
step in the process of preanalytic, analytic, and postanalytic perfor-
mance must be documented for acceptability. Best practices in
quality control management include having an overall quality assur-
ance plan along with the records of procedure and instrument
validations, reagents used in testing, daily instrument performance
and control results, assay records of specimen analysis that include
the gating diagrams, and any proficiency testing results. For a
summary of records that should be readily accessible and up to
date see the Notes section of this chapter.

3.8 Conclusion Quality control (QC) and quality assurance (QA) in immunophe-
notyping spans all three phases of laboratory testing: preanalytic,
analytic, and postanalytic procedures. As part of an overall quality
 Quality Control of Immunophenotyping 271

program performing and documenting each result of these controls

helps ensure the best outcome of an experiment. Good laboratory
practice requires that every aspect of the assay should have measur-
able methods for monitoring the quality of the reagents, the instru-
ment’s performance, and the assay’s expected results. Records of
QC for each procedure should be well maintained and up to date
and include the signature of the person or persons who oversee the
testing.

4 Notes

4.1 Reagent Recipes 1. PBS buffer, pH 7.2, with sodium azide. Some laboratories
Used in Immuno- include bovine calf serum in their PBS buffer as an additional
phenotyping stabilization agent for the cells. This recipe does not include
that additive.
Caution: Sodium azide is an irritant and a poison. Avoid all
direct skin contact, wear gloves, and wash hands thoroughly
after handling.
Materials:
(a) PBS, 1, Mediatech, Inc., Cat. # 21-040-CM.
(b) Sodium azide, Fisher Scientific, #S227-500.
Working solution:
(a) Add 1 g sodium azide to 1 l of plain PBS.
(b) Mix until dissolved.
(c) Check that the pH is between 7.2 and 7.4.
(d) Label and date (date made and date expired).
(e) Store at 2–8  C when not in use. Expiration ¼ 3 months.
2. Ammonium chloride (NH4Cl) RBC Lysing Agent:
Materials:
(a) Ammonium chloride, Fisher Scientific, #A661-500.
(b) Potassium bicarbonate, Fisher Scientific, P184-500.
(c) Disodium EDTA, Fisher Scientific, #S311-500.
Working solution:
(a) Place the following into a 1-l volumetric flask:
l 8.26 g ammonium chloride.
l 1.0 g potassium bicarbonate.
l 0.037 g (37 mg) disodium EDTA.
(b) Add deionized water to 1-l line on flask. Mix until all
ingredients are dissolved.
(c) Check pH: 7.2–7.4. Discard, if >7.4.
(d) Dispense into 100 ml bottles. Fill bottles to near the top
so that there is no air space∗.
272 Bruce Greig

(e) Cap tightly. Put Parafilm around the cap to prevent air
exposure.
(f) Label and Date (date made, and date expired). Expira-
tion ¼ 1 week. Store at 4  C until ready to use.
∗NOTE: There should be as little air on top of the
lysing solution as possible. Air will cause a decrease in the
pH over time and lower the effectiveness of the lyse
solution.
3. 7-AAD Viability Dye.
Materials:
(a) 7-Aminoactinomycin D, Molecular Probes, Cat.#A-
1310, 1 mg.
(b) Dimethyl sulfoxide (DMSO), Fisher Scientific, Cat.
#D128-500.
(c) 500 ml Flow PBS, stock solution.
7-AAD Stock Solution:
(a) Add 1 ml DMSO to 1.0 mg 7-AAD.
(b) Store at room temperature or in the dark for up to
1 month.
7-AAD Working Dilution
(a) Use 5 μl of 7-AAD stock solution/1–two million cells.

4.2 Four-Color T-Cell

Subsets Cocktail for a Antibody Volume/test∗ Antibody total Buffer volume
10- Test Aliquots CD3 FITC: 10 μl/test 100 μl 650 μl
Recipe CD4 PE 10 μl/test 100 μl
CD8 PerCPcy5.5 10 μl/test 100 μl
CD45 APC 5 μl/test 50 μl

∗Volume/test is based on titration results of initial antibody quality control

4.3 Instrument 1. Prepare to use the instrument by turning it on and allowing

Quality Control sufficient time to warm up the lasers and electronics, usually
Workflow 10–15 min.
2. Check or Refill Fluids for Operation: Sheath tank, cleaning
solutions, etc. and empty waste containers; add bleach or
other neutralizing hazardous waste solutions to a waste
container.
3. Perform any scheduled maintenance following the manufac-
turer’s guidelines: clean interior of flow cell using an approved
medical device cleaning solution, change fluidics filters, once or
twice a month perform an extended whole instrument clean
(“long clean”).
 Quality Control of Immunophenotyping 273

4. Open Operation Software and initiate instrument fluidics

startup if applicable.
5. Prepare QC material: beads, stained cells, and controls.
6. Open the Quality Control software program and run beads or
some other vendor-suggested material according to the Instru-
ment Setup procedure to check the: alignment, sensitivity,
precision, accuracy, electronics/PMT settings, laser power out-
put, and fluidics performance.
7. (Optional step, see step 12 for details) Optimize the instrument
settings using stained cells to produce the highest signal–noise
ratio for each PMT as well as the forward versus side light
scatter to insure good population separation and resolution.
Include this step whenever a new QC bead lot number is put
into use or baseline performance is set up, or when there has
been major maintenance such as a scheduled PM, laser or PMT
replacement, or laser/flow cell alignment. If you add new
parameters to your assay optimizing those detectors is
suggested.
8. Perform compensation monitoring daily or weekly using either
beads or stained control cells that includes all of the fluoro-
chromes used in that day’s experiments to assure proper com-
pensation for running cell samples.
(a) If three or more parameters need adjustment to the com-
pensation settings, then the full Compensation Setup pro-
cedure (see step 13) should be performed using either
single-stained beads or single-stained cells.
9. Run quantitative commercial controls if those assays are per-
formed in your laboratory. Two levels including a normal and
an abnormal are required for regulatory purposes [30]. Package
insert ranges must be validated before putting into daily use.
10. Record QC results and comment on any out-of-range results.
If three or more parameters are out of range or fail specifica-
tions or if a single parameter is out three consecutive times,
proceed to troubleshooting.
11. If troubleshooting is required, document all steps taken to
resolve the issue. Follow the instructions found in the trouble-
shooting section of the instrument operation procedure. If
results are still out, further action such as contacting the instru-
ment operational support, or the vendor service engineer may
be required. Document any troubleshooting or service per-
formed along with records of the instrument performance
after any repairs are made or otherwise addressed.
12. Optional steps: Optimize the Instrument Settings.
(a) Optimization is a procedure for adjusting the instrument
settings to achieve maximum performance and is not
required daily but should follow the initial installation
274 Bruce Greig

and/or any major maintenance such as a scheduled PM or

repair. Optimize the instrument PMT settings using
stained controls.
(b) If any instrument PMT settings are changed by more than
5 volts, compensation setup should also be performed and
should follow the instrument optimization procedure as
well as:
l After any major changes in PMT settings as part of
bead QC.
l Whenever a new antibody panel is introduced.
l If there are color-conjugate changes or additions in an
established panel. For example, changing CD45 FITC
to CD45 APC or adding PE-cy7 to replace PC7.
13. Setting up compensation.
(a) The Compensation Setup procedure uses either single-
stained beads or cells with a high-density antibody such
as CD8, gated on lymphocytes that are acquired with the
optimized PMT settings and no compensation applied.
(b) Run one tube/single color for each conjugate used in the
experiment as well as a tube with unstained cells. Use
label-specific tubes if more than one conjugate is used
for a single detector. For example, a panel that includes
FITC in one tube and Alexa 488 in a different tube would
require separate compensation settings for each color
used for the FL1 detector. Label-specific settings are also
recommended for different lot numbers of tandem
conjugates.
(c) After acquiring the panel of tubes, the compensation
software program creates a matrix and calculates the
amount of compensation to apply to each PMT signal to
correct for any spectral overlap between fluorochromes to
ensure signal fidelity.
(d) Compensation controls such as FMO (fluorescence minus
one color) or FMX (fluorescence minus all but one color)
should follow the compensation setup procedure to fine
tune if necessary any overcompensation or undercompen-
sation observed with the calculated values.
(e) The compensation settings are then saved with the PMT
settings and used to run the experiment samples.
(f) After an initial setup, compensation must then be moni-
tored and demonstrated to work daily using controls
to ensure the settings are still valid for use.
14. For clinical laboratories that perform quantitative testing, at
least two levels of controls with known acceptable ranges must
be run every day the flow cytometer is used. Controls with
 Quality Control of Immunophenotyping 275

both normal and abnormal results demonstrate that the instru-

ment’s accuracy and precision limits are still valid.
15. For laboratories that perform qualitative testing, running
known positive samples should be included daily to insure
reliability of the antibodies and cocktails being used.
16. Document all QC results in the laboratory records and com-
ment on any results that are out of range and any corrective
action taken.
17. Perform scheduled maintenance on any ancillary lab equip-
ment such as centrifuges, fume hoods, water baths, refrigera-
tors and freezers, pipettes, or other miscellaneous laboratory
equipment used for sample testing. Temperature-dependent
devices should be monitored for meeting the acceptable tem-
perature range as determined by the laboratory.
18. The laboratory supervisor or designee should review and sign
off on all QC results monthly.
19. Keep records of all QC results, maintenance, troubleshooting,
and repairs for at least 2 years.
20. Proceed with testing samples.

4.4 Instrument Set up single tubes and stain with one antibody-fluorochrome for
Optimization each detector on the instrument and add an extra tube for an
Procedure Using unstained sample. Follow your normal surface staining procedure.
Single-Stained Cells A typical 8-color Optimization Panel would include the following
tubes:

Tube 1 Unstained sample

Tube 2 CD4 FITC
Tube 3 CD4 PE
Tube 4 CD4 PerCP cy5.5 or PE-cy5
Tube 5 CD4 PE-cy7 or PC7
Tube 6 CD4 APC
Tube 7 CD4 APC-cy7 or APC-H7
Tube 8 CD4 BV421 or Pacific blue™
Tube 9 CD4 BV510 or Krome™ Orange

1. Start with the PMT settings from the bead QC. Be sure to turn
off the Compensation for this exercise. Compensation will be
set up after the Optimization procedure.
2. Run the unstained tube-1 and adjust forward and side scatter
light signals to display all three WBC populations clearly.
Adjust each PMT setting so that the leading edge of the nega-
tive signal extends just beyond the background or noise, as
performed in the previous method.
276 Bruce Greig

Fig. 28 Optimized Gating setup that demonstrates CD4 gated on Lymphocytes and Monocytes (R1). The bright
population are CD4+ lymphocytes, the intermediate population are monocytes, the negative population are
CD4- lymphocytes and granulocytes. Using this setup is a way to optimize PMT settings to clearly show all
three populations.

3. After running the unstained sample run the first tube of the
stained sample. If using CD4 draw a gate that includes both the
lymphocytes and monocytes. See Fig. 28 : Optimized Gating
setup that includes lymphocytes and monocytes.
4. Next, draw either a single-parameter histogram of the conju-
gate being acquired, that is, CD4 FITC ¼ FL1, etc. or a
two-parameter dot plot of SSC/FL1. Adjust the PMT voltages
for each detector until the three separate populations can be
clearly seen like that of the CD4 histogram and dot plot of the
figure.
5. Continue acquiring each single-color tube adjusting the PMT
voltages to achieve the best separation between the three popu-
lations until each PMT is optimized.
6. Save the now optimized PMT settings. Once the PMT set-
tings are established the instrument is ready to set up the
compensation.

4.5 Compensation 1. Set up Single Stained Controls:

Setup Procedure Using (a) If using the stained-cell method, select a whole blood
Cells or Beads specimen that has a normal WBC and a normal distribu-
tion of lymphocytes.
2. Antibodies and fluorochromes.
(a) Use one for each fluorochrome (CD8 FITC, CD8 PE,
etc.) in the experiment.
(b) Use one for each tandem lot number.
(c) Include an unstained tube that will have sample only.
Below is an example of an eight-color compensation
panel that uses CD8 with a different conjugate in each
tube.
 Quality Control of Immunophenotyping 277

Tube 1 Unstained sample

Tube 2 FITC, Alexa 488, BB 515, or other green wavelength conjugate
Tube 3 PE
Tube 4 PerCP, PE-cy5, or other yellow/orange conjugate
Tube 5 PE-cy7, PC7, or other
Tube 6 APC
Tube 7 APC-cy7, APC-H7™, or other
Tube 8 Pacific blue, V450, BV421, or other
Tube 9 Krome Orange™, V500, BV512, or other

(d) Set up the panel using the same volume/titer of each

antibody that is used in the experiment and follow the
laboratory staining procedure incubation times, RBC-lyse
time (cell method only), wash cycles, and resuspension
media such as PBS buffer.
(e) Follow the procedure for whatever Compensation pro-
gram your instrument is designed to use and after-
wards save the updated calculated compensation settings.
(f) Continue to the Compensation Controls procedure found
in Section 3.3.3.

4.6 Quality Control 1. The Immunophenotyping Laboratory should have a Quality

Documents Assurance Plan—a statement that explains all aspects of quality
and Records control including preanalytic, analytic, and postanalytic proce-
dures, and results of those procedures.
2. Preanalytic Records.
(a) Reagent Usage—a written procedure consistent with the
manufacturer’s instructions OR records of method accu-
racy evaluation if alternative procedures are used such as
for a lab developed test. These records should be kept for
at least 2 years.
(b) Antibody Validation—Antibodies used are validated on
the cell subpopulation of interest in the context of the
antibody combination used in an assay.
(c) New Reagent Lot/Confirmation testing for suitability.
(d) Procedure Validation—written records of any procedure
validation including signature approval. These records
should be kept as long as the procedure and/or instru-
ment stay in use.
278 Bruce Greig

3. Analytic Records.
Instrument Maintenance—records of all maintenance pro-
cedures as well as any maintenance performed by a service
engineer including scheduled maintenance.
Instrument Performance including all daily QC results and
signed monthly summary should be kept for at least 2 years.
4. Postanalytic Records.
(a) Gated dot plots and histograms should be saved either
electronically or as a hard copy. Clinical laboratories are
required to keep these records for at least 10 years.

References
1. CAP all common checklist and CAP flow cyto- 10. Greig B, Stetler-Stevenson M, Lea J (2014)
metry checklist, College of American Patholo- Stabilization media increases recovery in pauci-
gists, 325 Waukegan Rd. Northfield, IL cellular cerebrospinal fluid specimens submit-
60053, www.cap.org ted for flow cytometry testing. Cytometry B
2. Banfi G, Salvagno GL, Lippi G (2007) The role Clin Cytom 86(2):135–138
of ethylenediamine tetraacetic acid (EDTA) as 11. Hulspas R (2010) Titration of fluorochrome-
in vitro anticoagulant for diagnostic purposes. conjugated antibodies for labeling cell surface
Clin Chem Lab Med 45(5):565–576 markers on live cells. Curr Protoc Cytom.
3. Bowen RAR, Remaley AT (2014) Interferences Chapter 6:Unit 6.29. doi: https://doi.org/
from blood collection tube components on 10.1002/0471142956.cy0629s54
clinical chemistry assays. Biochem Med. Pub- 12. Wood BL, Arroz M, Barnett D et al (2007)
lished online 2014 Feb 15. https://doi.org/ 2006 Bethesda international consensus recom-
10.11613/BM.2014.006 mendations on the immunophenotypic analysis
4. Davis BH, Dasgupta A, Kussick S, on behalf of of hematolymphoid neoplasia by flow cytome-
ICSH/ICCS working group et al (2013) Vali- try: optimal reagents and reporting for the flow
dation of cell-based fluorescence assays: prac- cytometric diagnosis of hematopoietic neopla-
tice guidelines from the ICSH and ICCS- part sia. Cytometry B Clin Cytom 72B:S12–S22
II- preanalytical issues. Cytometry B Clin 13. Department of Health and Human Services,
Cytom 84B:286–290 Centers for Medicare and Medicaid Services
5. Heinlein C, Deppert W, Braithwaite A et al (1992) Clinical laboratory improvement
(2010) A rapid and optimization-free proce- amendments of 1988; final rule. Fed Register.
dure allows the in vivo detection of subtle cell 7146 [42CFR493.1256]
cycle and ploidy alterations in tissues by flow 14. Department of Health and Human Services,
cytometry. Cell Cycle 9. https://doi.org/10. Centers for Medicare and Medicaid Services
4161/cc.9.17.12831 (2003) Clinical laboratory improvement
6. Novelli P, Savoia I, Cambieri R et al (2000) amendments of 1988; final rule. Fed Register.
Collagenase digestion and mechanical disag- 7166 [42CFR493.1256(f)]
gregation as a method to extract and immuno- 15. Bagwell CB, Adams EG (1993) Fluorescence
phenotype tumour lymphocytes in cutaneous spectral overlap compensation for any number
T cell lymphomas. Clin Exp Dermatol of flow cytometry parameters. Ann N Y Acad
25:423–431 Sci 677:167–184
7. Becton Dickinson MedimachineTM brochure. 16. D Biosciences, technical bulletin: an introduc-
Becton Dickinson Biosciences, 2350 Qume tion to compensation for multicolor assays on
Dr., San Jose, CA, 95131 digital flow cytometers (2009) https://www.
8. Moriarty AT, Wiersema L, Snyder W et al bdbiosciences.com/documents/Compensa
(1993) Immunophenotyping of cytologic spe- tion_Multicolor_TechBulletin.pdf
cimens by flow cytometry. Diagn Cytopathol 9 17. Maecker HT, Trotter J (2008) Flow cytometry
(3):252–258 controls, instrument setup, and the determina-
9. Caldwell CW (1998) Analyte-specific reagents tion of positivity. Cytometry A 69A
in the flow cytometry laboratory. Arch Pathol (9):1037–1042
Lab Med 122:861–864
 Quality Control of Immunophenotyping 279

18. Landay A et al (1992) Clinical applications of Laboratory Standards Institute, 940 West Val-
flow cytometry: quality assurance and immuno- ley Road, Suite 1400, Wayne, Pennsylvania
phenotyping of peripheral blood lymphocytes, 19087–1898 USA
Document H42-T, Vol. 12, No. 6, pp. 6–8 26. Schmid I, Krall WJ, Uittenbogaart CH et al
19. Keeney M, Gratama JW, Chin-Yee IH et al (1992) Dead cell discrimination with
(1998) Isotype controls in the analysis of lym- 7-aminoactinomycin D in combination with
phocytes and CD34+ stem and progenitor cells dual color immunofluorescence in single laser
by flow cytometry – time to let go! Cytometry flow cytometry. Cytometry 13(2):204–208
343:280–283 27. CAP Policy PP Retention of laboratory records
20. Lindmo T, Fundingsrud K (1981) Measure- and materials. College of American Patholo-
ment of the distribution of time intervals gists, 325 Waukegan Rd. Northfield, IL
between cell passages in flow cytometry as a 60053, www.cap.org
method for the evaluation of sample prepara- 28. CD-Chex Plus™ Immunology Control for
tion procedures. Cytometry 2:151–154 Cluster Designation package insert. Streck.
21. Sphero TM Rainbow calibration particles. 7002 S. 109 Street, La Vista, NE 68128
Spherotech, Inc. 27845 Irma Lee Circle, Lake 29. All Common Checklist, pp 26-29. College of
Forest, IL 60045 American Pathologists, 325 Waukegan Road,
22. BD Biosciences (2014) Multicolor flow cyto- Northfield, IL 60093-2750. Aug 22, 2018
metry: beyond the basics, course workbook. 30. Flow Cytometry Checklist. College of Ameri-
BD Biosciences, San Jose, CA can Pathologists, 325 Waukegan Rd., North-
23. Darzynkiewicz Z, Halicka HD, Zhao H (2010) field, IL, 60093-2750. Aug 22, 2018.
Analysis of cellular DNA content by flow and 31. Sutherland DR, Anderson L, Keeney M, et al.
laser scanning cytometry. Adv Exp Med Biol (1996) The ISHAGE Guidelines for CD34+
676:137–147 cell determination by flow cytometry. Interna-
24. Nunez R (2001) DNA measurement and cell tional Society of Hematotherapy and Graft
cycle analysis by flow cytometry. Curr Issues Engineering. J Hematother. 5(3):213–26
Mol Biol 3(2):67–70 32. Davis
25. Clinical and Laboratory Standards Institute 33. Biolegend Blog. The Stain Index: What is it
(CLSI) (2007) Clinical flow cytometric analysis and what does it tell you. http://www.biole-
of neoplastic hematolymphoid cells; approved gend.com/newsdetail/1245/
guideline—Second Edition. CLSI document
H43-A2 (ISBN 1-56238-635-2). Clinical and
 Chapter 15

Immunophenotyping of Acute Myeloid Leukemia

Pallavi Kanwar Galera, Chunjie Jiang, and Raul Braylan

Abstract
Immunophenotyping by multiparameter flow cytometry is a rapid and efficient technique to simultaneously
assess and correlate multiple individual cell properties like size and internal complexity along with antigen
expression in a population of cells. This method is utilized for rapid characterization of the blasts and
classification of acute myeloid leukemia (AML), in both the peripheral blood (PB) and bone marrow (BM).
This technique is not only useful in the initial diagnosis but also in monitoring and determining prognosis
of the disease through minimal residual disease (MRD) testing. This chapter provides an overview of
procedures for specimen processing, staining, and immunophenotyping of AML and describes the princi-
ples of data analysis for AML classification and MRD testing.

Key words Immunophenotype, Acute myeloid leukemia, Myelodysplastic syndrome, Minimal resid-
ual disease, Flow cytometry

1 Introduction

AML is a clonal myeloid precursor neoplasm characterized by an

increased number of blasts or blast equivalents in the BM and
PB. Myeloid blasts include myeloblasts, monoblasts, erythroblasts,
megakaryoblasts, and blast equivalents include promyelocytes and
promonocytes. AML is currently being subdivided into six cate-
gories by the 2017 Revision of WHO classification [1], including
AML with recurrent genetic abnormalities; AML with
myelodysplasia-related changes (AML-MRC); therapy-related
myeloid neoplasms; AML, not otherwise specified (AML, NOS);
myeloid sarcoma and myeloid proliferations associated with Down
syndrome. These categories are based on morphological, immuno-
phenotypic, genetic, and clinical data with immunophenotype of
the blasts being an important component. Historically, AML was
defined by the presence of greater than 20% blasts or blast equiva-
lents in a BM or PB specimen; however, this prerequisite is not

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_15, © Springer Science+Business Media, LLC, part of Springer Nature 2019

281
282 Pallavi Kanwar Galera et al.

required anymore for some of the genetically defined AML cate-

gories such as cases harboring the t(15;17), t(8;21), or inv(16).
Flow cytometric immunophenotyping can be utilized for rapid
identification and characterization of the blasts and blasts equiva-
lents, in both the PB and BM [2, 3] and, at the same time, it allows
us to assess the maturation patterns of the various BM elements,
further refining the classification of AML [4, 5]. With adequate
population selection (gating) strategies and appropriate antibody
panels, flow cytometry can also be very useful in the enumeration of
blasts, especially in PB. However, while flow cytometry is an excel-
lent means for detection and characterization of blasts, it may not
be as accurate in quantitating blasts in BM specimens primarily due
to uncontrolled hemodilution [6], leading to underestimation of
blast numbers and making morphology/immunohistochemistry
still the gold standard for blast counting in BM samples. Other
confounding factors like loss of nucleated red blood cells (RBC)
during lysis and processing, may artificially increase the blast per-
centage [7, 8]. These issues do not apply to blast counts in PB.
Flow cytometry is also of paramount importance in monitoring
disease and prognosticating it through minimal residual disease
(MRD) testing [9–11]. MRD is the presence of leukemic blasts at
levels lower than the limit of detection by conventional morpho-
logic evaluation. Multiple studies have demonstrated that detection
of leukemic blasts following induction and consolidation therapy
identifies patients who are at a higher risk of relapse and thus a
worse outcome [12–15]. Due to its affordability, wide spread avail-
ability and quick turnaround time, flow cytometry is extensively
used for MRD testing [16–18].
This chapter outlines the protocol used in our laboratory,
providing an approach for diagnosing, and classifying AML using
a 10-parameter (eight-color) measuring flow cytometer (FACS-
Canto II) with FACSDiva data acquisition software (both from
Becton Dickinson (BD)). It also describes the various strategies
utilized for data analysis using FCS Express software (De Novo).
It should be emphasized that there are plethora of approaches and
guidelines that can be utilized for these purposes [19, 20], and
other laboratories may use different procedures to obtain equiva-
lent results.

2 Materials

2.1 Reagents 1. Phosphate buffered saline (PBS) (KD Medical)/bovine serum

albumin (BSA) (EMD Millipore Corp.) buffer: 500 mL 1
PBS and 0.5 g bovine serum albumin, fraction V (fatty acid
 Immunophenotyping of Acute Myeloid Leukemia 283

free). Mix well and filter through a 45 μm Millipore filter. Store

at 4  C for up to 1 month.
2. FACSLyse Buffer (BD): Prepare a 1:10 dilution of the FAC-
SLyse buffer using deionized water. Store at room temperature
in a glass bottle for up to 1 month.
3. Stabilizing Fixative (BD): Prepare a 1:3 dilution of the Stabiliz-
ing Fixative using deionized water. The solution can be kept for
1 month when stored in a polystyrene container at room
temperature.
4. Erythrocyte Lysis Buffer (Qiagen).
5. Fix Perm Reagent A and B (Caltag).
Precautions
All PB and BM products should be considered biohazardous
and treated with proper precautions as if they were capable of
transmitting infection. The Stabilizing Fixative contains 3% para-
formaldehyde. Appropriate disposal methods should be used.

2.2 Monoclonal Table 1 describes the antibodies utilized in our laboratory for
Antibody Panels detection, characterization, and classification of AML with appro-
priate corresponding fluorochromes (see Note 1).

2.3 Specimens 1. PB (anticoagulated in EDTA or Heparin).

2. BM Aspirate (anticoagulated in EDTA or Heparin).

Table 1
Antibodies and fluorochromes utilized in our laboratory*

APC- Alexa Pacific

FITC PE PerCP-Cy5.5 PE-Cy7 APC 750 APC-H7 Blue V500
CD16 CD10 CD5 CD19 CD10 CD34 CD14 HLA-DR CD45
CD42b + CD61 CD13 CD11c CD56 CD56 CD38 CD2
CD36 CD19 CD117 CD34 CD11b CD71 CD4
cMPO# CD64 CD34 CD15 CD7 CD3 CD19
CD5 CD33 CD123 cCD3#
cCD79a# CD3 cTDT#
Abbreviations: APC Allophycocyanin, APC-Alexa 750 Allophycocyanin Alexa Fluor 750, APC-H7 Allophycocyanin H7,
FITC Fluorescein isothiocyanate, MPO Myeloperoxidase, PB Pacific blue, PE Phycoerythrin, PE-Cy7 (PC7) Phycoery-
thrin cyanine-7, PerCP-Cy5.5 Peridinin chlorophyll protein cyanine 5.5, TDT Terminal deoxynucleotidyl transferase,
V500 BD Horizon dye V500
*
Other laboratories may utilize different antibody and fluorochrome combinations
#
c: Intracytoplasmic antibodies
284 Pallavi Kanwar Galera et al.

3 Methods

3.1 Labeling It is also critical to appropriately titrate the amount of antibody to

with Monoclonal get an optimal antibody dilution for maximum signal-to-noise ratio
Antibodies Against [21, 22].
Surface Antigens 1. Label 12  75 mm disposable polystyrene tubes and add the
antibodies.
2. Wash the PB or BM sample at least twice as follows: To wash
the sample, in a 15 mL disposable conical tube add at least
2 mL of PBS/BSA to the sample and spin at 311  g for 8 min.
Aspirate the supernatant and repeat again by adding at least
2 mL of PBS/BSA, mixing, and spinning at 311  g for 8 min.
Aspirate the supernatant. Resuspend by adding PBS/BSA to
the sample to make a final white cell concentration of
5000–10,000 cells/μL.
3. Add 100 μL of cell suspension (white cell concentration
5000–10,000 cells/μL) to each 12  75 mm disposable poly-
styrene tube with antibodies.
4. Incubate the tubes that contain the antibodies and the sample for
15 min to 1 h at room temperature in the dark (see Note 2).
5. Add 2 mL 1 FACSLyse Buffer to each tube to lyse the red
blood cells. Vortex.
6. Incubate for 10 min at room temperature in the dark.
7. Spin at 216  g for 5 min.
8. Decant supernatant. Mix.
9. Add 2 mL PBS/BSA to each tube.
10. Repeat steps 7–9 (see Note 3).
11. Add 400 μL 1 Stabilizing Fixative to each tube (see Note 4).
12. The cells are now ready for analysis on the flow cytometer.

3.2 Labeling This method allows the detection of cytoplasmic antigens, along
with Monoclonal with the assessment of expression of surface antigens. It utilizes a
Antibodies Against cross-linking fixative to stabilize cellular contents prior to labeling
Combined Surface with antibodies against cytoplasmic antigens. The procedure outlined
and Intracellular here utilizes Erythrocyte Lysis Buffer (Qiagen) and Fix Perm Reagent
Antigens A and B (Caltag), but other products may be used. It is also critical
to appropriately titrate the amount of antibody to get an optimal
antibody dilution for maximum signal-to-noise ratio [21, 22].
1. Lyse RBCs from sample using Erythrocyte Lysis Buffer: To lyse
the RBCs add 2 mL of the Erythrocyte Lysis Buffer to 200 μL
of cell suspension in a 12  75 mm disposable polystyrene
tube. Allow to lyse for approximately 10 min. Spin at 216  g
for 5 min. Decant supernatant.
 Immunophenotyping of Acute Myeloid Leukemia 285

2. Wash the sample: To wash the sample add 2 mL of PBS/BSA

and then spin at 216  g for 5 min. Decant supernatant.
3. Add appropriate antibodies to the 12  75 mm disposable
polystyrene tube containing the cells for cell surface staining.
4. Incubate for 15 min.
5. Wash the sample: To wash the sample add 2 mL of PBS/BSA
and then spin at 216  g for 5 min. Decant supernatant.
6. Add 100 μL of Fix Perm Reagent A.
7. Incubate for 10 min.
8. Wash the sample: To wash the sample add 2 mL of PBS/BSA
and then spin at 216  g for 5 min. Decant supernatant.
9. Add 100 μL of Fix Perm Reagent B.
10. Add intracellular antibodies.
11. Incubate for 15 min.
12. Wash the sample: To wash the sample add 2 mL of PBS/BSA
and then spin at 216  g for 5 min. Decant supernatant.
13. Resuspend in 400 μL of Stabilizing Fixative.

3.3 Instrument A variety of instruments are available and different laboratories may
Settings use different instruments to perform flow cytometric analysis. Set-
tings and calibrations are instrument-dependent, and manufac-
turer’s guidelines should be employed. The number of cells
acquired for analysis per tube may vary depending on the level of
detection sensitivity desired. In our laboratory we acquire at least
100,000 cells/tube for routine analysis; however, higher numbers
are necessary for MRD testing.

3.4 Data Analysis The initial evaluation of results consists of an assessment of the
quality of data and viability of the specimen. One strategy to
3.4.1 Data Quality
evaluate quality and consistency of the acquired data is to observe
and Viability
data over time to identify air aspiration artifacts due to sample
exhaustion and fluidic instabilities (see Note 5) (Fig. 1). The second
step in ensuring the quality of data collected is to make certain that
the data analyzed is an actual measure of a single cell at a time.
Frequently, high cell concentration, high flow rates or nonspecific
antibody bindings can induce cell aggregates and clump forma-
tions. To analyze exclusively single cells, a combination of forward
light scatter area (FSC-A) and forward light scatter width (FSC-W)
and/or height (FSC-H) is utilized to exclude cell doublets and
aggregates from the data (Fig. 1).
Finally, to evaluate the viability of the specimen and to exclude
debris and degenerated cells, viability dyes (propidium iodide,
7AAD) can be utilized. Alternatively, for practical purposes, nonvi-
able cells may be excluded utilizing light scatter characteristics
[19, 23, 24].
286 Pallavi Kanwar Galera et al.

Fig. 1 Normal bone marrow. Data quality: Data is observed over time to identify and exclude fluidic instabilities
such as air aspiration artifacts (upper left graph). A combination of forward light scatter area (FSC-A) and
forward light scatter width (FSC-W) (lower left graph), and/or forward light scatter height (FSC-H) (lower
middle graph) is utilized to exclude doublets and to ensure that the data analyzed is a measure of a single cell
at a time. Initial gating (right graph): Based on the side light scatter (SSC-A) and CD45 expression the marrow
populations can be broadly selected as red blood cells (brown), granulocytes (green), lymphocytes (blue),
monocytes (aqua), and cells in the blast gate (red)

3.4.2 Initial Cell The initial gating, after excluding all elements that are not single
Population Gating viable cells, is performed to define broad categories of cell popula-
tions. This is done by examining side light scatter (SSC-A) and
CD45 signals (see Note 6) [25]. Using this combination, the
following populations can be broadly identified: erythroid (low
SSC-A and negative CD45), granulocytes (moderate to high
SSC-A and moderate CD45), lymphocytes (low SSC-A and bright
CD45), monocytes (intermediate SSC-A between lymphocytes and
granulocytes and bright CD45) and blasts (low SSC-A and dim
CD45) (Fig. 1) [26]. Though an effective initial gate, the “blast
gate” based on SSC-A and CD45 may be inaccurate [27] since
there may be other populations that fall in this gate, namely baso-
phils, hypogranular myeloid cells, plasmacytoid dendritic cells, and
immature monocytes. Also, there would be instances where the
blasts may have a brighter or dimmer CD45 and a varied SSC-A.
Other populations may also, to a certain extent, be relatively
ill-defined using this initial gating approach. Therefore, a panel of
additional markers is required to further delineate the blasts as well
as other distinct cell populations.
 Immunophenotyping of Acute Myeloid Leukemia 287

Fig. 2 Bone marrow with increased, phenotypically abnormal myeloblasts, comprising 23% of total cells. The
blasts (red dots) express CD45, CD34, CD13, CD117, CD33 (bright), CD38 (dim/negative), HLA-DR, CD64
(dim/negative), and CD123 (dim). Blasts are negative for CD14, CD15, CD16, CD11b, CD7, and CD2

3.4.3 Strategies for Blast To assess the presence of myeloblasts, commonly used antigens
Identification include CD34 (Fig. 2) and CD117 since blasts usually express
these markers. However, some myeloblasts may be negative for
CD34 and/or CD117 and not even fall within the typical location
in the CD45 and SSC-A graphs, making their recognition difficult.
In these instances, other reagents and strategies are necessary for
their detection. For example, blasts in acute promyelocytic leuke-
mia are CD34 dim or negative and may mimic normal granulocytes
but they usually have distinct phenotypic features [28]. Also,
monoblasts usually express no or dim CD34, requiring the use of
reagents for monocyte-associated markers for their recognition.

3.4.4 Strategies Myeloid lineage markers commonly used include CD13, CD15,
for Lineage Assignment CD33, MPO, and CD16 along with markers expressed during
monocytic differentiation such as CD11b, CD64, CD14, and
CD4 (Fig. 2). Erythroid precursors express CD71, CD105,
CD117, Glycophorin A (CD 235a), and CD36, and megakaryo-
cytic precursors express platelet-associated glycoproteins such as
CD61 and CD42b. Assessment of lymphoid markers (CD3,
CD2, CD4, CD5, CD7, CD19, CD56, CD22, and CD79a)
should be included at the time of initial diagnosis, to assign lineage
or to look for aberrant antigen expression. Non-lineage-specific
markers (CD38, CD56, CD123, and HLA-DR) are also helpful
to further delineate blasts and aberrancies.
288 Pallavi Kanwar Galera et al.

Fig. 3 Bone marrow involved by AML with FLT3 mutation and phenotypically abnormal myeloblasts comprising
95% of total cells. The blasts (red dots) express CD45, CD34 (dim to negative), CD13, CD117 (variable), CD64
(variable), CD38 (partial), HLA-DR, CD123, CD33, CD7 (partial aberrant). Blasts are negative for CD16, CD14,
and other T-cell and B-cell markers (not shown)

Immunophenotypic aberrancies include over or under expres-

sion of antigens, or absence of antigen expression (e.g., dim or
absent CD38, CD13, or CD33 expression on myeloblasts), asyn-
chronous antigen expression (expression of markers of mature ele-
ments in early precursors such as CD11b, CD64, or CD15 in
myeloblasts), abnormally homogenous expression of certain mar-
kers (e.g., of CD38 or CD33) and expression of nonmyeloid anti-
gens on myeloblasts (CD2, CD7, CD56, and CD19) (Figs. 3, 4,
and 5) [7, 29].

3.4.5 Background In acute myeloid leukemias, it may be useful to evaluate the back-
Myelopoiesis ground myelopoiesis to determine the presence of possible dysplas-
tic changes. For this purpose, knowledge of normal maturation
patterns utilizing various combinations of antigens is essential.
Patterns generated by CD45 and SSC-A (SSC-A reflecting granu-
larity), CD13 and CD16, CD11b and CD64, and CD11b and
CD16 have been well studied for assessment of myeloid maturation
and are reproducible [4, 5] (Fig. 5). Alterations of these patterns
and other aberrancies like loss of CD10 on mature granulocytes, or
expression of CD56 on myeloid and monocytic cells are helpful
clues into probable underlying myelodysplasia [7, 19, 30].

3.4.6 AML Immunophenotyping can be helpful as an adjunct in classifying

with Recurrent Genetic AML with recurrent genetic abnormalities; however, cytogenetic
Abnormalities, Special and molecular studies remain the mainstay in the diagnostic algo-
Considerations rithm. Although the association between the immunophenotype
and the recurrent genetic abnormality is not absolute, a few of the
AML with recurrent genetic abnormalities show a strong correla-
tion with the immunophenotype including AML with t(8;21)(q22;
 Immunophenotyping of Acute Myeloid Leukemia 289

Fig. 4 A patient with history of autoimmune lymphoproliferative syndrome with FAS mutation. Bone marrow
aspirate shows two populations of myeloblasts (red dots) based on CD34 expression with the minority of the
blasts (approximately 25%) expressing CD34 and the majority of the blasts are negative for CD34. The blasts
also express CD45, CD117, CD13, HLA-DR (bright), CD56 (aberrant), and CD33. A small subset with dim
expression of CD19 is noted. The blasts are negative for CD7, CD14, and CD64. They were also negative for
CD10, CD20, CD3, and CD5 (not shown)

Fig. 5 A patient with history of BRCA1 and BRCA2 mutations and therapy for breast cancer 7 years prior to
presentation. Bone marrow aspirate shows abnormal myeloblasts (red dots), comprising 13% of total cells.
The blasts express CD45, CD34, CD13, CD33, CD117 (dim to negative), HLA-DR, CD7 (partial & aberrant) and
are negative for CD38 (aberrant), CD2, CD16, CD14, CD64, and CD123. Granulocytes (green dots) reveal
disrupted maturation pattern based on the expression of CD13 and CD16

q22.1), acute promyelocytic leukemia, t(15;17)(q22;q11-12), and

AML with t(9;11)(p21.3;q23.3). Some salient immunophenotypic
features of AML with recurrent genetic abnormalities are described
below:
290 Pallavi Kanwar Galera et al.

1. AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1: Blasts

usually show high expression of CD34, HLA-DR, MPO, and
CD13, and weak expression of CD33. Neutrophilic differenti-
ation and maturation as assessed by expression of CD15
and/or CD65, maturation asynchrony (coexpression of
CD34 and CD15) and frequent coexpression of B-cell markers
(CD19, and cytoplasmic CD79a) [31, 32] are noted (see Note
7). CD56 expression is seen in a subset of cases, and this may
portend an adverse prognosis [33, 34]. The adverse prognosis
could be secondary to the fact that cases with KIT mutations
have higher CD56 expression [35].
2. AML with inv(16)(p13.1q22) or t(16;16)(p13.1;q22); CBFB-
MYH11: Blasts express CD34 and CD117 (KIT) intensely.
Blast populations differentiating toward granulocytic lineage
(positive for CD13, CD33, CD15, CD65, and MPO) and the
monocytic lineage (positive for CD14, CD4, CD11b, CD11c,
CD64, and CD36) are seen. Maturation asynchrony (coexpres-
sion of CD34 and CD15) and coexpression of CD2 with
myeloid markers are frequently noted [1, 36].
3. Acute promyelocytic leukemia (APL) with t(15;17)(q22;q11-
12); PML-RARA: Abnormal promyelocytes (blast equivalents)
demonstrate relative high SSC-A, homogeneous bright expres-
sion of CD33, heterogeneous expression of CD13, CD117
and MPO, and low or absent expression of HLA-DR, CD34,
CD11a, and CD11b. CD64 expression is common. In the
microgranular variant there is frequent expression of CD34
and CD2 by at least some cells [37]. CD2 expression in APL
has been associated with FLT3-ITD [38]. CD56 expression is
seen in a subset of cases and has been associated with a worse
outcome [39].
4. AML with t(9;11)(p21.3;q23.3); KMT2A-MLLT3: Most
adult AML cases reveal monocytic differentiation (CD14,
CD4, CD11b, CD11c, CD64, and CD36,) with a variable
expression of markers of immaturity (CD34, CD117), and
CD56 [40] (Fig. 6). In children, there is a high expression of
CD33, CD65, CD4, and HLADR, with a lower expression of
CD13, CD34, and CD14 [41].
5. AML with t(6;9)(p23;q34.1); DEK-NUP214: The blasts have
a nonspecific myeloid immunophenotype, with expression of
MPO, CD9, CD13, CD33, CD38, CD123, and HLA-DR.
Most of the cases show an expression of CD34, CD117, and
CD15. A subset of cases is TdT positive. Associated basophilia
is often present and can be detected with basophils uniformly
expressing CD123, CD33, and CD38 but no HLA-DR
[42–45].
 Immunophenotyping of Acute Myeloid Leukemia 291

Fig. 6 AML with t(9;11)(p21.3;q23.3); KMT2A-MLLT3. Bone marrow aspirate reveals over 90% blasts, with
two different immunophenotypically abnormal populations: The first population is of immature myeloblasts
(red dots), representing approximately 81% of total cells and expressing CD45(dim), CD34 (partial, 72% of
total cells), CD117, CD13, HLA-DR, CD33 (bright), CD123, and CD64 (dim to negative). These blasts are
negative for CD14, CD16, and CD11b. They are also negative for CD15, CD61/CD42b, T-cell and B-cell
markers (not shown). The second population is of monoblasts (aqua dots), representing approximately 10% of
total cells with intermediate SSC-A and expressing CD45 (brighter than first population), CD13, HLA-DR
(bright), CD38, CD33 (bright), CD64 (bright), CD11b (variable), CD36 (not shown). These are negative for CD34,
CD117, CD16, and CD14

6. AML with inv(3)(q21.3q26.2) or t(3;3)(q21.3;q26.2);

GATA2, MECOM: The blasts express CD34, CD33, CD13,
CD117, HLA-DR, and CD38 (commonly), and a frequent
aberrant CD7. A subset of cases may express megakaryocytic
markers such as CD41 and CD61 [46, 47].
7. Acute megakaryoblastic leukemia with t(1;22)(p13.3;q13.1);
RBM15-MKL1: The blasts express one or more of the platelet
glycoproteins: CD41 (glycoprotein IIb/IIIa), CD61 (glyco-
protein IIIa), and CD42b (glycoprotein Ib) (see Note 8) along
with variable expression of myeloid markers CD13 and CD33.
They also express CD36. CD34, CD45, and HLA-DR are
often negative [48–50].
8. AML with mutated NPM1: The blasts show high expression of
CD33 and variable (often low) CD13 expression along with
CD117 and CD123 [51]. CD34 and HLA-DR are often neg-
ative (see Note 9). The blasts can either have an immature
myeloid or a monocytic immunophenotype with expression
of CD36, CD64, and CD14 [52]. A blast population with
expression of CD34, CD25, CD123, and CD99 is associated
with FLT3-ITD mutations [53].
292 Pallavi Kanwar Galera et al.

9. AML with biallelic mutation of CEBPA: The blasts usually

express HLA-DR and CD34. Expression of CD7 is also com-
monly seen. The cases with biallelic mutation are reported to
have a higher frequency of HLA-DR, CD7, and CD15 expres-
sion and a lower frequency of CD56 expression in comparison
to cases with a single mutation [54, 55]. Monocytic markers
such as CD14 and CD64 are usually absent.
10. AML with mutated RUNX1: The blasts usually express CD13,
CD34, and HLA-DR, with variable expression of CD33,
monocytic markers, and MPO [56, 57].

3.4.7 Minimal Residual The detection limit of leukemic blasts by flow cytometry ranges
Disease from 0.1% to 0.01% of leukocytes. The detection limit is dependent
on various factors, including the number of cells collected, anti-
body panel, flow cytometer, and distinct immunophenotypic prop-
erties of the leukemic blasts. For a detection sensitivity of 0.01%, at
least 1,000,000 cells should be analyzed.
Two strategies are commonly employed in MRD testing for
AML by flow cytometry. The first strategy is identifying a combina-
tion of aberrant markers and their expression levels on leukemic
blasts that are usually absent in normal hematopoietic precursors.
This strategy is widely known as “leukemia-associated immunophe-
notype” or LAIP. At diagnosis a broad, extensive antigen panel is
utilized to define the LAIP. Subsequently after treatment, the
selected antibody panel is used to identify blasts with pretreatment
LAIP. The advantages of this approach are that the data analysis is
relatively simple, and sensitivity for each LAIP (defined based on the
background expression in nonleukemic specimens) is known. The
disadvantages include the requirement to know the complete phe-
notype of the pretreatment leukemic sample, the need for custo-
mized antibody panels for subsequent evaluations after the
evaluation of the initial diagnostic specimen and, finally, the immu-
nophenotype of the leukemic blasts might change after treatment
[3, 9, 11, 13, 20, 23].
The second strategy known as “difference from normal” iden-
tifies leukemic blasts based on immunophenotypic differences
between the leukemic blasts and normal hematopoietic precursors.
There is no requirement of defining or having prior knowledge of
LAIPs at diagnosis (which is a hindrance in many tertiary or referral
centers that do not have this information available); the same
standardized panel can be used for all evaluations and since this
method is based on differences between blasts and normal myeloid
precursors, immunophenotypic shifts and change of LAIPs after
treatment do not interfere with the assessment. However, extensive
knowledge of antigen expression patterns of normal hematopoietic
precursors and their maturation patterns is essential for an accurate
evaluation [3, 9, 11, 20, 23]. In practice, a combination of the two
strategies is often utilized.
 Immunophenotyping of Acute Myeloid Leukemia 293

4 Notes

1. A variety of antibodies with a variety of fluorochrome conju-

gates are available by various manufacturers. The antibody and
fluorochrome selections and combinations may vary depending
upon the instrument, number of detectors and the approach of
the particular lab. Additional antibodies that are not frequently
used in our laboratory such as CD99, CD25, and CD41 can
also be added to aid in the diagnosis of AML [46–50, 53].
2. Prolonged fixation and exposure to fluorescent light may
degrade fluorochromes. To minimize fluorochrome degrada-
tion, labeling of the specimen with antibodies prior to red cell
lysis along with limited fixation and decreased light exposure is
recommended.
3. Addition of the wash step reduces background fluorescence
due to unbound antibody.
4. The use of a fixative helps in better preservation of the light
scatter properties of the cells and provides a better separation
between viable and nonviable cells.
5. In case of air bubble artifact secondary to aspiration of air due
to sample exhaustion, the error can be corrected by simply
excluding the data during those time periods.
6. The advantage of using SSC-A and CD45 for initial gating is
that by just adding a single common antibody across all tubes
the various populations can be compared with ease across all
tubes.
7. AML with t(8;21) can be misclassified as mixed phenotype
acute leukemia due to frequent expression of B-cell markers
(CD19, and cytoplasmic CD79a) on the blasts. Presence of
CD56 and CD19 positive myeloid blasts should raise a possi-
bility of AML with t(8;21). Cases of acute monoblastic leuke-
mia, especially with t(8;16), may show a higher expression of
CD14 [58].
8. In AML (megakaryoblastic) with t(1;22) cytoplasmic expres-
sion of CD41 or CD61 is more specific and sensitive than
surface staining.
9. Cases of AML with NPM1 mutations are often negative for
CD34 and HLA-DR; similar to blasts equivalents seen in APL;
however, they have a lower SSC-A and fall in the blast gate.
CD34 positive cases do occur and have been associated with an
adverse prognosis [59, 60].
294 Pallavi Kanwar Galera et al.

Acknowledgments

This research was supported by the Intramural Research Program

of the National Institutes of Health.

References

1. Swerdlow SH, Campo E, Harris NL et al (eds) 12. Dohner H, Estey E, Grimwade D et al (2017)
(2017) WHO classification of tumours of hae- Diagnosis and management of AML in adults:
matopoietic and lymphoid tissues, 4th edn. 2017 ELN recommendations from an interna-
IARC, Lyon tional expert panel. Blood 129(4):424–447
2. Craig FE, Foon KA (2008) Flow cytometric 13. Terwijn M, van Putten WL, Kelder A et al
immunophenotyping for hematologic neo- (2013) High prognostic impact of flow cyto-
plasms. Blood 111:3941–3967 metric minimal residual disease detection in
3. Chen X, Cherian S (2017) Acute myeloid leu- acute myeloid leukemia: data from the
kemia immunophenotyping by flow cytometric HOVON/SAKK AML 42A study. J Clin
analysis. Clin Lab Med 37:753–769 Oncol 31(31):3889–3897
4. van Lochem EG, van der Velden VH, Wind HK 14. Venditti A, Buccisano F, Del Poeta G et al
et al (2004) Immunophenotypic differentia- (2000) Level of minimal residual disease after
tion patterns of normal hematopoiesis in consolidation therapy predicts outcome in
human bone marrow: reference patterns for acute myeloid leukemia. Blood 96
age-related changes and disease-induced shifts. (12):3948–3952
Cytometry B Clin Cytom 60(1):1–13 15. Zhu HH, Zhang XH, Qin YZ et al (2013)
5. Gorczyca W, Sun ZY, Cronin W et al (2011) MRD-directed risk stratification treatment
Immunophenotypic pattern of myeloid popu- may improve outcomes of t(8;21) AML in the
lations by flow cytometry analysis. Methods first complete remission: results from the
Cell Biol 103:221–266 AML05 multicenter trial. Blood 121
6. Loken MR, Chu SC, Fritschle W (2009) Nor- (20):4056–4062
malization of bone marrow aspirates for hemo- 16. Freeman SD, Virgo P, Couzens S et al (2013)
dilution in flow cytometric analyses. Cytometry Prognostic relevance of treatment response
B Clin Cytom 76(1):27–36 measured by flow cytometric residual disease
7. Wood BL (2007) Myeloid malignancies: mye- detection in older patients with acute myeloid
lodysplastic syndromes, myeloproliferative dis- leukemia. J Clin Oncol 31(32):4123–4131
orders, and acute myeloid leukemia. Clin Lab 17. Walter RB, Gooley TA, Wood BL et al (2011)
Med 27(3):551–575. vii Impact of pretransplantation minimal residual
8. Wood BL, Arroz M, Barnett D et al (2007) disease, as detected by multiparametric flow
2006 Bethesda international consensus recom- cytometry, on outcome of myeloablative hema-
mendations on the immunophenotypic analysis topoietic cell transplantation for acute myeloid
of hematolymphoid neoplasia by flow cytome- leukemia. J Clin Oncol 29(9):1190–1197
try: optimal reagents and reporting for the flow 18. Getta BM, Devlin SM, Levine RL et al (2017)
cytometric diagnosis of hematopoietic neopla- Multicolor flow cytometry and multigene next-
sia. Cytometry B Clin Cytom 72(Suppl 1): generation sequencing are complementary and
S14–S22 highly predictive for relapse in acute myeloid
9. Wood BL (2016) Principles of minimal residual leukemia after allogeneic transplantation. Biol
disease detection for hematopoietic neoplasms Blood Marrow Transplant 23(7):1064–1071
by flow cytometry. Cytometry B Clin Cytom 19. van Dongen JJ, Lhermitte L, Böttcher S, Euro-
90(1):47–53 Flow Consortium (EU-FP6, LSHB-CT-2006-
10. Xu J, Jorgensen JL, Wang SA (2017) How do 018708) et al (2012) EuroFlow antibody
we use multicolor flow cytometry to detect panels for standardized n-dimensional flow
minimal residual disease in acute myeloid leu- cytometric immunophenotyping of normal,
kemia? Clin Lab Med 37:787–802 reactive and malignant leukocytes. Leukemia
26(9):1908–1975
11. Zhou Y, Wood BL (2017) Methods of detec-
tion of measurable residual disease in AML. 20. Schuurhuis GJ, Heuser M, Freeman S et al
Curr Hematol Malig Rep 12:557–567 (2018) Minimal/measurable residual disease
in AML: a consensus document from the
 Immunophenotyping of Acute Myeloid Leukemia 295

European LeukemiaNet MRD working party. expression of immature B-cell antigen CD19
Blood 131(12):1275–1291 together with stem cell antigen CD34. Blood
21. Reilly JT (1996) Use and evaluation of leuko- 80:470–477
cyte monoclonal antibodies in the diagnostic 33. Baer MR, Stewart CC, Lawrence D et al
laboratory: a review. Clin Lab Haematol (1997) Expression of the neural cell adhesion
18:1–5 molecule CD56 is associated with short remis-
22. Owens MA, Loken MR (1995) Quality control sion duration and survival in acute myeloid
of flow cytometer and reagents. In: Owens leukemia with t(8;21)(q22;q22). Blood
MA, Loken MR (eds) Flow cytometry: princi- 90:1643–1648
ples for clinical laboratory practice. Quality 34. Iriyama N, Hatta Y, Takeuchi J et al (2013)
assurance for quantitative immunophenotyp- CD56 expression is an independent prognostic
ing. Wiley, New York, pp 45–72 factor for relapse in acute myeloid leukemia
23. Wood BL (2013) Flow cytometric monitoring with t(8;21). Leuk Res 37:1021–1026
of residual disease in acute leukemia. Methods 35. De J, Zanjani R, Hibbard M et al (2007)
Mol Biol 999:123–136 Immunophenotypic profile predictive of KIT
24. Chen W, Luu HS (2017) Immunophenotyping activating mutations in AML1-ETO leukemia.
by multiparameter flow cytometry. Methods Am J Clin Pathol 128:550–557
Mol Biol 1633:51–73 36. Ouyang J, Goswami M, Peng J et al (2016)
25. Stelzer GT, Shults KE, Loken MR (1993) Comparison of multiparameter flow cytometry
CD45 gating for routine flow cytometric anal- immunophenotypic analysis and quantitative
ysis of human bone marrow specimens. Ann N RT-PCR for the detection of minimal residual
Y Acad Sci 677:265–280 disease of core binding factor acute myeloid
26. Borowitz MJ, Guenther KL, Shults KE et al leukemia. Am J Clin Pathol 145(6):769–777
(1993) Immunophenotyping of acute leuke- 37. Exner M, Thalhammer R, Kapiotis S et al
mia by flow cytometric analysis. Use of CD45 (2000) The “typical” immunophenotype of
and right-angle light scatter to gate on leuke- acute promyelocytic leukemia (APL-M3):
mic blasts in three-color analysis. Am J Clin does it prove true for the M3-variant? Cytome-
Pathol 100:534–540 try 42:106–109
27. Harrington AM, Olteanu H, Kroft SH (2012) 38. Takenokuchi M, Kawano S, Nakamachi Y et al
A dissection of the CD45/side scatter “blast (2012) FLT3/ITD associated with an imma-
gate”. Am J Clin Pathol 137:800–804 ture immunophenotype in PML-RARα leuke-
28. Orfao A, Chillón MC, Bortoluci AM et al mia. Hematol Rep 4(4):e22
(1999) The flow cytometric pattern of CD34, 39. Montesinos P, Rayón C, Vellenga E et al
CD15 and CD13 expression in acute myelo- (2011) Clinical significance of CD56 expres-
blastic leukemia is highly characteristic of the sion in patients with acute promyelocytic leu-
presence of PML-RARalpha gene rearrange- kemia treated with all-trans retinoic acid and
ments. Haematologica 84(5):405–412 anthracycline-based regimens. Blood
29. Kussick SJ, Wood BL (2003) Using 4-color 117:1799–1805
flow cytometry to identify abnormal myeloid 40. Muñoz L, Nomdedéu JF, Villamor N et al
populations. Arch Pathol Lab Med (2003) Acute myeloid leukemia with MLL
127:1140–1147 rearrangements: clinicobiological features,
30. van de Loosdrecht AA, Alhan C, Béné MC et al prognostic impact and value of flow cytometry
(2009) Standardization of flow cytometry in in the detection of residual leukemic cells. Leu-
myelodysplastic syndromes: report from the kemia 17:76–82
first European LeukemiaNet working confer- 41. Creutzig U, Harbott J, Sperling C et al (1995)
ence on flow cytometry in myelodysplastic syn- Clinical significance of surface antigen expres-
dromes. Haematologica 94(8):1124–1134 sion in children with acute myeloid leukemia:
31. Hurwitz CA, Raimondi SC, Head D et al results of study AML-BFM-87. Blood
(1992) Distinctive immunophenotypic fea- 86:3097–3108
tures of t(8;21)(q22;q22) acute myeloblastic 42. Alsabeh R, Brynes RK, Slovak ML et al (1997)
leukemia in children. Blood 80:3182–3188 Acute myeloid leukemia with t(6;9) (p23;q34):
32. Kita K, Nakase K, Miwa H et al (1992) Pheno- association with myelodysplasia, basophilia,
typical characteristics of acute myelocytic leu- and initial CD34 negative immunophenotype.
kemia associated with the t(8;21)(q22;q22) Am J Clin Pathol 107:430–437
chromosomal abnormality: frequent
296 Pallavi Kanwar Galera et al.

43. Lai YY, Li Y, Shi Y et al (2012) Characteristics 52. Liu YR, Zhu HH, Ruan GR et al (2013)
of 11 patients with acute myeloid leukemia NPM1-mutated acute myeloid leukemia of
accompanied with karyotype aberration t monocytic or myeloid origin exhibit distinct
(6;9). Zhongguo Shi Yan Xue Ye Xue Za Zhi immunophenotypes. Leuk Res 37:737–741
20:1293–1296 53. Angelini DF, Ottone T, Guerrera G et al
44. Oyarzo MP, Lin P, Glassman A et al (2004) (2015) A leukemia-associated CD34/
Acute myeloid leukemia with t(6;9)(p23;q34) CD123/CD25/CD99+ immunophenotype
is associated with dysplasia and a high fre- identifies FLT3-mutated clones in acute mye-
quency of flt3 gene mutations. Am J Clin loid leukemia. Clin Cancer Res 21:3977–3985
Pathol 122:348–358 54. Lin LI, Chen CY, Lin DT et al (2005) Charac-
45. Slovak ML, Gundacker H, Bloomfield CD et al terization of CEBPA mutations in acute mye-
(2006) A retrospective study of 69 patients loid leukemia: most patients with CEBPA
with t(6;9)(p23;q34) AML emphasizes the mutations have biallelic mutations and show a
need for a prospective, multicenter initiative distinct immunophenotype of the leukemic
for rare ‘poor prognosis’ myeloid malignancies. cells. Clin Cancer Res 11:1372–1379
Leukemia 20:1295–1297 55. Hou HA, Lin LI, Chen CY et al (2009) Reply
46. Medeiros BC, Kohrt HE, Arber DA et al to ‘heterogeneity within AML with CEBPA
(2010) Immunophenotypic features of acute mutations; only CEBPA double mutations,
myeloid leukemia with inv(3)(q21q26.2)/t but not single CEBPA mutations are associated
(3;3)(q21;q26.2). Leuk Res 34:594–597 with favorable prognosis’. Br J Cancer
47. Raya JM, Martı́n-Santos T, Luño E et al (2015) 101:738–740
Acute myeloid leukemia with inv(3)(q21q26.2) 56. Tang JL, Hou HA, Chen CY et al (2009)
or t(3;3)(q21;q26.2): clinical and biological fea- AML1/RUNX1 mutations in 470 adult
tures and comparison with other acute myeloid patients with de novo acute myeloid leukemia:
leukemias with cytogenetic aberrations involving prognostic implication and interaction with
long arm of chromosome 3. Hematology other gene alterations. Blood 114:5352–5361
13:1607845415Y0000000003 57. Schnittger S, Dicker F, Kern W et al (2011)
48. Bernstein J, Dastugue N, Haas OA et al (2000) RUNX1 mutations are frequent in de novo
Nineteen cases of the t(1;22)(p13;q13) acute AML with noncomplex karyotype and confer an
megakaryblastic leukaemia of infants/ children unfavorable prognosis. Blood 117:2348–2357
and a review of 39 cases: report from a t(1;22) 58. Haferlach T, Kohlmann A, Klein HU et al
study group. Leukemia 14:216–218 (2009) AML with translocation t(8;16)(p11;
49. Carroll A, Civin C, Schneider N et al (1991) p13) demonstrates unique cytomorphological,
The t(1;22) (p13;q13) is nonrandom and cytogenetic, molecular and prognostic features.
restricted to infants with acute megakaryoblas- Leukemia 23:934–943
tic leukemia: a pediatric oncology group study. 59. Dang H, Chen Y, Kamel-Reid S et al (2013)
Blood 78:748–752 CD34 expression predicts an adverse outcome
50. Inaba H, Zhou Y, Abla O et al (2015) Hetero- in patients with NPM1-positive acute myeloid
geneous cytogenetic subgroups and outcomes leukemia. Hum Pathol 44:2038–2046
in childhood acute megakaryoblastic leukemia: 60. Chen CY, Chou WC, Tsay W et al (2013)
a retrospective international study. Blood Hierarchical cluster analysis of immunopheno-
126:1575–1584 typed classify AML patients with NPM1 gene
51. Nomdedéu J, Bussaglia E, Villamor N et al mutation into two groups with distinct prog-
(2011) Immunophenotype of acute myeloid nosis. BMC Cancer 13:107
leukemia with NPM mutations: prognostic
impact of the leukemic compartment size.
Leuk Res 35:163–168
 Chapter 16

Immunophenotyping of Acute Lymphoblastic Leukemia

Joseph A. DiGiuseppe and Jolene L. Cardinali

Abstract
Immunophenotyping by flow cytometry is an important component in the diagnostic evaluation of patients
with acute lymphoblastic leukemia. This technique further permits the detection of minimal residual disease
after therapy, a robust prognostic factor that may guide individualized treatment. We describe here
laboratory methods for both the initial characterization of lymphoblasts at diagnosis, and the detection
of rare leukemic lymphoblasts after treatment. In addition to antibody combinations suitable for diagnosis
and detection of minimal residual disease, we describe procedures for peripheral blood and bone marrow
sample preparation, procedures for labeling of cell-surface and intracellular proteins with fluorochrome-
conjugated antibodies, and approaches to analysis of immunophenotypic data.

Key words Immunophenotyping, Flow cytometry, Acute lymphoblastic leukemia, Minimal residual
disease, Antibodies

1 Introduction

Acute lymphoblastic leukemia (ALL) comprises a group of malig-

nant neoplasms of committed B-lineage or T-lineage precursor
lymphoid cells termed lymphoblasts [1]. Diagnosis and classifica-
tion of ALL require an integrated approach incorporating mor-
phology, immunophenotype, cytogenetics, and molecular genetics
[1, 2]. Immunophenotyping by flow cytometry is an important
component in the diagnostic evaluation of patients with ALL [3],
enabling its rapid distinction from morphologic mimics such as
acute myeloid leukemia (AML) [4] and acute leukemia of ambigu-
ous lineage [5], including mixed phenotype acute leukemia
(MPAL). In addition, immunophenotyping by flow cytometry per-
mits the detection and quantification of rare leukemic blasts that
may persist after treatment of ALL (reviewed in [3, 6, 7]). The
presence of such minimal residual disease (MRD) following therapy
is among the strongest prognostic factors in ALL [8], and may
potentially inform subsequent individualized therapies. Here we
describe methods for immunophenotyping of ALL by flow

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_16, © Springer Science+Business Media, LLC, part of Springer Nature 2019

297
298 Joseph A. DiGiuseppe and Jolene L. Cardinali

cytometry. We describe antibody combinations suitable for diagno-

sis and detection of MRD, procedures for peripheral blood and
bone marrow sample preparation, procedures for labeling of cell-
surface and intracellular proteins with fluorochrome-conjugated
antibodies, and approaches to analysis of immunophenotypic data.

2 Materials

2.1 Reagents 1. PBS/azide (Phosphate-buffered saline with 0.1% sodium azide):

Add 120.0 g NaCl, 3.0 g KCl, 17.25 g anhydrous Na2HPO4,
3.0 g KH2PO4, and 15.0 g of sodium azide to a 15 L container,
and fill to 15 L with distilled water. Mix thoroughly, allow to sit
overnight, and verify that pH is between 7.2 and 7.4. Store at
room temperature for up to 3 months.
2. 0.5% PBS/BSA/azide (Phosphate-buffered saline with 0.5%
Bovine serum albumin (w/v) and 0.1% sodium azide): Add
7.5 g bovine serum albumin (BSA) to 1500 mL PBS/azide
(Reagent 1) in a 1500-mL bottle. Mix thoroughly until the
BSA is completely dissolved. Store at 2–8  C for up to 3 months.
3. 2.5% PBS/BSA/azide (Phosphate-buffered saline with 2.5%
Bovine serum albumin (w/v) and 0.1% sodium azide): Add
5.0 g BSA to 200 mL PBS/azide (Reagent 1) in a 250-mL
Erlenmeyer flask. Mix thoroughly and store at 2–8  C for up to
3 months.
4. 5% PBS/BSA/azide (Phosphate-buffered saline with 5% Bovine
serum albumin (w/v) and 0.1% sodium azide): Add 10.0 g BSA
to 200 mL PBS/azide (Reagent 1) in a 250-mL Erlenmeyer
flask. Mix thoroughly and store at 2–8  C for up to 3 months.
5. RPMI/FCS (Roswell Park Memorial Institute 1640 medium
with 10% fetal calf serum (v/v)): Incubate 50 mL of fetal calf
serum at 37  C for 30 min, and add to 450 mL of RPMI. Store
at 2–8  C for up to 1 month.
6. 2% Formaldehyde solution: Add 200 mL of 10% methanol-free
formaldehyde to a 1000-mL volumetric flask. Bring to volume
with 0.5% PBS/azide, and adjust pH to 7.2–7.4. Store at
2–8  C for up to 6 months.
7. FACS™ Lysing Solution (BD Biosciences): Add 20 mL of 10
FACS™ Lysing Solution to a 200-mL volumetric flask. Bring
to volume with distilled water. Store in a light-resistant bottle
at 2–8  C for up to 1 month.

2.2 Antibody Antibody combinations suitable for the diagnosis and detection of
Combinations minimal residual disease (MRD) in acute lymphoblastic leukemia
(ALL) are presented in Table 1. These combinations were devel-
oped for use on flow cytometers equipped with three lasers (excita-
tions at 405 nm, 488 nm, and 633 nm), and capable of detecting at
 Immunophenotyping of Acute Lymphoblastic Leukemia 299

Table 1
Antibody combinations suitable for diagnosis and detection of minimal residual disease in acute
lymphoblastic leukemia

PB, V450,
Tube FITC PE PerCP-Cy5.5 PE-Cy7 APC APC-AF750 or BV421 V500
1 Kappa Lambda CD20 CD19 CD10 CD38 CD5 CD45
2 CD20 CD22 CD34 CD19 CD13 + CD33 CD38 CD10 CD45
3 CD20 CD49f CD34 CD19 CD58 CD38 CD10 CD45
4 CD24 CD304 CD34 CD19 CD86 CD38 CD10 CD45
5 CD16 CD56 CD5 CD3 CD7 CD8 CD4 CD45
6 CD7 CD1a CD3 CD2 CD5 CD8 CD4 CD45
7 cyMPO cyCD3 CD34 – CD7 – HLA-DR CD45
8 cyMPO cyCD22 cyCD79a CD19 CD34 – HLA-DR CD45
9 nTdt cyCD3 cyCD79a CD19 CD34 – HLA-DR CD45
FITC fluorescein isothiocyanate, PE phycoerythrin, PerCP-Cy5.5 peridinin-chlorophyll-Cy5.5, APC allophycocyanin,
APC-AF750 APC-Alexa Fluor 750, PB Pacific Blue, V450 BD Horizon™ V450, BV421 Brilliant Violet™ 421, V500 BD
Horizon™ V500

least eight different fluorochromes. Not all the antibody combina-

tions in Table 1 are required in each case. Tubes 1, 2, and 7–9
would suffice to establish a diagnosis of B lymphoblastic leukemia,
and tubes 5–9 would be sufficient for a diagnosis of T lymphoblas-
tic leukemia; in both instances, the assessment of intracellular anti-
gens (tubes 7–9) permits exclusion of MPAL [5]. For MRD
assessment in B-lineage ALL (tubes 2–4) and T-lineage ALL
(tubes 5, 6), more focused panels are appropriate.

3 Methods

Cell suspensions can be prepared from peripheral blood or bone

marrow samples anticoagulated with either EDTA or heparin (see
Note 1). Universal precautions should be observed in handling
specimens.

3.1 Preparation 1. Pipet an anticoagulated peripheral blood specimen into a

of Cell Suspension 10-mL specimen tube labeled with two patient identifiers,
from Peripheral Blood and denote the meniscus with a permanent marker.
2. Dilute the peripheral blood with 0.5% PBS/BSA/azide to
10 mL.
3. Centrifuge at 800 RCF for 5 min.
4. Remove the supernatant, and reconstitute to 10 mL with 0.5%
PBS/BSA/azide.
300 Joseph A. DiGiuseppe and Jolene L. Cardinali

5. Centrifuge at 800 RCF for 5 min.

6. Repeat steps 4 and 5.
7. Resuspend in 5.0% PBS/BSA/azide to designated initial
volume.
8. Adjust to a WBC concentration of 10,000/μL or less (see Note
2). For experiments in which cell-surface immunoglobulin
light chains will not be labeled (see Tubes 2–9, Table 1), this
cell suspension is used in Subheadings 3.3 and 3.4.
9. If cell-surface immunoglobulin light chains will be labeled (see
Tube 1, Table 1), aliquot 1 mL of the cell suspension prepared
above into an empty 10-mL specimen tube labeled with two
patient identifiers, denote the meniscus with a permanent
marker, dilute to 10 mL with FCS/RPMI, and incubate at
37  C for 30 min (see Note 3). Centrifuge, wash twice, and
resuspend the incubated specimen as in steps 3–8, prior to use
in Subheadings 3.3 and 3.4.

3.2 Preparation 1. Filter anticoagulated bone marrow specimen through a 53-μM

of Cell Suspension nylon mesh into a labeled specimen cup.
from Bone Marrow 2. Gently dislodge cells from marrow particles collected on the
mesh using the bottom of a polystyrene tube or bulb of a
transfer pipette, and wash any freed cells through the mesh
with 2.5% PBS/BSA/azide.
3. Transfer the filtered specimen to an appropriately labeled
10-mL specimen tube. Cap and mix well. Denote the meniscus
with a permanent marker. Obtain a cell count of the
suspension.
4. Centrifuge at 800 RCF for 10 min.
5. Remove the supernatant, resuspend in 0.5% PBS/BSA/azide,
centrifuge at 800 RCF for 5 min, and decant the supernatant.
Wash once more in this manner.
6. Resuspend in 5.0% PBS/BSA/azide to achieve a final WBC
concentration of 10,000/μL. For experiments in which cell-
surface immunoglobulin light chains will not be labeled (see
Tubes 2–9, Table 1), this cell suspension is used in Subheadings
3.3 and 3.4.
7. If cell-surface immunoglobulin light chains will be labeled (see
Tube 1, Table 1), aliquot 1 mL of the washed cell suspension
prepared above into an empty 10-mL specimen tube labeled
with two patient identifiers, denote the meniscus with a perma-
nent marker, add 10 mL of RPMI/FCS, and incubate at 37  C
for 30 min (see Note 3).
8. Centrifuge, wash twice, and resuspend the incubated specimen
as in steps 3–8 of Subheading 3.1, prior to use in Subheadings
3.3 and 3.4.
 Immunophenotyping of Acute Lymphoblastic Leukemia 301

3.3 Labeling 1. Label 12  75-mm polystyrene tubes with two patient identi-
Procedure for Cell- fiers and the antibody combination to be added.
Surface Antigens 2. Deliver 100 μL of the appropriate antibody combination (cock-
tail) to each tube (see Note 4).
3. Deliver 100 μL of the cell suspension (or preincubated cell
suspension in the case of immunoglobulin light chain studies)
as prepared in Subheadings 3.1 and 3.2 to each tube and vortex
(see Note 5).
4. Incubate the tubes for 15 min at room temperature in the dark.
5. Pipet 2 mL of FACS™ Lysing Solution into each tube and
vortex.
6. Incubate for 10 min at room temperature in the dark.
7. Centrifuge for 5 min at 800 RCF, and remove the supernatant.
8. Dilute to a volume of 2 mL with 0.5% PBS/BSA/azide and
vortex.
9. Centrifuge for 5 min at 800 RCF and remove the supernatant.
10. Fix with 500 μL of 0.5% PBS/BSA/azide and 500 μL of 2%
formaldehyde.
11. Acquire the labeled cells on the flow cytometer (see Note 6).
Acquisition of 100,000 viable cells is sufficient for initial diag-
nosis; for MRD detection, acquisition of 1,000,000 cells is
attempted (see Note 7). To mitigate carryover during acquisi-
tion for MRD detection, tubes containing distilled water
should be positioned between sample tubes.

3.4 Combined Abnormal patterns of cell-surface antigen expression (including

Labeling Procedure expression of antigens that are normally absent, loss of antigens
for Cell-Surface that are normally expressed, and expression of antigens or combi-
and Intracellular nations of antigens at inappropriate developmental stages) are
Antigens common among the acute leukemias, including B-lineage and
T-lineage acute lymphoblastic leukemia. By comparison, cytoplas-
mic expression of certain antigens confers improved lineage speci-
ficity; incorporation of intracellular studies into the diagnostic
panel is therefore useful. The following procedure, which permits
simultaneous detection of cell-surface and intracellular antigens,
should be used for antibody combinations 7–9 in Table 1 (see
Note 8). We use the Fix & Perm cell fixation and cell permeabiliza-
tion kit (ThermoFisher), but similar products are manufactured by
others.
1. Label 3 12  75-mm polystyrene tubes with two patient iden-
tifiers, and add 100 μL of cocktail containing antibodies
directed against the cell-surface antigens designated in Tubes
7–9 (Table 1).
2. Add 100 μL of cell suspension to each tube, and vortex.
302 Joseph A. DiGiuseppe and Jolene L. Cardinali

3. Incubate for 15 min at room temperature in the dark.

4. Add 100 μL of Reagent A to each tube, and vortex.
5. Incubate for 15 min at room temperature in the dark.
6. Add 2 mL of 0.5% PBS/BSA/azide and centrifuge for 5 min at
800 RCF.
7. Decant the supernatant, add 100 μL of Reagent B to each tube,
and vortex until the cell pellet is thoroughly resuspended.
8. Add the appropriate volumes of antibodies against the intracel-
lular epitopes designated in Tubes 7–9 (Table 1), and vortex.
9. Incubate for 15 min at room temperature in the dark.
10. Add 2 mL of 0.5% PBS/BSA/azide and centrifuge for 5 min at
800 RCF.
11. Decant the supernatant, and add 500 μL of 0.5% PBS/BSA/
azide and 500 μL of 2% formaldehyde.
12. Acquire the labeled cells on the flow cytometer (see Note 6).
Acquisition of 100,000 viable cells is sufficient for initial diag-
nosis; for MRD detection, acquisition of 1,000,000 cells is
attempted (see Note 7). To mitigate carryover during acquisi-
tion for MRD detection, tubes containing distilled water
should be positioned between sample tubes.

3.5 Data Analysis At both diagnosis and assessment of MRD after therapy, the goal of
data analysis is the identification of normal and abnormal cell
populations on the basis of their immunophenotypic and light-
scatter properties. Data analysis is an iterative process, which
requires a thorough understanding of normal and abnormal matu-
ration patterns [6, 9], and familiarity with the technical artifacts
that may produce spurious immunophenotypic patterns [10]. The
magnitude of this challenge is underscored by a recent study of
MRD detection in B lymphoblastic leukemia, in which analytic
precision among a cohort of experienced practitioners using iden-
tical data sets was disappointing [11]. Although the development of
algorithms capable of analyzing complex multiparametric data sets
holds the promise of improving analytic precision [12], such
approaches have yet to be widely adopted in routine practice.

3.5.1 Exclusion Assessment of immunophenotypic patterns is facilitated by exclu-

of Artifacts sion of artifacts due to fluidic perturbations during acquisition, and
focusing the analysis on viable singlet cells. A display of all events
acquired over time (Fig. 1) may reveal sample exhaustion or fluidic
perturbations; if present, such compromised data may be excluded
from further analysis. Coincident events may yield unexpected
combinations of antigen expression, and thereby simulate abnor-
mal populations. A display of peak area vs. peak height for a single
parameter enables gating on singlet events, and exclusion of
 Immunophenotyping of Acute Lymphoblastic Leukemia 303

Fig. 1 Gating strategies to exclude overt data artifacts. (Left) Assessment of acquisition using event number to
detect potential fluidic perturbations. (Center) Exclusion of coincident events (cellular doublets or higher-
ordered aggregates) by gating on singlets defined by FSC-A and FSC-H. (Right) Exclusion of debris and
nonviable cells by gating on cells with the FSC and SSC properties of viable cells

Fig. 2 Identification of blasts at diagnosis in B lymphoblastic leukemia (peripheral blood). The B-lymphoblasts
(blue) have dim CD45 relative to mature B cells (red), and low SSC. In addition to B-lineage antigens (CD19,
CD22), the blasts express CD10 and CD34 (All viable singlets are shown)

coincident events from subsequent analysis (Fig. 1). Finally, since

degenerating or nonviable cells may be bound nonspecifically by
fluorochrome-conjugated antibodies, the analysis may be facilitated
by gating on cells with the light-scatter properties of viable cells
(Fig. 1; see Note 9).

3.5.2 Identification The next step in data analysis is identification of the abnormal
of the Abnormal population. A display of CD45 vs. SSC is very useful in this regard,
Population: Diagnosis as lymphoblasts commonly occupy a region defined by dim CD45
and low SSC (Figs. 2 and 3). In the case of B lymphoblastic
leukemia, CD45 expression may be virtually undetectable
(Fig. 1), while in the case of T lymphoblastic leukemia, overlap of
CD45 expression between the blasts and mature T cells is more
common (Fig. 2). Bivariate plots of antigen vs. light scatter, or
antigen vs. antigen are then used to refine phenotypic
304 Joseph A. DiGiuseppe and Jolene L. Cardinali

Fig. 3 Identification of blasts at diagnosis in T lymphoblastic leukemia (periph-

eral blood). The T-lymphoblasts (blue) express CD45 at levels slightly dimmer
than (but partially overlapping with) those of the normal CD4+ (red) and CD8+
(green) T cells. The blasts are phenotypically heterogeneous, with absent or dim
surface CD3, and partial, variable CD1a. Subsets of the blasts are CD4–/CD8–,
CD4–/CD8(dim)+, CD4(dim)+/CD8–, or CD4+/CD8+ (All viable singlets are
shown in the upper left panel; only T cells are shown in the remaining panels)

characterization of the blast population (Figs. 2 and 3). Identifica-

tion of the abnormal population is usually straightforward at the
time of diagnosis, since blasts typically comprise a substantial frac-
tion of all cells.

3.5.3 Identification By comparison, identification of a rare persistent blast population

of the Abnormal after treatment may be challenging. In many cases, knowledge of
Population: MRD the patient’s diagnostic immunophenotype may facilitate
subsequent MRD detection. However, the patient’s diagnostic
studies may not be available at the time of MRD assessment.
Moreover, alterations in phenotype during therapy are well
described in both B-lineage [13] and T-lineage ALL [14]. In con-
trast, identification of populations whose composite immunophe-
notypes differ from those of normal cells is not subject to these
 Immunophenotyping of Acute Lymphoblastic Leukemia 305

limitations. An important caveat in the identification of MRD using

either approach is the recognition of normal precursor lymphoid
cells. In the case of MRD detection in B lymphoblastic leukemia, in
particular, the presence of normal B-cell precursors [15] in bone
marrow may complicate the analysis [11].
In analyzing the data for potential abnormal populations after
therapy, it is helpful to draw an initial gate using CD19 vs. SSC for
B lymphoblastic leukemia (Fig. 4; see Note 10) or CD7 vs. SSC for
T lymphoblastic leukemia, as these antigens are expressed in virtu-
ally all cases. This initial gate is then refined in an iterative fashion
using various combinations of antigen expression to isolate a cluster
of events whose multiparametric phenotypic properties differ from
those of normal precursor cells (Figs. 4 and 5). In the illustrated
example of MRD detected in a patient treated for B lymphoblastic
leukemia (Fig. 4), the initial CD19+ gate includes a preponderance
of normal B-cell precursors (blue), with dim CD45, little or no
CD20, and substantial CD34, all of which are immunophenotypic
properties shared by the patient’s MRD (red, highlighted).

Fig. 4 Identification of MRD in B lymphoblastic leukemia (bone marrow). The initial CD19+ gate includes a
preponderance of normal B-cell precursors (blue), with dim CD45, little or no CD20, and substantial CD34; in
contrast, the MRD (red, highlighted) differs from the normal B-cell precursors in its expression one or more
myeloid antigens (CD13 + CD33), lack of CD10 and CD38, and overexpression of CD49f and CD58 (All viable
singlets are shown in the upper left and center panels; only B cells are shown in the remaining panels)
306 Joseph A. DiGiuseppe and Jolene L. Cardinali

Fig. 5 Identification of MRD in T lymphoblastic leukemia (bone marrow). An

abnormal T-cell population (blue, highlighted) is identified with dim CD5 and
CD45, absent surface CD3, and coexpression (albeit weak) of CD4 and CD8.
These phenotypic features differ from those of the normal CD4+ and CD8+ T
cells (red and green, respectively), but resemble those of a subset of blasts seen
in the patient’s diagnostic specimen (Lower Right) (All viable singlets are shown
in the upper left panel; only T cells are shown in the remaining panels)

However, the MRD in this case differs from the normal B-cell
precursors in its expression of one or more myeloid antigens
(CD13 + CD33), lack of CD10 and CD38 expression, and over-
expression of CD49f (see Note 11) and CD58 (Fig. 4). In the
example of MRD detected in a patient treated for T lymphoblastic
leukemia (Fig. 5), iterative gating discloses an abnormal T-cell
population (blue, highlighted) with dim CD5 and CD45, absent
surface CD3, and coexpression (albeit weak) of CD4 and CD8.
These phenotypic features differ from those of the normal CD4+
and CD8+ T cells (red and green, respectively) but resemble those
of a subset of blasts seen in the patient’s diagnostic specimen
(Fig. 5, lower right panel).
 Immunophenotyping of Acute Lymphoblastic Leukemia 307

4 Notes

1. In general, samples should be processed as soon after collection

as is feasible to minimize loss of cell viability. However, if a
delay in preparation of the cell suspension cannot be avoided,
the specimen may be diluted 1:2 with RPMI and held over-
night at 2–8  C.
2. Ideally, the final WBC concentration of the cell suspension will
be 10,000/μL, enabling the delivery of 1,000,000 cells for
labeling upon pipetting a volume of 100 μL.
3. Preincubation at 37  C allows for Fc-receptor-mediated endo-
cytosis of bound immunoglobulin molecules, thereby reducing
background staining in studies with fluorochrome-conjugated
antibodies directed against immunoglobulin light chains.
4. The individual fluorochrome-conjugated antibodies that com-
prise previously characterized antibody combinations (see
Table 1) may be pooled to produce “cocktails” with reagents
sufficient for use in multiple (e.g., 50) assays. These are
prepared such that the appropriate final concentration of each
antibody will be achieved upon addition of 100 μL of cell
suspension to 100 μL of cocktail. However, the laboratory
should verify that the performance of the pooled reagents
does not differ significantly from that observed with individu-
ally added reagents, and determine the stability of each cocktail.
We aliquot cocktails into light-resistant glass vials, and have
found these aliquots to be stable for up to 4 weeks at 2–8  C.
To verify that the correct reagents have been added during its
preparation, each new cocktail is tested with an appropriate
normal specimen. Fluorescence intensities and patterns of anti-
gen expression observed using the new cocktail should be
visually indistinguishable from those seen with the previous
cocktail (i.e., lot-to-lot comparison).
5. For MRD analysis, in which the goal is to collect up to
1,000,000 labeled cells, 200 μL of washed cell suspension
(containing a total of 2,000,000 cells) should be added to
200 μL of cocktail.
6. Immediate acquisition of fixed samples is desirable, since break-
down of tandem fluorochromes is enhanced by exposure to
heat, light, and fixatives. In our experience with the antibody
combinations listed in Table 1, storage of fixed samples over-
night at 2–8  C in the dark prior to acquisition on the flow
cytometer does not result in substantial tandem degradation.
However, tandem dyes and their conjugates vary widely in their
propensity to degrade, and each laboratory should verify that
storage of fixed, labeled samples prior to acquisition does not
yield significant tandem degradation.
308 Joseph A. DiGiuseppe and Jolene L. Cardinali

7. Visual recognition of a rare abnormal population comprising

fewer than 20 cellular events is difficult, and the threshold for
positivity in most studies demonstrating the clinical signifi-
cance of MRD in ALL has been 0.01%, or 1 in 10,000 cells
[8]. At a minimum, then, 200,000 viable cells would be
required to recognize a cluster of 20 abnormal cells at a fre-
quency of 0.01%. The ability to recognize a potentially abnor-
mal cluster as a distinct cellular population is also influenced by
the magnitude of antigenic deviation from normal, the extent
to which antigen expression is homogeneous, and the number
of admixed normal events; in certain instances, populations
comprising greater than 20 cells may not be apparent. We
therefore attempt to acquire 1,000,000 events for each anti-
body combination in MRD assays.
8. By enabling the simultaneous detection of a relatively myeloid-
specific antigen (myeloperoxidase) and T-lineage (cytoplasmic
CD3-ε; Tube 7) or B-lineage (cytoplasmic CD22 and CD79a;
Tube 8) antigens, Tubes 7 and 8 facilitate the identification of
mixed phenotype acute leukemia (MPAL), T/myeloid and
B/myeloid, both of which should be distinguished from
ALL. Tube 9 includes the nuclear antigen, TdT, in conjunction
with T-lineage and B-lineage antigens; TdT expression (though
not specific for ALL) distinguishes precursor lymphoid neo-
plasms from their mature counterparts.
9. By eliminating much of the background/nonspecific staining
associated with degenerating or nonviable cells, selective gating
on events with the light scatter properties of viable cells may
facilitate recognition of normal and abnormal cellular popula-
tions during data analysis. However, care must be taken to
ensure that cells of interest are not inadvertently excluded
from subsequent analysis.
10. Caution is advised in using CD19 as a gating target for MRD
detection in B lymphoblastic leukemia because of the increas-
ing use of therapeutic agents directed against this molecule
(e.g., blinatumomab [16], anti-CD19 chimeric antigen recep-
tor (CAR)-T cells [17]), which may result in loss of the epitope.
In these cases, antibody combinations that do not rely on
detection of CD19 should be used [18].
11. In our experience [19, 20], overexpression of CD49f is
among the most common recurrent phenotypic aberrations
in B lymphoblastic leukemia, though we have observed an
apparent diminution in CD49f expression in samples held
for extended periods (e.g., greater than a day) prior to proces-
sing. Extended delays in sample processing should therefore
be avoided.
 Immunophenotyping of Acute Lymphoblastic Leukemia 309

Acknowledgments

The authors gratefully acknowledge Dr. Brent Wood, University of

Washington, Seattle, for his generous gift of WoodList 2.7.8
software.

References
1. Swerdlow SH, Campo E, Harris NL et al Darzynkiewicz Z, Roederer M, Tanke HJ (eds)
(2017) WHO classification of tumours of hae- Cytometry. New developments, Methods in
matopoietic and lymphoid tissues, 4th edn. cell biology, vol 75, 4th edn. Elsevier, San
IARC, Lyon Diego, CA, pp 559–576
2. Borowitz MJ, DiGiuseppe JA (2014) Acute 10. Wood B (2006) 9-color and 10-color flow
lymphoblastic leukemia/lymphoblastic lym- cytometry in the clinical laboratory. Arch
phoma. In: Orazi A, Weiss LM, Foucar K et al Pathol Lab Med 130:680–690
(eds) Knowles’ neoplastic hematopathology. 11. Keeney M, Wood BL, Hedley BD et al (2017)
Lippincott Williams & Wilkins, Philadelphia, A QA program for MRD testing demonstrates
pp 1019–1029 that systematic education can reduce discor-
3. DiGiuseppe JA (2016) Acute lymphoblastic dance among experienced interpreters. Cyto-
leukemia/lymphoma: diagnosis and minimal metry B Clin Cytom 94:239–249
residual disease detection by flow cytometric 12. DiGiuseppe J, Tadmor MD, Pe’er D (2015)
immunophenotyping. In: Detrick B, Detection of minimal residual disease in B lym-
Hamilton R, Schmitz J (eds) Manual of molec- phoblastic leukemia using viSNE. Cytometry B
ular and clinical laboratory immunology, 8th Clin Cytom 88B:294–304
edn. ASM Press, Washington, D.C., pp 13. Gaipa G, Basso G, Aliprandi S et al (2008)
207–216 Prednisone induces immunophenotypic mod-
4. Wood B, Soma L (2016) Acute myeloid leuke- ulation of CD10 and CD34 in nonapoptotic
mia: diagnosis and minimal residual disease B-cell precursor acute lymphoblastic leukemia
detection by flow cytometry. In: Detrick B, cells. Cytometry B Clin Cytom 74:150–155
Schmitz J, Hamilton RG (eds) Manual of 14. Roshal M, Fromm JR, Winter S et al (2010)
molecular and clinical laboratory immunology, Immaturity associated antigens are lost during
8th edn. ASM Press, Washington, D.C. induction for T cell lymphoblastic leukemia:
5. Borowitz MJ, Bene M-C, Harris NL et al implications for minimal residual disease detec-
(2017) Acute leukaemias of ambiguous line- tion. Cytometery B Clin Cytom 78:139–146
age. In: Swerdlow SH, Campo E, Harris NL 15. McKenna RW, Washington LT, Aquino DB
et al (eds) WHO classification of tumours of et al (2001) Immunophenotypic analysis of
haematopoietic and lymphoid tissues, 4th edn. hematogones (B-lymphocyte precursors) in
IARC, Lyon, pp 180–187 662 consecutive bone marrow specimens by
6. Wood BL (2016) Principles of minimal residual 4-color flow cytometry. Blood 98:2498–2507
disease detection for hematopoietic neoplasms 16. Kantarjian H, Stein A, Gokbuget N et al
by flow cytometry. Cytometry B Clin Cytom (2017) Blinatumomab versus chemotherapy
90:47–53 for advanced acute lymphoblastic leukemia. N
7. Gaipa G, Basso G, Biondi A et al (2013) Detec- Engl J Med 376:836–847
tion of minimal residual disease in pediatric 17. Park JH, Riviere I, Gonen M et al (2018)
acute lymphoblastic leukemia. Cytometry B Long-term follow-up of CD19 CAR therapy
Clin Cytom 84:359–369 in acute lymphoblastic leukemia. N Engl J
8. Borowitz MJ, Devidas M, Hunger SP et al Med 378:449–459
(2008) Clinical significance of minimal residual 18. Cherian S, Miller V, McCullouch V et al (2018)
disease in childhood acute lymphoblastic leu- A novel flow cytometric assay for detection of
kemia and its relationship to other prognostic residual disease in patients with
factors: a Children’s oncology group study. B-lymphoblastic leukemia/lymphoma post
Blood 111:5477–5485 anti-CD19 therapy. Cytometery B Clin
9. Wood B (2004) Multicolor immunophenotyp- Cytom 94B:112–120
ing: human immune system hematopoiesis. In:
310 Joseph A. DiGiuseppe and Jolene L. Cardinali

19. DiGiuseppe JA, Fuller SG, Borowitz MJ acute lymphoblastic leukemia: efficiency of
(2009) Overexpression of CD49f in precursor minimal residual disease (MRD) detection by
B-cell acute lymphoblastic leukemia: potential flow cytometric immunophenotyping com-
usefulness in minimal residual disease detec- pared with commonly used markers. Mod
tion. Cytometry B Clin Cytom 76:150–155 Pathol 89:388A
20. Shinoda-Matsuoka H, DiGiuseppe JA (2009)
Overexpression of CD49f in precursor B-cell
 Chapter 17

Clinical Flow-Cytometric Testing in Chronic Lymphocytic

Leukemia
Dalia A. Salem and Maryalice Stetler-Stevenson

Abstract
Flow-cytometric demonstration of the typical chronic lymphocytic leukemia (CLL) immunophenotype is
vital for diagnosis. CLL has a characteristic immunophenotype, expressing CD5, CD19, dim CD20, dim
CD22, CD23, bright CD43, dim CD45, dim to negative CD79b, dim CD81, CD200, and dim monoclo-
nal surface immunoglobulin. This characteristic immunophenotype allows a definitive diagnosis and the
ruling out of another leukemia or lymphoma. Flow cytometry also provides important prognostic informa-
tion and accurate assessment of response to therapy. Here we describe optimal specimen collection, red cell
lysis, appropriate panel, cell staining, acquisition on a flow cytometer, and analysis for CLL specimens.

Key words Flow cytometry, Immunophenotype, Chronic lymphocytic leukemia, Minimal residual
disease

1 Introduction

Chronic lymphocytic leukemia/small lymphocytic lymphoma

(CLL/SLL) is a neoplasm of mature B-lymphocytes involving
peripheral blood (PB), bone marrow (BM), and secondary lym-
phoid tissues (spleen, lymph nodes). CLL is the most common
leukemia of adults in western countries and it accounts for 7% of
non-Hodgkin lymphomas. The diagnosis is established by blood
count, morphology, and immunophenotyping by flow cytometry
(FC) of circulating B-lymphocytes [1].
FC demonstration of the typical CLL immunophenotype is
vital for diagnosis. CLL typically displays a characteristic immuno-
phenotype, expressing CD5, CD19, dim CD20, dim CD22, CD23
(see Note 1), bright CD43, dim CD45, dim-to-negative CD79b,
dim CD81, CD200, and dim monoclonal surface immunoglobulin
(Ig) (Fig. 1) [2] but negative for CD10, CD103, and CD123 as
well as other T-cell and myeloid antigens [3, 4]. Diagnosis of CLL
requires
5  103/μl circulating monoclonal B-lymphocytes with
a CLL immunophenotype in the PB. The term SLL designates the

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_17, © Springer Science+Business Media, LLC, part of Springer Nature 2019

311
312 Dalia A. Salem and Maryalice Stetler-Stevenson

a
5 5 5
10 10 10

4 4 4
10 10 10
CD5 V450

CD5 V450

CD5 V450
3 3 3
10 10 10

2 2 2
10 10 10
1 1 1
-10 -10 -10

-101 102 103 104 105 -101 102 103 104 105 -100 102 103 104 105
CD79b PE CD20 AH7 CD22 PerCP Cy55

5
10 105 105

Lambda-m PE
4 4
10 10
4
10
CD38 V450
CD81 FITC

3
10 3 10
3
10

2
10 2
10
102
2
10
1 2 -10
-10
2 3 4 5 2 3 4 5 2
-10 10 10 10 -10 10 10 10 -10 0 102 10
3
10
4
10
5

CD43 APC CD5 APC Kappa-m FITC

b
0.01%
105 10
5
5.91% 10
5

104 104
CD200 PC7

CD38 V450

104
CD49d PE

103 10
3
3
10

2
10 102
102
2
-10 -102 0.00% 94.08% 2
-10
1 2 3 4 5 2 2 3 4 5 1 2 3 4 5
10 10 10 10 10 -10 10 10 10 10 10 10 10 10 10
CD19 PerCP Cy55 CD5 V450 CD5 APC

Fig. 1 (a) Gated on CD19 positive lymphocytes. Abnormal CLL cells (black) with normal B-cells (gray) in
background. CLL cells are positive for CD5, have bright CD43, and have dim CD79b, CD20, CD22, and CD81.
CD38 is negative. The CLL cells are monoclonal, dim positive for lambda surface light chain and negative for
kappa surface light chain. (b) Gated on abnormal CLL cells. Cells are CD200 positive. Prognostic markers
CD49d (less than 30% positive) and CD38 are negative

cases with a circulating CLL cell count <5  103/μl and recog-
nized nodal, splenic, or other extramedullary involvement. The
differential diagnosis of CLL/SLL primarily includes monoclonal
B-cell lymphocytosis (MBL) and mantle cell lymphoma (MCL).
 Flow Cytometry in CLL 313

MBL is the presence of <5  103/μl circulating monoclonal

B-lymphocytes in absence of any associated lymphadenopathy,
organomegaly, extramedullary involvement, or any other feature
of B-cell lymphoproliferative disorders [5]. MBL CLL-type (75% of
MBL cases) has an immunophenotype identical to CLL. The
expression of CD200 and CD23 in combination with dim CD20,
CD22, CD45, CD79b, and CD81 as well as bright CD43 differ-
entiates CLL from MCL [4]. Infrequently CLL cases may have an
atypical immunophenotype such as lack of CD5 and CD23, normal
intensity of CD20, CD22, and CD79b, aberrant expression of
T-cell, myeloid or other B-cell antigens [6–8].
In addition to its role in diagnosis of CLL, FC also provides
important prognostic information. CD38 expression in 30% or
greater of the CLL cells was found to be associated with aggressive
clinical course and to correlate with Ig heavy chain mutational
status [9–11]. CD49d expression by 30% or greater of the CLL
cells is an independent indicator of CLL prognosis and is superior
to CD38 in predicting clinical progression in CLL patients. CD49d
in CLL is associated with unfavorable cytogenetic profile and is
particularly associated with trisomy12 [12, 13]. CD49d expression
is stable over time and determination of a positive or negative status
is straightforward [14, 15]. FC detection of minimal residual dis-
ease (MRD) levels during and after therapy has also been shown to
be an independent predictor of progression-free and overall survival
in CLL [16, 17].

2 Materials

2.1 Lysing Solution 1. Weigh 8.29 g ammonium chloride (NH4Cl) and 1 g potassium
(1) bicarbonate (KHCO3) and place them into 1000 ml glass or
clear plastic bottle.
2. Add 4 ml 0.5 M EDTA solution, and fill to a volume of
1000 ml with distilled water (H2O). Cap and mix well.
3. Store at room temperature for up to 6 months.

2.2 Wash Solution: 1. Weigh 4 g BSA and add to 4 l of 1 PBS with 0.1% sodium
Phosphate Buffered azide. Cap and mix well by inverting several times. Allow wash
Saline (PBS) with 10% solution to sit at room temperature for 30 min, mix again, and
Bovine Serum Albumin check to be sure the BSA has dissolved.
(BSA) and Sodium 2. Perform pH testing after BSA is in solution and while the
Azide solution is at room temperature. Adjust pH to 7.3–7.5 by
adding small amounts of concentrated HCL (if the pH is
>7.5) or concentrated NaOH (if the pH is <7.3) (see
Notes 2 and 3).
3. Store at 4  C for 2 months from preparation.
314 Dalia A. Salem and Maryalice Stetler-Stevenson

2.3 Fixative: 0.5% 1. Mix 50 ml of 10% buffered formaldehyde with 950 ml of PBS
Formalin Fixative in a bottle, cap and mix well (see Note 4).
Solution 2. Wrap bottle with foil and tape to protect from light. Check PH
and adjust to 7.3–3.5 (see Subheading 2.2).
3. Store at room temperature for up to 1 year.

3 Method

3.1 Cocktails The panel in Table 1 is an example of an 8 color panel that can be
and Panels used to diagnose CLL and monitor disease post therapy. Each
cocktail is made by addition of the appropriate volumes of different
Antibody (Ab) as recommended by the vendor and/or titration.
The amount of the cocktail to be added during staining tube
(Subheading 3.6, step one below) is based on the sum of the
different Ab volumes in this cocktail. The cocktail should be stored
at 4  C and light exposure minimalized. The stability of these
cocktails should be determined by each clinical laboratory.

3.2 Specimen PB and BM are the primary samples to be examined, although

Collection lymph node (LN) biopsy or fine needle aspiration (FNA) may be
indicated. PB and BM specimens collected in sodium heparin are
stable at room temperature for up to 72 h and 24 h respectively
while EDTA PB and BM specimens are stable for 24 and 12 h
respectively. Improper or lengthy specimen storage can negatively
impact results. Overnight shipment of specimens with the possibil-
ity of exposing the specimen to temperature extremes has been
shown to alter antigen expression [18]. An LN biopsy must be
made into a cell suspension using mechanical tissue disaggregation
(using commercial devices or manual tools). FNA samples usually
do not require additional disaggregation. FNA or cell suspension
can be stored in RPMI with 10% fetal bovine serum for up to 18 h
at 4  C.

Table 1
Antibody panel for CLL diagnosis/follow-up

FITC PE PerCP PC7 APC AH7 v450 v500

1 CD103 CD25 CD123 CD19 CD23 CD20 CD11c CD45
2 CD81 CD79b CD22 CD19 CD43 CD20 CD5 CD3
3 Kappa-m Lambda-m CD20 CD19 CD5 CD14 CD38 CD45
 Flow Cytometry in CLL 315

3.3 Pre-lysis, Cell Minimizing sample treatment before staining and timely sample
Count, and Viability processing greatly reduces the risk of losing cells of interest (espe-
Assessment [18, 19] cially in case of CLL MRD) and maintains cell viability and
integrity.
1. Transfer 5 ml of either PB or BM into labelled 50 ml conical
tube(s).
2. Add ammonium chloride lysing Solution (see Subheading 2.1)
to fill each conical tube. Cap and invert several times to mix.
Incubate 10 min room temperature, inverting to mix occasion-
ally and at end.
3. Centrifuge tubes at 500  g for 5 min, check for pellet and
aspirate or decant supernatant into waste container (see Note 5).
4. Cap and vortex tube(s) gently for less than 5 s to dislodge pellet
(s). Fill the tube with PBS, cap, and invert to mix. Repeat
step 3 above (see Note 6).
5. Cap and vortex tube gently to dislodge the entire cell pellet.
Add PBS to the cell pellet according to the pellet size (0.2–2 ml
for small pellets and 3–5 ml for larger pellets), and flick or
vortex gently to mix. Record the final cell suspension volume
to be used later to adjust the cell count.
6. Assess viability and perform a cell count by using a viability dye
that is excluded from viable cells and automated cell counters
or a hemocytometer [20]. Samples with less than 75% viability
should be rejected unless the sample is irreplaceable.

3.4 Cell 1. Divide the total WBCs count (obtained by multiplying the cell
Concentration count in the “count” tube by the dilution factor multiplied by
Adjustment the preparation volume in Subheading 3.3, step 5) by the
target WBCs count to get the amount of the required final
volume. Example: if the total WBCs count ¼ 15.0  106 and
the target count is 20  106/ml, then the final volume should
be 0.75 ml.
2. Adjust the final volume of the cell suspension to reach the
target cell concentration of 20  106 cells/ml (see Note 7).

3.5 Staining Cells 1. Deliver the appropriate volume of the premade cocktail (Sub-
with Antibodies heading 3.1) to the bottom of the labeled 12  75 mm tubes
[18, 19 ] using automatic pipette.
2. Add 100 μl patient cell preparation to each tube and vortex
briefly to mix. Incubate 30 min at room temperature protected
from the light.
3. Wash stained tubes by adding 3–4 ml wash solution to each
tube; vortex to mix then centrifuge at 500  g for 5 min, check
for pellet and aspirate or decant tube into waste container, flick
or vortex tube to dislodge entire cell pellet.
316 Dalia A. Salem and Maryalice Stetler-Stevenson

4. Repeat wash step 3 above.

5. Add 200–300 μl formalin fixative and vortex to mix and store
stained and fixed cells at 4  C until acquisition (see Note 8).

3.6 Acquisition Data is acquired on a cytometer. Care should be taken not to

exclude low FSC events. Acquisition of ungated data is stored as
list mode data. 1,000,000 to 2,000,000 events (dependent on level
of MRD detection required) are acquired when cellular content is
adequate. In specimens with low cell count (e.g., FNA) acquire as
many events as possible. Acquiring a high number of events is
crucial for MRD detection. At the minimum all fluorescence chan-
nels, time, FCS-A, FSC-H and SSC-A should be acquired.

3.7 Analysis Sequential gating of the list mode data is performed to distinguish
abnormal cells of interest from normal cells based on light scatter
and antigen profile. An example of a typical sequence of hierarchical
analysis gating follows:
1. The FSC-A vs. time plot of ungated cells is evaluated for
discontinuity indicating fluidic disturbances (e.g. bubbles,
clogs, or in the case of low cell number the tube being acquired
dry). A time gate is drawn to exclude such events (Fig. 2).
2. A viable gate is drawn on the time gated SSC-A vs. FSC-A plot
to exclude debris with very low scatter properties (Fig. 2) and is
placed under the time gate.
3. A singlets gate is drawn on the viable gated FSC-H vs. FSC-A
plots to exclude doublets (Fig. 2) and is placed under the
viable gate.
4. Antigen vs. SSC gates are drawn to define T-cells (CD3+) and
B-cells (CD19+) (Fig. 2) and placed under the singlet gate.
5. The antigen based gates are displayed on a FSC-A vs. SSC-A
plot to create a scatter gate that includes all lymphocytes
(Fig. 2). This lymphocyte gate is placed under the singlet gate.
6. The antigen profile of the lymphoid cells is evaluated under the
lymphocyte and B cell antigen gates to identify the CLL immu-
nophenotype (Fig. 1).
7. For MRD assessment, low level CLL involvement must be
detected among a predominately polyclonal B-cell back-
ground. There are two general analysis approaches.
(a) Method 1: Based upon European Research Initiative in
CLL (ERIC) method [2, 16, 21]. The ERIC method
utilizes cells stained with cocktail 2 in Table 1 and a
Boolean gating strategy in which CD19 positive B-cells
are evaluated and a series of analysis gates are created in
areas where normal B-cell populations are not observed
(Fig. 3): A lymphocyte scatter gate that includes all CD19
 Flow Cytometry in CLL 317

Fig. 2 Basic analysis for CLL: (a) Time gate. (b) Time gate displayed. Viable gate to exclude debris. (c) Time
gate displayed. Singlets gated to exclude doublets. (d) Gated on Time, Viable, and Singlets gates.
Antigen vs. SSC gates to define T-cells (CD3+) (e) Antigen vs. SSC gates to define B-cells. (f) Gated on
time, viable, and singlets gates. Antigen-based gates are displayed on a FSC-A vs. SSC-A plot to ensure
lymphocyte gate includes all lymphocytes. (g) Final analysis gate: Gated on time, viable, singlets, and
lymphocyte gates

positive cells is the first gate. A gate of CD19 positive and

CD3 negative cells is placed under this gate. Then four
gates are created: gate 1; CD5 positive CD79b dim to
negative, gate 2; CD5 positive CD20 dim to negative,
gate 3; CD5 positive CD22 dim to negative, and gate 4;
CD43 positive and CD81 dim to negative. The CLL
analysis gate consists of all cells in the lymphocyte scatter
gate, CD19 positive, CD3 negative, and present in gates
1, 2, 3, and 4. Normal B-cells do not meet all of these
criteria.
(b) Method 2: Hierarchical gate based upon abnormal anti-
gen expression to define a monoclonal B cell population
(Fig. 4). Steps 1 through 5 are performed. A CD19
318 Dalia A. Salem and Maryalice Stetler-Stevenson

Fig. 3 CLL MRD Method 1: (a) Viable and Lymphocyte gates. (b) CD19+ gate. (c) CD19+CD3– gate. (d–g)
Gated on viable, singlet, lymphocyte and CD19+CD3– gates. The four gates of this method. (h) Minimal
residual CLL. Gated on Viable, Singlet, Lymphocyte, CD19+CD3– gates and gates 1–4

Fig. 4 CLL MRD analysis method 2: (a) Viable and lymphocyte gates. (b) CD19+ gate. (c) Gated on viable,
singlet, lymphocyte, and CD19+ gates. Abnormal CLL cells CD20 dim and CD5+ in ellipse. (d) Gated on viable,
singlet, lymphocyte and CD19+ gates. Abnormal CLL cells CD5+ and homogeneously CD38 negative in
ellipse. (e) Gated on viable, singlet, lymphocyte and CD19+ gates. Abnormal CLL cells in ellipse masked by
polyclonal B cells. (f) Gated on viable, singlet, lymphocyte, CD19+, and abnormal CD5, CD20, and CD38 gates.
Abnormal CLL cells in ellipse are monoclonal, dim positive for kappa and negative for lambda
 Flow Cytometry in CLL 319

positive gate is drawn within the lymphocyte scatter gate.

The B-cells in the CD19 gate are interrogated and light
chain expression is evaluated in the cells expressing CD5,
dim CD20, and atypical CD38.

4 Notes

1. CD23 is labile and can decrease with specimen storage prior to

processing.
2. The wash solution should be checked with a fresh PB sample
before use in patients (see Subheading 3). The supernatant from
the lysed control blood should be clear. If there is evidence of
hemolysis, discard solution and prepare a new one.
3. As this solution contains sodium azide, its vapor should not be
breathed, and it should not be poured down the sink where
explosive conditions may develop.
4. Formaldehyde is a known carcinogen and acutely hazardous;
do not inhale and upon skin contact flush with water.
5. Post removal of supernatant during red cell lysis, gauze or a
cotton swap can be used to carefully wipe debris from the
insides of the tube(s); do not disturb the pellet before wiping
inside the tube.
6. Washing cells with room temperature PBS prior to staining
removes cytophilic antibody (bound immunoglobulin from
the serum). As cytophilic antibody may be still bound to NK
and T-cells, kappa and lambda expression is analyzed in associ-
ation with a B-cell marker, such as CD19, CD20, or CD22,
(besides exclusion of CD3 and or CD14 positive events) in
order to remove the monocytes, T and NK cells with bound
extrinsic Ab from the analysis.
7. Preparation of target cell concentration:
(a) If greater than 20  106/ml (Increase the final PBS
volume).
l Divide WBC concentration by 20, multiply by cell
volume to get the final PBS volume.
l Add PBS up to the new volume and note on Flow QA
sheet.
l Example 1: If cell count is 20  106/ml in 5 ml,
20/20 ¼ 1  5 ¼ 5 ml; add PBS to 5 ml.
l Example 2: If cell count is 80  106/ml in 3 ml,
80/20 ¼ 4  3 ¼ 12.0 ml; add PBS to 12.0 ml.
320 Dalia A. Salem and Maryalice Stetler-Stevenson

(b) If less than 10  106/ml (Re-concentrate to decrease the

final PBS volume):
l Centrifuge specimen at 300  g for 10 min.
l Divide WBC concentration by 20, multiply by cell
volume to get the final PBS volume.
l Remove supernatant down to the new final.
l Example: If cell count is 3.0  106/ml in 3 ml,
3/20 ¼ 0.15 ml  3 ¼ 0.45 ml final volume.
8. For best results, acquire stained fixed cells within 24 h.

References
1. Campo E, Ghia P, Montserrat E et al (2017) mature B cell neoplasms. Cytometry B Clin
Chronic lymphocytic leukemia/small lympho- Cytom 50:243–248
cytic lymphoma. In: Swerdlow SH, Campo E, 8. Kampalath B, Barcos MP, Stewart C (2003)
Harris NL, Pileri SA, Jaffe ES, Stein H, Thiele J Phenotypic heterogeneity of B cells in patients
et al (eds) WHO classification of tumours of with chronic lymphocytic leukemia/small lym-
hematopoietic and lymphoid tissues, 4th edn. phocytic lymphoma. Am J Clin Pathol 119
IRAC, Lyon, pp 216–221 (6):824–832. https://doi.org/10.1309/
2. Rawstron AC, Villamor N, Ritgen M et al 4agut3lkeurd7t7k
(2007) International standardized approach 9. Damle R, Wasil T, Fais F et al (1999) Ig V gene
for flow cytometric residual disease monitoring mutation status and CD38 expression as novel
in chronic lymphocytic leukaemia. Leukemia prognostic indicators in chronic lymphocytic
21(5):956–964. https://doi.org/10.1038/sj. leukemia. Blood 94:1840–1847
leu.2404584 10. Durig J, Naschar M, Schmucker U, Renzing-
3. Muller-Hermelink HKME, Catovsky D, Kohler K, Holter A, Huttmann T, Duhrsen U
Campo E, Harris NL, Stien H (2007) Chronic (2002) CD38 expression is an important prog-
lymphocytic leukemia/small lymphocytic lym- nostic marker in chronic lymphocytic leukae-
phoma. WHO classification of hematopoietic mia. Leukemia 16:30–35
and lymphoid tissues. IARC, Lyon 11. Hamblin TJ, Orchard JA, Ibotson RE et al
4. Alapat D, Coviello-Malle J, Owens R et al (2002) CD38 expression and immunoglobulin
(2012) Diagnostic usefulness and prognostic variable region mutations are independent
impact of CD200 expression in lymphoid prognostic variables in chronic lymphocytic
malignancies and plasma cell myeloma. Am J leukemia but CD38 expression may vary dur-
Clin Pathol 137(1):93–100. https://doi.org/ ing the course of the disease. Blood
10.1309/ajcp59uorcyzevqo 99:1023–1029
5. Marti G, Rawstron AC, Ghia P, Hillmen P, 12. Bulian P, Shanafelt TD, Fegan C et al (2014)
Houlston RS, Kay N, Aurran Schleinitz T, CD49d is the strongest flow cytometry-based
Caporaso N, on behalf of The International predictor of overall survival in chronic lympho-
Familial CLL Consortium (2005) Diagnostic cytic leukemia. J Clin Oncol 32(9):897–904.
criteria for monoclonal B-cell lymphocytosis https://doi.org/10.1200/jco.2013.50.8515
(MBL). Br J Haematol 130:325–332 13. Chiorazzi N (2012) Implications of new prog-
6. Ahmad E, Steinberg SM, Goldin L, Hess CJ, nostic markers in chronic lymphocytic leuke-
Caporaso N, Kreitman RJ, Wiestner A, mia. Hematology Am Soc Hematol Educ
Wilson W, White T, Marti G, Stetler-Stevenson Program 2012:76–87. https://doi.org/10.
M (2008) Immunophenotypic features distin- 1182/asheducation-2012.1.76
guishing familial chronic lymphocytic leukemia 14. Gattei V, Bulian P, Del Principe MI et al (2008)
from sporadic chronic lymphocytic leukemia. Relevance of CD49d protein expression as
Cytometry B Clin Cytom 74:221–226 overall survival and progressive disease prog-
7. Kingma DW, Imus P, Xie XY, Jasper G, nosticator in chronic lymphocytic leukemia.
Sorbara L, Stewart C, Stetler-Stevenson M Blood 111(2):865–873. https://doi.org/10.
(2002) CD2 is expressed by a sub-population 1182/blood-2007-05-092486
of normal B cells and is frequently present in
 Flow Cytometry in CLL 321

15. Gooden CE, Jones P, Bates R et al (2018) H43-A2, 2nd edn. Clinical and Laboratory
CD49d shows superior performance character- Standards Institute, Wayne
istics for flow cytometric prognostic testing in 19. Stelzer GT, Marti G, Hurley A et al (1997) U.
chronic lymphocytic leukemia/small lympho- S.-Canadian consensus recommendations on
cytic lymphoma. Cytometry B Clin Cytom 94 the immunophenotypic analysis of hematologic
(1):129–135. https://doi.org/10.1002/cyto. neoplasia by flow cytometry: standardization
b.21384 and validation of laboratory procedures. Cyto-
16. Rawstron AC, Kennedy B, Evans PA et al metry 30(5):214–230
(2001) Quantitation of minimal disease levels 20. Cadena-Herrera D, Esparza-De Lara JE,
in chronic lymphocytic leukemia using a sensi- Ramirez-Ibanez ND et al (2015) Validation of
tive flow cytometric assay improves the predic- three viable-cell counting methods: manual,
tion of outcome and can be used to optimize semi-automated, and automated. Biotechnol
therapy. Blood 98:29–35 Rep (Amst) 7:9–16. https://doi.org/10.
17. Moreton P, Kennedy B, Lucas G et al (2005) 1016/j.btre.2015.04.004
Eradication of minimal residual disease in 21. Rawstron AC, Bottcher S, Letestu R et al
B-cell chronic lymphocytic leukemia after (2013) Improving efficiency and sensitivity:
alemtuzumab therapy is associated with pro- European research initiative in CLL (ERIC)
longed survival. J Clin Oncol 23 update on the international harmonised
(13):2971–2979. https://doi.org/10.1200/ approach for flow cytometric residual disease
jco.2005.04.021 monitoring in CLL. Leukemia 27
18. Wayne P (2007) Clinical flow Cytometric anal- (1):142–149. https://doi.org/10.1038/leu.
ysis of neoplastic Hematolymphoid cells; 2012.216
approved guideline, 2nd ed. CLSI document
 Chapter 18

Immunophenotyping of Paroxysmal Nocturnal

Hemoglobinuria (PNH)
Andrea J. Illingworth, Iuri Marinov, and D. Robert Sutherland

Abstract
Paroxysmal nocturnal hemoglobinuria (PNH) is a rare but often debilitating disease which may lead to
death in up to 35% of patients within 5 years if unrecognized and untreated. Detection of PNH and
assessment of PNH clone size in RBC and WBC lineages by flow cytometric analysis has increased in
importance due to the availability of novel therapies. These therapies typically block the hemolysis of red
blood cells and thus significantly lower the morbidities and mortality associated with this disease. This
chapter describes validated, state-of-the-art, high-sensitivity flow cytometric methodologies based on latest
published testing guidelines for PNH.

Key words Paroxysmal nocturnal hemoglobinuria (PNH), Glycophosphatidylinositol (GPI)-

anchored protein, Aplastic anemia (AA), Myelodysplastic syndrome (MDS), CD59, FLAER

Reagents and Materials

Flow cytometer capable of at least five-color analysis
Test tubes
Antibodies (see Tables 1, 2, 3, and 4 for recommended antibody combinations).
Lysing solutions: BC Immunoprep, BC OptiLyse, BC VersaLyse, BD FACSLyse, or other
validated lysing solution.
Buffers: Phosphate Buffered Saline (PBS), PBA (PBS and 0.1% sodium azide and 0.1%
bovine serum albumin) refrigerated.

1 Introduction: Clinical Utility of GPI-AP Deficiency Testing in PNH and BMFS

Glycophosphatidylinositol-anchored protein (GPI-AP)-deficient

hematopoietic stem cells (HSC) occur following somatic mutations
of the X-linked phosphatidylinositol glycan complementation Class
A gene (PIG-A), [1, 2]. Mutated HSC are able to escape immune
attack mediated by CD4+ Th1, Th17 and autoreactive CD8+
cytotoxic T cells, resulting in clonal expansion followed by addi-
tional genetic or epigenetic events [3]. The proinflammatory bone

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6_18, © Springer Science+Business Media, LLC, part of Springer Nature 2019

323
324 Andrea J. Illingworth et al.

marrow environment is characterized by reduction of CD4+CD25

+FOX3 regulatory T cells with an altered cytokine profile, expan-
sion of CD4+ helper Th1 and effector Th17 cells involved in organ-
specific autoimmunity and expansion of CD8+ cytotoxic T cells.
Such responses represent a well-documented, common pathophys-
iological mechanism for disorders comprising paroxysmal noctur-
nal hemoglobinuria (PNH) and bone marrow failure syndromes
(BMFS). Such disorders present with clonal hematopoiesis, exhibit
overlapping clinical symptoms and propensity to develop myelo-
dysplastic syndrome (MDS) and/or acute myeloid leukemia
(AML) [4–6]:
1. The distinction between the rare inherited forms of bone mar-
row failure syndromes (IBMFS) such as congenital neutropenia
(CN), Diamond–Blackfan anemia (DBA), Schwachman–Dia-
mond Sy (SDS), amegakaryocytic thrombocytopenia (AT),
reticular dysgenesis (TAR), Fanconi anemia (FA), dyskeratosis
congenita (DC) and the more frequent acquired forms of bone
marrow failure syndromes (ABMFS) such as aplastic anemia
(AA) and hypoplastic MDS (h-MDS) is not usually straightfor-
ward: approximately 5% of patients younger than 40 years with
diagnosed AA have been reported to carry germline mutations
[7]. GPI-AP-deficient clones are never present in IBMFS, thus
high sensitivity flow cytometry (HS FCM) in addition to chro-
mosome breakage studies, lymphocyte telomere length analysis
and molecular methods is useful for the differential diagnosis,
due to high negative predictive value [8].
2. The presence of GPI-AP-deficient clones in ABMFS has been
reported in 25–45% of patients with AA, 5.5–9.8% of patients
with MDS, 5.7% of patients with unexplained cytopenia, 6% of
patients with pancytopenia [9–11]. The GPI-AP-deficient
clone size can vary from <0.1% (BMFS, subclinical PNH, and
aplastic PNH) to >90% in hemolytic and/or thrombotic PNH
forms. The clinical significance of middle to large PNH clones
is related to hemolysis and/or thrombosis in PNH patients,
while the presence of rare GPI-AP-deficient cells (<0.1%) in
AA and h-MDS is contributive for exclusion of IBMFS, predic-
tive for better response to immunosuppressive therapy (IST)
and lower incidence of transformation to MDS/AML [12].
3. GPI-AP-deficient clones can grow with time: approximately
15–20% of patients with BMFS could develop subclinical or
clinically relevant PNH clone and 20–30% of patients with
classical PNH could develop pancytopenia [13–15].
4. Current international recommendations for GPI-AP deficiency
testing by HS FCM comprise all cases with unexplained
isolated anemia, unclear iron deficiency anemia, unexplained
isolated neutropenia or thrombocytopenia, cytopenias, any
 Immunophenotyping of PNH 325

subtype of MDS with evidence of hemolysis, venous or arterial

thrombosis with atypical location (portal vein, hepatic vein,
mesenteric vein, splenic vein, sinus vein) and signs of hemolysis,
cytopenia or in young patients as well as cases with acquired
DAT-negative hemolysis, intravascular hemolysis or recurrent
abdominal pain or dysphagia. For patients with AA and no
detectable PNH phenotypes the current guidelines recom-
mend reassessment every 6 months for 2 years and yearly
thereafter as long as they remain negative for PNH phenotypes.
For patients with AA and detectable PNH phenotypes, re-
assessment is recommended every 3 months for 2 years, with
subsequent yearly testing if level of PNH phenotypes remains
stable [16, 17].
While the ability to rapidly detect GPI-deficient cells by flow
cytometry has led to improved diagnosis, patient management, and
prognosis in PNH and related disorders, early, simple CD55/
CD59 based approaches were neither accurate nor sensitive below
the 1–4% clone size, rendering them inadequate to detect small
PNH clones present in PNH+ AA and MDS cases [16, 18]. Since
40% of samples positive for the presence of PNH cells contain
GPI-deficient cells at a level of 1% or less (see Fig. 1 in [19]), the
development and validation of sensitive, standardized methodolo-
gies is essential to reliably detect small populations of GPI-deficient
PNH phenotypes. Once established, the frequencies of PNH phe-
notypes in red and white blood cell lineages in multiple normal
samples must be determined [18–20]. Additionally, the sensitivities
of the assays have to be determined by titrating (“spiking”) a PNH
sample into a normal sample [18–24]. Furthermore, there is much
variability between different antibody clones/conjugates used in
the analysis of PNH in RBC and WBC lineages. To successfully
develop highly sensitive, accurate and reproducible assays, careful
selection and titration of antibody clones/conjugates for lineage-
specific gating (RBCs, neutrophils, and monocytes) and specific
GPI-antigen detection within each cell lineage is required
[18, 23, 24].
The detection of PNH phenotypes by flow cytometry represents
a unique challenge in that the assays are designed to detect “nega-
tives,” that is, PNH phenotypes have “lost” the antigens of interest,
namely, glycophosphatidylinositol (GPI)-linked structures. This has
considerable implications for how instruments are set up and com-
pensated because the target GPI-negative PNH cell populations
(both RBCs and WBCs) need to be visible and fully on scale.
326 Andrea J. Illingworth et al.

2 Materials

1. Antibody clone/conjugate selection for high-sensitivity PNH

RBC assay
(a) Gating antibody: CD235a (glycophorin A) is the only
RBC-specific gating reagent available for delineating
RBCs from other events (WBCs, platelets, debris, etc.).
Since CD235a (in particular) conjugates can cause mas-
sive aggregation of RBCs when used at anything close to
saturating levels, careful selection and extensive titration is
required to minimize aggregation while retaining good
signal-to-noise characteristics of the assay. Screening of a
large number of CD235a clones/conjugates showed that
negatively charged FITC conjugates generated signifi-
cantly less aggregation than their PE-conjugated counter-
parts [18, 25].
(b) GPI-linked antibodies: While loss of the GPI-linked CD55
and CD59 structures was traditionally used to detect
PNH RBCs, CD55 is inferior to CD59 due to its dim
expression and inability to delineate Type II and Type III
PNH RBCs [16, 18, 22]. Thus, CD59PE conjugates offer
the best separation of Type I, Type II, and Type III RBCs.
Screening and titration of a large number of CD59PE
conjugates was used to identify the best-performing
CD59 reagents [18, 24]. A summary of recommended
CD235a and CD59 clones/conjugates is shown in
Table 1. To avoid the generation of false negatives, it is
highly recommended that “cocktails” of the chosen
CD235aFITC and CD59PE reagents be premixed and
diluted with clean PBS to aid the accurate addition of
reagents for this assay [18, 24].

Table 1
Recommended CD235a-FITC and CD59-PE conjugates for high-sensitivity PNH RBC assay

Target Antibody conjugates Purpose Clone and vendor

RBC CD235a-FITC Gating on RBC 10F7MN (eBio)
YTH 89.1 (Cedarlane)
KC16 (BC)
JC159 (DAKO)
CD59-PE GPI-linked for RBC OV9A2 (eBio)
MEM-43 (Invitrogen)
MEM-43 (EXBIO/Cedarlane)
 Immunophenotyping of PNH 327

2. Antibody clone/conjugate selection for WBC analysis

(a) Gating antibodies:
l CD45: The inclusion of a pan-CD45 conjugate in the
WBC staining cocktail is highly recommended. Along
with a light scatter plot, CD45 is highly effective at
removing subcellular debris, platelets and unlysed
RBCs and stains peripheral blood samples in a highly
predictable manner allowing for “pattern recognition.”
l Lineage-gating reagents—neutrophils and monocytes:
The development of high sensitivity assays is critically
dependent on the use of good lineage-specific “gating”
reagents for neutrophils and monocyte. In the past,
CD33 was widely used to “gate” both neutrophils
and monocytes [25] but for a variety of reasons was a
suboptimal choice for high sensitivity detection of
PNH neutrophils and monocytes [18]. As detailed
elsewhere [18, 23, 24, 26], appropriately selected, vali-
dated, and titrated conjugates of CD15 and CD64
outperform the ability of CD33 conjugates to accu-
rately delineate neutrophils and monocytes. Occasion-
ally, CD64 may be upregulated in activated neutrophils
making the separation between monocytes and neutro-
phils challenging in such cases.
l GPI-linked reagents—neutrophils and monocytes: As
outlined previously [16, 18, 22–24], two GPI-linked
structures must be analyzed per WBC lineage assessed.
Combinations of FLAER and CD24, or FLAER and
CD157 represent the most tested combination of
reagents to detect GPI deficient neutrophils, while
FLAER in combination with either CD14 or CD157
represent the most validated combination of reagents
to detect GPI-deficient monocytes [23, 24].

After much empirical testing of multiple clones and

instrument-specific conjugates, recommended clones/conjugates
of GPI-specific reagents for use on instruments equipped with 4-,
5-, and 6- (or more) PMTs are shown in Tables 2, 3 for Beckman
Coulter and BD Biosciences platforms respectively [24], and in
Table 4 for recently developed cross-platform assays [27].

3 Methods

3.1 Instrument 1. Standard setup instructions from both major manufacturers are
Setup; Key Points optimized for specific applications and do not transfer directly
to PNH testing in either red or white cell assays.
328 Andrea J. Illingworth et al.

Table 2
Recommended clones/conjugates for high-sensitivity detection of PNH WBC on Beckman Coulter
cytometers

Target Antibody conjugates Purpose Clone and vendor

WBC FLAER-Alexa 488 GPI-linked NA (Cedarlane)
(Neuts + Monos)
CD24-PE GPI-linked (Neuts) SN3 (eBio), ALB9 (BC)
CD24-APC SN3 (eBio, EXBIO)
CD14-PE GPI-linked (Monos) 61D3 (eBio), RMO52 (BC)
CD14-APC700 Tuk4 (Invitrogen) RMO52 (BC)
CD157-PE GPI-linked SY11B5
(Neuts + Monos) (eBio, EXBIO, BD, BC, Sysmex)
CD64-PC5 Gating on Monocytes 22 (BC)
CD64-ECD 22 (BC)
CD64-PC7 22 (BC), 10.1 (EXBIO)
CD15-PC5 Gating on Neutrophils 80H5 (BC)
CD15-PerCP-eF710 MMA (eBio)
CD15-PerCPCy5.5 MEM158 (EXBIO)
CD45-PC7 Debris/unlysed RBC exclusion + J33 (BC)
CD45-KO pattern recognition J33 (BC)
CD45-eF450 2D1 (eBio)

Table 3
Recommended clones/conjugates for high-sensitivity detection of PNH WBC on Becton Dickinson
cytometers

Target Antibody Conjugates Purpose Clone and vendor

WBC FLAER-Alexa 488 GPI-linked (Neuts + Monos) NA (Cedarlane)
CD24-PE GPI-linked (Neuts) SN3 (eBio), ML5 (BD)
CD24-APC SN3 (eBio, EXBIO)
CD14-PE GPI-linked (Monos) 61D3 (eBio),
CD14-APC Tuk4 (Invitrogen)
MoP9 (BD)
CD157-PE GPI-linked (Neuts + Monos) SY11B5 (eBio, EXBIO, BD, BC,
Sysmex)
CD64-APC Gating on Monocytes 10.1 (BD, eBio)
CD64-PECy7 10.1 (EXBIO), 22 (BC)
CD15-APC Gating on Neutrophils HI98 (BD)
CD15-PerCP-eF710 MMA (eBio)
CD15-PerCPCy5.5 MEM 158 (EXBIO)
CD45-eF450 Debris/unlysed RBC exclusion 2D1 (eBio)
CD45-PerCP + pattern recognition 2D1 (BD)
CD45-APC-H7 2D1 (BD)
 Immunophenotyping of PNH 329

Table 4
Recommended clones/conjugates for cross-platform three-laser B-C Navios/BD Canto II High
sensitivity PNH WBC 5-, 6-, and 7-color assays

Target Antibody conjugates Purpose Clone and vendor

WBC FLAER-Alexa 488 GPI-linked (Neuts + Monos) NA (Cedarlane)
CD157-PE (all) GPI-linked (Neuts + Monos) SY11B5
(eBio, EXBIO, BD, BC, Sysmex)
CD24-APC (6C, 7C) GPI-linked (Neuts) SN3 (eBio)
CD14-APCeF780 (7C) GPI-linked (Monos) 61D3 (eBio),
CD64-APC (5C) Gating on Monocytes 10.1 (BD, eBio)
CD64-PECy7 (6C, 7C) 10.1 (EXBIO), 22 (BC)
CD15-PerCP-eF710 (all) Gating on Neutrophils MMA (eBio)
CD15-PerCPCy5.5 (all) MMA (eBio)
CD15-PerCPCy5.5 (all) MEM 158 (EXBIO)
CD45-eF450 (all) Debris/unlysed RBC exclusion + 2D1 (eBio)
pattern recognition

2. FLAER is an Alexa 488 conjugate and specific compensation

settings are required for this conjugate. Settings optimized for
FITC must not be used because of the different spillover of
Alexa 488 and FITC.
3. On instruments equipped with biexponential or logicle displays
it is important to ensure “negative cell populations,” the very
targets of testing in PNH (see below) are fully on scale with
median channels above the zero axis. When optimally set up for
the detection of PNH phenotypes, such populations should
appear similar when viewed in either logarithmic or biexponen-
tial/logicle mode.
4. The appropriate target population must be used for PMTv
settings. For the red cell assay, unstained red blood cells are
appropriate; for the white cell assay, it is important to ensure
the unstained lymphocyte population is clearly on scale.

3.2 Sample For instruments that do not automatically save “TIME” as a param-
Preparation eter, this parameter must be collected so that fluidics can be moni-
and Instrument Setup tored if needed. Instrument setup for the RBC assay is performed
for RBC Analysis using a fresh normal blood sample, diluted 1:100 with clean Phos-
phate buffered saline (PBS). Light scatter voltages, photomultiplier
tube voltages (PMTv) and compensation are optimized as detailed
in several previous publications [18, 23, 24, 28], and recent exam-
ples can be viewed in Figs. 1 and 2 of [24]).

3.3 Recommended Anti-coagulated (EDTA is preferred, but heparin can also be used)
Staining Protocol peripheral blood samples should be less than 48 h old. If possible,
for RBC these samples should be kept refrigerated for this may extend the
stability of these samples.
330 Andrea J. Illingworth et al.

1. Prepare a 1:100 dilution of the sample with clean PBS.

2. Pipet 100 μl of diluted blood sample into the bottom of a test
tube using reverse-pipetting to avoid aerosol generation, and
remove pipette carefully to avoid leaving blood trails on the
inside of the tube.
3. Prepare CD235aFITC/CD59PE antibody cocktail as
described in previous publications [18, 24] Note that specific
clones should be validated/titrated by each laboratory to con-
firm optimal performance.
4. Pipet appropriate cocktail amount directly into the diluted
blood sample and mix by gently swirling the sample using a
low speed vortex. Note: For the “immature RBC (iRBC)”
staining protocol, 1 μl of CD71-APC (clone MEM-75,
EXBIO) is also added at this point.
5. Incubate the sample in the dark for 20 min at room tempera-
ture (incubation up to 60 min has been validated).
6. Wash twice with PBS by centrifugation and resuspend in
0.5–1.0 ml of PBS.
7. “Rack” the sample (drag across a hard plastic or metal test tube
rack several times) to disrupt any RBC aggregates generated by
the staining/washing procedure immediately before acquisi-
tion on the cytometer.
8. It is generally good to acquire sample immediately, as the
binding of some CD235a conjugates has been shown to be
unstable after washing/racking. Delay can also allow for the
reformation of some RBC aggregates.
9. Acquire a minimum of 100,000 RBCs (gated on FS log versus
CD235a. If two or more events were displayed in the Type III
PNH RBC region, data acquisition should be continued until
one million events are acquired.

3.4 Sample For instruments that do not automatically save “TIME” as a param-
Preparation eter, this parameter must be collected so that fluidics can be moni-
and Instrument Setup tored if needed. To analyze white blood cells, 100 μl of
for WBC Analysis anticoagulated peripheral blood is stained with the reagent cocktail
of choice and subject to an RBC lysis step. Regardless of lysing agent
used, the cells need to be washed thereafter with phosphate buffered
saline supplemented with 1% bovine serum albumin before acquisi-
tion. A multistep Boolean gating approach has been adopted to
efficiently gate neutrophils and monocytes that begins with setting
light scatter voltages in linear mode to ensure that all cells of interest
(neutrophils and monocytes for PNH “clone” detection and lym-
phocytes used as internal staining controls) are properly on-scale and
above the threshold/discriminator, used to remove subcellular
debris. Light scatter voltages and PMTv are optimized as detailed
in several previous publications [18, 23, 24, 28].
 Immunophenotyping of PNH 331

3.5 Establishing Light scatter voltages, photomultiplier tube voltages (PMTv) and
the Compensation compensation for the PNH WBC assay are optimized as detailed in
Matrix several previous publications [18, 23, 24, 28] and examples are
shown in Figs. 6 and 7 for a FLAER-based 6-color assay on a
10-color Navios cytometer.

3.6 Recommended All individual antibodies were verified for appropriate reactivity
Staining Protocol with target cells and titrated to optimize specific staining perfor-
for Five- and Six-Color mance prior to being cocktailed for use in the single tube five- or
WBC Assays six-color assays. It is recommended that samples be stained fresh or
(Neutrophils after a maximum storage time of 48 h at 4
C [19, 23, 24].
and Monocytes) 1. Using reverse-pipetting, pipet 100 μl of fresh peripheral blood
carefully into the bottom of a test tube without touching the
side of the tube.
2. Add an appropriate volume of premixed and validated cocktail
directly to the blood aliquot in the bottom of the tube and mix
very gently but thoroughly as described above for the RBC
assay.
3. Incubate for 20 min in the dark (incubation times for up to
60 min have been validated).
4. Lyse the RBCs using an appropriate lysing agent. Immuno-
Prep, VersaLyse (BC), FACSLyse (BD) or ammonium
chloride-based lysing agents following the manufacturer’s
recommendations. While all are acceptable, those containing
fixatives may help retain cellular integrity better than those that
do not.
5. After lysing, wash cells once with PBS, resuspend in 1 ml of
PBS and acquire on the cytometer. Note: Samples should be
acquired immediately, as delays can cause light-scatter changes,
especially when fixative-free lysing agents are used.
6. Acquire a minimum of 50,000 neutrophils and 5000 mono-
cytes in listmode for clinical test samples.
7. If small numbers of GPI-deficient cells are observed, acquisi-
tion times should be increased until a statistically reliable num-
ber of “PNH phenotypes” are acquired (20–50 cells).

4 Analysis

A correct diagnosis of paroxysmal nocturnal hemoglobinuria

(PNH) is essential for effective patient management and possible
treatment if clinically indicated [10, 11, 16]. Although flow cyto-
metry has become the gold standard for the diagnosis and moni-
toring of patients with PNH and related diseases [16, 23], it is
important to emphasize that flow cytometry by itself does not
establish the clinical diagnosis of overt/clinical PNH as this is
332 Andrea J. Illingworth et al.

determined by a number of other clinical factors in addition to the

finding and quantification of glycosylphosphatidylinositol (GPI)-
deficient cells by flow cytometry [1–3]. HS FCM analysis can
identify the presence of PNH phenotypes and is able to quantify
the PNH clone size in RBCs and WBCs (both neutrophils, and
monocytes) down to a lower limit of quantification (LLOQ) of
0.01% for RBC and 0.05–0.1% for neutrophils. The PNH clone size
in WBCs is typically used to determine the extent of disease. His-
torically, the neutrophil clone size was used for this purpose based
on the absence of at least two GPI-linked structures from neutro-
phils. However, more recently many laboratories have noted the
additional value of assessing and reporting the PNH clone size in
the monocytes as well [19].

4.1 RBC Analysis For high-sensitivity RBC analysis, most previously published
Guidelines recommend the use of CD235a-FITC for lineage-
specific RBC gating and CD59-PE to detect GPI-deficient cells
[18, 24]. RBCs are analyzed by a series of gating dot plots and
diagnostic dot plots:
1. TIME (x axis) versus SS log (y axis) dot plot. TIME is collected
as a parameter (as is recommended for all clinical flow assays)
and monitored during acquisition so that if fluidics problems
are encountered, the sample can be reacquired if possible, or if
not, data acquired prior to the fluidics hiatus can be “gated”
and only that portion of the data file subsequently analyzed
(Fig. 1, top row left).
2. FS log (x axis) versus SS log (y axis) dot plot to gate out debris
(Fig. 1, top row middle). It is important to adjust the threshold
(discriminator) for the forward scatter so that no RBCs are
excluded from acquisition. Make sure that gate includes the
RBC cluster and excludes any cellular debris on the left.
3. CD235a-FITC (x axis) versus FS (y axis) dot plot to gate singlet
RBCs and to quantify and exclude any remaining RBC aggre-
gates (Fig. 1, top row right).
4. CD59-PE (x axis) versus CD235a-FITC (y axis) dot plot (Fig. 1,
bottom row left). Check the Type III area for cells, which are
CD235a brightly positive/negative for CD59 and record the
percentage. Check the Type II area for cells, which are also
brightly positive for CD235a with intermediate CD59 expres-
sion and record the percentage.
5. CD59-PE (x axis) versus CD235a-FITC (y axis) density plot
(Fig. 1, bottom row middle). Use this plot if the separation is
not clear based on the previous dot plot.
6. CD59-PE single parameter histogram (Fig. 1, bottom row
right). Bivariate dot plots and/or density plots are typically
recommended over single-parameter histograms, especially
 Immunophenotyping of PNH 333

Fig. 1 Normal patient with no PNH RBC clone. Sample stained with CD235a-FITC/CD59-PE

for samples containing small numbers of PNH phenotypes, for

identifying poorly stained samples that need to be restained,
and for detecting media contamination and troubleshooting
instrumentation issues (see Supplementary Data of [18]). How-
ever, while data regarding clone sizes comes predominantly
from the two-dimensional plots, in which the gating regions
are linked across the dot plot and density plots, the single
parameter histogram can also be useful in some situations,
and when comparing RBC reagent cocktails lot-to-lot. All
three plots work in concert for optimal adjustment of the
regions for Type III PNH cells and Type II PNH cells (see
Fig. 2).

4.2 Separation In most cases the PNH clusters are evident and the initial region
Between Normal Type settings can be used for the quantification of the Type II and Type
I RBC, PNH Type II III PNH RBC based on a combination of bivariate dot plots,
and PNH Type III RBCs density plots and the single parameter CD59 histogram (Fig. 2).
However, in some patients the regions may need to be adjusted
(dynamic region setting) based on slight patient-specific differences
in the CD59 expression rather than strictly adhering to the initial
region setting. Occasionally there is no clear distinction (clustering)
between PNH Type II and PNH Type III or between PNH Type II
and normal Type I cells which renders the interpretation
334 Andrea J. Illingworth et al.

Fig. 2 Patient with PNH RBC Type II and PNH RBC Type III clone. Sample stained with CD235a-FITC/CD59-PE

challenging and the PNH clone size assessment and reporting

potentially problematic. In such cases it may be appropriate to
include a descriptive comment in the report and state that the
percentages provided are an estimate only. In cases with no separa-
tion between Type III PNH cells and Type II PNH cells, it is
recommended to report a combined population of RBCs with
CD59 deficiency rather than setting arbitrary regions for Type II
and Type III. Patients with poor separation between PNH Type II
and normal Type I RBCs are especially challenging as the region
setting between these two populations will affect the total com-
bined PNH clone size. Reasons for this poor separation may
include the presence of microcytes (smaller RBCs) in certain
patients, transfusion (transfused RBCs typically have slightly
decreased CD59 expression) and other unexplained decrease in
CD59 expression. In these cases, a static region setting (based on
the initial region setting) would not be optimal and a more flexible
region setting based on the delineation of cell clusters is recom-
mended. Considering the challenges of setting gating regions for
accurate assessment of normal Type I RBC, PNH Type II, and
PNH Type III populations, the addition of CD71, that stains
immature RBCs (iRBC) allows for a much more objective assess-
ment (see below).
 Immunophenotyping of PNH 335

4.3 Assessment We have recently performed a study of over 25 large-clone PNH

of PNH cases with the regular two-color RBC assay supplemented with
in immature RBCs CD71-APC [29]. This assay can be performed on any cytometer
(iRBCs) equipped with blue and red lasers. While mature RBCs can be
assessed for PNH clone content, the CD71 is used to gate imma-
ture RBCs (iRBC). The latter can then be assessed for PNH iRBC
clone content. The results demonstrated that the CD71-based
RBC assay is better able to delineate Type III, Type II and normal
iRBCs in PNH cases with WBC clones of 5% or larger. Total iRBC
clone sizes (Type II plus Type II) were in almost all cases, very
similar to the WBC clone sizes (whether measured in neutrophils or
monocyte lineages). The assay retained the ability to detect minor
PNH RBC clones in other BMFS cases in which the lack of suffi-
cient reticulocytes does not allow for the assessment of the latter.
1. Figure 3 shows the advantage of the iRBC assay in a case in
which delineating PNH Type III and Type II RBCs from
normal cells was impossible. However, the use of CD71 to
gate iRBCs (bottom left plot) allowed for the delineation of
the Type III (7.4%) and Type II PNH RBCs (83.7%) from the
normal RBCs. The WBC clone size in this sample was 94.2%
and 94.8% for neutrophils and monocytes respectively (data
not shown) compared to the total iRBC PNH content of
91.1%.
2. In the second example (Fig. 4), from another patient with a
large WBC clone who was receiving regular RBC transfusions,
there were relatively easy-to-delineate populations of Type III
and Type II PNH RBCs although the Type II cells were not
well separated from the normal RBCs (middle left plot). How-
ever, analysis of the CD71+ iRBC population demonstrated
virtually all the PNH iRBCs (lower left) to be of the Type III
phenotype (93.5%) with very few (0.7%) Type II iRBCs. Thus,
even in some cases in which the PNH content of mature RBCs
can apparently be relatively easy to analyze with confidence, the
assay is in fact generating totally misleading data.
3. In a third example shown in Fig. 5 (top row), large WBC clones
were detected (99.25% and 99.03% in neutrophils and mono-
cytes respectively) while no discernable PNH clones could be
detected in mature RBCs (lower left plot). Only when the Type
II gate was moved (lower middle plot) to include the major
population of RBCs (that expressed slightly lower levels of
CD59 staining than bona fide normal Type I cells, data not
shown), was it apparent that in this very unusual case, the Type
II RBCs expressed almost as much CD59 as normal RBCs in a
non-PNH sample. This interpretation was confirmed by the
analysis of CD71+ iRBCs (lower right plot) in which the iRBCs
also expressed CD59 at slightly lower levels than those from
non-PNH samples (not shown). This example again highlights
336 Andrea J. Illingworth et al.

Fig. 3 PNH sample containing 94.2% PNH neutrophils and 94.8% PNH monocytes (data not shown) stained
with CD235aFITC/CD59-PE/CD71-APC

the utility of the CD71-based RBC assay and also underscores

the necessity of performing flow analysis on both WBC and
RBC lineages in the detection of bona fide PNH phenotypes.
Overall, this study demonstrated (Figs. 3, 4, 5) that in PNH
patients with large PNH clones, it is not possible to derive accurate
Type III and Type II clone sizes from the mature (CD235a+) RBC
component. Attempts to do so even by experts are often futile and
can lead to the generation of completely misleading data. Mislead-
ing Type II RBC data is the biggest culprit and even when it appears
straightforward to “gate” Type II RBCs (Fig. 4), such populations
are usually artifacts with prior transfusions representing a major
source of “false” Type II RBC phenotypes. In the majority of
 Immunophenotyping of PNH 337

Fig. 4 PNH sample containing 97.3% PNH neutrophils and 99.6% PNH monocytes (data not shown) stained
with CD235aFITC/CD59-PE/CD71-APC

large-clone PNH cases, Type III PNH iRBCs are the predominant
PNH population although striking examples of predominantly
Type II PNH iRBCs can also be observed (Fig. 5).
This study has also shown that if analysis of the red cell lineage
is to be clinically useful, it needs to be accurate. For large-clone
PNH cases, accurate data cannot be derived from mature RBCs
(due to transfusion and hemolysis). For large-clone PNH cases,
accurate data can be derived only from immature RBCs (iRBCs).
For small clone PNH cases (clone sizes <1%) and BMFS cases
containing small numbers of PNH RBC phenotypes, it is difficult
to collect enough iRBCs for statistically precise data. However, in
the latter, sufficient mature PNH RBCs can usually be identified to
generate reliable and precise data.
338 Andrea J. Illingworth et al.

Fig. 5 PNH sample containing 99.25% PNH neutrophils (top right plot) and 99% PNH monocytes (top left plot)
stained with CD235aFITC/CD59-PE/CD71-APC700

4.4 WBC Analysis High-sensitivity methodologies to detect PNH phenotypes in neu-

trophils and monocytes have been published extensively [18, 22,
23]. Initial methods were based on a four-color neutrophil tube
(FLAER, CD24, CD15, and CD45) with a reflex monocyte tube
(FLAER, CD14, CD64, and CD45). Newer flow cytometers with
5, 6, or more PMTs allow for the simultaneous detection and
quantification of both neutrophils and monocytes in a single tube.
Currently recommended WBC panels based on established PNH
guidelines include a FLAER/CD24/CD14-based panel and a
FLAER/CD157-based panel [19, 24].
1. 6-color FLAER/CD24/CD14-based single tube Assay for Neu-
trophils and Monocytes: Fig. 6 shows the sequence of bivariate
gating dot plots from a Navios-specific reagent set comprising
FLAER-Alexa 488, CD24-PE (clone ALB9), CD15-PC5
(clone 80H5), CD64-PC7 (clone 22), CD14-APCA700
(clone RMO52), and CD45-KO (clone J33). In this panel
(Figs. 6 and 7) neutrophils are gated on CD45 and CD15
and a PNH clone can be detected based on the absence of the
GPI markers FLAER and CD24. Monocytes are gated on
CD45 and CD64 and a PNH clone can be detected based on
the absence of FLAER and CD14.
 Immunophenotyping of PNH 339

Fig. 6 FLAER/CD24/CD14-based WBC assay with no PNH clone in neutrophils and monocytes. Sample stained
with FLAER-Alexa 488/CD24-PE/CD15-PC5/CD64-PC7/CD14-APC700/CD45-KrO

(a) TIME (x axis) versus CD45 or SS (not shown) on y axis dot

plot with a “TIME” gate to confirm that no fluidics irre-
gularities are present. This time gate can be adjusted
horizontally to include only the events without fluidics
issues if necessary.
(b) Optional dot plot containing FS INT/FS W (x axis) versus
FS peak/FS H (y axis) with a singlet gate (not shown) to be
gated on in subsequent plots. (This is not mandatory but
may be useful if a large amount of cellular aggregates is
present in the sample.)
(c) FS linear (x axis) versus SS linear (y axis) dot plot allows for
confirmation of all WBC clusters and debris exclusion. All
WBC subsets should be clearly visible and optimally sepa-
rated. There are two approaches, either a “live” gate to be
340 Andrea J. Illingworth et al.

Fig. 7 FLAER/CD24/CD14-based WBC assay with PNH clone present in neutrophils and monocytes. Sample
stained with FLAER-Alexa 488/CD24-PE/CD15-PC5/CD64-PC7/CD14-APC700/CD45-KrO

gated on in subsequent plots (shown in Fig. 8) or a

“debris” exclusion gate (shown in Figs. 6 and 7) which
can be used to gate out the debris in subsequent plots).
Make sure that the debris gate excludes the debris (plate-
lets and degenerating granulocytes).
(d) CD45 (x axis) versus SS (y axis) dot plot with a “CD45+”
gate. Set gate around all CD45+ cell populations (lym-
phocytes, monocytes, and granulocytes) and exclude
CD45-negative events. Gate all subsequent plots on this
CD45+ gate. The CD45 vs. SS plot is not only useful for
pattern recognition but also useful for excluding any
unlysed RBCs and other debris not removed by the debris
exclusion gate in the light scatter plot.
 Immunophenotyping of PNH 341

Fig. 8 CD157 and FLAER-based cross-platform assay with PNH clone present in neutrophils and monocytes.
Sample stained with FLAER-Alexa 488/CD157-PE/CD15PerCPeFluor710/CD64-PECy7/CD45-eFluor450

(e) CD15 (x axis) versus SS (y axis) dot plot is gated on the

CD45+ populations and includes a gate drawn around the
CD15++ neutrophils excluding as well as possible the
CD15 dim+ eosinophils visible to the left of the neutro-
phil population.
(f) FLAER (x axis) versus CD24 (y axis) dot plot is gated on
the CD15++ neutrophils, and a region is drawn to encom-
pass the FLAER-negative/CD24-negative cells, which
represent the PNH neutrophils. In most cases, the PNH
clone in neutrophils is composed of the so-called Type III
neutrophils (complete absence of GPI markers FLAER
and CD24), but occasionally there may be “Type II”
neutrophils (intermediate staining for FLAER and
CD24). It is recommended to include both Type III
and Type II neutrophils into one combined PNH
neutrophil gate.
(g) CD64 (x axis) versus SS (y axis) dot plot is gated on the
CD45+ cells, and a region is drawn around the CD64++
monocytes.
342 Andrea J. Illingworth et al.

(h) FLAER (x axis) versus CD14 (y axis) dot plot is gated on

the CD64++ monocytes, and a region is drawn to delin-
eate the FLAER-negative/CD14-negative cells, which
represent the PNH monocytes. As previously described
for the neutrophils, typically the PNH clone in monocytes
is composed of the so-called Type III monocytes (com-
plete absence of GPI markers FLAER and CD14), but
occasionally there may be “Type II” monocytes (interme-
diate staining for FLAER and CD14). It is recommended
to include both Type III and Type II monocytes into one
combined PNH monocyte gate. Make sure to exclude the
monocytes with bright FLAER but scattered decreased
CD14 expression; these events may represent dendritic
cells or immature monocytes and should not be included
into the PNH clone gate for monocytes.
2. Internal controls for six-color FLAER/CD24/CD14-based
panel: Optimal instrument setup and standardization in terms
of PMT voltage optimization, daily monitoring and computer
assisted spectral overlap compensation is a prerequisite for a
validated flow cytometry assay. The use of a normal control
sample is necessary as an external control to verify the assay
performance for PNH RBC and WBC testing. Lymphocytes
are not a suitable target population for the PNH clone quanti-
fication due to their long life span, but they serve as internal
controls for verification of antibody staining and compensation
settings in the patient sample. For protocols using FLAER in
combination with CD24 for neutrophils and CD14 for mono-
cytes as GPI-specific markers, the following plots should be
created and used for analysis:
(a) FLAER (x axis) versus CD24 (y axis) dot plot, which is
gated on lymphocytes based on lymphocyte gate set on
the CD64 vs. SS plot. This allows for the clear distinction
of T lymphocytes (FLAER++/CD24), NK cells
(FLAER+/CD24), and B lymphocytes (FLAER++/
CD24+). Figure 6 demonstrates these internal controls
and also confirms that no PNH lymphocytes (FLAER/
CD24) are seen in this normal sample. Figure 7 shows
the same Boolean gating strategy on a PNH patient with
the diagnostic plots clearly demonstrating the presence of
PNH phenotypes in both neutrophil and monocyte
lineages. The lymphocytes (internal controls) also show a
much smaller PNH clone in the dual-negative region,
thus verifying that the instrument settings and compensa-
tion are optimally set.
(b) FLAER (x axis) versus CD14 (y axis), CD15 (y axis) and
CD64 (Y axis) plots (Figs. 6 and 7, bottom row) verifies
the instrument settings with all subsets visible on scale and
clustered in the expected locations.
 Immunophenotyping of PNH 343

3. 5-color FLAER/CD157 based single tube cross-platform assay for

neutrophils and monocytes: In this panel neutrophils are gated
on CD45 and CD15 and monocytes are gated on CD45 and
CD64. A PNH clone can be detected in both the neutrophils
and monocytes based on the absence of the GPI markers
FLAER and CD157. CD157 is highly expressed on both
mature neutrophils and monocytes [30] leading to the possi-
bility that CD157 could replace both CD24 and CD14, allow-
ing for the development of a single tube, high sensitivity 5C
assay to identify and quantify both PNH neutrophils and PNH
monocytes on cytometers with five or more PMTs [31]. It is
important to note that several CD157-negative, non-PNH
cases have been observed in the authors’ laboratories
[23, 24]. For these rare cases, the inclusion of the second
GPI reagent (FLAER) as part of the built-in robustness of
the assay prevents the misinterpretation of the data as a PNH
clone-containing sample; the loss of two GPI-linked structures
is required to identify bona fide PNH phenotypes. Further-
more, in keeping with current state-of-the-art guidelines
[19, 22–24] the RBC lineage should also be analyzed on
every sample tested for the presence of PNH WBCs. As these
rare CD157-negative non-PNH samples only contain normal
(Type I) RBCs, there is even less chance of misinterpretation.
For protocols using FLAER in combination with CD157 for
neutrophils and monocytes as GPI-specific markers, the follow-
ing plots should be created and used for analysis:
(a) TIME (x axis) versus CD45 or SS (not shown) on y axis with
a “TIME” gate to confirm that no fluidics irregularities
are present. This time gate can be adjusted horizontally to
include only the events without fluidics issues if necessary.
(b) Optional dot plot containing FS INT (x axis) versus FS peak
(y axis) with a singlet gate (not shown) to be gated on in
subsequent plots. (This is not mandatory but may be
useful if a large amount of cellular aggregates is present
in the sample.)
(c) FS linear (x axis) versus SS linear (y axis) dot plot allows for
confirmation of all WBC clusters and debris exclusion. All
WBC subsets should be clearly visible and optimally sepa-
rated. There are two approaches, either a “live” gate to be
gated on in subsequent plots (shown for this panel) or a
“debris” exclusion gate (as previously shown in Figs. 6
and 7) which can be used to gate out the debris in
subsequent plots). Make sure that the debris gate excludes
the debris (platelets and degenerating granulocytes).
(d) CD45 (x axis) versus SS (y axis) dot plot with a “CD45+”
gate. Set gate around all CD45+ cell populations
344 Andrea J. Illingworth et al.

(lymphocytes, monocytes, and granulocytes) and exclude

CD45-negative events. Gate all subsequent plots on this
CD45+ gate. The CD45 vs. SS plot is not only useful for
pattern recognition but also useful for excluding any
unlysed RBCs and other debris not removed by the debris
exclusion gate in the light scatter plot.
(e) CD15 (x axis) versus SS (y axis) dot plot is gated on the
CD45+ populations and includes a gate drawn around the
CD15++ neutrophils excluding as well as possible the
CD15 dim+ eosinophils visible to the left of the neutro-
phil population.
(f) FLAER (x axis) versus CD157 (y axis) dot plot is gated on
the CD15++ neutrophils, and a region is drawn to encom-
pass the FLAER-negative/CD157-negative cells, which
represent the PNH neutrophils (Fig. 8, bottom left). In
most cases, the PNH clone in neutrophils is composed of
the so-called Type III neutrophils (complete absence of
GPI markers FLAER and CD24), but occasionally “Type
II” neutrophils (intermediate staining for FLAER and
CD157) may also be seen (as shown in Fig. 8). It is
recommended to include both Type III and Type II neu-
trophils into one combined PNH neutrophil gate.
(g) CD64 (x axis) versus SS (y axis) dot plot is gated on the
CD45+ cells, and a region is drawn around the CD64++
monocytes.
(h) FLAER (x axis) versus CD157 (y axis) dot plot is gated on
the CD64++ monocytes, and a region is drawn to delin-
eate the FLAER-negative/CD157-negative cells, which
represent the PNH monocytes (Fig. 8, bottom middle).
As previously described for the neutrophils, typically the
PNH clone in monocytes is composed of the so-called
Type III monocytes (complete absence of GPI markers
FLAER and CD157), but occasionally there may be
“Type II” PNH monocytes (intermediate staining for
FLAER and CD157). It is recommended to include
both Type III and Type II monocytes into one combined
PNH monocyte gate. In some cases, the FLAER/CD157
expression of Type II monocytes may be almost as bright
as the staining for normal monocytes and careful interpre-
tation and knowledge of the position/staining of normal
monocytes is necessary to separate PNH Type II mono-
cytes from normal monocytes.
4. Internal controls for FLAER-CD157-based panel: For protocols
including FLAER in combination with CD157 for both neu-
trophils and monocytes as GPI specific markers, the FLAER
versus CD157 dot plot, gated on lymphocytes can be less
 Immunophenotyping of PNH 345

informative, since CD157 is not expressed on mature T and B

lymphocytes. However, basophils are usually present in gated
lymphocyte populations with this assay and they provide an
internal control in situations where PNH clone sizes in neu-
trophils and monocytes approach 100%. Meanwhile, the delin-
eation of PNH and non-PNH lymphocytes based on FLAER
staining still provides reliable evidence of proper instrument
setup (PMT voltage and compensation) and antibody specific-
ity (Fig. 8). Gated lymphocytes are displayed on control plots
of FLAER versus CD157, CD15 and CD64 (far right plots) to
demonstrate optimal instrument setup and compensation has
been established. Technical details of the assay design and
reagent selection can be found in [31], and for cross-platform
variants in [27].

5 Minor PNH Clones

Regardless of lineage, when less than 1% of the target population is

GPI-deficient (see Fig. 9), PNH clone sizes have been historically
referred to as “minor” [22]. Such small populations of
GPI-deficient cells are often encountered in patients with AA and
some subsets of MDS [16]. Clinically, these patients do not usually
show symptoms of hemolysis but the presence of minor clones in
AA has been associated with better response to immunosuppressive
therapy [15]. AA patients must be monitored at specific intervals
for possible clone size expansion to identify patients, who progress
to clinical PNH [17]. From a technical perspective, the same PNH
RBC and WBC reagent panels can be used to screen all patient
samples. However, for those containing only small numbers of
PNH phenotypes, the number of events acquired will need to be
increased depending on the sensitivity of the assay as established by
the laboratory. The generally accepted smallest number of events
required to reproducibly detect a PNH population and determine

Fig. 9 Minor clone in all three lineages showing a quantifiable PNH clone (>50 PNH cells)
346 Andrea J. Illingworth et al.

the limit of detection (LOD) is 20 PNH events, the generally

accepted smallest number of events required to reproducibly quan-
tify a PNH population and determine lower limit of quantification
(LLOQ) is 50 PNH events. Lower levels should be validated in
each laboratory.
The characteristics of quasi-quantitative assays and the different
relevant parameters are described in more detail in the ICCS/
ESCCA PNH Consensus Guidelines, data analysis [19] and valida-
tion section [20].

6 Reporting

Based on the recent ICCS/ESCCA PNH consensus guidelines

[19], the following components are recommended for a PNH
report:
1. Report if a PNH clone present or absent. It is important to be
clear and to avoid potentially misleading ambiguous terminol-
ogy. A report stating that a CD59 test is “negative” may imply
to some providers that the target population is negative for the
GPI marker CD59 (thus indicating a PNH clone) or that no
CD59 absence is seen (thus indicating the absence of a PNH
clone).
2. Report the PNH clone size in the RBCs (total PNH clone size as
well as the percentages for Type II and Type III PNH popula-
tions). There is a clinical significance associated for Type II and
Type III RBCs. Type I RBCs are normal red blood cells with
bright CD59 expression and a lifespan of approximately
120 days. Type III PNH RBCs have complete CD59 defi-
ciency, which results in no protection from complement-
mediated lysis and a shortened lifespan of 10–15 days. Type
II PNH RBCs have partial CD59 deficiency resulting in partial
protection from complement mediated lysis. Just as the expres-
sion of CD59 on Type II RBCs varies considerably from
patient to patient, the lifespan of Type II cells reflects this
being intermediate between Type I normal RBCs and Type
III PNH RBCs. Since the clinical significance of Type II PNH
RBCs and Type III PNH RBCs is well established, it is recom-
mended to report them separately and combined as the total
PNH RBC clone.
3. Report the PNH clone size in both lineages for the WBCs (neu-
trophils and monocytes). The PNH monocyte clone is often
larger than the neutrophils PNH clone and reporting only the
PNH neutrophil clone may underestimate the PNH clone size
in the WBCs. Neutrophils and monocytes may also show the
presence of Type II populations but the clinical and biological
significance of these populations has not been established at
 Immunophenotyping of PNH 347

this time. It is therefore recommended to report only the total

PNH clone size in the neutrophils and monocytes.
4. Interpretive terminology of reporting PNH clones based on CSLI
H52-A2 [21]:
(a) PNH population >1%: “PNH clone”.
(b) PNH population 0.1–1%: “minor population of PNH
cells” or “minor PNH clone”.
(c) PNH population <0.1%: “rare cells with GPI deficiency”
or “rare cells with PNH phenotype.
5. List all gating and diagnostic markers used for the PNH assay.
6. State the lower limit of quantification (LLOQ) for the neutro-
phil/granulocyte assay and the RBC assay on the report, stating
the recommended LLOQ of 0.05% or better for RBCs
(100,000 gated cells) and 0.1% or better for neutrophils
(50,000 gated cells). It is important to include this information
to the provider as an LLOQ of 1% means that the possibility of
a minor clone (less than 1%) cannot be excluded based on
this LLOQ.
7. Retesting recommendations: see current recommendations and
requirements based on International Guidelines for retesting
frequencies in PNH [16] and related diseases [17]. Examples
of informative PNH reporting templates are shown in reference
19 and Appendix A of its Supplementary Data. An interactive
reporting template that contains the above recommendations
has recently been developed under the auspices of the Canadian
PNH Network and can be downloaded at http://www.
pnhnetwork.ca/pnh-resources.

7 Assay Validation

The high-sensitivity PNH assay is a quasi-quantitative assay and the

intended use is to evaluate for the presence or absence of a PNH
clone with respect to LOD as well as to quantify the PNH clone size
if present with respect to LLOQ. The validation process should
meet criteria for accuracy, specificity, sensitivity, repeatability, repro-
ducibility and stability [32–34]:
1. The accuracy of a measurement, which refers to the closeness of
agreement between the average value of a large number of test
results and the true or accepted reference value [33] cannot be
directly determined for PNH assays, because of unavailable
reference sample. External quality assessment and/or inter-
laboratory comparison represent the only available option for
validation and mandatory step for ISO accreditation [35].
348 Andrea J. Illingworth et al.

2. The analytical specificity of the assay reflects the validation of all

antibodies/reagents and corresponding fluorochromes, which
was discussed in previous section.
3. The clinical specificity represents the ability to exclude abnor-
mal specimen, defined by true negatives/true negatives + false
positives and should be determined by assay of a series of
samples and scoring for abnormality in comparison to a suitable
reference method, such as clinical diagnosis [36].
4. The analytical sensitivity of the assay is determined by the limit
of blank (LOB) defined by the highest apparent signal detected
in replicates of a sample containing no measurand and the LOD
defined by the lowest level of measurand that can be reliably
distinguished from the LOB [37]. For the high-sensitivity
PNH assays the LOB could be determined by measuring a
few replicates of a few negative specimens run over a few
separate days and calculating the mean and standard deviation
(SD) according to the formula: LOB ¼ mean of blank + 1.645
SD of blank assuming, that 95% of negative values will be below
this limit. Typically, the LOB for well-established PNH assays is
<0.001% (<10 PNH phenotypes out of 1000,000 acquired
events). LOD is closely related but usually greater than LOB
and could be determined by measuring a few replicates of a few
negative specimens run over a few separate days and calculating
the mean and SD according to the formula: LOD ¼ mean of
blank + 2SD (3SD) of blank or by measuring a few replicates of
a few low positive specimens run over a few separate days and
calculating the SD according to: LOD ¼ LOB + 1.645 SD of
low positive. It is highly recommended for each lab to perform
a “spiking experiment” in order to evaluate the sensitivity of
the assay. This is performed by making a serial dilution of PNH
blood “spiked” into normal blood which provides various
levels of PNH clone sizes down to very low numbers of PNH
cells [21]. Alternatively, target LOD could be estimated by
measuring a few replicates of a few low positive specimens run
over a few separate days and calculating the reproducibility
(inter-assay imprecision) expressed as coefficient of variation
(CV%) or by confirming, that no more than 5% of the values
for a target LOD fall beyond the LOB. The generally accepted
smallest number of events required to reproducibly detect a
PNH population and determine LOD is 20 PNH events, lower
levels should be validated individually.
5. The functional sensitivity of the assay is determined by the
LLOQ, representing the lowest level of measurand that can
be reliably detected at predefined levels of bias and imprecision
[37]. LLOQ is usually greater than LOD and for PNH assays
could be determined by measuring a few replicates of a few
positive (near the expected LLOQ) specimens run over a few
 Immunophenotyping of PNH 349

separate days and calculating the reproducibility (inter-assay

imprecision) expressed as CV%, which should be acceptable at
levels below 10%. The generally accepted smallest number of
events required to reproducibly quantify a PNH population
and determine LLOQ is 50 PNH events, lower levels should
be validated in each laboratory.
6. The clinical sensitivity or the ability to detect an abnormal
specimen and distinguish from normal specimens, defined as
true positive/true positive + false negative should be deter-
mined by assay of a series of abnormal samples and scoring
for abnormality in comparison to a suitable reference method,
such as clinical findings [37].
7. The assay performance characteristics are determined by
repeatability (intra-assay imprecision) and reproducibility
(inter-assay imprecision). It is recommended to assay a few
replicates from at least 5 samples within a single analytical run
for repeatability and a few replicates from at least 5 samples in
separate analytical runs for reproducibility [38]. For confirma-
tion of good performance characteristics CV% below 10%
should be obtained for samples with more than 1% target
PNH cells, and below 20% for samples with minor clones
(<1%).
8. The validation of specimen, processed specimen and reagent
stability has been reviewed in previous sections.

8 Additional Options for PNH Testing

8.1 Reagent Since publication of the 2010 Guidelines [16] a large number of
Selection specific clones/conjugates have been assessed for lineage-specific
gating and detection of GPI-deficient neutrophils and monocytes.
1. Optimal titration ranges must be established for selected con-
jugates and validated both individually and in combination
with other reagents on both normal and PNH samples.
2. Using this information, high-sensitivity assays were developed
using four-color cocktails based on FLAER, CD24, CD15 and
CD45 (for neutrophils) and FLAER, CD14, CD64 and CD45
(for monocytes) and (Supplementary Data, 23, 24).
3. Platform-specific versions of these robust four-color assays
were developed for clinical cytometers with four or more
PMTs [18, 23] and were capable of successful deployment
beyond the laboratories of original developers [39]. These
assays and reagent sets were found to be highly reliable across
FC500 and FACSCalibur instrument platforms in Interna-
tional studies performed under the auspices of the UK
NEQAS organization using whole blood stabilized
samples [40].
350 Andrea J. Illingworth et al.

4. Subsequently, five-color single tube assays capable of simulta-

neous high-sensitivity detection of both PNH neutrophils and
PNH monocytes based on FLAER, CD157, CD15, CD64,
and CD45 were developed for a variety of clinical cytometers
with 5 or more PMTs [23, 31, 41], and again independently
validated [42].
5. Instrument-specific CD24/CD14-based 6-color [19, 24, 28,
39, 41] assays have been developed for clinical instruments
with six or more PMTs (Beckman Coulter Navios and BD
FACSCanto).
6. Most recently, 7-, 6-, and 5-C reagent sets have been developed
that can be analyzed on either B-C Navios or BD Biosciences
Canto II cytometers or on both [27]. Thus, for large reference
flow laboratories that are often equipped with instruments
from multiple manufacturers, a single reagent set can be
employed and analysis performed on whichever instrument
platform is available at the time. Because these reagent sets
utilize all three lasers in Navios and Canto II platforms, instru-
ment setup and compensation are extremely simple with such
an approach, and maximal separation of PNH WBC pheno-
types from their normal counterparts is thereby optimized. As
indicated above, it is advisable to evaluate both neutrophils as
well as monocytes as the latter may reflect the size of the PNH
clone more accurately in some cases (see Fig. 12 in [19]).

8.2 Instrument- 1. Note that cocktailing of reagents is very important for WBC
Specific Reagent assays, just as it is for the RBC assay. If reagents are added one
Cocktails for Four-, at a time, there is increased risk of failing to add individual
Five-, and Six-Color reagents, or adding too much or too little, especially where
WBC Assays very small volumes of some individual reagents need to be
accurately pipetted. Such errors run the risk of generating
“laboratory-induced PNH” if the GPI-specific reagents are
inadvertently omitted. For laboratories that run the assay infre-
quently, there is an increased requirement to ensure that the
reagents used in the assay are working as specified. In this
setting, it is recommended that the laboratory stain a normal
sample at least once per month to validate the performance of
all reagents used in the assay [18, 22].
2. As detailed elsewhere, the Practical Guidelines [18] defined
two 4-color reagent cocktails for the high sensitivity detection
of PNH neutrophils and PNH monocytes for both Beckman
FC500 and BD FACSCalibur platforms. Subsequent studies
using these assays/reagent sets have shown them to be repro-
ducible and robust using fresh samples across multiple labora-
tories [39] and in International studies performed with
stabilized whole blood samples under the auspices of the UK
NEQAS organization [40].
 Immunophenotyping of PNH 351

3. As also detailed elsewhere [23, 24, 31], the availability of a

bright PE conjugate of the SY11B5 CD157 clone allowed for
the development of a five-color (FLAER, CD157-PE, CD64-
ECD, CD15PCy5, CD45-PC7), single tube, high-sensitivity
assay for both neutrophils and monocytes that could be per-
formed on the Beckman FC500 and Navios instruments. A
variant taking advantage of all three lasers of the BD FACS-
Canto II (FLAER, CD157-PE, CD15-PerCPeFluor710,
CD64-APC, and CD45eFluor450) was also developed that
could also be deployed on the three-laser Navios [23]. The
five-color assays were also robust and reproducible as evidenced
by other studies comparing a six-color (FLAER, CD24, CD14,
CD15, CD64, and CD45) cocktail with the CD157-based
five-color assay [42]. It should be noted that the neutrophils
and monocytes of occasional non-PNH patient samples ana-
lyzed with CD157/FLAER based assays fail to express CD157
[23, 24]. While a polymorphism in the CD157 gene has been
shown to be responsible for some of the CD157-negative cases
[43], this explanation does not account for all of them
[44]. While the biological cause(s) of this remains under
study, the mandatory inclusion of the second GPI reagent
(FLAER) is part of the built-in robustness of the assay and
prevents the misinterpretation of the data as a “PNH clone-
containing sample”; the loss of two GPI-linked structures per
lineage is required to identify bona fide PNH phenotypes.
Furthermore, in keeping with current state-of-the-art guide-
lines [19, 22–24, 41], the RBC lineage was also analyzed and in
every CD157-negative case there was no evidence of any PNH
phenotypes confirming the finding as an isolated single
(CD157) GPI deficiency. An example of a CD157-negative
case using a seven-color combination (FLAER, CD157-PE,
CD15-PerCPeFluor710, CD64-PECy7, CD24-APC, CD14-
APCeFluor780, and CD45-eFluor450) is shown (Case 14) in
the Supplementary data from [19]. Such CD157-negative,
non-PNH cases should simply be reported as “No evidence
of PNH in RBCs and WBCs” [24].
4. A number of six-color, single tube, high-sensitivity assays have
been developed whose gating strategies are also derived from
the original four- and five-color assays. As shown below, a
Navios-specific cocktail comprising FLAER, CD24-PE,
CD15-PC5, CD64-PC7, CD14-APC700, and CD45-KO
was used to stain a normal (and PNH sample (Figs. 6 and 7).
Neutrophils and monocytes are gated by a combination of light
scatter, CD45 expression and either CD15 staining (neutro-
phils) or CD64 staining (monocytes). PNH phenotypes are
detected and/or quantified based on their loss of FLAER and
CD24 binding (neutrophils) or FLAER and CD14 binding
352 Andrea J. Illingworth et al.

(monocytes) [41]. Similar six-color reagent sets have also been

developed previously for the BD FACSCanto platform
[39, 42].

8.3 Reagent 1. Recently, six-color cocktails comprising FLAER, CD157-PE,

Cocktails for Cross- CD15-PerCP-eFluor710 (or CD15-PerCPCy5.5) CD64-
Platform Assays PECy7, CD24-APC, and CD45-eFluor450 have been devel-
oped [27] that can be performed on both Navios and Canto II
instruments (see Fig. 16 of [24]).
2. Samples can be stained with this 6-C reagent set and be ana-
lyzed on either Navios or three-laser Canto variants or on both.
3. In addition, a five-color version of this cross-platform assay
(FLAER, CD157-PE, CD15PerCPCy5.5, CD64-APC, and
CD45eFluor450) has also been developed [27]. Specific
reagent cocktails, recommended clones and conjugated forms
used in these 5C, 6C, and 7C cross-platform (3-laser Navios
and Canto II instruments) assays are detailed in [27] and
shown in Table 4.

References

1. Parker CJ (2007) The pathophysiology of par- 9. Sugimori C, Mochizuki K, Qi Z et al (2009)

oxysmal nocturnal hemoglobinuria. Exp Origin and fate of blood cells deficient in
Hematol 35:523–533 glycosylphosphatidylinositol-anchored protein
2. Oni SB, Osunkoya BO, Luzzatto L (1970) among patients with bone marrow failure. Br J
Paroxysmal nocturnal hemoglobinuria: evi- Haematol 147:102–112
dence for monoclonal origin of abnormal red 10. Raza A, Ravandi F, Rastogi A et al (2014) A
cells. Blood 36:145–152 prospective multicenter study of paroxysmal
3. Young NS, Calado RT, Scheinberg P (2006) nocturnal hemoglobinuria cells in patients
Current concepts in the pathophysiology and with bone marrow failure. Cytometry B Clin
treatment of aplastic anaemia. Blood Cytom 86:175–182
108:509–519 11. Morado M, Freire Sandes A, Colado E et al
4. Kordasti S, Marsh J, Al-Khan S et al (2012) (2017) Diagnostic screening of paroxysmal
Functional characterization of CD4+ T cells in nocturnal hemoglobinuria: Prospective multi-
aplastic anemia. Blood 119:2033–2043 centric evaluation of the current medical indi-
5. Solomou EE, Rezvani K, Mielke S et al (2007) cations. Cytometry B Clin Cytom
Deficient CD4+CD25+FOXP3+ T regulatory 92B:361–370
cells in acquired aplastic anemia. Blood 12. Dezern AE, Borowitz MJ (2018) Consensus
110:1603–1606 guidelines for the flow cytometry testing of
6. Kordasti SY, Afzali B, Lim Z et al (2009) patients with suspected paroxysmal hemoglo-
IL-17-producing CD4 (1) T-cells, proinflam- binuria (PNH) and related disorders part 1—
matory cytokines and apoptosis are increased in clinical utility. Cytometry B Clin Cytom
low-risk myelodysplastic syndrome. Br J Hae- 94:16–22
matol 145:64–72 13. Hill A, Rother RP, Wang X et al (2010) Effect
7. Mufti GJ, Marsh J (2018) Somatic mutations of eculizumab on haemolysis-associated nitric
in aplastic anemia. Hematol Oncol Clin N Am oxide depletion, dyspnoea, and measures of
32:595–607 pulmonary hypertension in patients with par-
oxysmal nocturnal haemoglobinuria. Br J Hae-
8. DeZern AE, Symons HJ, Resar LS et al (2014) matol 149:414–425
Detection of paroxysmal nocturnal hemoglo-
binuria clones to exclude inherited bone mar- 14. Socie G, Mary JY, de Gramont A et al (1996)
row failure syndromes. Eur J Haematol Paroxysmal nocturnal haemoglobinuria: long-
92:467–470 term follow-up and prognostic factors. French
Society of Haematology. Lancet 348:573–577
 Immunophenotyping of PNH 353

15. Sugimori C, Chuhjo T, Feng X et al (2006) assay in the primary screening of PNH clones.
Minor population of CD55-CD59- blood cells Am J Clin Pathol 132:564–572
predicts response to immunosuppressive ther- 26. Dalal BI, Khare NS (2013) Flow cytometric
apy and prognosis in patients with aplastic ane- testing for paroxysmal nocturnal hemoglobin-
mia. Blood 107:1308–1314 uria: CD64 is better for gating monocytes than
16. Borowitz MJ, Craig FE, DiGiuseppe JA et al CD33. Cytometry B Clin Cytom 84:33–36
(2010) Guidelines for the diagnosis and moni- 27. Sutherland DR, Ortiz F, Quest G et al (2018)
toring of paroxysmal nocturnal hemoglobin- High-sensitivity 5-, 6-, and 7-color PNH WBC
uria and related disorders by flow cytometry. Assays for both Canto II and Navios Platforms.
Cytometry B Clin Cytom 78B:211–230 Cytometry B Clin Cytom 94:637–651
17. Killick SB, Bown N, Cavenagh J et al (2016) 28. Keeney M, Illingworth A, Sutherland DR
Guidelines for the diagnosis and management (2017) Paroxysmal Nocturnal Hemoglobin-
of adult aplastic anaemia. Br J Haematol uria assessment by flow cytometric analysis.
172:187–207 Clin Lab Med 37:855–867
18. Sutherland DR, Keeney M, Illingworth A 29. Sutherland DR, Richards SJ, Ortiz, F et al.
(2012) Practical guidelines for the high- Addition of CD71 to high-sensitivity
sensitivity detection and monitoring of parox- CD235a/CD59 PNH RBC assay allows the
ysmal nocturnal hemoglobinuria clones by flow additional accurate delineation of PNH Type
cytometry. Cytometry B Clin Cytom III, PNH Type II and normal (Type I) nucle-
82:195–208. and Supplementary Data ated RBCs: Cytometry B 2019, in press
19. Illingworth A, Marinov I, Sutherland DR et al 30. Hernández-Campo PM, Almeida J, Sánchez
(2018) ICCS/ESCCA consensus guidelines to ML et al (2006) Normal patterns of expression
detect GPI-deficient cells in PNH and related of glycosylphosphatidylinositol-anchored pro-
disorders part 3—data analysis, reporting and teins on different subsets of peripheral blood
case studies. Cytometry B Clin Cytom cells: A frame of reference for the diagnosis of
94:49–66. and Supplementary Data paroxysmal nocturnal hemoglobinuria. Cyto-
20. Oldaker T, Whitby L, Saber M et al (2018) metry B Clin Cytom 70:71–81
ICCS/ESCCA Consensus Guidelines for the 31. Sutherland DR, Acton E, Keeney M et al
Flow Cytometry Testing of Patients with sus- (2014) Use of CD157 in FLAER-based assays
pected Paroxysmal Hemoglobinuria (PNH) for high-sensitivity PNH neutrophil and PNH
and related Disorders Part 4 - Assay Validation monocyte detection. Cytometry B Clin Cytom
and quality assurance. Cytometry B Clin 86:44–55
Cytom 94:67–81 32. Owens MA, Vall HG, Hurley AA, Wormsley SB
21. Illingworth A, Oldaker T, Sutherland DR et al.: (2000) Validation and quality control of immu-
Verification of PNH assay sensitivity through nophenotyping in clinical flow cytometry. J
spiking experiment. (2018). https://www. Immunol Methods 243:33–50
cytometry.org/web/modules/Module10.pdf 33. International Organization for Standardization
22. Davis BH, Keeney M, Brown R et al (2014) (ISO) Statistics-Vocabulary and Symbols. ISO
CLSI H52-A2 red blood cell diagnostic testing 3534-1. Geneva; 1993
using flow cytometry; approved guideline, 2nd 34. https://www.iso.org
edn. Clinical and Laboratory Standards Insti-
tute, Wayne, PA. ISBN Number: 1-56238- 35. Libeer J-C (2001) Role of external quality
957-2 assurance schemes in assessing and improving
quality in medical laboratories. Clin Chim Acta
23. Sutherland DR, Richards SJ, Keeney M et al 309:173–177
(2015) High-sensitivity detection of PNH red
blood cells, red cell precursors and white blood 36. CLSI EP12-A2 (2008) User Protocol for Eval-
cells. Curr Protoc Cytom 72(Unit 6.37):1–29 uation of Qualitative Test Performance- Sec-
ond Edition – Approved Guideline. National
24. Sutherland DR, Illingworth A, Marinov I et al Committee for Clinical Laboratory Standards,
(2018) ICCS/ESCCA Consensus Guidelines Wayne, PA
to detect GPI-deficient cells in Paroxysmal
Nocturnal Hemoglobinuria (PNH) and related 37. Clinical and Laboratory Standards Institute
disorders part 2—reagent selection and assay (2004) Protocols for Determination of Limits
optimization for high-sensitivity testing. Cyto- of Detection and Limits of Quantitation
metry B Clin Cytom 94:23–48 Approved Guideline. CLSI document EP17.
CLSI, Wayne, PA
25. Sutherland DR, Kuek N, Azcona-Olivera J et al
(2009) Use of FLAER-based white blood cell 38. New York State Department of Health (2011)
Clinical Laboratory Evaluation Program, Assay
354 Andrea J. Illingworth et al.

Approval in Cellular Immunology. New York 42. Marinov I, Kohoutová M, Tkáčová V et al

State Department of Health, Albany, NY (2015) Clinical relevance of CD157 for rapid
39. Marinov I, Kohoutova M, Tkacova V et al and cost-effective simultaneous evaluation of
(2013) Intra- and inter-laboratory variability PNH granulocytes and monocytes by flow
of paroxysmal nocturnal hemoglobinuria test- cytometry. Int J Lab Hematol 37:231–237
ing by flow cytometry following the 2012 Prac- 43. Blaha J, Schwarz K, Fischer C et al (2018) The
tical Guidelines for high-sensitivity paroxysmal Monoclonal Anti-CD157 Antibody Clone
nocturnal hemoglobinuria testing. Cytometry SY11B5, used for High Sensitivity detection
B Clin Cytom 84:229–236 of PNH Clones on WBCs, fails to Detect a
40. Fletcher M, Sutherland DR, Whitby L et al Common Polymorphic Variant Encoded by
(2014) Standardizing leucocyte PNH clone BST-1 published in line. Cytometry B Clin
detection: An international study. Cytometry Cytom 94:652
B Clin Cytom 86B:311–318 44. Sutherland DR, Musani R (2019) Re Blaha J
41. Illingworth A, Keeney SDR (2016) High- et al.: Re: Blaha J et al.: The monoclonal anti-
sensitivity detection of PNH Red and White CD157 antibody clone SY11B5, used for high
Blood Cells by Multiparameter Flow Cytome- sensitivity detection of PNH clones on WBCs,
try. In: Detrick B, Schmitz JL, Hamilton RG fails to detect a common polymorphic variant
(eds) Manual of Clinical Laboratory Immunol- encoded by BST-1 Cytometry B Clin Cytom
ogy, 8th edn. ASM Press, Washington DC, pp 96(1):16-18
168–181. Chapter 18
 INDEX

A E
Aberrant marker expression................................. 287, 289 Effector Tregs (eTreg) ........................142, 146, 152, 156
Acoustic cell enrichment............................................... 214 Endothelial ........................................................... 203–210
Acute leukemia ....................................293, 297, 301, 308 Expansion ............................................................ 116, 183,
Acute lymphoblastic leukemia (ALL) ................. 297–308 186, 324, 339
Acute myeloid leukemia (AML)..................................254,
281–293, 297, 324 F
Adipose tissue (AT).............................................. 115–125 Fetal calf serum (FCS) ............................ 71, 77, 182, 298
Analysis ................................................................ 2, 31, 53, FLAER................................................................. 327–329,
69, 81, 106, 122, 130, 143, 185, 194, 204, 213,
331, 336, 339–345, 349–351
228, 282, 298, 316, 324 Flow cytometric analysis ....................................... 53, 106,
Antibodies ........................................................... 2, 3, 6, 7, 130, 193–200, 285
16, 19, 20, 26, 31, 53, 69, 89, 105, 116, 131, 144,
Flow cytometry (FC) .................................................1, 31,
181, 195, 203, 215, 229, 282, 298, 317, 325 53, 69, 81, 106, 123, 130, 142, 185, 194, 203,
Aplastic anemia (AA) .................................. 324, 325, 339 215, 228, 282, 297, 311, 324
Fluorescence ........................................................ 1–27, 53,
B
64, 71, 82, 85, 87, 88, 91, 93, 95, 97, 98, 106,
Body mass index (BMI)...............................115, 123–125 107, 122–124, 143, 144, 152, 154, 156–158,
160–164, 166, 169, 172, 173, 182, 185, 190,
C 200, 209, 210, 223, 235, 240, 244, 249, 259,
Cardiovascular (CV) ..................................................... 115 260, 264, 267, 274, 293, 307, 316
Fluorescence minus one (FMO)
CD59 ................................................................... 325, 326,
330, 332–334, 336–338, 346 controls ........................................... 16, 19, 21, 23,
CD127...............................................................5, 90, 142, 24, 26, 97, 154–157, 200, 224, 240, 259
144–148, 152, 154, 156, 157, 160–162, 165, Fluorescent cell barcoding (FCB) ............................53–68
FoxP3.................................................... 90, 142, 144–148,
168, 181, 186, 191
Chronic lymphocytic leukemia (CLL) ................ 311–320 152, 154–156, 159–163, 165, 168, 182
Circulating tumor cells (CTCs) .......................... 213–225 Free fatty acids (FFAs) .................................................. 116
Color compensation.....................................................247,
G
252–258, 262, 276
Compensation .............................................. 3, 32, 59, 71, Gating .................................................................. 2, 45, 55,
93, 107, 112, 122, 133, 143, 198, 205, 229, 329 77, 81, 107, 122, 136, 144, 185, 197, 204, 218,
Cytokines ..............................................39, 84, 91, 94, 97, 265, 282, 303, 316, 332
112, 116, 141, 180, 182–183, 186, 188, 191, 324 Germinal center (GC).......................................... 118, 120
Cytometry by time of flight (CyTOF)......................1, 44, Glycophosphatidyl inositol-anchored protein
107, 110 (GPI-AP) ........................................................... 323

D H
Dendritic cells (DCs) .............................................. 70, 91, Hanks’ balanced salt solution (HBSS)................ 117, 118
98, 193–200, 286, 342 Harmonization ....................................................... 70, 146
Dried antibody cocktails ...........................................70, 77 High-dimensional ..........................................1–27, 44, 97
Dulbecco’s modified Eagle’s medium Human................................................................. 4, 39, 70,
(DMEM).......................................... 117, 118, 120 81, 106, 116, 130, 142, 180, 205

J. Philip McCoy, Jr (ed.), Immunophenotyping: Methods and Protocols, Methods in Molecular Biology, vol. 2032,
https://doi.org/10.1007/978-1-4939-9650-6, © Springer Science+Business Media, LLC, part of Springer Nature 2019

355
 IMMUNOPHENOTYPING: METHODS AND PROTOCOLS
356 Index
I Phenotyping ............................................................ 53, 56,
70, 105–113, 120, 213–225
Immunofluorescence .................................................... 110 Plasmacytoid dendritic cells (pDCs) ...........193–200, 286
Immunology.............................................. 1, 12, 101, 130 Polychromatic................................. 19, 20, 130, 242, 244
Immunophenotype ................................................. 32, 39, Proficiency testing ..........................................83, 264, 270
194, 281, 288–291, 297, 304, 311, 313, 316
Immunophenotyping............................................. v, 1–27, Q
31–50, 53–78, 89, 107, 115–125, 141–173,
179–191, 203–210, 227–278, 281–293, Quality assurance (QA).......................228, 259, 270, 277
297–308, 311, 323–351 Quality control (QC)............................................... v, 2, 3,
Immunoprofiling........................................................... 106 6, 20, 85, 87, 101, 154, 221–223, 227–278
Induced regulatory T cells (iTregs).............................. 142
R
Innate lymphoid cells (ILCs) .............................. 179–191
Insulin resistance (IR).......................................... 115, 116 Reactive oxygen species (ROS) .................................... 116
Isolation ................................................................. 55, 111, Red blood cells (RBCs) .......................................... 40, 55,
117–121, 181, 183–184, 189, 205, 207, 213, 214 111, 130, 133, 147, 149, 150, 159, 190, 195,
208, 214, 215, 217, 218, 224, 229, 230, 232,
L 233, 241, 271, 277, 282, 284, 286, 325–337,
Levey–Jennings (L-J) ........................................... 246, 247 339, 340, 342, 343, 346, 347, 350, 351
Lyophilization ........................................v, 70–73, 75, 146 Regulatory T cells (Tregs) .......................................70–72,
90, 141–147, 151, 152, 154–158, 160–162, 165
M Room temperature (RT) .................................... 6, 18, 20,
39, 40, 42, 43, 55–60, 120, 121, 124, 130, 133,
Magnetic depletion .............................................. 214, 218 181, 184–186, 190, 191, 194–196, 204–208,
Mass cytometry (MC)...................................................v, 1, 216–223, 230, 231, 272, 283, 284, 298, 301,
31–50, 81, 106, 107, 110 302, 313–315, 317, 330
Memory T cells..................................................... 131, 137
Microparticles (MPs) .....................................55, 203–210 S
Minimal residual disease (MRD)........................ 267, 282,
285, 289, 297–299, 301, 302, 304–308, 313, Spillover spreading error matrix (SSM) .......................... 6,
315, 316, 318 12–17, 26
Molecular indexing ....................................................... 223 Staining ................................................................ 4, 33, 53,
Monoclonal antibodies (mAb) ................................ v, 132, 69, 82, 108, 118, 131, 143, 182, 197, 206, 214,
134, 147, 149–151, 155, 160–164, 167–169, 231, 285, 307, 314, 327
173, 181–183, 195–198, 229, 233, 283–285 Stromal vascular fraction (SVF) .......................... 117–123
Multicolor flow cytometry.......................... 153, 198, 262 Subcutaneous adipose tissue (SAT) .................... 116–124
Multiplexing .................................................................. 110 Suppressor T cells.......................................................... 251
Myelodysplastic syndrome (MDS)............. 324, 325, 339
T
N T cells4, 67, 70, 83, 117, 129, 142, 179, 193, 251, 303,
Naive T cells ................................................ 129, 131, 137 311, 323
Natural killer (NK) cells...................................... 4, 70, 93, Toll-like receptor (TLR) ............................. 118, 122, 123
98, 105–113, 116, 117, 120, 137, 180, 193, 251, Transcription factors (TF) .................................. 117, 125,
265, 317, 342 142, 163, 180, 182, 185, 186, 189, 191
Natural Tregs (nTregs) ........................................ 141, 153 Type-2 diabetes (T2D) ................................................. 115

P W

Panel design............................................................. 2, 4, 6, Whole blood................................................. 6, 12, 40, 48,

12–17, 24, 36–38, 82, 160, 161, 169, 172, 237 55, 56, 59, 90, 101, 106, 111, 133, 137, 148,
Paroxysmal nocturnal hemoglobinuria 158, 162, 165, 168, 205, 207, 217, 231, 234,
(PNH) .................................................83, 323–351 248–250, 262, 265, 276, 349, 350