CONTENTS

Table of Contents

INFORMATION
Continuing Medical Education..................................................................................................5
Guidelines for Speakers...........................................................................................................6
Guidelines for Poster Presentations..........................................................................................8

SPEAKER ABSTRACTS
Abstracts................................................................................................................................9

POSTER ABSTRACTS
Basic Research (Location – Room 101)................................................................................63
Clinical (Location – Room 8)...............................................................................................141

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 3
4 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
 EACCME
European Accreditation Council for Continuing Medical Education

2018 Joint Global Neurofibromatosis Conference

Paris, France, 02/11/2018 – 06/11/2018

has been accredited by the

European Accreditation Council for Continuing Medical Education (EACCME®)
for a maximum of 27 European CME credits (ECMEC®s).

Each medical specialist should claim only those credits that he/she actually spent in the educational activity.

The EACCME® is an institution of the European Union of Medical Specialists (UEMS), www.uems.net.
Through an agreement between the European Union of Medical Specialists and the American Medical
Association, physicians may convert EACCME® credits to an equivalent number of AMA PRA Category 1
Credits™. Information on the process to convert EACCME® credits to AMA credits can be found at
www.ama-assn.org/education/earn-credit-participation-international-activities.

Live educational activities occurring outside of Canada, recognised by the UEMS-EACCME® for ECMEC®
credits are deemed to be Accredited Group Learning Activities (Section 1) as defined by the Maintenance of
Certification Program of the Royal College of Physicians and Surgeons of Canada.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 5
 IMPORTANT NOTES
Guidelines for Speakers
The Organizing Committee of the JOINT GLOBAL NEUROFIBROMATOSIS Conference will do its utmost to help authors for their presentations and to facilitate
their arrival and stay at the conference.

Please take a few minutes to read the following guidelines regarding the on-site organisation of the meeting for the smooth running of the sessions.

A speakers’ preview room will be installed in room 103 on the 1st floor of the Maison de la Chimie, Paris (follow signs on site). Its role will be to manage the following:

• Coordinate and ensure the overall smooth running of the conference sessions,
• Follow the general schedule of the sessions,
• Assist speakers for any requests they may have onsite.

PLATFORM PRESENTATIONS
In order to match with the most recent technology, the conference meeting room will be equipped with a video-projector and a laptop at the lectern.

All speakers are requested to use the PREVIEW ROOM (follow signs on site), which can be accessed once you have picked up your badge at the Welcome desk.

Each speaker should also check in the final Programme that the time of his/her session has not been modified.

We kindly ask you to keep the time limit in mind and remember to save time for discussion!

Qualified personnel will act as liaison between speakers and projectionists: speakers will not have access to the projection rooms; therefore speakers must go to
the PREVIEW ROOM to hand in their computer assisted presentations that will be transferred to the conference room on time.

Speakers are entirely responsible for the order, the loading and the pre-projection of their computer assisted presentation, using the equipment made available by
the organisers.

PRESENTATION (POWER-POINT STYLE)

To avoid delays caused by switching on computers on the platform, booting up computers and potential compatibility problems, the Organising Committee has
made available to speakers the standard A/V system used in the convention sector. There will be a master computer in the meeting room and to ensure smooth
transition between speakers and appropriate technical support, the Organisers request that speakers do not connect their own laptop. Every speaker has to go to
the Preview room beforehand to provide his/her PowerPoint presentation.

OFFICIAL LANGUAGE

The official language of the Conference is English, which means that all presentations and questions must be delivered in ENGLISH.

FORMAT – PRESENTATION

Only Presentations for PC’s (Windows latest versions) and PC’s compatible
(to avoid problems of compatibility between PC’s and MAC, please use fonts Presentations must be displayed in 16:9.
compatible with both PC’s and MAC) will be accepted, (no UNIX).

Your presentation should be saved in the .pdf or .pptx format only. Other formats
will not be accepted.

• If you have pictures, they must be under the following format: .jpg,
.png, or .gif, format (.pict prohibited).
• If you have video files attached to your power point presentation, they
must be in the following format: .mpg, .mpeg, .avi, .wmv, mp4 or .mov
and must be saved in the same folder as the presentation.

When saving your final presentation on a USB stick, make sure to include your video files and all links to these multimedia files.

6 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
IMPORTANT NOTES
Guidelines for Speakers
DEPOSITING OF FILE

• Your computer file must be handed over to the staff of the PREVIEW ROOM, either on a memory stick or a hard drive, as far in advance as possible and
ONE AND A HALF hour BEFORE the beginning of each session AT THE LATEST. The presentation for an early morning session should be handed
over the evening before.
• In the PREVIEW ROOM, you will be assisted by a technician, who will help you to transfer your presentation to the internal network. You will also be
able to review your presentation and to verify that it has been transferred correctly to the network.
• The opening hours of the PREVIEW ROOM during the Conference will be:

Friday, 2 November 2018 from 09:00 to 19:00

Saturday, 3 November 2018 from 07:30 to 18:30
Sunday, 4 November 2018 from 08:00 to 18:30
Monday, 5 November 2018 from 07:30 to 18:30
Tuesday, 6 November 2018 from 07:30 to 14:00

• All presentations will be considered as confidential by our staff and removed from the system at the end of the conference.

IN THE MEETING ROOM

• Your presentation will be sent directly to the meeting room through the internal computer network of the venue. The PC on the lectern is programmed in
16/9 and is linked to a video-projector.
• Once the presentation is launched, you, the speaker, will control the program. By clicking on the mouse, your computer assisted slides will go on as usual.
• Please, do NOT come at the last minute with your own laptop to the meeting room: you will NOT be able to connect it. Go to the PREVIEW ROOM
in time beforehand.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 7
 IMPORTANT NOTES
Guidelines for Poster Presentations
The Organizing Committee of the JOINT GLOBAL NEUROFIBROMATOSIS Conference will do its utmost to help authors for their presentations and to facilitate
their arrival and stay at the conference.

The Poster exhibition will be accessible from Friday to Monday in the exhibition area of the Maison de la Chimie, Paris.

SET-UP/ REMOVAL INSTRUCTIONS

Mounting Friday, 2 November 2018 From 15:30 to 20:00

Removing Monday, 5 November 2018 From 16:45 to 18:45

POSTER SESSIONS

The posters will be discussed in front of the poster boards during the dedicated sessions on Saturday and Sunday.

In the dedicated session slot, the presenting author must be available next to his/her dedicated board to present and discuss the poster.

BASIC RESEARCH POSTERS Saturday, 3 November 2018 From 18:00 to 19:30 Room 101 – level 1
CLINICAL RESEARCH POSTERS Sunday, 4 November 2018 From 17:25 to 18:55 Room 8 – ground floor

POSTER DIMENSION AND FORMAT

The poster must not exceed 120 cm in height and 90 cm in width, PORTRAIT ONLY.

LANGUAGE

Your poster must be written in ENGLISH.

POSTER CONTENTS

Each poster should contain, in addition to the scientific elements, the title of the selected
abstract as submitted, and the full names and affiliation of contributing authors.

Text, tables and drawings for figures should be large enough to be seen from a 2 meters distance. Illustrations should be used to convey important points;
diagrams, graphs, bar charts, scatter grams, pie charts and photographs will enhance your presentation.

Make short statements, avoid long explanatory sentences.

MATERIALS

Your poster should not exceed 120 cm high and 90 cm wide (Portrait only) in order to fit the poster board. Prepare your material beforehand so that it will fit
neatly into the space available and can be easily attached to the board. Suitable fixing materials will be provided by the Conference organizers. There will be a
poster helpdesk close to the poster area, where staff will be happy to assist you.

Thin cardboard is more suitable than paper. Use a computer or enlarge a typed text by photographic methods. If you can have your poster produced by your
institution, the final effect is more professional.

8 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
ABSTRACTS
Speaker Abstracts

EDUCATIONAL SYMPOSIUM (PART 1) – CLINICAL

Chairs: Dusica Babovic-Vuksanovic, MD, Mayo Clinic, US; Mimi Berman, MD, University of Melbourne, Australia; Karin Cunha, PhD, Federal University of
Brazil

A Review of the Complex and Unique Testing Approaches Available for the Neurofibromatoses
Friday, 2 November, 13:30 – 13:55

Alicia Gomes, MS, CGC, University of Alabama at Birmingham Genomics Laboratory

This talk will review the complexities and unique testing approaches that may be needed to confirm a molecular diagnosis of neurofibromatosis type 1, Legius
syndrome, and schwannomatosis. A review of the benefits and limitations of both germline and somatic testing for each condition will be provided as well as
specific case reports to highlight unique testing approaches that can be utilized in several clinical scenarios.

The RASopathies: Distinct Syndromes with Shared Phenotypic Findings

Friday, 2 November, 13:55 – 14:20

Karen Gripp, MD, Nemours/Alfred DuPont Hospital for Children, US

The RASopathies are syndromic disorders caused by germline derived mutations in genes affecting the RAS/MAPK signaling pathway. The individual conditions
were delineated as distinctive syndromes based on their characteristic findings and include von Recklinghausen disease (now known as neurofibromatosis type
1), Noonan syndrome, cardio-facio-cutaneous syndrome, Costello syndrome and others. These syndromes share central nervous system, cardiac and skin
abnormalities, relative macrocephaly and short stature. Tumor predisposition is present in most RASopathies. It can be difficult to differentiate between these
conditions, particularly in young patients, because the presentation of each syndrome is variable and may be age dependent. The notable overlap in phenotypes
and presenting findings is best understood in the context of their shared underlying biology of dysregulation of the RAS/MAPK signaling pathway.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 9
 Don’t Forget Constitutional Mismatch Repair Deficiency (CMMRD)
Friday, 2 November, 14:20 – 14:45

Katharina Wimmer, PhD, Division of Human Genetics of the Medical University Innsbruck, Austria

Constitutional mismatch repair (MMR) deficiency (CMMRD) is a rare, autosomal recessively inherited childhood cancer susceptibility syndrome resulting from
biallelic germline loss-of-function mutations in one of the MMR genes. Individuals with CMMRD have a high risk to develop a broad spectrum of malignancies
and frequently display features reminiscent of neurofibromatosis type 1 (NF1), mainly in the form of multiple café-au-lait macules (CALMs), but also other NF1
signs. The implications of the phenotypic overlap of CMMRD and NF1 are manifold.

(i) To avoid misdiagnosing paediatric cancer patients with CMMRD as having NF1, which impedes adequate management of the patients and their families,
it is important to increase awareness that a clinical phenotype reminiscent of NF1 in addition to any malignancy other than an MPNST, JMML or pilocytic
astrocytoma, well known to be associated with NF1, in a child should raise a high level of suspicion of CMMRD syndrome.

(ii) The spectrum of CMMRD-associated childhood malignancies includes high-grade glioma, acute myeloid leukaemia and rhabdomyosarcoma. Reported
associations between NF1 and these malignancies are to a large extent based on studies that neither proved the presence of an NF1 germline mutation nor
ruled-out CMMRD in the affected. Hence, these associations are challenged by our current knowledge of the phenotypic overlap between NF1 and CMMRD
and should be re-evaluated in future studies.

(iii) Awareness that CALMs and occasionally other NF1 signs may be present in a child with CMMRD prior to tumour onset leads to the conclusion that
CMMRD is a legitimate differential diagnosis in a healthy child with signs reminiscent of NF1/Legius syndrome without a detectable underlying NF1/SPRED1
germline mutation. In this situation, a diagnosis of CMMRD may provide an opportunity for cancer surveillance of a highly penetrant childhood cancer
syndrome prior to onset of the first malignancy. It will also allow predictive genetic testing and may impact family planning. Therefore, physicians and
geneticists have begun to discuss if and when to counsel and test for CMMRD in suspected NF1 patients. Potential harm associated with CMMRD counselling
and testing in this setting needs to be considered in this discussion. It should be outweighed by expected benefits for both the index patient and at-risk
relatives. Appropriate counselling and testing strategies that serve to minimize the potential harm of testing should be available.

Overview of Prenatal Testing and Preimplantation Genetic Diagnosis – NF1/NF2 and Schwannomatosis
Friday, 2 November, 14:45 – 15:20

Myra Wick, MD, PhD, Mayo Clinic, Rochester Minnesota, Departments of Medical Genomics and OBGYN
Kristi Jones MBBS, PhD, Clinical Geneticist, Sydney Children’s Hospitals Network, Genea Fertility, Sydney Australia

Background: Prenatal testing and screening has evolved rapidly over the last decade.

Cell free fetal DNA (cfDNA), which was first described more than 2 decades ago is now used routinely in Europe for determination of fetal RhD status in Rh
negative mothers. Cell free fetal DNA for aneuploidy screening is utilized worldwide. Several other potential applications for cfDNA have been reported, including
analysis for single gene disorders. Prenatal testing is also evolving rapidly, with routine use of chromosome microarray and whole exome sequencing (WES).

Preimplantation genetic diagnosis (PGD) is a technology utilized by individuals affected by or carrying a monogenic disorder, to prevent transmission of the
disorder to future offspring.

Case reviews: All of these technologies are applicable to family planning in NF1/NF2 and schwannomatosis – and these conditions present some specific
challenges.

Mosaicism or segmental presentation of NF1 and NF2, including germline mosaicism, are well documented in the literature. A patient with segmental NF1 who
desires preimplantation diagnosis is presented.

The primary underlying genetic cause may remain elusive in familial schwannomatosis, and testing may rely on indirect linkage methods. We present the
experience of PGD in one such family.

Conclusions: Preimplantation and prenatal screening and testing options are changing rapidly. It is important for obstetric providers and medical geneticists to
have a working knowledge of these options and their limitations, as well as societal guidelines regarding these test options.

10 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Less Well Known Dermatological Manifestations in NF1
Friday, 2 November, 15:20 – 15:45

Sirkku Peltonen, MD, PhD, University of Turku and Turku University Hospital, Turku, Finland

The most common diagnostic signs of NF1 are on skin: café au lait macules, smaller pigmented spots in skinfolds, generally but misleadingly called as freckles, and
cutaneous neurofibromas. In addition to these, NF1 is associated with other less well known skin manifestations.

Melanocytes are cells which synthesize melanin pigment and thus contribute to the skin color. Mutations of the NF1 gene lead to up-regulated melanin production, as
seen in café au lait macules and freckles where melanocytes harbor second hit mutations in the NF1 gene. In addition to skinfolds, small pigment spots often cover
large skin areas especially in trunk. The number and area of spots increase by age. In addition to localized pigmentation, patients with NF1 may have a slightly darker
general skin color than their family members without NF1. Single hypopigmented spots may also be seen.

Pale spots in skin may be part of nevus anemicus, which is a congenital vascular anomaly of the skin. Some anemic nevi are visible only after elicitation of
vasodilatation in surrounding skin by gentle friction, while some are detectable without friction. Nevus anemicus is most commonly located on upper chest or back.

Juvenile xanthogranulomas are orange papules commonly seen children with NF1. They may be solitary or multiple, and the number varies from single to 20-30. Head
and neck are the most common sites. Juvenile xanthogranulomas are predominant in infancy: 40% to 70% appear during the first year of life and they also disappear
during childhood. Case reports and molecular studies have connected juvenile xanthogranulomas, NF1 and juvenile myelomonocytic leukemia (JMML). However,
JMML is a very rare type of childhod leukemia, and it has been estimated that only 2-3 % of all childhood leukemias are of this type. Recent large cohort studies
did not demonstrate any leukemia cases in children with NF1. Since juvenile xanthogranulomas are common in children with NF1, but leukemia is very rare, finding
juvenile xanthogranulomas should not lead to active follow-up of blood tests.

Glomus tumours are benign but painful neoplasms of the glomus body of fingers and toes. Glomus bodies are thermoregulatory shunts which regulate capillary blood
flow according to temperature. Glomus tumors in the digits are located under nails or in the pulp, usually invisible to eye, but cause localised tenderness, paroxysmal
pain, and sensitivity to cold. Glomus tumor can be imaged by MRI and operated by hand surgeon.

EDUCATIONAL SYMPOSIUM (PART 2)

Chairs: Dusica Babovic-Vuksanovic, MD; Mimi Berman, MD; Karin Cunha, PhD

A Holistic Model to Discuss the Impact of Muscular Phenotype on Lifestyle of Individuals with NF1
Friday, 2 November, 16:10 – 16:35

Juliana Souza MD, PhD, Neurofibromatosis Outpatient Reference Center, Hospital das Clínicas da Universidade Federal de Minas Gerais, Brazil

Over the last decades it has been observed that individuals with NF1 can present with significant physical fitness impairment perceived as reduced muscle
mass, low muscle tone, global muscular weakness and decreased exercise capacity. These features have been attributed to central nervous system dysplasia
and cognitive deficits (due to intrinsic neurofibromin deficiency), clinically observed as poorer motor proficiency and coordination. Nevertheless, recent pre
clinical studies have shown a primary peripheral role for neurofibromin in muscle growth and metabolism, suggesting that a lipid storage disease phenotype
may underlie muscle weakness in NF1.

Although we still lack a precise understanding of the whole mechanisms involved in the pathophysiology of NF1 muscular phenotype, we can observe its
influence on lifestyle and also foresee the potential impact on NF1 shorter life expectancy, poorer quality of life and more frequent and earlier cardiovascular
involvement, when compared to the general population. It looks like enough material to propose a holistic model that can propel the NF community towards
discussing, studying, proposing and implementing interventions looking forward to approaching such a relevant and potentially modifiable matter.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 11
 Lessons Learnt from Breast Cancer Screening in High Risk Women—Screening in Women with NF1, When
and How?
Friday, 2 November, 16:35 – 17:00

Eva Trevisson, MD, PhD, University of Padova, Italy

Neurofibromatosis type 1 (NF1) (MIM#162200) is one of the most common autosomal dominant tumor predisposition syndromes, affecting one in 3500
live births. Clinical diagnosis of this condition relies on specific diagnostic criteria (Ach Neu 1988). The disease is caused by germline loss-offunction
mutations in the NF1 gene encoding neurofibromin, which acts on the RAS pathway by negatively regulating the GTPase activity of p21RAS. Given its role
as a tumor suppressor, somatic mutations of NF1 are found in a number of sporadic tumors (Ratner 2015). NF1 is a multisystem disease displaying various
clinical manifestations, including an increased risk of tumor development. The most frequent malignancies are those originating in the central nervous system
(particularly optic pathway glioma OPG) and malignant peripheral nerve sheath tumors (MPNST) that usually arise on a pre-existing plexiform neurofibroma.
Other neoplasms, such as gastrointestinal stromal tumors (GIST) and pheochromocytomas, have been reported more frequently in NF1 individuals.

Breast cancer (BC) is the most frequent tumor in women and previous studies have reported an higher risk among NF1 patients (Madanikia 2012, Wang
2012, Seminog 2015, Uusitalo 2016), particularly at younger age (< 50 yrs). While standard care for NF1 patients is well established (consisting of periodical
detailed clinical evaluations, ophthalmologic examinations and blood pressure measurement), the utility of a specific screening program for BC in NF1 patients
has not been investigated so far and is still debated (Sharif 2007, Evans 2012, Howell 2017).

Different BC screening programs have been developed in different populations: standard screening in Italy consists of mammography every two years
starting from 50 years of age. BRCA1/2 mutated patients or other high risk women undergo clinical evaluations every 6 months since 25 years and annual
mammography/breast MRI since 35 years.

In this presentation I will show the results of our study, focusing on the incidence of BC in a cohort of NF1 Italian women, on possible associated risk factors
and on the effects of an earlier BC screening. Our preliminary data suggest that a precocious “high-risk” screening in NF1 women is effective and should be
offered to all patients starting at 40 years. Multicenter studies are warranted to evaluate BC risk in NF1 women and the need for a specific screening program.

Oral Manifestations of Neurofibromatosis Type 1

Friday, 2 November, 17:00 – 17:25

Karin Cunha, PhD, Fluminense Federal University, Brazil

A variety of oral manifestations may occur in individuals with neurofibromatosis 1. They are very common and their prevalence is estimated to be around
70-90%. Oral manifestations in neurofibromatosis 1 include alterations in soft tissues, jaws, teeth, and salivary glands, such as neurofibromas in oral mucosa,
intraosseous neurofibromas, enlargement of the fungiform papillae of the tongue, enlargement of the mandibular canal, mandibular foramen and mental
foramen, alterations in craniofacial morphology, hyposalivation, beyond other alterations. Therefore, it is important that physicians and dentists are aware of
oral manifestations in patients with neurofibromatosis 1 so that preventive measures can be introduced, and the proper diagnosis and management of these
alterations and their consequences can be performed. The aim of this talk is to present our current knowledge of oral manifestations in neurofibromatosis 1,
including their prevalence, diagnosis, and management in the clinical setting.

12 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Beyond Esthetic Surgery: Electrodessication for Cutaneous Neurofibromatosis
Friday, 2 November, 17:25 – 17:50

Hubert Weinberg, MD, FACS, Mt. Sinai Hospital, US

One of the most devastating consequences of NF1 is the psychosocial impact of the cutaneous lesions. Removal of these tumors can be of immeasurable
benefit to the patient. Approximately half of NF1 patients have cutaneous tumors that number well into the hundreds or thousands. Lesions that occur around
the face, neck and scalp account for roughly 48% of the total and are highly visible and may be distressing to the affected individuals. Those on the remainder of
the body are most common on the trunk (64%), followed by the upper (54%) and lower (31%) extremities. Although more easily concealed, these lesions can
also cause these patients to feel awkward and embarrassed in social situations. As a result, these individuals can become socially withdrawn. Their cutaneous
neurofibromas serve as a constant visible reminder of their perceived disease.

There are several established techniques for removal of the lesions in NF1: surgical excision and/or laser ablation. Surgical excision is a time-tested method that
yields a fairly predictable scar, but removal of hundreds of lesions is impractical due to time constraints. This reality forces patients to choose a small subset of
their lesions to be removed at one time, producing suboptimal patient satisfaction. Lasers may be excellent for small surface lesions, but the treatment can be
time consuming and requires specialized equipment.

I have pioneered a method for the treatment of multiple cutaneous neurofibromas that has numerous benefits. Electrosurgical destruction of cutaneous
neurofibromas (often called electrodessication) with needle-point tip mono-polar cautery has proven to be a rapid, effective method of treating large numbers of
cutaneous neurofibromas in an efficacious manner. The instrumentation is readily available in all outpatient and in-patient surgical centers. Using a low current
provides instant haemostasis with minimal thermal damage to surrounding tissue. The technique is able to treat 500-1000 lesions at any one sitting, in part
due to the possibility of a surgeon and an assistant using two or more cautery devices simultaneously. Cosmesis is excellent with only a few instances of scar
hypertrophy. Results are quite gratifying to patients and surgeon alike.

A Swine Model for Neurofibromatosis Type 1

Friday, 2 November, 17:50 – 18:15

David A. Largaespada, PhD, Masonic Cancer Center, University of Minnesota, Minneapolis, MN

We’ve developed a variety of model systems for studying benign and malignant tumor formation in Neurofibromatosis Type 1 (NF1) syndrome. These include
models based on genetically engineered mice, human cells and most recently Ossabaw breed swine. Using transcription activator-like effector nucleases
(TALENs) and homology-dependent repair (HDR) constructs, we generated NF1+/- (NF1) minipigs harboring a premature termination codon found recurrently
in human patients. We have shown that these NF1+/- minipigs spontaneously develop café-au-lait macules (CALMs) and dermal neurofibromas (NFs) with
spontaneous loss of heterozygosity (LOH), and retention of the mutant NF1 allele, in melanocytes and Schwann cells respectively. One NF1+/- animal developed
an optic pathway glioma-like lesion. We’ve successfully administered 3 different MEK inhibitors to these pigs, achieving pharmacologically relevant levels. Our
results suggest that NF1 minipigs, and the cell lines generated from their tissues, will be useful in answering prevailing questions in the field of NF1 and may
facilitate development of new therapies for NF1-associated nervous system tumors in humans. Potential uses of the NF1 minipig model will be described in the
context of ongoing work in the lab using other model systems.

Full List of Authors: David A. Largaespada1, Sara H. Isakson1, Tony Rizzardi2, Alex Coutts2, Daniel F. Carlson2, Mark N. Kirstein1,3, James Fisher3, Jeremie Vitte4, Nancy Ratner5, Scott C.
Fahrenkrug6, Marco Giovannini4, Christopher L. Moertel1, and Adrienne L. Watson2

1
Masonic Cancer Center, University of Minnesota, Minneapolis, MN 55455
2
Recombinetics Inc., St. Paul, MN 55104
3
Department of Experimental and Clinical Pharmacology, College of Pharmacy, University of Minnesota, Minneapolis, MN 55455
4
Department of Head and Neck surgery, University of California, Los Angeles, Los Angeles, CA 90095
5
Cincinnati Children’s Hospital, Department of Pediatrics, Cincinnati, OH 45229
6
Department of Clinical and Diagnostic Sciences, University of Alabama, Birmingham, AL 35294

Funding: Children’s Tumor Foundation, American Cancer Society, Children’s Cancer Research Fund

Conflict of Interest: DL is the co-founder and co-owner of several biotechnology companies, specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc. (recently acquired
by Immunsoft, Inc.), and B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc., (a subsidiary of Recombinetics, Inc.) and Genentech, Inc. is funding some of his research.
This abstract describes collaborative research between Surrogen and The University of Minnesota. This work was supported by the Children's Tumor Foundation Synodos for NF1
Award, the National Institutes of Health under Award Number T32OD0100993 (SHI), the American Cancer Society Research Professor Award (DAL), and the Children's Cancer
Research Fund.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 13
 Platform Talk: SPRINT: Phase II Study of the MEK 1/2 Inhibitor Selumetinib (AZD6244, ARRY-142886) in
Children with Neurofibromatosis Type 1 (NF1) and Inoperable Plexiform Neurofibromas (PN)
Saturday, 3 November, 8:40 – 9:00

Andrea M. Gross, MD, National Cancer Institute, Pediatric Oncology Branch, National Institutes of Health, Bethesda

Background: PN in NF1 can cause substantial morbidity, and there are no approved medical therapies. In a phase I trial of selumetinib, 17/24 (71%) patients
(pts) had a partial response (PR)1. This open-label phase II study (NCT01362803) determines the PR rate of PN treated with selumetinib and changes in PN
related morbidities.

Methods: Patients 2-18 years old with NF1, inoperable PN and ≥ 1 PN related morbidity received selumetinib at the recommended phase II dose (25 mg/m2
PO BID) with continuous dosing (1 cycle = 28 days). Response was evaluated with volumetric MRI analysis (PR = target PN volume decrease ≥20%) and PN
related morbidities (pain, disfigurement, functional morbidities) assessed with standardized evaluations after every 4 cycles. Wilcoxon matched-pairs signed rank
tests were used to evaluate for change between baseline and pre-cycle13 morbidity assessments.

Results: Fifty children (30 male, median age 10.2 years, range 3.5, 17.4) enrolled. Disfigurement (n=44), motor dysfunction (n=33) and pain (n=28) were the
most frequent PN related morbidities. As of November 5, 2017: median cycle number 19.5 (range 0, 29) with 38 pts remaining on treatment; median change in
PN volume from baseline -27.7% (range -50.6%, 2.2%); best response PR (36 pts, 72%), stable disease (12 pts, 24%); 2 subjects (4%) had no re-staging. Of
the 36 total PR, 32 were confirmed on ≥ two consecutive restaging scans and 22 continue to have a PR ≥ 1 year after it was first achieved. Between baseline
and end of year 1 evaluations, parent and child-reported pain intensity and pain interference scores significantly improved (p <0.01), as did strength (0-5 scale)
and range of motion (degrees) of affected muscle groups/joints (p < 0.01). The most frequent toxicities were nausea/vomiting, diarrhea, asymptomatic creatine
kinase increase, acneiform rash and paronychia. Selumetinib dose was reduced in 12 pts, of which 4 were removed from treatment.

Conclusion: The response rate from this study (72%) confirms our previously observed response rate (71%). Most responses were sustained ≥6 months.
Improvements in PN related pain and motor impairment demonstrate that selumetinib can provide clinical benefit. Data validation and further analyses are ongoing.

1
Dombi, et al. N Engl J Med 2016; 375:2550-2560

Full List of Authors: Andrea M. Gross*1, Pamela Wolters1, Andrea Baldwin1, Eva Dombi1, Michael J Fisher2, Brian Weiss3, Aerang Kim4, Jaishri Blakely5, Patricia Whitcomb1, Marielle
Holmblad6, Staci Martin1, Marie Claire Roderick1, Scott M Paul1, Janet Therrien1, Kara Heisey6, Laurence Doyle7, Malcolm Smith7, John Glod1, Seth Steinberg8, Brigitte C Widemann1
1
National Cancer Institute, Pediatric Oncology Branch, National Institutes of Health, Bethesda, 2Section of Neuro-Oncology, Children's Hospital of Philadelphia, Philadelphia, 3Cancer
and Blood Diseases Institute, Cincinnati Children's Hospital Medical Center, Cincinnati, 4Center for Cancer and Blood Diseases, Children's National Medical Center, Washington,
5
Department of Neurology/Neuro-Oncology, Johns Hopkins University, Baltimore, 6Leidos Biomedical Research, National Cancer Institute, 7Cancer Therapy Evaluation Program,
8
National Cancer Institute, National Institutes of Health, Bethesda, United States

KEYNOTE LECTURE: Epigenetics: One Genome, Multiple Phenotypes

Saturday, 3 November, 9:00 – 10:00

Danny Reinberg, PhD, Howard Hughes Medical Institute, NYU Langone School of Medicine at Smilow Research Center, Department of Biochemistry
and Molecular Pharmacology, New York, USA

Epigenetics encompasses changes in gene expression profiles that occur without alterations in the genomic DNA sequence of a cell. This arises from the dynamic
processes that structure regions of chromosomal DNA through a range of compaction in eukaryotes. The altered pattern of gene expression is pivotal to cellular
differentiation and development and is inherited by daughter cells thereby maintaining the integrity, specifications, and functions for a given cell type. Aberrancies
in this epigenetic process give rise to perturbations that are also inherited and disruptive to normal cellular properties. The histone proteins that package DNA into
chromatin are subject to post-translational modifications generating different chromatin structures. The polycomb repressive complexes play pivotal functions in
maintaining cellular identity through alteration of chromatin domains. Functions of these complexes will be discussed.

14 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PLENARY SESSION: EPIGENETICS
Chairs: Eric Pasmant, PhD, Universite Paris-Descartes, Hopital Cochin, France; Thomas De Raedt, PhD, Children’s Hospital of Philadelphia, US;
Karlyne Reilly, PhD, National Cancer Institute, US

Exploiting Epigenetic Vulnerabilities to Improve Immunotherapy

Saturday, 3 November, 10:30 – 10:55

Thomas De Raedt, PhD, Genetics, Harvard Medical School, Boston; Pediatrics, CHOP/UPenn, Philadelphia

We have shown that MPNSTs are hyper sensitive to treatment of a MEK and BRD4 inhibitor combination. Interestingly, upon treatment we observed a very
rapid change in tumor immune micro-environment with a dramatic increase in cytotoxic T-cells. We hypothesized that we could activate the infiltrating T-cells
with a PD1-checkpoint inhibitor. Indeed, the combination of MEK and BRD4 inhibitors with PD1-checkpoint blockade caused dramatic tumor shrinkage (on
average 80%). Importantly, we show that the cytotoxic T-cells are crucial for the observed tumor shrinkage. Additionally, we show that the observed increase
in T-cells is caused by a loss of M2 macrophages. M2 macrophages secrete factors that prevent T-cell infiltration. Interestingly, treatment with the MEK/BRD4
inhibitor combination dramatically reduces the expression of CCL2, a cytokine that attracts monocytes and macrophages, in the tumor. Finally, we show that
we are able to shrink tumors in a genetically engineered mouse model for MPNSTs by direct inhibition of the macrophages.

Full List of Authors: Thomas De Raedt*1, 2, Vikram Juneja3, Alessandra Gavin1, Chloe Emmerson1, Arlene Sharpe3, Karen Cichowski1
Genetics, Harvard Medical School, Boston, 2Pediatrics, CHOP/UPenn, Philadelphia, 3Harvard Medical School, Boston, United States
1

Learning from Rare Tumors: Tumor Immunogenicity Beyond Somatic Mutation Burden
Saturday, 3 November, 10:55 – 11:20

Josh Waterfall, PhD, Institut Curie, Paris, France

Rare cancers are common, collectively accounting for roughly one quarter of all cancer diagnoses and deaths. Because such homogeneous clinical phenotypes
often result from shared molecular mechanisms, the study of rare cancers has been a rich source of discoveries of fundamental oncogenic pathways. While high
somatic mutation burden can drive immune recognition of certain tumors, mechanisms of immunogenicity in lowly-mutated tumors is relatively uncharacterized.
We will describe how investigating the immune response to malignant rhabdoid tumor, combining patients and animal models as well as bulk and single-cell
genomics, has revealed mechanisms of immune recognition in this context as well as opportunities for combination immunotherapy development.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 15
 Platform Talk: The Genomic Landscape of Schwannomatosis Schwannomas
Saturday, 3 November, 11:20 – 11:40

Sheila Mansouri, Princess Margaret Cancer Center, University Health Network, Toronto, Canada

Background: Schwannomas are characteristic manifestations of NF2 and schwannomatosis syndromes. However, the majority of schwannomas are solitary
and sporadic. It is unclear whether and to what extent sporadic and syndrome-associated schwannomas or their histologic subtypes represent distinct biological
groups. Clinically, although schwannomatosis schwannomas are considered benign, the majority of patients experience unmanageable pain; however, the
underlying mechanism of this pain is not well understood. There is increasing evidence for DNA methylation profiling being able to distinguish biologically
relevant tumor subgroups, even within the same cellular lineage and histopathologically similar tumors.

Methods: In this study, we used Illumina Methylation EPIC arrays for methylome-based characterization of 88 schwannomatosis schwannomas, in comparison
to 90 sporadic schwannomas and 14 NF-2 schwannomas. We performed unsupervised hierarchical clustering selecting 30,000 probes that showed the highest
median absolute deviation (MAD) across all beta values.

Results: Three different clustering sets were utilized to obtain the most refined differentiation. Schwannomatosis schwannomas formed 3 distinct methylome-
based subgroups, which were fully distinct from sporadic schwannomas and NF-2 schwannomas. Additionally, we performed copy number analysis using
the DNA methylation data to infer gross chromosomal deletions or gains among the sporadic and syndromic schwannomas, in addition to schwannomas with
hybrid features. Methylation subgroups were further correlated with clinical parameters including age, gender, anatomic location, tumor size, germline mutation
status (LZTR1/SMARCB1), and 22q LOH, in addition to the histopathologic features associated with each tumor and pain. Furthermore, RNA sequencing was
performed to examine gene expression profiles associated with the 3 methylome subgroups and the data was integrated with DNA methylation profiles to
establish the biological relevance of hypo- and hyper-methylation of the top varying CpG sites.

Conclusions: Our findings suggest that schwannomatosis schwannomas are epignenetically distint from sporadic schwannomas. furhtermore,
schwannomatosis schwannomas form 3 distinct epigenetic subgroups that are significantly associated with tumor anatomic location.

Full List of Authors: Sheila Mansouri*1, Irene Paganini2, Shirin Karimi3, Suganth Suppiah3, Olivia Singh3, Anat Stemmer-Rachamimov4, Kenneth Aldape5, Laura Papi2, Gelareh Zadeh3
1
Princess Margaret Cancer Center, University Health Network, Toronto, Canada, 2Biomedical Sciences, University of Florence, Florence, Italy, 3Neurosurgery, University Health Network,
Toronto, Canada, 4Pathology, Massachusetts General Hospital, Boston, 5Pathology, National Cancer Institute, Bethesda, United States

16 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Cutaneous and Plexiform Neurofibromas Maintain Distinct Epigenetic Profiles with Predicted
Impact on Tumor Biology
Saturday, 3 November, 11:40 – 12:00

Matthew Steensma, MD, Cancer and Cell Biology, Van Andel Research Institute, Byron Center, Grand Rapids, United States

Background: Although prior work confirms a shared cell of origin, there are biological differences between cutaneous (cNFs) and plexiform neurofibromas
(pNFs) that are not fully explained by genomic sequence or structural alterations. We hypothesized that epigenetic alterations correlate with established clinical
features. Prior work questions whether the epigenetic alterations in cNF and pNF tumors are distinct.

Methods: We performed a clinical trial (n=54 subjects, 85 samples) to establish a collection of fully annotated cNF and pNF specimens from NF1 subjects.
Samples were profiled for site-specific methylation differences (EPIC array; Infinium) examining >850,000 loci among cNF’s and pNF’s of varied sizes. Data
were analyzed using a modified workflow similar to the ChAMP EPIC array analysis procedure. Differentially methylated region (DMR) analysis was carried out
using the DMRcate package in “R.”

Results: In this stringent, unbiased analysis we we identified ~100,000 significant DMPs (q < 0.05) and >3000 significant DMRs between cutaneous and
plexiform neurofibromas, as well as significant methylation changes in cutaneous and plexiform neurofibromas relative to their respective normal tissues.
The majority of the differentially methylated regions overlap promoter space (~86%). Additionally, there were approximately 18% of DMPs that overlapped
known CpG islands. QA analysis demonstrated distinct methylation profiles between cNFs and pNFs with a high degree of concordance between cNF tumors
of varying size. Unsupervised hierarchical cluster analysis of global methylation status revealed strong associations based on histopathologic diagnosis (cNF
vs pNF). These differences were not attributable to gender. Chromatin states in hyper- versus hypo-methylated regions differed significantly and the distribution
of differentially methylated regions (DMRs) relative to transcription start sites also differed. Pathway enrichment analysis (Top 25 GO Terms; Top 10 KEGG
Pathways) revealed several strong associations between epigenetic alterations and key signaling pathways. Notable findings include a link to inflammation, pain
receptor pathways, ras signaling, and cytoskeleton regulation.

Conclusions: Our study conclusion is that epigenetic differences between cNFs and PNFs do exist. The biological significance of these profiling differences is
unclear, but the predicted effects strongly implicate differential methylation events in cNF and pNF progression.

Full List of Authors: Matthew Steensma*1, Ben Johnson1, Jamie Grit1, Patrick Dischinger1, Scott Rothbart1, Heidi DeVries2, Megan Bowman3, Carrie Graveel3
1
Cancer and Cell Biology, Van Andel Research Institute, Byron Center, 2Spectrum Research Department, Spectrum Health, 3Cancer and Cell Biology, Van Andel Research Institute,
Grand Rapids, United States

KEYNOTE LECTURE: Schwann Cell Biology

Saturday, 3 November, 13:00 – 14:00

Kristjan R. Jessen, PhD, University College London

The Schwann cell lineage is characterized by a striking phenotypic plasticity. This is seen in the retention of surprisingly broad development options even late in
development, and in the persistent instability of the Schwann cell phenotype in adult nerves. Schwann cell plasticity may play a role in the development of Schwann
cell related tumours and, in particular, dispose to demyelinating diseases. It does, however, provide a striking advantage in one important situation, namely peripheral
nerve injury. In this case, Schwann cells, which in uninjured nerves function to accelerate electrical transmission and maintain nerve homeostasis, are reprogrammed
to cells specialized to deal with injury and promote regeneration, repair Schwann cells. This is a key reason for the strong regenerative potential of peripheral nerves.
Developmentally, myelin and non-myelin (Remak) Schwann cells originate from the neural crest in three main transitions. The first of these is the generation of
Schwann cell precursors (SCP). Unlike crest cells, SCPs are tightly embedded among the axons of embryonic nerves, are acutely dependent on axonal neuregulin1
for survival, and express Schwann cell-related gens. While SCPs therefore unambiguously have a glial phenotype, SCPs also retain one notable feature of the neural
crests, namely a broad developmental potential. Thus, during development SCPs give rise to melanocytes, endoneurial fibroblasts and neurons, in addition to Schwann
cells. The second transition in the Schwann cell lineage is the generation of immature Schwann cells, while the third transition is the formation of the myelin and
Remak cells found in the adult. In injured nerves, the generation of repair cells from myelin and Remak Schwann cells can be considered the fourth main transition in
the lineage. This shows many similarities with injury responses of other tissues, including the process of adaptive cellular reprogramming and activation of epithelial
mesenchymal transitions/stemness genes. Repair cells activate a sequence of supportive functions that engineer myelin clearance, prevent neuronal death, and help
axon growth and guidance. Among the signals that drive Schwann cell development are, Sox10, Zeb2, Krox20 (Egr2), Oct6, Gpr126, and ERK1/2, while the generation
of repair cells is subject to major regulation by c-Jun and Merlin, but involves also other signals including Notch, Gpr126, and ERK1/2. A more detailed knowledge of
these pathways will lead to a more constructive understanding of Schwann cell pathology and allow the manipulation of these signals to enhance the repair supportive
functions of Schwan cells in injured nerves.

Full List of Authors: Kristjan R. Jessen and Rhona Mirsky, University College London

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 17
 PLENARY SESSION: SCHWANN CELLS AND NEUROFIBROMAS
Chairs: Lu Le, MD, PhD, University of Texas, Southwestern, US; Juha Peltonen, MD, PhD, University of Turku, Finland; Eduard Serra, PhD, ICTP, Spain

Developmental Origin and Spatiotemporal NF1 Loss of Heterozygosity Leads to Different Types of Cutaneous
Neurofibroma
Saturday, 3 November, 14:00 – 14:25

Lu Le, MD, PhD, Dermatology, University of Texas Southwestern Medical Center, Dallas; Neurofibromatosis Clinic; Simmons Comprehensive Cancer
Center; Hamon Center for Regenerative Science and Medicine, University of Texas Southwestern Medical Center, Dallas, United States

The neurocutaneous tumor predisposition disorder Neurofibromatosis Type I (NF1) has a wide spectrum of clinical presentations, including developmental,
pigment and neoplastic aberrations of the skin, nervous system, bones, endocrine organs, blood vessels and the eyes. Dermal or cutaneous neurofibromas
(cNF), a Schwann cell tumor in the skin, affect most adults with NF1 and are a major source of emotional, physical and social distress as NF1 patients can
have thousands of these tumors covering most of their skin. Thus, patients with NF1 often identify these tumors as their greatest burden. To date, there is
no available medical treatment for cNF, no known way to prevent them from developing. The major barriers that impede progress in this field are the lack of
accurate models of these common cNF tumors for drug evaluation and a limited understanding of their pathogenesis as well as the identity of specific cell of
origin that directly gives rise to cNF. Here, we uncovered that a Homeobox B transcription factor serve as the lineage marker to trace the developmental origin
of cNF neoplastic cells that completely recapitulates human neurofibromatosis, generating a novel mouse model that spontaneously develops both cutaneous
and plexiform neurofibroma. In addition to providing insights into the developmental origin of cNF, this novel model also demonstrates genetic modifier for
tumorigenesis, yielding vital clues to neurofibroma pathogenesis and now opens the doors for deciphering the evolution of neurofibromagenesis to identify and
testing therapeutic targets.

Full List of Authors: Zhiguo Chen1, Juan Mo1, Jean-Philippe Brosseau1, Tracey Shipman1, Yong Wang1, Chung-Ping Liao1, Jonathan Cooper1, Robert Allaway2, Sara Gosline2, Justin
Guinney2, Thomas Carroll3, Lu Le*1, 4, 5, 6
1
Dermatology, University of Texas Southwestern Medical Center, Dallas, 2Sage Bionetworks, Seattle, 3Molecular Biology, 4Neurofibromatosis Clinic, 5Simmons Comprehensive
Cancer Center, 6Hamon Center for Regenerative Science and Medicine, University of Texas Southwestern Medical Center, Dallas, United States

Funding: This work was supported by funding from Children’s Tumor Foundation, the National Cancer Institute, the US Department of Defense, the Giorgio Foundation, the
Neurofibromatosis Therapeutic Acceleration Program and the NF1 Research Consortium Fund.

Tumour Progression and Pathogenic Mechanisms of Cutaneous Neurofibromas Revealed by Mice Carrying
NF1 Loss in Prss56-Expressing Boundary Cap Cells
Saturday, 3 November, 14:25 – 14:50

Piotr Topilko, PhD, INSERM, France

Background: Patients carrying an inactive NF1 allele develop nerve sheath tumours of Schwann cell origin called neurofibromas (NFs). Genetically engineered
mouse (GEM) models have significantly enriched our understanding of plexiform forms of NFs (pNFs). However, this has not been the case for cutaneous
neurofibromas (cNFs), observed in all NF1 patients, as no previous model recapitulates their development.

Methods: Here, we present a novel GEM model carrying conditional Nf1 inactivation in neural crest derived, Prss56-positive boundary cap cells and their
derivatives.

Results: Targeted NF1 loss leads to bona fide paraspinal and subcutaneous plexiform NFs and cutaneous diffuse NFs. Their analysis allowed us to identify
sub-epidermal glia as a likely candidate for the cell type at the origin of the cNFs and provides insights on disease mechanisms, revealing a long, multistep
pathological process involved in the formation of these slowly-evolving tumors.Moreover, we show that development of cNFs: (i) is associated with abnormal
skin innervation and (ii) can be dramatically accelerated by skin injury.

Conclusions: This new mouse model is an important asset for future clinical and therapeutic investigations of NF1-associated neurofibromas.

Full List of Authors: Katarzyna Radomska1, Fanny Coulpier1, Alain Schmitt2, Amal Debbiche1, Pierre Wolkenstein3, Jean-Michel Vallat4, Patrick Charnay1, Piotr Topilko*1
1
Biology, IBENS, 2Institut Cochin, Paris, 3Dermatology, Hôpital Henri Mondor, Creteil, 4Neurology, Centre Hospitalier Universitaire de Limoges, Limoges, France

18 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Tonic ATP-Mediated Growth Suppression in Peripheral Nerve Glia Requires Arrestin-PP2 and
is Evaded in NF1
Saturday, 3 November, 14:50 – 15:10

Robert Coover, PhD, Experimental Hematology and Cancer Biology, Cincinnati Children’s Hospital, Cincinnati, United States

Background: Mutations in RAS-MAPK pathway genes cause Rasopathies including Neurofibromatosis Type 1 (NF1), which results from inactivating mutations in
the NF1 gene, and hyperactive RAS-GTP. While normal Schwann cells (SCs) are largely quiescent in adult nerves, NF1 mutant SCs proliferate and form benign SC
tumors (neurofibromas).

Methods: We investigated nerve activity dependent quiescence signaling in vivo pharmacologically with nerve activity inhibiting agents, as well as apyrase (an
enzyme that rapidly degrades ATP). We also used a genetically engineered mouse model of NF1 to evaluate possible therapeutic strategies. In vitro analyses were
conducted with wt and NF1 deficient SCs, both immortalized patient derived, and primary murine models. Combinations of shRNA, pharmacological intervention,
and biochemistry techniques to measure extracellular ATP induced phenotypes were used.

Results: We show that nerve activity–dependent ATP maintains normal SC quiescence. ATP activates the G-protein coupled receptor (GPCR) P2Y2. Downstream
of P2Y2, beta-arrestin-mediated signaling results in PP2-mediated de-phosphorylation of AKT, and PP2 activity is required for growth suppression. Elevating ATP
levels in vivo reduces neurofibroma cell proliferation. NF1 deficient SCs are resistant to the effects of beta-arrestin-mediated signaling and PP2-mediated de-
phosphorylation of AKT, and show reduced growth suppression by ATP.

Conclusions: We conclude that beta-arrested-mediated signaling is a critical component of nerve quiescence signaling to SCs. Also that when NF1 is lost, SCs
have aberrant beta-arresting signalling (and by extension GPCR signaling) which results in their ability to evade growth suppressive signals. Our results identify a
potential route that SCs utilize to evade normal growth suppression.

Full List of Authors: Robert Coover*1, Li Guo1, Tabitha Healy2, Katherine Chaney2, Robert Hennigan2, Craig Thomson2, Lindsey Aschbacher-Smith2, Michael Jankowski3, Nancy Ratner2
1
Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital, 2Experimental Hematology and Cancer Biology, 3Department of Anesthesia, Cincinnati Children’s
Hospital, Cincinnati, United States

Platform Talk: Trametinib in Pediatric Patients with Neurofibromatosis Type 1 (NF-1)–Associated Plexiform
Neurofibroma: A Phase I/IIa Study
Saturday, 3 November, 15:10 – 15:30

Christopher L. Moertel, MD, Pediatrics, University of Minnesota, Minneapolis, United States

Background: NF-1 loss-of-function alterations are associated with development of plexiform neurofibromas (PNs). NF-1–associated PNs can arise early in
life in different locations, with variable and significant morbidity. Many patients (pts) progress following surgery, and currently there are no approved systemic
therapies. The MEK inhibitor trametinib is being evaluated in pediatric pts across a spectrum of tumor types in a dose-escalation cohort of a phase I/IIa study
(NCT02124772) to determine a recommended dose; disease-expansion cohorts include pts with NF-1 PN. Here we present an interim analysis of safety and
clinical benefit of trametinib in pediatric pts with NF-1–associated PN.

Methods: Pts aged 1 mo to < 18 y with medically significant, unresectable NF-1–associated PN were treated with trametinib 0.025 to 0.040 mg/kg/d. The
primary objective was safety, and secondary objectives included tumor response assessed by independent review (IR) using published MRI volumetric criteria.

Results: Twenty-six pts received trametinib (0.025 mg/kg/d, n = 21; 0.032 mg/kg/d, n = 1; 0.040 mg/kg/d, n = 4). Presented here are results from the
disease-expansion cohort (n = 10). Median duration of exposure was 408 d (range, 360-429 d), and 8 pts (80%) had treatment ongoing at the data cutoff
(September 2017). Median age was 5.5 y (range, 1-16 y), and prior therapies included surgery (n = 5), biologics (n = 1), and targeted therapy (n = 1).
Treatment-related AEs (TRAEs) were reported in 9 of 10 pts (90%), and 1 pt discontinued due to a TRAE. The most frequent TRAEs were paronychia (50%)
and rash (40%). No deaths occurred on treatment. Analysis of the full NF-1 PN cohort (n = 26) is ongoing; across this cohort, 12 of 26 pts (46%) had a PR
(≥ 20% volume reduction) by IR, and 10 of the 12 responses (83%) were ongoing.

Conclusions: Trametinib demonstrated a manageable safety profile in pediatric pts with NF-1–associated PN. Using volumetric criteria for response
determination, the objective responses observed with trametinib support continued investigation in this pt population.

Full List of Authors: Christopher L. Moertel*1, Geoffrey McCowage2, Sabine Mueller3, Christine Pratilas4, Darren Hargrave5, James Whitlock6, Elizabeth Fox7, Pooja Hingorani8, Mark
Russo9, Kohinoor Dasgupta10, Lillian Tseng9, Bijoyesh Mookerjee9, Birgit Geoerger11
1
Pediatrics, University of Minnesota, Minneapolis, United States, 2Children’s Hospital at Westmead, Westmead, Australia, 3University of California San Francisco, San Francisco,
4
Johns Hopkins University School Of Medicine, Baltimore, United States, 5Pediatrics, Great Ormond Street Hospital for Children, London, United Kingdom, 6Pediatrics, Hospital
for Sick Children, Toronto, Canada, 7Pediatrics, Children's Hospital of Philadelphia, Philadelphia, 8Pediatrics, Phoenix Children's Hospital, Phoenix, 9Novartis Pharmaceuticals
Corporation, East Hanover, United States, 10Novartis Healthcare Pvt. Ltd., Hyderabad, India, 11Pediatrics, Gustave Roussy Cancer Center, Villejuif, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 19
 PLENARY SESSION: PSYCHOSOCIAL IMPACT OF NEUROFIBROMATOSIS
Chairs: Marie-Laure Armand, PhD, Hopital Henri-Mondor, France; Staci Martin, PhD, National Institutes of Health, US; Hilda Crawford, PhD, University of
Sydney, Australia

Supporting the Psychosocial Needs of Individuals with Neurofibromatoses: International Perspectives

Saturday, 3 November, 16:00 – 16:25

Belinda Barton, PhD, The Children’s Hospital at Westmead and University of Sydney, Australia

The neurofibromatoses (NF1, NF2 and schwannomatosis) are the most common neurological suppressor syndromes characterised by nerve sheath tumours
such as neurofibromas and schwannomas. While these tumours are generally benign, they can grow anywhere in the body and are often associated with
significant morbidity including disfiguring cutaneous and plexiform neurofibromas (NF1), unilateral/bilateral deafness and poor balance (NF2), cognitive
impairments (NF1) and chronic pain. These complications can have a significant impact on the quality of life (QoL) of individuals with the disorder with
research generally indicating poorer QoL in all aspects (i.e. physical function, pain, mental health, social functioning) when compared to the general population,
as well as higher rates of depression and anxiety. In addition, adults are often concerned about transmitting the disorder to their children and uncertain
disease progression. Despite the potential impact of neurofibromatoses on the wellbeing on individuals, little is known about how to support the psychosocial
needs of children and adults with the disorder. This presentation will overview the most common psychosocial concerns reported by children and adults with
neurofibromatoses and summarize the clinical practices implemented in different counties that aim to address these concerns. Recommendations on how we
can improve on the psychosocial care provided will also be made.

Acceptance and Commitment Therapy (ACT) for Individuals with NF1, Plexiform Neurofibroma Tumors (PNs),
and Chronic Pain: Results from a Randomized Controlled Trial
Saturday, 3 November, 16:25 – 16:50

Staci Martin, PhD, National Cancer Institute

Background: Plexiform neurofibroma (PN) tumors are associated with mild to severe pain. A randomized-controlled trial was conducted to evaluate the effects of
an ACT intervention among patients 16 to 59 years of age with NF1, PNs, and chronic pain. It was hypothesized that pain interference would decrease from pre-
to post-intervention. Exploratory analyses also examined the relationship between engagement in ACT-based skills and pain outcomes.

Methods: Individuals were recruited from CTF’s NF registry and from NF1 clinics around the country. Participants were randomized to receive the ACT
intervention immediately (ACT group) or after 8 weeks (Wait List group). The intervention was delivered through two, 2-hour in-person sessions followed by
weekly email assignments and biweekly video chats over the course of 8 weeks. Questionnaires assessing pain, pain-related coping, and disease-specific
quality of life (QOL) were administered before and after the 8-week intervention. At follow-up, participants reported how often they practiced ACT skills at home.

Results: Forty-six participants completed the intervention (M age = 28.5 years, 46% female). Pain acceptance increased more in the ACT group (n=24)
compared to the Wait List group (n=22). Mean pain interference (t = 3.1, p < .01) and pain intensity (t = 3.0, p < .01) significantly decreased from pre- to
post-intervention. Finally, engagement in home-based ACT practice had a significant indirect effect on pain interference changes at follow-up, which was
mediated by adaptive changes in pain inflexibility post-intervention (F(2,38)=5.60, p<.01).

Conclusions: Results indicate that ACT reduces pain interference and improves physical and social-emotional wellbeing. Future ACT interventions delivered
remotely may make this intervention accessible to more patients with NF1.

Full List of Authors: Staci Martin, PhD1, Taryn Allen, PhD2, Kari Struemph, PhD1, Mary Anne Tamula, MA2, Andrea Baldwin, CRNP1, Pam Wolters, PhD1, Brigitte Widemann, MD1
1
National Cancer Institute, National Institutes of Health; 2Clinical Research Directorate/Clinical Monitoring Research Program, Leidos Biomedical Research Inc.

Funded by the Neurofibromatosis Therapeutics Acceleration Program, and support was provided by the Intramural Program of the National Cancer Institute and Leidos Biomedical
Research Inc.

20 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Improving Quality of Life (QoL) in Internationally Diverse Patients with NF via Secure Live
Video: Results of 3 Randomized Controlled Trials (RCTs) in Adults, Adolescents and Patients Who Are Deaf
Saturday, 3 November, 16:50 – 17:10

Ana-Maria Vranceanu, PhD, Psychiatry/ Integrated Brain Health Clinical and Research Program, Harvard Medical School/Massachusetts General
Hospital, Boston, United States

Background: Patients with NF are at high risk for poor QoL relative to the general population and to cancer patients. Despite biomedical advances that have
improved outcomes, there remains an enormous need for interventions to reduce the emotional toll of an NF diagnosis. Our team has adapted a mind body
program – The Relaxation Response Resiliency Program (3RP) – that teaches mindfulness, coping and positive psychology skills, for the specific needs of:
1) adults with NF1, NF2 and schwannomatosis (3RP-NF); 2) adolescents with NF1 and NF2 (a3RP-NF); 3) adults with NF2 who are deaf (3RP-NF-CART).
The individually tailored programs were developed iteratively via focus groups and pilot work with each of the 3 patient populations. The 3 programs teach
the same core skills but differ in how the skills are presented, the language used, and the type of stressors addressed. We tested, in 3 individual RCTs, the
feasibility, acceptability and preliminary effect of each version of the program versus an attention placebo control on physical and psychological QoL (co-primary
outcomes) and anxiety, depression, pain and pain interference (secondary outcomes).

Methods: Patients (63 adults, 51 adolescents, 45 patients who are deaf) were recruited through an international NF registry. Screening and consent occurred
via live video, with CART for deaf patients. Data was collected electronically. The intervention programs and control (8 sessions; 90 minutes for adults, 45 for
adolescents, and 60 for deaf adults) were delivered by a psychologist. Participants in the intervention received a patient manual and age/symptom tailored
meditation recordings for home practice.

Results: Most patients who started the program finished (100% adults, 74.5% adolescents, and 90% deaf patients) and reported high satisfaction. Participants
in the 3RP-NF, a3RP-NF and 3RP-NF-CART improved significantly more than control on psychological (p=.04, .03, .045) and physical QoL (p=.05, .03, .05).
A similar pattern was observed for anxiety, depression and pain (p<.05). Only adolescents improved in pain interference (p = .038). All improvements were
clinically meaningful and maintained over time.

Conclusions: A live video, mind body program tailored for the unique needs of adults, adolescents and deaf patients is feasible, accepted and associated
with more improvement in outcomes compared to control. Details on program adaptations, results comparison by patient populations, future directions and
implications for NF care will be discussed.

Full List of Authors: Ana-Maria Vranceanu*1, Eric Ricklin1, Vanessa Merker2, Eric Macklin3, Justin Jordan2, Scott Plotkin2
1
Psychiatry/ Integrated Brain Health Clinical and Research Program, 2Neurology, 3Biostatistics, Harvard Medical School/Massachusetts General Hospital, Boston, United States

Disclosure of Interest: A.-M. Vranceanu has a conflict with: Department of Defense grant W81XWH-17-1-0121 to Ana-Maria Vranceanu; Childrens' Tumor Foundation, 3 clinical
awards to Ana-Maria Vranceanu, E. Ricklin: None Declared, V. Merker: None Declared, E. Macklin: None Declared, J. Jordan: None Declared, S. Plotkin: None Declared

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 21
 Platform Talk: Prospective Patient-Reported Outcomes (PROs) Document Clinical Benefit in Children with
Neurofibromatosis Type 1 (NF1) and Inoperable Plexiform Neurofibromas (PNs) on SPRINT: a Phase II Trial
of the MEK 1/2 Inhibitor Selumetinib (AZD6244,ARRY-142886)
Saturday, 3 November, 17:10 – 17:30

Pamela L. Wolters, PhD, Pediatric Oncology Branch, National Cancer Institute, Bethesda, MD

Background: PNs may cause substantial morbidity, including pain, functional impairment, and disfigurement, which can negatively impact quality of life (QOL).
The FDA emphasizes the importance of documenting clinical benefit in addition to PN shrinkage. We prospectively assessed clinical benefit in children with NF1
and PNs using PROs in a phase II trial of selumetinib (NCT01362803).

Methods: Children 2-18 years old with NF1, inoperable PNs, and ≥1 PN-related morbidity were enrolled. Children (>8) and parents (of children >5)
prospectively completed PROs assessing tumor pain intensity (Numeric Rating Scale-11), pain interference in daily life (Pain Interference Index), QOL (PedsQL),
and global impression of change (GIC) every 2-4 cycles (cy); 1 cy=28 days. Wilcoxon Signed Rank tests evaluated changes in PRO scores from baseline to cy
12, and inductive qualitative analysis identified themes in reported cy 12 changes. Volumetric MRI analysis assessed PN response.

Results: Fifty children (60% male, median age=10.2 years, range=3.5-17.4) enrolled at four sites. As of November 2017, 36 children (72%) had a partial
response (PR; >20% decrease in target PN volume). At baseline, 70% rated having tumor pain (31% mild; 24% moderate; 15% severe). From baseline to cy 12,
child ratings of worst pain in the past week for their physician-selected target tumor and parent/child mean ratings of pain interference improved significantly (each
p<0.01); 74% had clinically significant decreases of >2 points in their target tumor pain of which 79% had a PR; the other 3 had stable disease. Parent total QOL
scores, and physical, emotional, and social (but not school) domain scores improved significantly (each p<0.01) while child physical domain scores improved
significantly (p<0.05). Parent/child median cy 12 GIC ratings both indicated “much improved” change in tumor pain and other tumor-related problems. Cy 12
qualitative responses described predominantly positive changes (parent=91%; child=78%); most frequent themes noted by parents and children were improved
appearance, better function, and decreased pain; less frequently reported negative changes (parent=7%; child=11%) mainly described adverse events.

Conclusions: This is the first trial to document significant improvements in PN-related pain and QOL by prospective standardized PRO evaluations, indicating that
selumetinib provides clear clinical benefits in the setting of PN volume reduction in children with NF1 and PNs.

Full List of Authors: Pamela L. Wolters*1, Andrea Gross1, Staci Martin1, Marie Claire Roderick1, Mary Anne Toledo-Tamula2, Jonas Okafor1, Kari Struemph1, Eva Dombi1, Andrea
Baldwin1, Marielle Holmblad1, Patricia Whticomb1, Michael Fisher3, Victoria Collier3, Brian Weiss4, Beth Stockman4, AeRang Kim5, Karin Walsh5, Seth Steinberg6, Brigitte Widemann1
1
Pediatric Oncology Branch, National Cancer Institute, Bethesda, MD, 2Clinical Research Directorate/Clinical Monitoring Research Program, Frederick National Laboratory for Cancer
Research sponsored by the National Cancer Institute, Frederick, MD, 3Children's Hospital of Philadelphia, Philadelphia, PA, 4Cincinnati Children's Hospital Medical Center, Cincinnati,
OH, 5Children's National Health System, Washington, DC, 6National Cancer Institute, Bethesda, MD, United States

NF Clinical Trials Consortium

Saturday, 3 November, 17:30 – 17:50

Michael Fisher, MD, Children’s Hospital of Philadelphia

The Neurofibromatosis Clinical Trials Consortium (NFCTC, http://www.uab.edu/nfconsortium) was established by the Department of Defense Neurofibromatosis
Research Program (NFRP) to develop and perform clinical trials for the treatment of neurofibromatosis (NF) complications in children and adults. The
Consortium is composed of fifteen clinical sites, nine collaborating affiliate sites, and an Operations Center at the University of Alabama at Birmingham. The
purpose of the Operations Center is to provide administrative, data management, and statistical support to the NFCTC. Each of the clinical and collaborating sites
has expertise in the treatment and management of NF and an established patient population available for clinical trials. The objectives of the Consortium are to
develop innovative biologically-based therapies for both children and adults with NF, and to implement trials for multiple clinical manifestations of NF in order to
rapidly improve the outcome for children and adults with NF.

22 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PLENARY SESSION: MUTATIONS IN OTHER CANCERS
Chairs: Meena Upadhyaya, PhD, University of Cardiff, Wales, UK; Ophelia Maertens, PhD, Brigham and Women’s Hospital/Harvard University, US

The Expanding Role of NF1 in Sporadic Cancer and Resistance to Targeted Therapies
Sunday, 4 November, 10:00 – 10:25

Ophelia Maertens, PhD, Brigham and Women’s Hospital, Harvard Medical School

RAS GTPase Activating Proteins (RAS GAPs) have emerged as an expanding class of tumor suppressors that, when inactivated, provide an alternative
mechanism of activating RAS. The most extensively studied RAS GAP is neurofibromin, the protein product of the NF1 gene. Despite the fact that germline
mutations in NF1 are known to underlie the cancer predisposition syndrome Neurofibromatosis Type 1, until recently a role for NF1 in sporadic cancers had not
been appreciated. Through genomic, cellular, and mouse modeling studies it has now become evident that the NF1 gene/protein plays a much broader role in
cancer than previously recognized. Moreover, accumulating evidence suggests that NF1 plays an important role in mediating resistance to targeted therapies.
For example, NF1 loss has been shown to drive resistance to kinase inhibitors in a number of clinical settings, including BRAF-ˇdriven melanoma and EGFR-
ˇdriven lung cancer. My presentation will highlight mechanistic insights into how NF1 cooperates with other genetic events in sporadic cancer and how this
can be exploited to develop effective therapies.

The Landscape of Molecular Alterations and Mechanisms of Progression NF1 Gliomas

Sunday, 4 November, 10:25 – 10:50

Antonio Iaravone, MD, Columbia University

Neurofibromatosis type 1 (NF1) is a common tumor predisposition syndrome in which glioma is one of the most prevalent tumor. Gliomagenesis in NF1
results in heterogeneous spectrum of low to high-grade tumors occurring during the entire patients lifespan. The pattern of genetic and epigenetic alterations
characterizing glioma that develops in NF1 patients and the similarities with sporadic glioma remain unknown. Here, we present the molecular landscape of low-
and high-grade glioma in patients affected by NF1 (NF1-glioma). We found that the predisposing germline mutation of the NF1 gene was frequently converted
to homozygosity and the somatic mutational load of NF1-glioma was influenced by age and grade. High-grade tumors harbored genetic alterations of TP53 and
CDKN2A, frequent mutations of ATRX associated with Alternative Lengthtening of Telomere and were enriched in genetic alterations of transcription/chromatin
regulation and PI3 kinase pathways. Low-grade tumors exhibited fewer mutations over-represented in genes in the MAP kinase pathway. Approximately 50% of
low-grade NF1-glioma displayed an immune signature, T lymphocyte infiltrates and increased neo-antigen load. DNA methylation assigned NF1-glioma to LGm6,
a poorly defined IDH wild type subgroup enriched with ATRX mutations. Thus, the profiling of NF1-glioma defined a distinct landscape that recapitulates a subset
of sporadic tumors.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 23
 Platform Talk: PEDF Dysregulation Promotes Proliferation and Invasion in Malignant Melanoma with NF1
Driver Mutation
Sunday, 4 November, 11:20 – 11:40

Girish K. Patel, PhD, European Cancer Stem Cell Research Institute

Background: Neurofibromatosis Type 1(NF) and and Legius syndrome are both inherited conditions, caused by germline inactivating mutations in the NF1
and SPRED1 genes respectively, with similar clinical phenotypes that exhibit pigmentary abnormalities including intertriginous freckling and flat localised
hyperpigmented regions of skins termed café-au-lait macules (CALMs). Consistent with activation of the Ras/Raf/MAPK signalling in CALM, we hypothesised
that the mechanism for melanocyte activation would be similar in melanoma with NF1 driver mutations.

Methods: CALMs and unaffected skin biopsies from patients with classical NF (n=2), 3-bp deletion NF (n=3) and Legius syndrome (n=1) were obtained after
regulatory approval by NHS and Ethics committees. The tissue was dissociated and melanocytes selected in culture prior to DNA and RNA extraction for somatic
mutation screening and gene expression profiling respectively. Subsequent gene expression was validated by RT-PCR on normal and NF melanocytes, as well
as metastatic melanoma cell lines. Pigment epithelium derived factor (PEDF) dysregulation was evaluated by cell viability, proliferation, apoptosis and invasion
assays.

Results: All CALMs demonstrated loss of heterozygosity and as expected activation of the RAS-MAPK pathway. Gene enrichment and pathway analysis of the
differentially expressed genes identified a number of significantly enriched gene clusters and pathways including 13 pathways, 37 Biological Process and 23
Molecular Function of Gene Ontology functional annotations. 859 genes were significantly differentially expressed (≥1.5 fold & corrected p<0.05) of which
61 genes which were consistently differentially expressed and were validated by qPCR. Intriguingly SerpinF1, which encodes PEDF, was downregulated in all
samples and discordant with RAS dependent activation of MITF (the melanocyte maturation factor). PEDF loss was observed in melanoma and only in NF1
mutant melanoma cell lines was responsible for increased proliferation, migration and invasiveness.

Conclusions: CALM melanocytes through dysregulation of PEDF acquire a proliferative potential, which is kept in check by cellular senescence. NF1 mutations
are common in malignant melanoma, where loss of senescence combined with dysregulation of PEDF is associated with a more aggressive phenotype.

Full List of Authors: Charlotte Lovatt1, Huw Morgan2, Carlotta Olivero1, Eric Legius3, Girish K. Patel1, Meena Upadhyaya4 and Cardiff University
1
European Cancer Stem Cell Research Institute, 2European Cancer Research Institute, Cardiff University, Cardiff, United Kingdom, 3Department of Human Genetics, KU Leuven, Leuven,
Belgium, 4Medical Genetics, Cardiff Univeristy, Cardiff, United Kingdom

24 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Molecular and Functional Consequences of Somatic NF1 Mutations in Non- Small Cell Lung Cancers
Sunday, 4 November, 11:40 – 12:00

Eric Pasmant, PhD, Service de Génétique et Biologie Moléculaires, Hôpital Cochin, HUPC, AP-HP; EA7331, Université Paris Descartes, Faculté de
Pharmacie de Paris

Background: Driver molecular alterations of RAS-MAPK pathway are found in >40% of non-small cell lung cancers (NSCLC). NF1 is a tumor suppressor gene
that encodes neurofibromin, an inhibitor of the RAS-MAPK pathway. According to The Cancer Genome Atlas data, NF1 somatic mutations are found in ~15% of
NSCLC. However, NF1 mutations are not extensively explored in NSCLC to date.

Methods: We performed NF1 analysis using next generation sequencing in NSCLC surgical specimens. We evaluated the molecular and clinical specificities
of NF1 mutated NSCLC. Then, we established NF1-mutated cellular models with different NF1 wild-type (WT) cell lines. We chose two NSCLC (A549 and
NCI H-1703, ATCC) and one non-tumorigenic human bronchial epithelial cell lines (HBE4-E6/E7-C1, ATCC). Mono- and bi-allelic NF1 mutations models were
established using CRISPR-Cas9 and nickase technologies. In vitro functional tests were performed using these isogenic cell models. In vivo pharmacological
tests were performed in patient derived xenografts (PDX) mice models.

Results: In our series of 138 lung adenocarcinoma specimens, 18% showed NF1 mutations and 8% showed NF1 deletions. Most of patients with NF1 alterations
were males (72%) and smokers (75%). Overall survival and disease-free survival were statistically better in patients with NF1 alterations patients (N=35)
than in KRAS mutated patients (N=30). Then, we established cellular models of NF1-mutated NSCLC. Loss of NF1 expression was confirmed by western blot
(WB): partial and total loss-of-expression of neurofibromin was found in mono-allelic and bi-allelic NF1 mutated cell lines respectively. Bi-allelic NF1 mutations
increase proliferation and migration capacity in our cell models. Using WB, we showed that pERK/ERK ratio was higher in NF1-mutated cell lines versus WT cell
lines, confirming that NF1 loss-of-function triggered RAS-MAPK pathway activation. Transcriptome analysis confirmed this activation. Pharmacological screen
(including MEK inhibitors) in our cell and PDX mice models is ongoing.

Conclusions: Our results confirm that NF1 is frequently mutated and represents a distinct molecular and clinical subtype of NSCLCs. Homozygous NF1 mutated
cells seem more aggressive compared to heterozygous and WT mutated cells. Moreover, NF1 loss-of-function triggered RAS-MAPK pathway activation in our
NF1 mutated models. A better comprehension of functional consequences of NF1 mutations may open new avenues for NSCLC therapy.

Full List of Authors: Camille Tlemsani1, 2, Michel Wassef3, Aurélia Gruber2, Armelle Luscan2, Jennifer Varin2, Doriane Saintemarie2, Ingrid Laurendeau2, Benoit Rousseau4, Franck Letourneur5,
Benjamin Saintpierre5, Didier Decaudin6, Elodie Montaudon6, Fariba Nemati6, Pierre Laurent-Puig7, Nicolas Pécuchet7, Hélène Blons7, Ivan Bièche2, Michel Vidaud1, 2, Eric Pasmant1, 2
1
Service de Génétique et Biologie Moléculaires, Hôpital Cochin, HUPC, AP-HP, 2EA7331, Université Paris Descartes, Faculté de Pharmacie de Paris, 3PSL Research University INSERM
U934, Institut Curie, Paris, 4Institut Mondor de recherche biomédicale, Inserm U955 équipe 18, Faculté de médecine de Créteil, Créteil, 5GENOMI'C Platform, Institut Cochin, 6PSL
Research University, Translational Research Department, Preclinical Investigation Laboratory, Institut Curie, 7INSERM UMR-S1147, Université Sorbonne-Paris-Cité, Paris, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 25
 PARALLEL SESSION: NF2 AND SCHWANNOMATOSIS STATE OF THE ART
Chairs: Michel Kalamarides, MD, PhD; Marco Giovannini, MD, PhD, UCLA, US

An Update on Sensitivity of Pathogenic Variant Testing and Mosaicism in NF2 Using Next Generation Sequencing
Sunday, 4 November, 13:30 – 14:00

D Gareth Evans, MD, Genomic Medicine, University of Manchester, Manchester, United Kingdom

Background: Since the identification of the NF2 gene in 1993, over 4,000 genetic tests have been carried out on unrelated NF2 affected kindreds worldwide
(1108 in Manchester). The pathogenic variant detection rate in leukocyte DNA depends on which generation is tested in a family.

Methods: We used sequence analysis latterly with NGS and a test for large rearrangements (MLPA) to identify pathogenic variants in DNA from leukocytes and
tumour if available from individuals meeting NF2 criteria.

Results: When testing a member of the second generation of families with NF2 we identified a pathogenic variant in 143/154 (93%), 31/154 (20%) had
deletions or duplications detectable on MLPA. In the 11 without a variant on standard testing three have been identified with inversions affecting the 5’ end
(one family) and 3’ end (two families) that required additional genomic analysis (both were missed on standard RNA testing) a further two had chromosome
translocations identified on cytogenetic analysis and a final case had a 5’ UTR variant (c.-66 -65insT) that causes an early initiation codon. Thus additional
testing identifies a pathogenic variant in 149/154 (97%) of NF2 second generation families. In simplex cases (i.e., a single occurrence in a family) the mutation
detection rate is around 50-60% (516/951). About 30% of mutations are not detected due to mosaicism. We have currently identified 217 mosaic patients,
40% by analysis of tumour. When the mutation is not detectable in blood, we can now detect below 1% with NGS particularly when a variant is identified on
tumour analysis. However, variants may still not be detected after targeted analysis. NGS detected the pathogenic variant at 0.8-9% allele frequency in 26/54
(48%) of leukocyte DNA samples where Sanger had previously failed to confirm this.

Conclusions: A combined approach with NGS sequencing and MLPA alongside additional tests to assess the 5' and 3' ends of the NF2 gene is a highly
sensitive technique and implies that there is no other gene that causes classical NF2 in multiple family members. NGS is also a major advantage in identifying
mosaic NF2 and can provide estimates for offspring transmission.

Full List of Authors: D Gareth Evans*1, Miriam Smith1, Andrew Wallace2, George Burghel3
1
Genomic Medicine, University of Manchester, 2Genomic Medicine, Manchester Foundation Trust, 3Genomic Medicine, MFT, Manchester, United Kingdom

Disclosure of Interest: D. G. Evans has a conflict with: Astrazeneca travel paid, M. Smith: None Declared, A. Wallace: None Declared, G. Burghel: None Declared

26 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
The Management of Vestibular Schwannomas in Neurofibromatosis Type 2
Sunday, 4 November, 15:00 – 15:20

Michael J Link, MD, Mayo Clinic

Introduction: The treatment of vestibular schwannomas (VS) in patients with Neurofibromatosis Type 2 (NF2) is challenging and in most cases, controversial. Most
patients with this disorder will eventually become functionally deaf. The goal of treatment is to provide tumor control, and maintain or restore cranial nerve function.

Methods: We reviewed our experience with surgery and radiosurgery for the treatment of NF2 associated VS.

Results: Between Sept, 1999 and Sept, 2018 I have operated 711 VS, of which 33 (5%) were in patients with NF2. Tumor size ranged from intracanalicular
(IC) to > 5.0 cm, (mean: 2.65 cm, median 2.8 cm). Mean clinical and radiographic follow up was 76 months (median: 64 months, range: 14-195 months).
Two patients had severe facial weakness at presentation, one of which underwent facial reanimation surgery. Two patients were House Brackman (HB)
grade 3 at presentation; one remained HB gr 3 and one was HB gr 4 at last f/u. The remaining 29 patients were HB 1 or 2 at presentation. Sixteen patients
(55%) had good (HB gr 1-2) facial nerve function at last f/u, 8 (28%) had moderate (HB gr 3) weakness and the remaining 5 (17%) had severe (H B gr 4-6)
facial weakness. These results are significantly worse compared to size and age matched non-NF2 patients. Ten ears had AAO-HNS Class A or B hearing
preoperatively. Postoperatively, 3 (30%) maintained useful hearing. Seven patients underwent placement of a cochlear implant and 15 Auditory Brainstem
Implants were inserted, one of which had to be explanted for infection.

Between 1990 and 2010, 26 patients with 32 NF2-related VSs underwent SRS using the Gamma Unit (Elekta, Stockholm, Sweden). Median marginal dose and
tumor volume were 14 Gy and 2.7 cm, respectively. Twenty-seven tumors (84%) showed no growth (median follow-up, 7.6 years). Kaplan-Meier estimates for
5- and 10-year progression-free survival were 85% and 80%, respectively. Cox proportional hazards demonstrated a significant inverse association between
higher marginal doses and tumor progression (hazard ratio, 0.49; 95% confidence interval, 0.17-0.92; P = .02). Audiometric data were available in 30 ears,
with 12 having class A/B hearing before SRS. Only 3 maintained serviceable hearing at the last follow-up. Four underwent cochlear implantation. Initially, 3
achieved open-set speech recognition, although only 1 experienced long-term benefit. Facial nerve function remained stable in 50% of cases.

Conclusions: The management of NF2 associated VS is very challenging. There are many strategies available and the treatment strategy in our practice is
extremely nuanced and individualized. The primary goals remain tumor control to avoid life-threatening complications from progressive mass effect, and the
maintenance or restoration of cranial nerve function.

Full List of Authors: Michael J Link, MD, Colin LW Driscoll, MD, Matthew L Carlson, MD, Brian A Neff, MD, Bruce E Pollock MD, Marina L Castner RN, Mayo Clinic

NF2 Management in China

Sunday, 4 November, 16:40 – 17:00

Hao Wu, MD, PhD, Shanghai Ninth People’s Hospital Affiliated Shanghai Jiao Tong University School of Medicine; Ear Institute Shanghai Jiao Tong
University School of Medicine; Shanghai Key Laboratory of Translational Medicine on Ear and Nose Diseases, Shanghai, China

NF2 (neurofibromatosis type 2, NF2) is a heritable syndrome that leads to the development of multiple intracranial tumors, including schwannomas,
meningiomas and gliomas. Bilateral VSs are identified in 90–95% of NF2 patients. There is about 35000 NF2 patients in China. The treatment options for NF2-
associated VSs mainly include observation, stereotactic radiosurgery, microsurgery and drug therapy. Previously in China, NF2 is treated by single department
and the primary treatment strategy is microsurgery. Total resection of bilateral VSs usually results in bilateral hearing loss and facial paralysis, which can
seriously affect the quality of life. Nowadays in China, multidisciplinary management of NF2 is conducted in few centers. Observation is the most common
treatment strategy, that is useful for patients who are poor candidates for hearing preservation surgery. The goal of treating NF2-associated VSs is to maximize
the duration of useful hearing, while minimizing morbidity to the brainstem and facial nerve. However, nowadays the guideline for appropriate management
of NF2-associated VSs are still controversial in China. The treatment for NF2 should be individualized according to clinical symptoms and general condition.
Nowadays, in China, there is few studies about factors influencing growth rate of VSs, drug therapy, Gamma knife surgery, ABI, and so on. Further studies are
needed in order to find out the best treatment for NF2 patients with VSs in China. In our center, MDT is routinely conducted in treating NF2 patients and several
studies, involving 500 NF2 patients, about drug therapy and treatment strategy of NF2 are carried out.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 27
 PARALLEL SESSION: GLIOMA – IN/OUT OF NF
Chairs: Michael Fisher, MD, Children’s Hospital of Philadelphia, US; David Walker, MD, Children’s Brain Tumour Center, UK; Antonio Iavarone, MD,
Columbia University, US

New Insights into the Molecular and Cellular Pathogenesis of Optic Glioma
Sunday, 4 November, 13:00 – 13:25

David Gutmann, MD, PhD, Washington University, St. Louis

Children and adults with neurofibromatosis type 1 (NF1) are prone to the development of benign and malignant tumors of the brain. During the first decade of life,
the most common brain tumor is a low-grade glioma involving the optic pathway (optic pathway glioma; OPGs). Since OPGs are rarely surgically removed, the
Gutmann laboratory has generated a large series of Nf1 mutant mouse strains that develop optic gliomas. Using these informative and authenticated genetically
engineered mouse (GEM) lines, his team has defined the cells and signals important for gliomagenesis, tumor growth, and glioma-induced vision loss. One of
these non-cancerous cell types is a brain macrophage-like cell (microglia), which helps mediate glioma formation, maintenance, and vision loss. In his talk, Dr.
Gutmann will present new data on the immunological circuitry that orchestrates glioma development and progression.

Platform Talk: Synodos for NF1 Glioma: In-Depth Genomic Characterization of Pediatric NF1-Associated
Gliomas Reveals Recurrent Cooperating Mutations and an Aggressive Tumor Subset
Sunday, 4 November, 13:25 – 13:45

David T. W. Jones, PhD, German Cancer Research Center (DKFZ); Hopp Children's Cancer Center at the NCT Heidelberg (KiTZ), Heidelberg, Germany

Background: Low-grade gliomas arise in 15-25% of children with Neurofibromatosis Type 1 (NF1); however, most of these tumors are treated without a prior
tissue diagnosis. As such, very few molecular studies have been performed previously, and little is known about cooperating genetic alterations or other
molecular features that may correlate with their heterogeneous clinical behaviour. We therefore initiated a comprehensive, in-depth profiling of a large series of
pediatric NF1-associated gliomas.

Methods: Tumor and matched germline DNA from 35 pediatric NF1-associated gliomas were subjected to whole-genome sequencing at an average of >70x
coverage, to give sufficient power to detect subclonal alterations in these tumors (which can display low tumor cell purity). Transcriptome profiling by RNAseq
was also performed. DNA methylation microarrays were used for molecular classification of NF1 tumors relative to sporadic gliomas. An independent expansion
cohort of 40 additional tumors was profiled by DNA methylation and targeted panel sequencing.

Results: Next-generation sequencing identified the germline NF1 alteration in all cases, confirming the power of this technique. Truncating alterations were
far more frequent than missense mutations. Copy-neutral loss of heterozygosity (CN-LOH) was a frequent mechanism of inactivating the second NF1 allele,
occurring in approximately half of all tumors examined. Recurrently mutated genes in addition to NF1 included the FGFR1 and PIK3CA genes. While most
tumors showed similarities to subsets of sporadic low-grade gliomas based on DNA methylation, about 10% harbored additional cell cycle gene alterations
(CDKN2A/B loss, TP53/PPM1D mutation) and more closely resembled a recently described subset of IDH-wildtype anaplastic astrocytoma (Reinhardt et al., Acta
Neuropathol. 2018).

Conclusions: This first large-scale genomics study of pediatric NF1 glioma revealed several novel insights, including the secondary mechanisms of NF1 loss, the
presence of additional cooperating oncogenes, and the existence of a subset of tumors with more aggressive molecular features. Our findings have important
clinical relevance in terms of optimizing diagnosis, prognosis and treatment for children with NF1-related glioma.

Full List of Authors: David T. W. Jones*1, 2, Natalie Jaeger1, 2, Joanna J. Philipps3, William A. Weiss3, Angela J. Waanders4, Adam C. Resnick4, Till Milde1, 2, 5, Olaf Witt1, 2, 5, Stefan M.
Pfister1, 2, 5, David H. Gutmann6, Michael J. Fisher4 and The Synodos for NF1 Glioma Consortium
1
German Cancer Research Center (DKFZ), 2Hopp Children's Cancer Center at the NCT Heidelberg (KiTZ), Heidelberg, Germany, 3University of California, San Francisco, San
Francisco, 4Children's Hospital of Philadelphia, Philadelphia, United States, 5Heidelberg University Hospital, Heidelberg, Germany, 6Washington University, St. Louis, United States

28 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Larger Tumor Volume is Associated with Visual Acuity Loss and Axonal Degeneration in
Children with Optic Pathway Gliomas Secondary to Neurofibromatosis Type 1
Sunday, 4 November, 13:45 – 14:05

Robert Avery, DO, Children’s Hospital of Philadelphia, Philadelphia

Background: Approximately 50% of optic pathway gliomas secondary to Neurofibromatosis type 1 (NF1-OPG) cause vision loss. It has been proposed that
larger NF1-OPGs are more likely to be associated with visual acuity (VA) loss and axonal loss, measured using optical coherence tomography (OCT). In this
cross-sectional study, we determined whether MRI measures of tumor dimension (diameter and width) and volume can identify children who experienced VA
loss and axonal loss from their NF1-OPG.

Methods: Children with NF1-OPGs enrolled in an ongoing prospective study of VA were eligible if they had undergone 3-Tesla MRI that included a T1-weighted
volumetric sequence and concurrent OCT measures of the circumpapillary retinal nerve fiber layer (cpRNFL) thickness, a biomarker of axon integrity. The
maximum diameter of the optic nerves and optic tracts along with the width and height of the optic chiasm were measured from the T1 sequence using our
semi-automated algorithm. The individual and combined volume (ml) of these components comprising the proximal anterior visual pathway (AVP) were also
measured. VA and cpRNFL thickness were reported on a per-eye basis.

Results: Fifty-two study eyes (26 children) with NF1-OPGs met inclusion criteria, of which 40% (N=21) had abnormal VA and 60% (N = 31) had normal VA.
Total AVP volume > 1.75 ml detected VA loss with 86% sensitivity and 92% specificity. Linear regression demonstrated that for every 1 ml increase in optic
chiasm volume or AVP volume, the cpRNFL declined by 6 or 10 microns, respectively. In children with OPGs involving the optic chiasm with or without optic
tract involvement, an optic chiasm width greater than 17mm demonstrated a 89% and 86% positive predictive value for VA loss and abnormal cpRNFL thickness
(<80 microns), respectively.

Conclusions: Volumetric measures of NF1-OPGs accurately identified children with VA loss and axonal degeneration. Larger tumor volumes and dimensions
were associated with more severe axonal injury and greater VA loss. Both tumor volume and cpRNFL thickness may be helpful in making treatment decisions in
children with NF1-OPGs that present with abnormal VA or are uncooperative with VA testing.

Full List of Authors: Robert Avery*1, Awais Mansoor2, Grant Liu1, Carmelina Trimboli-Heidler1, Gui-Shuang Ying3, Cameron Centrella1, Roger Packer2, Michael Fisher1, Marius Linguraru2
1
Children's Hospital of Philadelphia, Philadelphia, 2Children's National Health System, Washington, 3University of Pennsylvania, Philadelphia, United States

NF1 Optic Pathway Glioma: Treat or Observe? A Proposal for a Randomised Selection of Treatment Versus
Observation
Sunday, 4 November, 14:05 – 14:30

Amedeo Azizi, MD, PhD, Department of Pediatrics and Adolescent Medicine, Medical University of Vienna, Vienna, Austria

Background: While 15-20% of children with Neurofibromatosis type 1 (NF1) develop optic pathway glioma (OPG), only half of these patients are symptomatic
and only about 20% receive therapy for their OPG. Different risk factors for visual deterioration have been characterised such as age, involvement of the posterior
optic tracts, optic disc pallor and female sex. Yet, data on the natural history of OPG and information on the actual impact of therapy on vision remains scarce
and data are almost exclusively of retrospective nature. The selection criteria for treatment in previous trials of NF1 OPG were visual deterioration, tumour growth
or an ill-defined “threat to vision”.

Present challenges: While novel agents such as MEK inhibitors push forward in upcoming trials other questions still remain to be answered: Do we select the right
patients for treatment? Do treatment strategies result in visual improvement? When is the best time point to treat? Do we know who will profit from treatment?

Treating patients that present with visual deterioration or optic disk pallor may be at high risk for further visual loss, but might already have arrived at a stage of
disease that may not be salvaged. Treating children early in the course of disease might perhaps increase the chances of visual improvement, but would risk to
treat patients that might not show further progression.

A recent survey among international experts (Walker et al., unpublished data) underlined the ongoing dissent on whom or when to treat. Future trial designs
should therefore not only compare different treatment arms but also investigate patient selection criteria. Whereas some patient groups, such as young patients
with bilateral vision loss would immediately receive treatment, others, such as young patients with extensive OPG involving to the posterior tracts but with minor
or no visual loss might be considered for randomisation between treatment and observation.

Conclusion: Identifying the correct selection criteria for treatment initiation in children with NF1 associated OPG may help to in improve visual outcome and
should be considered in the design of future trials.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 29
 PARALLEL SESSION: FUNCTIONAL GENETICS AND INTERPRETATION/CLASSIFICATION OF
VARIANTS
Chairs: Akihiko Yoshimura, PhD, Keio University, Japan; Eric Legius, MD, PhD, University of Leuven, Belgium; Frans Verheijen, PhD, Erasmus MC, Netherlands

Splice Effect Assessment: The Important First Step of NF1 Variant Interpretation
Sunday, 4 November, 13:00 – 13:25

Katharina Wimmer, PhD, Division of Human Genetics of the Medical University Innsbruck, Innsbruck, Austria

The NF1 gene is one of the largest human genes spanning 280kb of genomic DNA and a coding sequence of ~8.4 kb in 57 constitutively and 3 alternatively
spliced exons. Already in the early days of NF1 mutation analysis it was suggested that an RNA-based assay may overcome diagnostic challenges resulting from
the complexity of this large gene lacking mutational hotspots. A seminal paper published in 2000 showed that indeed mutation analysis strategies including RNA-
based mutation analysis protocols as core assays reach the highest mutation detection rates. The two main obstacles of effective mutation detection at transcript
level, i.e. illegitimate splicing and non-sense mediated decay, are effectively circumvented in the currently applied direct cDNA sequencing assay by using RNA
extracted from puromycin treated short-term lymphocyte cultures.

The systematic application of this assay for nearly 20 years reveals that ~30% of NF1 patients have mutations affecting mRNA splicing and two thirds of these
mutations either completely elude genomic DNA (gDNA)-based mutation analysis protocols or they defy ready classification as (likely) pathogenic mutations
without knowledge of splice effect. This explains to a large extend the success of RNA-based mutation analysis in the NF1 gene.

Evaluation of the splice effect in patients’ tissues plays in several ways an important role in studies aimed at assessing genotype-phenotype correlations for NF1
mutations.

The large number of non-canonical NF1 mutations fully characterized at mRNA level represent a highly suited data set to study molecular mechanisms of splice site
definition and disruption and by this to improve tools to predict splicing effects of genomic variants also in other genes.

Although the effect on mRNA splicing will be described for a very large number of NF1 mutations in near future, it can reasonably be expected that private
mutations (i.e. not reported in any other patient) are still identified in a significant percentage of patients. Hence, direct cDNA sequencing to assess a potential
splice effect directly in patients’ tissues will remain an important tool in NF1 diagnostics and research even when massive parallel sequencing becomes the core
assay of the comprehensive mutation analysis protocols. Individual examples show that replacement of these methods by ex-vivo splicing assays using mini-gene
constructs should be critically evaluated as they man not (fully) reflect the natural situation.

Functional Assessment of NF1 Missense Variants Identified in Individuals Suspected of Neurofibromatosis

Type 1
Sunday, 4 November, 13:25 – 13:50

Rick van Minkelen, PhD, Department of Clinical Genetics, Erasmus MC, Wytemaweg 80, 3015CN Rotterdam, The Netherlands

Neurofibromatosis type 1 (NF1) is caused by inactivating mutations in the NF1 tumour suppressor gene. The broad phenotypic spectrum and age-dependent
symptoms associated with NF1 makes clinical diagnosis challenging, particularly in young individuals. Mutation detection is therefore an useful part of the
clinical work-up. However, the large size and high mutation rate at the NF1 locus makes not only mutation detection, but also variant interpretation, challenging.
NF1 encodes neurofibromin (NF1), a GTPase activating protein (GAP) for RAS. To obtain insight into the pathogenicity of NF1 variants identified in individuals
suspected of NF1, we have applied assays that interrogate NF1 function. We have investigated the effects of 26 NF1 variants on NF1 RAS GAP activity and on the
interaction between NF1 and the SPRED1 gene product. In 9 cases (35%), we obtained evidence for an effect of the variant on NF1 function.

Full List of Authors: Mark Nellist, Hannie Douben, Marian Kroos-de Haan, Jacqueline Boonman, Babeth van Ommeren, Marianne Hoogeveen-Westerveld, Anja Wagner, Margreeth van
Vliet, Rick van Minkelen
Department of Clinical Genetics, Erasmus MC, Wytemaweg 80, 3015CN Rotterdam, The Netherlands

30 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Effects of Mutations in SPRED1 on Legius Syndrome and Relationship to NF1
Sunday, 4 November, 13:50 – 14:15

Akihiko Yoshimura, PhD, Department of Microbiology and Immunology, Keio University School of Medicine, Tokyo, Japan

SPRED (Sprouty-related protein with an EVH1 domain) family proteins, initially discovered as c-kit- and c-fms-binding proteins, have been shown to suppress
the Ras-ERK pathway. SPREDs form a subfamily of the Sprouty/Spred family, which is characterized by the Sprouty-related C-terminal cysteine-rich (SPR)
domain. SPREDs also contain N-terminal EVH1 domain and central c-kit binding domain (KBD). Constitutional heterozygous loss-of-function mutations
in the SPRED1 gene cause a phenotype known as Legius syndrome, which consists of symptoms of multiple café-au-lait macules, axillary freckling,
learning disabilities and macrocephaly. It has been demonstrated that SPRED1 functions as a negative regulator of the RAS-ERK pathway and interacts with
neurofibromin, the NF1 gene product. We found that the SPRED1 EVH1 domain interacts with the N-terminal 16 amino acids (aa) and the C-terminal 20 aa of
the GTPase-Activating Protein (GAP)-related domain (GRD) of neurofibromin, which form two crossing α-helix coils outside the GAPdomain. These regions
have been shown to be dispensable for GAP activity and are not present in p120GAP. Several mutations in these N- and C-terminal regions of the GRD in NF1
patients and pathogenic missense mutations in the EVH1 domain of SPRED1 in Legius syndrome reduced the binding affinity between the EVH1 domain and
the GRD. EVH1 domain mutations with reduced binding to the GRD also disrupted the ERK suppression activity of SPRED1. These data clearly demonstrate that
SPRED1 inhibits the Ras-ERK pathway by recruiting neurofibromin to Ras through the EVH1-GRD interaction, and this study also provides molecular basis for
the pathogenic mutations of NF1 and Legius syndrome.

We also characterized SPR domain of SPRED1. Mutations in the SPR (V408E, P415A, P425A/L, C416R, C418R, P422R) disrupt the suppressive function
of SPRED1 against the EGF-mediated ERK activation. SPR is rich in cysteine residues and has been shown to be palmitoylated through the palmitoyl
acyltransferase. We found that SPRED1 with these mutations in the SPR lacked palmitoylation in HEK293 cells, and localized as inclusion bodies but not in the
cell surface membrane. Substitution of cysteine residues suggest that multiple cysteine residues are palmitoylated, but palmitoylation at C426 seems to be most
important for membrane localization. Molecular characterization of mutations found in Legius syndrome will facilitate the understanding of SPRED1 function and
symptoms of the diseases.

Full List of Authors: Akihiko Yoshimura*1, Yasuko Hirata1, Hilde Brems2, Katharina Wimmer3, Ludwine Messiaen4, Eric Legius2
1
Department of Microbiology and Immunology, Keio University School of Medicine, Tokyo, Japan, 2Department of Human Genetics, Catholic University of Leuven, Leuven, Belgium,
3
Department für Medizinische Genetik, Molekulare und Klinische Pharmakologie, Innsbruck, Austria, 4Medical Genomics Laboratory, University of Alabama at Birmingham, Birmingham,
United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 31
 Platform Talk: Neurofibromin a New SUMO Target Involved in Neurofibromatosis Type 1 Disease
Sunday, 4 November, 14:15 – 14:35

Mohammed Bergoug, PhD, CNRS, Centre de biophysique moléculaire, Orléans, France

Background: Neurofibromin (Nf1), the protein responsible for neurofibromatosis type1 disease, has been identified as regulated by post-translational
modifications (PTM) such as phosphorylation and ubiquitination. in our team we are studying its regulation by another PTM called SUMOylation.
In a previous work our team showed that Nf1 colocalizes with PML (promyelocytic leukemia) nuclear bodies (PML-NBs) in the nucleus. PML-NBs are dynamic
proteic structures, and PML protein is their key organizer which localizes at their surface and recruits an ever growing number of proteins whose common
known features are SUMOylation or presence of SUMO interaction motifs. Furthermore, Nf1 carries 15 SUMOylation consensus sites. These observations
suggest that Nf1 might be modified by SUMOylation.

Our aims in studying Nf1 SUMOylation are:

Define whether Nf1 is SUMOylated
Identify on which amino acids the SUMOylation occurs
And how it regulates Nf1 functions

Methods: To detect the SUMOylated forms of Nf1 and its GRD (GAP related domain) and SecPH domains we used Immunoprecipitation, 6HIS Pulldown, and
Western blot.

We also used site directed mutagenesis to identify on which amino acids the Nf1 SUMOylation occurs.

And we are currently performing Ras-GTP and Phospho-ERK quantification to evaluate the Ras-GAP activity of an Nf1 mutant affected in SUMOylation.

Results: This work led us to demonstrate that endogenous Nf1 is SUMOylated. Its GRD and SecPH domains are modified by SUMO2 and they display different
forms of SUMOylation. One amino acid was identified as a SUMO acceptor. Our work is now focused on the functional role of SUMOylation at this site and at the
level of the entire Nf1 protein. We would also like to know if SUMOylation might play a role in NF1 disease, and for this purpose we will test if patient’s mutations
in Nf1 affect the protein SUMOylation profile.

Conclusions: Results already obtained open wide perspectives for studying Nf1 functions regulation by SUMOylation. Our work will focus on the role of
SUMOylation in RAS-GAP activity, Nf1 interaction with its partners, its subcellular localization and its stability.

Full List of Authors: Mohammed Bergoug*1, Aurélie Cosson1, Michel Doudeau1, Fabienne Godin1, Béatrice Vallée1, Hélène Béneditti1
1
CNRS, Centre de biophysique moléculaire, Orléans, France

32 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PARALLEL SESSION: LEARNING DISABILITIES
Chairs: Thijs van der Vaart, MD, PhD, Erasmus University Medical Center, Netherlands; Jonathan Payne, PsyD, Murdoch Children’s Hospital, Australia;
Nicole Ullrich, MD, PhD, Boston Children’s Hospital, Harvard University, US

Multi Parametric Imaging and Behaviour in an Early-Phase Mechanism Trial of Simvastatin for NF1-Autism
Sunday, 4 November, 15:00 – 15:50

Stavros Stivaros, PhD, Shruti Garg, PhD, University of Manchester, UK

Autism Spectrum Disorder is a strongly occurring comorbidity in NF1. Statins rescue the social and cognitive phenotype in animal knockout models, but
translational trials, with NF1 subjects >8yrs on general cognition and behaviour outcomes, have shown mixed results. This single-site triple-blind RCT of
simvastatin vs. placebo breaks new ground by studying for the first time: i) younger children (n=30, mean age 8.1yrs, SD1.8); ii) children with NF1 plus comorbid
autism; iii) multi-parametric imaging outcomes. Assessment (baseline and 12 week endpoint) included: peripheral MAPK assay; awake magnetic resonance
imaging spectroscopy (MRS; GABA and glutamate+glutamine (Glx)); arterial spin labelling (ASL); apparent diffusion coefficient (ADC); resting-state functional
MRI; and autism behavioural outcomes (Aberrant Behaviour Checklist and Clinical Global Impression). Simvastatin was well tolerated. Imaging was fully acquired
in 24/30 subjects (placebo N=15/16; simvastatin, N=11/14). Evidence for simvastatin treatment effect was seen in: i) increased frontal white matter MRS
GABA (t(12)=-2.12,p=.055), GABA/Glx ratio (t(12)=-2.78,p=.016), and reduced deep grey nuclei Glx (t(18)=-3.08, p=.006); ii) increased ASL perfusion
in ventral diencephalon (Mann-Whitney-p<0.01); iii) decreased ADC in cingulate gyrus (Mann-Whitney p<0.01). Machine-learning classification including all
imaging outcomes showed 79% (p<.05) accuracy in endpoint group allocation (placebo vs. simvastatin), compared to chance level at baseline. Autism symptom
response was seen in 3/12 (25%) simvastatin cases compared to none in placebo. Multiparametric imaging suggests evidence of a simvastatin effect towards
normalising function in brain areas previously associated with NF1 pathology and the social brain network. We show feasibility of peripheral MAPK assay and
measurement of autism symptom change.

Full List of Authors: Stavros Stivaros*, Shruti Garg*, Maria Tziraki, Ying Cai, Owen Thomas, Joseph Mellor, Andrew A. Morris, Carly Jim, Karolina Szumanska-Ryt, Laura Parkes, Hamied
A. Haroon, Daniela Montaldi, Nicholas Webb, John Keane, Francisco Castellanos, Alcino J. Silva, Sue Huson, Stephen Williams, D. Gareth Evans, Richard Emsley, Jonathan Green and
the ‘SANTA CONSORTIUM’

Social Behavior Deficits in a Spred1 Knockout Mouse Model of Legius Syndrome

Sunday, 4 November, 15:50 – 16:15

Hilde Brems, PhD, Department of Human Genetics, UZ/KU Leuven, Leuven, Belgium

Loss-of function mutations in the SPRED1 gene lead to a RASopathy disorder termed Legius syndrome. This syndrome presents with a clinical phenotype similar
to Neurofibromatosis type 1, however milder. The features include café-au-lait macules, axillary freckling, macrocephaly, along with frequent incidence of mild
learning disabilities and autism spectrum disorder (ASD). The SPRED1 protein is member of the SPROUTY/SPRED protein family, which acts as a negative
regulator of RAS-MAPK signaling. A mouse model for Legius syndrome, the Spred1 knockout mouse, recapitulates learning deficits seen in this syndrome.
Here, we investigated whether Spred1 knockout mice can model social behavior deficits analogous to the ASD symptoms seen in Legius syndrome. ASD
compromises of deficits in social and communicative behaviors, and restricted and repetitive patterns of behavior. Spred1 knockout mice exhibited deficits
across a range of social behavior tests compared with their wildtype littermate controls, including impairments in social dominance and social communication.
Spred1 knockout mice also exhibit abnormal response to novelty in several behavioral tasks. Treatment of Spred1 knockout mice in adulthood with the specific
MEK inhibitor PD325901 could rescue selected social behavior deficits. These results suggest that social behavior deficits relevant to RASopathy disorders can
be reliably modelled in Spred1 knockout mice, and that RAS-MAPK pathway over-activation plays a key role in mediating these social deficits.

Full List of Authors: Hilde Brems1, Sarah C. Borrie1, Ellen Plasschaert1, Ype Elgersma2, Steven Kushner2, Akihiko Yoshimura3, Eric Legius1.
1
Department of Human Genetics, UZ/KU Leuven, Leuven, Belgium, 2Department of Neuroscience, Erasmus Medical Center, Rotterdam, The Netherlands, 3Department of Microbiology
and Immunology, Keio University School of Medicine, Toyko, Japan

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 33
 Platform Talk: Genotype and Behavioral Phenotypes in Children and Adolescents with Neurofibromatosis Type 1
Sunday, 4 November, 16:15 – 16:35

Andre Rietman, PhD, Child and adolescent psychiatry/psychology, Erasmus Medical Centre, Rotterdam, Netherlands

Background: Neurofibromatosis type 1 (NF1) presents a variable phenotype and genetic factors are important determinants of this variation. The NF1 phenotype
can be determined by specific germ-line mutations and also by the timing and location of the somatic second-hit events that underlie many of the hallmark
somatic features. The cognitive deficits associated with NF1 are generally presumed to be highly variable, but studies addressing this variability and the
underlying factors are lacking. In this study, we describe the association between the NF1 genotype and cognitive and behavioral characteristics of a cohort of
children with NF1 and of monozygotic twin pairs with NF1.

Methods: In this prospective cohort study, the records of all pediatric NF1 patients born 1990-2013 and seen at our NF1 outpatient clinic were reviewed. Data
of neuropsychological assessment included intelligence (IQ) and parent-rated behavioral problems (global behavioral/emotional problems, ASD traits and
ADHD symptoms). Pathogenic NF1 variants were classified as likely to result in the absence of neurofibromin (group X) or as likely to result in the expression
of abnormal neurofibromin (group P). To obtain intra-class correlations (ICC) of IQ within monozygotic twin pairs, a linear mixed model was used. The effect
of mutation type on IQ was determined using Mann-Whitney U tests, one-way ANOVA, and linear regression analyses. All parts of the study were approved or
waived by the medical ethical committee.

Results: The variability in IQ (87.6±15.4) in a cohort of children with NF1 (n=359) resembles the variability in the general population. Monozygotic twin pairs
with NF1 (n=11) show the same high ICC as unaffected monozygotic twins (ICC=0.90). Individuals with group X mutations did not differ from individuals with
group P mutations in IQ or behavioral/emotional problems (p=0.34). However, the P group did show more ADHD symptoms (p=0.048) and more severe ASD
traits (p=0.02).

Conclusions: The variability of cognitive function in our cohort of individuals with NF1 was not significantly different from the reference population indicating
that the contribution of bi-allelic inactivation or specific genetic modifiers to IQ in NF1 is probably minimal. Notably, carriers of a group P mutation showed more
severe ADHD and ASD symptoms compared to those with a group X mutation. Our study adds to the limited knowledge regarding the causes of variability in the
behavioral NF1 phenotype.

Full List of Authors: Andre Rietman*1, Myrthe Ottenhof2, Sabine Mous1, Jeroen Legerstee1, Pieter De Nijs1, Rianne Oostenbrink3, Marie-Claire de Wit4, Rick van Minkelen5, Mark Nellist5,
Henriëtte Moll3, Gwendolyn Dieleman1, Ype Elgersma2 and ENCORE Expertise Center for Neurodevelopmental Disorders, Erasmus Medical Center Sophia Children’s Hospital
1
Child and adolescent psychiatry/psychology, 2Neuroscience, 3Pediatrics, 4Pediatric Neurology, 5Clinical Genetics, Erasmus Medical Centre, Rotterdam, Netherlands

34 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Risk of Psychosocial, Cognitive and Health Impairments in Adult Survivors of Childhood Glioma
with Neurofibromatosis Type 1
Sunday, 4 November, 16:35 – 16:55

Peter de Blank, MD, Pediatrics, Cincinnati Children's Hospital Medical Center, Cincinnati, OH

Background: Neurofibromatosis type 1 (NF1) is associated with an increased risk of tumors as well as non-malignant health conditions. The effect of NF1 on
incidence and severity of late outcomes in adult survivors of childhood glioma is poorly understood.

Methods: 147 >5yr survivors of childhood glioma from the Childhood Cancer Survivor Study (CCSS) were compared to 2,629 non-NF1 glioma survivors
and 5,051 siblings for neurocognitive impairment (CCSS neurocognitive questionnaire), emotional (Brief Symptom Inventory-18), socioeconomic outcomes,
chronic health conditions (CTCAE v4.0) and late mortality (occurring >5yrs from diagnosis) using logistic and Poisson regression adjusted for age, sex, race,
and treatment exposures (central nervous system radiation, surgery and chemotherapy). Specific chronic medical conditions commonly associated with NF1
(including epilepsy, vision loss, hypertension and headache) were also compared between groups.

Results: Compared to siblings, NF1 survivors (age at diagnosis [mean, range]: 6.8yr, 0.3-20.9yr; follow up: 14.4yr, 2.1-24.9yr) were more likely to report
anxiety (RR[95%CI]: 2.2[1.1-4.4]), difficulty with task completion (1.5[1.1-2.0]), not being married (1.8[1.1-2.9]), and not having attended college (1.8[1.2-
2.2]). NF1 glioma survivors reported more severe/life-threatening chronic health conditions (42.2% vs. 28.1%) than non-NF1 survivors at baseline, but the risk
of developing new chronic health conditions or conditions associated with NF1 >5yrs from diagnosis was not different between NF1 and non-NF1 survivors (RR
0.85[0.53-1.38]). However, NF1 survivors had significantly worse late mortality (41.1% 30yrs from diagnosis) compared to non-NF1 survivors (17.5% 30yrs
from diagnosis, p<0.001) and siblings (0.9% 30yrs from entry, p<0.001). Among specific causes of late mortality, death due to second neoplasm was more
likely in NF1 survivors than in non-NF1 survivors (15.7% vs 4.4% 30 years from diagnosis).

Conclusions: NF1 glioma survivors experience worse psychosocial and neurocognitive outcomes than non-NF1 glioma survivors but were not at increased risk for
developing late chronic health conditions. Nevertheless, the risk of late mortality is significantly higher in NF1 survivors compared to non-NF1 survivors. Risk of
late mortality due to second malignant neoplasm is an important consideration when choosing upfront treatment options for children with NF1-associated gliomas.

Full List of Authors: Peter de Blank*1, Nan Li2, Michael Fisher3, Nicole Ullrich4, Smita Bhatia5, Yutaka Yasui2, Charles Sklar6, Wendy Leisenring7, Rebecca Howell8, Kevin Oeffinger9, Kristina
Hardy10, M. Fatih Okcu11, Todd Gibson2, Leslie Robison2, Greg Armstrong2, Kevin Krull2
1
Pediatrics, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, 2St. Jude Children's Research Hospital, Memphis, TN, 3Children's Hospital of Philadelphia, Philadelphia, PA,
4
Harvard Medical School, Cambridge, MA, 5University of Alabama at Birmingham, Birmingham, AL, 6Memorial Sloan Kettering Cancer Center, New York, NY, 7Fred Hutchinson Cancer
Research Center, Seattle, WA, 8MD Anderson Cancer Center, Houston, TX, 9Duke Cancer Institute, Durham, NC, 10Children's National Medical Center, Washington, DC, 11Texas Children's
Hospital, Houston, TX, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 35
 Platform Talk: Understanding the NF1 Experience and the Priorities of NF1-Affected Individuals and Families
For Cognitive and Social-Emotional Research
Sunday, 4 November, 16:55 – 17:15

Tess Kennedy, Children’s National Helalth System, Washington, United States

Background: Patient engagement has become increasingly recognized as a valuable, essential aspect of NF research. Through active patient engagement,
clinicians and researchers can discover the primary concerns and research interests of individuals with NF and their families. Providing patients and families the
opportunity to share their concerns and interests is expected to enhance the development, methodology, and feasibility of clinical trials in NF.

Methods: Data has been collected through an IRB- and CTF Registry-approved REDCap survey developed to ascertain information on the respondents’
NF-related morbidities (neurologic, dermatologic, cognitive, social-emotional), priorities and interests regarding cognitive and social-emotional research, and
willingness to participate in such research. The survey was dispensed to 4,565 individuals consented to the CTF Registry including children with NF1 ages
10-17 years, adults age >18 years with NF1, and parents/caregivers of individuals with NF1 (with or without NF1 themselves). The email was opened by 35%
(N=1615).

Results: To date, 658 individuals have participated in the survey, with 76% being completed in full. Less than 10% of respondents have participated in cognitive
research, while upwards of 50% indicated that they have sought out opportunities for cognitive research. Over 80% of respondents believe that cognitive
research is very/extremely important. The top two areas that respondents indicated should be funded were learning/academics and emotional functioning. ADHD
was ranked last for funding. The respondents willingness to participate in certain areas of research mimicked their ranking of funding (learning/academic and
emotional), with some differences between respondents.

Conclusions: Initial analysis of this rich data highlights that most of the survey respondents believe cognitive and social-emotional research is very important in
NF, but a relatively small number have actually participated and this may be related to limited dissemination of information and opportunities to the broader NF
community. These results also highlight that the respondents consider academically based problems and emotional challenges to be research priorities, which
may or may not align with research foci in the scientific community. Continuing to engage patients and families with NF is expected to enhance the value and
engagement in cognitive research.

Full List of Authors: Tess Kennedy*1, Maureen Hussey2, Karin Walsh3

1
Children's National Health System, 2REiNS (Patient Representative Group), 3Neuropsychology, Children's National Helalth System, Washington, United States

36 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PARALLEL SESSION: GENOTYPE/PHENOTYPE OF NF1
Chairs: Ludwine Messiaen, PhD, University of Alabama at Birmingham, US; Hildegard Kehrer-Sawatzki, PhD, University of Ulm, Germany; Douglas
Stewart, MD, National Cancer Institute, US

Genotype-Phenotype Correlations in NF1: The Current State of Knowledge

Sunday, 4 November, 15:00 – 15:25

Ludwine Messiaen, PhD, genetics UAB, University of Alabama at Birmingham, Birmingham AL, United States

Introduction: As neurofibromatosis type 1 (NF1) is characterized by a highly variable and age-dependent clinical presentation and given the wide NF1 allelic
heterogeneity, genotype-phenotype correlations are exceptional in NF1. Although each specific genotype-phenotype correlation may affect only a small percentage
of individuals, together they impact counseling and management of a significant number of NF1 individuals, thus increased efforts towards identification of
additional clinically relevant genotype-penotype associations are needed.

Methods: Our research currently focuses on the evaluation of the clinical presentation of individuals heterozygous for recurrent NF1 missense or one-amino-acid
pathogenic variants.

Six recurrent missense or one-amino-acid deletion pathogenic variants occur with a prevalence of minimum 0.5% each in the UAB cohort comprising ~8,100
unrelated NF1 individuals, and these were included in this study.

We compare the aggregated phenotypes of the studied groups with cohorts of individuals with NF1 missense / one-amino-acid deletion pathogenic variants
affecting codons 1809, 844-848, 992 as well as with large-scale previously reported cohorts of individuals with “classic” NF1. We apply two-tailed Fisher's exact
test, adjusting P values for multiple comparisons using Benjamini-Hochberg procedure with false discovery rates at 0.05 and 0.01.

Results and relevance: The prevalence of the following features shows statistically significant results between cohorts carrying one of six different pathogenic
variants analyzed: Noonan-like features, pulmonic stenosis, externally visible plexiform neurofibromas, cutaneous neurofibromas, symptomatic spinal
neurofibromas and optic pathway gliomas. Our new and previously reported data demonstrate genotype-phenotype correlations involving these codons that may
be valuable for the management and genetic counseling of a significant number of NF1-affected individuals.

Full List of Authors: Ludwine Messiaen*1, Magdalena Koczkowska1 and clinical/laboratory collaborators
1
genetics UAB, University of Alabama at Birmingham, Birmingham AL, United States

Acknowledgements: This work was supported by the Children's Tumor Foundation by the Isaac and Sadie Fuchs Genotype-Phenotype Study and by internal funds from the Medical
Genomics Laboratory UAB.

Identification of the Factors Underlying the Cognitive Variance Associated with Neurofibromatosis Type 1
Sunday, 4 November, 15:25 – 15:50

Ype Elgersma, PhD, Erasmus University Medical Center

The common autosomal dominant syndrome Neurofibromatosis type 1 (NF1) (1:3000) presents a highly variable phenotype. This is largely triggered by somatic
second-hit mutations, as bi-allelic NF1 inactivation underlies the formation of (plexiform) neurofibromas, café-au-lait spots and malignant peripheral nerves
sheath tumors. In addition, NF1 severity is influenced by certain genotypes as the 17q11.2 microdeletion syndrome and a few specific missense mutations. The
cognitive deficits associated with NF1 are generally presumed to be highly variable as well, but detailed studies addressing the extent of the variability and the
underlying factors are lacking. Here we report that in a large, unselected cohort of children with NF1 (n=359), the variability in cognitive performance resembles
variability of the general population. Moreover, monozygotic twin pairs with NF1 (n=11) show the same high interclass correlation as unaffected monozygotic
twins, indicating that the variance in IQ is mainly genetically determined. We subsequently investigated if specific genetic determinants affect the cognitive
phenotype. Individuals with 17q11.2 microdeletions were found to have lower IQ-scores compared to other types of mutations (p<0.001). IQ of individuals with
missense mutations or small in-frame deletions in the NF1 gene, did not differ from individuals carrying mutations that would result in the inability to express full-
length Neurofibromin (p=0.34). We conclude that the variability of cognitive function in individuals with NF1 is largely determined by natural variation in genetic
background, and that the contribution of bi-allelic inactivation or genetic modifiers of NF1 is minimal. Only 17q11.2 microdeletions were found to worsen the
cognitive phenotype of NF1.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 37
 Looking Beyond the Lamppost: Exome Sequencing in >92,000 People in a Single US Healthcare System to
Investigate Prevalence, Penetrance, and Phenotype in NF1 and Legius Syndrome
Sunday, 4 November, 15:50 – 16:15

Jung Kim, PhD, Clinical Genetics Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD

Like many monogenic disorders, the clinical and research-study ascertainment of individuals with NF1 and Legius syndrome is typically triggered by syndrome
phenotype. The “phenotype first” approach is a traditional and proven one in clinical cancer genetics, and historically has been very productive in linking
phenotype with germline variation. However, risk estimates derived from phenotype-linked ascertainment likely 1) over-estimate syndrome severity and
penetrance, 2) miss non-penetrant risk-variant carriers, 3) miss rare or unknown manifestations of disease and 4) give an incomplete picture of the phenotypic
spectrum, especially at older ages. In this investigation, we used the genotype-first approach to evaluate the prevalence, penetrance, and phenotype of
individuals harboring pathogenic NF1 and SPRED1 germline variation in >92,000 exomes from the Geisinger cohort (98% white; 60% female; median age
59 years; age range 2-89 years). Germline DNA from Geisinger MyCode participants underwent exome sequencing as part of the DiscovEHR collaboration of
Geisinger (Danville, PA) and the Regeneron Genetics Center (Tarrytown, NY). NF1 and SPRED1 variants with minor allele frequency < 1% were classified as
pathogenic (P), likely pathogenic (LP), variant of unknown significance (VUS), likely benign (LB), and benign (B) per ACMG-AMP guidelines. We used ICD9/10
codes to query the electronic health record (EHR) for diagnoses related to malignancy. In individuals with NF1 and SPRED1 P/LP variation, cancer diagnoses
were confirmed by the Geisinger Cancer Registry and are exclusive of non-melanoma skin cancers. Since NF1 is a clinical diagnosis and not a genetic one,
caution is needed in estimating prevalence and penetrance based on genotype and EHR-derived phenotypes only. Our genotype first approach to ascertain NF1
and SPRED1 P/LP prevalence and penetrance in a large population-based, exome-sequenced cohort is novel. In participants with unusual or severe phenotypes,
the identification of P/LP variation in NF1 and non-NF1 genes shows how the genotype first approach can boost efforts to identify modifiers of NF1.

Full List of Authors: Jung Kim1, Uyenlinh Mirshahi2, David J. Carey2, Douglas R. Stewart1 on behalf of the DiscovEHR Collaboration
1
Clinical Genetics Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD, 2Weis Center for Research, Molecular & Functional Genomics,
Geisinger, Danville, PA

Platform Talk: Comprehensive Screening of NF1 Missense Mutations for Genotype-Phenotype Correlations
Using Drosophila and Mammalian Cell Models
Sunday, 4 November, 16:15 – 16:35

James Walker, PhD, Center for Genomic Medicine, Mass General Hosp, Boston, United States

Background: NF1 encodes a large ~320 kDa protein, termed neurofibromin, the only known function of which is to serve as a Ras-specific GTPase-activating
protein (GAP). However, pathogenic missense mutations are found throughout NF1, suggesting other parts of the highly conserved protein are also essential for
its function. We hypothesize that amino acid substitutions outside of the GAP domain may disrupt important novel functional domains, interactions with other
proteins, or alter the regulation, stability or subcellular localization of neurofibromin – any of which could perturb its activity and potentially contribute to the
varied clinical symptoms of NF1.

Methods: We have investigated the molecular and cellular consequences of NF1 missense mutations using both an invertebrate model of NF1 (Drosophila) and
human cells. Drosophila was used to rapidly assess the residual function of dNf1 transgenes bearing corresponding mutations from patients. The approach
relies on the fact that the majority of amino acids mutated in human neurofibromin are conserved in Drosophila, allowing us to correlate cellular and molecular
phenotypes to specific mutations in different regions of neurofibromin. The most informative NF1 mutations from our Drosophila screen are subsequently
modeled in human induced pluripotent stem cells (hiPSCs) using CRISPR/Cas9 gene editing.

Results: We have conducted a functional assessment of the ability of dNf1 transgenes bearing conserved pathogenic missense mutations from NF1 patients
to rescue established phenotypes in a fruit fly model of NF1: increased Ras signaling, reduced systemic developmental growth, aberrant circadian rhythms and
defective cognitive function. To determine whether mutations in transgenic neurofibromin alter protein interactions, we have used affinity purification and mass
spectrometry. Altered subcellular localization of mutant transgenic neurofibromin in fly neurons was assessed using confocal microscopy. Selected mutations of
interest were engineered into hiPSCs using CRISPR/Cas9 and used to derive neurons to create isogenic disease models enabling specific NF1 mutations to be
correlated with phenotypes.

Conclusions: This study, combining functional testing in vivo in Drosophila, with subsequent validation in human cells, allows us to correlate cellular and
molecular phenotypes to specific patient-derived NF1 mutations in different regions of neurofibromin. This knowledge may aid the discovery of new biomarkers
and therapeutic targets for NF1.

Full List of Authors: James Walker*1, Timothy O’Meara1, Brittany Leger1, Samuel Ching1, Rebecca Drake1, Theodosia Charitou1, Sayantanee Paul2, Sarah DuBois-Coyne2, Melissa Brown2,
Alexey Veraksa2
1
Center for Genomic Medicine, Mass General Hosp, 2University of Massachusetts, Boston, United States

38 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: NF1-Deficiency is Linked to Estrogen Signaling in Breast Cancer with Evidence of Genotype-
Phenotype Correlation
Sunday, 4 November, 16:35 – 16:55

Matthew Steensma, MD, Cancer and Cell Biology, Van Andel Research Institute, Byron Center, United States

Background: Breast cancer is an established phenotype of Neurofibromatosis Type 1 (NF1). NF1-related breast cancer survival is diminished regardless of age at
presentation, gender or tumor receptor status. The role of neurofibromin-mediated tumor suppression in breast cancer is unclear.

Methods: We developed NF1-deficient rats bearing mutations in the GRD of the NF1 gene orthologue. Germline Nf1-haploinsufficiency was induced using
CRISPR-Cas9 gene editing. Animal survival, hormone receptor status, kinase signaling (pERk, pMET, pSTAT3, pAKT) and Nf1 transcript variation (PCR) were all
assessed. We also assessed NF1 loss in sporadic human breast cancer using the METABRIC breast cancer dataset. To identify biological networks impacted by
NF1 deficiency, we constructed unbiased gene co-expression networks using weighted gene correlation network analysis (WGCNA).

Results: Germline, Nf1 deficient female rats exhibited early, penetrant mammary adenocarcinoma. The observed breast cancer phenotype was more penetrant
but not exclusive to the nonsense mutant lines. Nonsense mutations exhibited significantly diminished survival compared to missense mutant lines. ER+/
PR+/Her2+ expression was noted in the vast majority of tumors (IHC). Estrogen-dependence was verified by estrogen-ablation in Nf1 rats where rapid tumor
regression was observed (mean 4.7% size reduction/day (p<0.0001)). Conversely, Nf1-deficiency correlated with increased ER phosphorylation in mammary
epithelial cells, tissue and mammary adenocarcinomas. We also identified distinct neurofibromin protein isoforms in mammary tissue that were altered during
tumorigenesis. Alternative splice variants at mutant Nf1 loci correlated with diminshed survival among isogenic strains.

METABRIC data analysis revealed that NF1 shallow deletions occur in 25% of breast cancer patients and are associated with a significantly higher tumor
grade, stage, and size (p <0.0001). Moreover, NF1 shallow deletions were associated with a 1.4 fold increase in mortality (p = 0.001) relative to diploid NF1
status. Unbiased WGCNA revealed a network connected to ESR1 (estrogen receptor), androgen receptor (AR) and FOXA1. Unsupervised clustering of genes
demonstrated a strong association of NF1 shallow deletions with both ER+ and ER- tumor subsets.

Conclusions: These results demonstrate a significant role for NF1 in both NF1-related and sporadic breast cancer, and highlight a potential functional link
between neurofibromin and the estrogen receptor.

Full List of Authors: Matthew Steensma*1, Patrick Dischinger1, Curt Essenburg1, Elizabeth Tovar1, Jamie Grit1, Candace King1, Ashley Turner1, Anil Challa1, Robert Kesterson1, Carrie Graveel1
1
Cancer and Cell Biology, Van Andel Research Institute, Byron Center, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 39
 Platform Talk: Genetic Basis of Variable Clinical Expression in Neurofibromatosis 1: Multiple Loci Identified in
the First Genome-Wide Association Study
Sunday, 4 November, 16:55 – 17:15

Audrey Sabbagh, PhD, UMR 216 IRD MERIT

Background: Neurofibromatosis type 1 (NF1) is a Mendelian disease with full penetrance. Absence of genotype-phenotype correlation demonstrated the minor
role of the NF1 gene in the great variability of NF1 clinical expression with the rare exceptions of (i) missense mutations in codons 844-848 and Arg1809, one
short inframe deletion (p.Met992del), and (ii) the recurrent large deletions of the NF1 locus. Evidence of modifier genes was provided by animal models and
intra-familial phenotypic correlations. A large genotype-phenotype database dedicated to NF1 was established in France: a standardized phenotypic description
and the NF1 genotype were collected in more than 1,500 patients. We aim to carry out the first NF1 genome-wide association study (GWAS) in collaboration
with the National Genotyping Center and the Neurofibromatosis Reference Center.

Methods: The GWAS analysis was performed on the discovery cohort (997 patients) and the replication cohort (501 patients) using the Illumina
OmniExpressExome chip. The chip includes ~ 700,000 tag SNPs (capturing a large part of the common variation of the human genome) as well as >260,000
functional exon variants. Association tests with 14 clinical traits (5 quantitative and 9 binary traits collected with a standardized case report from) were
performed using a mixed-effects model, considering the age, gender, and inherited or de novo nature of the NF1 mutation as fixed effects.

Results: NF1 genotyping of the cohort allowed exclusion of patients with mutations associated with a known genotype-phenotype correlation (missense
mutations in codons 844-848 and Arg1809, the short inframe deletion p.Met992del, and large NF1 locus deletions). Our analysis confirmed the strong
heritability of most of the NF1 clinical traits. Moreover, we identified several loci showing significant genome-wide associations (p<10-8) with NF1 major
phenotypic traits in both discovery and replication cohorts.

Conclusions: The results of this study will contribute to a better understanding of the genetic basis of NF1 expression variability and its underlying
pathophysiological mechanisms. Our study will be completed by identification of the specific modifier genes and their functional studies in relevant cellular
models by genome editing approaches.

Full List of Authors: Audrey Sabbagh*1, Eric Pasmant2, Béatrice Parfait3, Anne Boland-Auge4, Delphine Bacq-Daian4, Ingrid Laurendeau3, Salah Ferkal5, Laurence Allanore5, Jean-François
Deleuze4, Michel Vidaud2, Dominique Vidaud2, Pierre Wolkenstein5
1
UMR 216 IRD MERIT, 2Genetics, 3Cochin Hospital and Paris Descartes University, PARIS, 4Centre National de Recherche en Génomique Humaine, CEA, Evry, 5Dermatology Department,
Hôpital Henri-Mondor, AP-HP, Créteil, France

KEYNOTE LECTURE: The Translational Mouse: From Genetic Models to the Mouse Hospital
and Co-Clinical Trials, for a Novel Understanding and Treatment of Cancer
Monday, 5 November, 8:30 – 9:30

Pier Paolo Pandolfi, PhD, Director, Cancer Center & Cancer Research Institute, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston,
MA, USA

The Co-Clinical Trial Project is a cutting-edge new platform for clinical trial optimization, which we have developed at our Cancer Center at Harvard. The platform
also rests on a “Mouse Hospital” infrastructure, which is equipped as a human hospital, if not better, to perform experimental clinical trials in mouse models of
disease, exactly as they would be run in the human hospital.

In the “Co-Clinical” Approach for Cancer Therapy optimization, mouse models of cancer, which are representative of the diversity of human cancer, are treated
with the same drug, and following the very same clinical protocol offered to human patients enrolled in experimental clinical trials in the human hospital. This
allows for “mice-to-human-to mouse” stratification and cross-validation of response and resistance to specific treatment modalities, and for the identification of
effective therapies that overcome such resistance.

Drugs can be tested on Organoids and “Immune-deficient” Patient Derived Xenograft: (PDX) models that are generated from biopsies from the same very same
patients enrolled in the trials. Importantly, however, “Co-Clinical” Trials can also be run by enrolling “immune-competent” genetically engineered mouse models
of cancer (GEMMs) bearing genetic mutations associated to human cancer to assess how the various cancer genetic make ups and therapeutic treatments
impact and are impacted by the immune microenvironment. Exciting new data from these platforms and on-going analyses will be presented.

40 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PLENARY SESSION: MICROENVIRONMENT
Chairs: Piotr Topilko, PhD, INSERM, France; D. Wade Clapp, MD, Indiana University, US; Alison Lloyd, PhD, University College, London, UK

Cellular and Extracellular Microenvironment of Cutaneous Neurofibromas

Monday, 5 November, 9:30 – 9:55

Juha Peltonen, MD, PhD, Institute of Biomedicine, University of Turku, Turku, Finland

The growth of cutaneous neurofibromas (cNFs) is a result of cell proliferation and deposition of extracellular matrix (ECM) which constitutes a significant proportion
of the tumor volume. Tumor growth requires permissive and enhancing, but also restricting factors since the size of the tumors typically does not exceed few
centimeters. Ultimately an equilibrium is reached and growth markedly slows or stops. The main framework of the extracellular matrix of cNFs comprises of
an abundant but relatively loose network of small diameter collagen fibrils organized in fine bundles. The majority of the neurofibroma cells can adhere to this
framework which in turn has the potential to influence their gene expression profiles. The ECM also may direct the migration of blood derived lymphocytes and
precursors of mast cells and macrophages. The non-fibrillar ECM effects the movement of soluble small molecular effectors in cNFs, such as growth factors and
cytokines. Most of the cNF cells express receptors of sex hormones explaining the apparent association of neurofibroma growth with puberty and pregnancy.
Previous studies by us and others have shown that the Schwann cells with a hit in both NF1 alleles (NF1-/-) are particularly sensitive to sex hormone regulation,
and on the other hand NF1-/- Schwann cells actively express stem cell factor (SCF) to which the NF1+/- mast cells are particularly sensitive. Our recent focus has
targeted the tumor immunology of cNFs. Immunity has a central and complex role in tumor growth, which is utilized when developing new treatment modalities
for tumors. Transformed cells are normally eradicated by cooperation of innate and adaptive immune cells. Immune surveillance could be expected to recognize
neurofibromin neoepitopes resulting from the second hit of the NF1 gene, but this is not taking place in cNFs. Mast cells are increasingly more interesting as
druggable targets in tumor biology and neurogenic itch. Thus, tumor immunology of cNFs has been our recent focus. In the first phase of the work, T cell
subpopulations of cNFs have been characterized with appropriate CD antibodies, and the two main mast cell populations by their enzyme profiles.

Cxcr3 Expressing Leukocytes Are Necessary for Neurofibroma Formation in Mice

Monday, 5 November, 9:55 – 10:20

Nancy Ratner, PhD, Divisions of Experimental Hematology and Cancer Biology, Cincinnati Children’s Hospital Medical Center, University of Cincinnati
College of Medicine

Plexiform neurofibroma is a major contributor to morbidity in Neurofibromatosis type I (NF1) patients. Macrophages and mast cells infiltrate neurofibroma,
and data from mouse models implicate these leukocytes in neurofibroma development. Anti-inflammatory therapy targeting these cell populations has been
suggested as a means to prevent neurofibroma development. Here, we compare gene expression in inflamed nerves from NF1 models which invariably form
neurofibroma to those with inflammation driven by EGFR overexpression which rarely progresses to neurofibroma. We find that the chemokine Cxcl10 is uniquely
up-regulated in NF1 mice that invariably develop neurofibroma. Global deletion of the Cxcl10 receptor Cxcr3 prevented neurofibroma development in these
neurofibroma-prone mice. Cxcr3 expression localized to T cells and dendritic cells (DCs) in both inflamed nerves and neurofibromas. These data support a
heretofore unappreciated role for T cells/DCs in neurofibroma initiation.

Full List of Authors: Jonathan S. Fletcher1,2, Jianqiang Wu1, Walter J. Jessen1,3, Jay Pundavela1, Jacob A. Miller1, Eva Dombi4, Mi-Ok Kim5, Tilat A. Rizvi1, Kashish Chetal6, Nathan
Salomonis6, and Nancy Ratner1*
1
Divisions of Experimental Hematology and Cancer Biology and 6Biomedicial Informatics, Cincinnati Children's Hospital Medical Center, University of Cincinnati College of Medicine,
3333 Burnet Ave., Cincinnati, OH, 45229, USA; 2Immunology Graduate Program, University of Cincinnati College of Medicine; 3Laboratory Corporation of America Holdings (LabCorp),
Burlington, NC, 27215, USA; 4Center for Cancer Research National Cancer Institute Building 10, Room 1-5750 Bethesda, MD 20892-1101; 5UCSF Helen Diller Family Comprehensive
Cancer Center Dept. of Epidemiology & Biostatistics, UCSF Box 0128 San Francisco, CA 94143-0128

This work was supported by 1 R01 NS28840; The Children’s Tumor Foundation; The Neurofibromatosis Therapeutic Acceleration Program and DOD W81XWH-11-1-0057 (to NR),
and 5 F30 NS096796-02 (to JSF).

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 41
 Platform Talk: Dermal Fibroblast-Derived Exosomes Enhance Angiogenesis in a Tissue-Engineered Skin
Model of Neurofibromatosis Type 1
Monday, 5 November, 10:20 – 10:40

Vincent Roy, Axe médecine regénératrice, Centre du CHU de Quebec/LOEX; Surgery, Université Laval, Quebec

Background: Neovascularization is a critical process for tumor formation in neurofibromatosis type 1 (NF1). In neurofibroma development, complete loss of
neurofibromin gene (NF1) function in Schwann cells is required. Plexiform neurofibromas are highly vascularized tumors that can progress to malignancy. It
is known that NF1 haploinsufficiency in endothelial cells exaggerates angiogenesis. Here we hypothesize that NF1-/+ human dermal fibroblasts secrete small
vesicles that communicate a pro-angiogenic signaling and play a crucial role in modifying tumor microenvironment. The purpose of this study is to evaluate the
protein content of exosomes derived from dermal fibroblasts and validate their effect on endothelial cells and angiogenesis.

Methods: Tissue-engineered skin (TES) was generated with dermal fibroblasts and keratinocytes, isolated from NF1 patients and control individuals, and
microvascular endothelial cells (MVECs) using the auto-assembly method. Micro-capillary networks were analyzed and compared to healthy control. Exosomes
produced for 72h from fibroblast sheets were isolated using a commercial kit and exosome concentration and size were evaluated with a nanosizer. Exosomal
proteins were also analyzed using an angiogenesis protein array. NF1 exosome’s direct influence on MVECs was evaluated by a tube formation assay on Matrigel®.

Results: MVEC seeded in NF1-TES formed a denser network with increased nodes, junctions and capillary branching frequencies compare to controls. No
variation in exosomal size and particles and proteins concentrations were shown between the studied populations. Analyses of NF1-derived exosomes reveals
pro-angiogenic profiles that significantly enhanced tube formation on Matrigel® after 24h of co-culture with MVECs. Finally, uptake of small vesicles by MVECs
was also confirmed by labelling exosomes with PKH26.

Conclusions: Our study suggests that NF1 haploinsufficiency alters dermal fibroblast function and creates a pro-angiogenic signal that increases capillary
formation. Our NF1-TES model could become a unique tool to better characterize the pathogenic mechanisms associated with skin tumor genesis. Ultimately, it
could provide better tools to develop new therapies for patients through the development of personalized medicine strategies.

Full List of Authors: Vincent Roy*1, 2, Rémy Lamontagne1, 2, Lydia Touzel-Deschênes1, 2, Peter Kannu3, Hélène T. Khuong1, 4, Nicolas Dupré1, 2, 5, François Gros-Louis1, 2
1
Axe médecine regénératrice, Centre du CHU de Quebec/LOEX, 2Surgery, Université Laval, Quebec, 3Pediatrics, University of Toronto, Toronto, 4Medicine, Université Laval, 5Axe
neurosciences, Centre du CHU de Quebec, Quebec, Canada

42 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Targeting the Hyaluronan-Rich Peripheral Nerve Sheath Tumor Microenvironment to Improve
Drug Efficacy and Delivery
Monday, 5 November, 10:40 – 11:00

Bryant J. Keller, University of Minnesota – Twin Cities, Minneapolis

Background: Malignant Peripheral Nerve Sheath Tumors are aggressive soft tissue sarcomas that manifest at a high rate in individuals with the genetic cancer
predisposition syndrome Neurofibromatosis Type 1. Although many therapeutic avenues have been explored, little improvement has been seen in the poor
prognosis. Surgical resection and nonspecific chemotherapeutics remain the only standard of care. Evidence is mounting that a desmoplastic reaction involving
high deposition of the glycosaminoglycan Hyaluronic Acid (HA) in the Extra-Cellular Matrix (ECM) can lead to poor drug penetrance and efficacy. Human MRI
examples display large regions of poor uptake of contrasting agent, and a human tissue microarray shows 100% positivity of HA deposition through all samples,
including plexiform neurofibromas. Breaking down this physical barrier is a promising potential avenue to improve drug penetrance, perfusion, and efficacy.

Methods: To accurately model high grade Peripheral Nerve Sheath tumors (PNST), we implemented our previously described genetically engineered mouse
(GEM)-PNST consisting of Dhh-Cre; Nf1fl/fl; PTEN fl/fl (Keng V., et al., Cancer Research. 2012). This model manifests multi-focal, 100% penetrant peripheral
nerve sheath tumors with a median lifespan of 18 days post birth. These tumors also display elevated HA levels in the ECM as well as collapsed and sparse
vasculature. This model is amenable for treatment with a human pegylated hyaluronidase (PEGPH20) from Halozyme (San Diego, CA) to deplete HA and improve
drug delivery.

Results: The GEM-PNST model, when treated with PEGPH20, displayed a dose-dependent decrease in tumor HA levels with no obvious toxicities to the animals.
Taking advantage of the natural fluorescence of the broad-spectrum chemotherapeutic doxorubicin, sections cut from animals that had received PEGPH20
treatment before doxorubicin showed a quantifiable increase in perfusion. Furthermore, CD31 staining suggests an increase in blood vessel patency post-
PEGPH20 treatment. Dual treatment of PEGPH20 with doxorubicin slightly improved longevity of these animals as compared to the monotherapy. Preliminary
results also show an exciting increase of life when PEGPH20 is combined with the targeted MEK inhibitor PD0325901.

Conclusions: PEGPH20 shows promising therapeutic benefit in targeting physical barriers in MPNSTs. Improved drug delivery and efficacy will open avenues to
further drug combinations in this currently incurable malignancy.

Full List of Authors: Bryant J. Keller*1, Adrienne Watson2, Kyle Williams1, Stephen Scully1, Rory Williams1, Leah Anderson1, Justin Knight1, Colleen Forster3, Kwangmin Choi4, Marjorie
Carlson1, Nancy Ratner4, Paolo Provenzano1, David Largaespada1
1
University of Minnesota -- Twin Cities, Minneapolis, 2Recombinetics Inc., St. Paul, 3BioNet, Minneapolis, 4Cincinnati Children's Hospital, Cincinnati, United States

Disclosure of Interest: B. Keller: None Declared, A. Watson: None Declared, K. Williams: None Declared, S. Scully: None Declared, R. Williams: None Declared, L. Anderson: None
Declared, J. Knight: None Declared, C. Forster: None Declared, K. Choi: None Declared, M. Carlson: None Declared, N. Ratner: None Declared, P. Provenzano: None Declared, D.
Largaespada has a conflict with: DL is the co-founder and co-owner of several biotechnology companies, specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc.
(recently acquired by Immunsoft, Inc.), and B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc., and Genentech, Inc. is funding some of his research. The business of all
these companies is unrelated to the contents of this abstract. Other authors have no conflict of interest to disclose. Supported by W81XWH-15-1-0114 DOD Grant awarded to D.A.L and
P.P.P

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 43
 PARALLEL SESSION: NF2/SCHWANNOMATOSIS STATE OF THE ART
Chairs: Michel Kalamarides, MD, PhD; Marco Giovannini, MD, PhD

Platform Talk: A Phase 0 Pharmacodynamic and Pharmacokinetic Study of Everolimus in Vestibular

Schwannoma (VS) and Meningioma Patients
Monday, 5 November, 9:30 – 9:50

Matthias A. Karajannis, MS, MD, Pediatrics, Memorial Sloan Kettering Cancer Center

Background: Inactivation of NF2 activates the mTORC1 signaling pathway. Inhibition of mTOR signaling has been shown to diminish growth of NF2 deficient
tumors in preclinical studies, and clinical data suggest that everolimus, an orally administered mTORC1 inhibitor, may slow tumor progression in a subset of
NF2 patients with VS. To date, no data exist regarding penetration of everolimus into tumor tissue or molecular target modulation in human VS or meningioma
patients. To assess the pharmacokinetics, pharmacodynamics and potential mechanisms of treatment resistance, we performed a pre-surgical (“phase 0”)
clinical trial of everolimus in patients undergoing surgery for VS or meningiomas.

Methods: Adult patients with meningioma or VS requiring tumor resection were eligible and received everolimus 10 mg daily for 10 days immediately prior
to surgery. Everolimus blood levels were determined immediately prior to and after surgery. Tumor samples were collected intraoperatively. Total everolimus
concentrations in tissue were determined using mass spectrometry. Primary molecular endpoint was complete inhibition of phospho-S6 in tumor tissue.
Phospho-S6 expression and related molecular markers were compared to control tissues from untreated patients matched by histology and NF2 status.

Results: Ten patients completed protocol therapy, including 5 patients with NF2-related meningioma, 3 patients with sporadic meningioma, and 2 patients
with NF2-related VS. Median pre- and post-operative plasma levels of everolimus were found to be in a high therapeutic range (17.4 ng/ml and 9.4 ng/ml,
respectively). Median tumor tissue drug concentration was 24.3 ng/g (range 9.2–169.2), and median tumor tissue to post-operative plasma drug concentration
ratio was 0.39. We observed only partial inhibition of phospho-S6 in the treated tumors, indicating incomplete target inhibition (p = 0.005). Consistent with prior
observations that inhibition of mTORC1 may lead to MAPK pathway activation through a PI3K-dependent feedback loop, we observed a statistically significant
increase of phospho-ERK (p < 0.03) versus untreated controls.

Conclusions: In patients with meningioma or VS, treatment with everolimus leads to incomplete inhibition of mTORC1 signaling and upregulation phospho-ERK.
These data may explain the limited anti-tumor effect of everolimus observed in clinical studies for NF2 patients, and identify upregulation of phospho-ERK as a
likely resistance mechanism that could be exploited with combination therapies.

Full List of Authors: Matthias A. Karajannis*1, Shiyang Wang2, Judith D. Goldberg3, J. Thomas Roland4, Chandranath Sen5, Dimitris G. Placantonakis5, John G. Golfinos5, Jeffrey C. Allen2,
Erin Dunbar6, Scott R. Plotkin7, Srivandana Akshintala2, Robert Schneider8, Jingjing Deng9, Thomas A. Neubert9, Filippo G. Giancotti10, David Zagzag11, Jaishri O. Blakeley12
1
Pediatrics, Memorial Sloan Kettering Cancer Center, 2Pediatrics, 3Population Health, 4Otolaryngology, 5Neurosurgery, NYU Langone Health, New York, 6Neuro-Oncology, Piedmont
Healthcare, Atlanta, 7Neurology, Massachusetts General Hospital, Boston, 8Microbiology, 9Cell Biology, NYU Langone Health, 10Cancer Biology, MD Anderson Cancer Center, 11Pathology,
NYU Langone Health, New York, 12Neurology, Johns Hopkins Medicine, Baltimore, United States

Disclosure of Interest: M. Karajannis has a conflict with: This study received grant support from the NIH/NCI (R01CA164295) and Novartis (CRAD001CUS205T), S. Wang: None
Declared, J. D. Goldberg: None Declared, J. T. Roland: None Declared, C. Sen: None Declared, D. G. Placantonakis: None Declared, J. G. Golfinos: None Declared, J. C. Allen: None
Declared, E. Dunbar: None Declared, S. R. Plotkin: None Declared, S. Akshintala: None Declared, R. Schneider: None Declared, J. Deng: None Declared, T. A. Neubert: None Declared, F.
G. Giancotti: None Declared, D. Zagzag: None Declared, J. O. Blakeley: None Declared

44 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: A Single Arm Phase 2 Study of the Dual mTORC1/mTORC2 Inhibitor Vistusertib Provided on an
Intermittent Schedule for Neurofibromatosis 2 Patients with Progressive or Symptomatic Meningiomas
Monday, 5 November, 9:40 – 9:50

Scott Plotkin, MD, PhD, Neurology, Massachusetts General Hospital, Boston, United States

Background: Meningiomas are the second most common tumor in NF2 patients, with a cumulative prevalence of 80% by age 70 and a high prevalence of
multiple meningiomas. Surgery remains standard of care for these tumors, yet outcomes remain suboptimal for many patients with multiple tumors. Loss of NF2
expression is associated with activation of the mTOR pathway, but treatment with the mTORC1 inhibitor everolimus is not associated with tumor shrinkage. We
hypothesized that inhibition of both mTORC1 and mTORC2 pathways would result in increased activity against meningiomas.

Methods: This single center, phase II, open label study evaluated adults (≥18 years) with NF2 and progressive meningiomas treated with vistusertib. Subjects
received vistusertib 125 mg BID two consecutive days per week. Radiographic response was defined as ≥20% decrease in tumor volume from baseline. The
primary endpoint was radiographic response rate in the target meningioma; secondary endpoints included radiographic response in non-target meningiomas and
vestibular schwannomas.

Results: We enrolled 18 subjects (5M;13F) with a median age of 40 years (range 18-60 years). Baseline volume of target meningioma was 14.2 ml (range,
3.3-69.2 ml) and median pretreatment growth rate was 33%/year. A radiographic response was seen in 1/18 (6%) target meningiomas and in 2/20 (10%) of
non-target meningiomas and in 2/21 (9.5%) of vestibular schwannomas. Three target meningiomas (17%) progressed during treatment. Seven subjects (39%)
discontinued treatment by choice due to tolerability and 8 remain on study. Adverse events included fatigue, nausea, vomiting, anorexia, rash, mucositis, and
hypophosphatemia.

Conclusions: Vistusertib treatment is associated with radiographic response rates of 5-10% in meningiomas and schwannomas in NF2 patients, with only a
minority of meningiomas progressing during treatment. Grade 2 toxicity led to treatment discontinuation in a significant minority of patients. Future analyses will
address the relationship between tumor shrinkage and activation of TORC1/2 pathways in archival tumor specimens.

Full List of Authors: Scott Plotkin*1, Justin Jordan1, Roberta Beauchamp2, Alona Muzikansky3, Anat Stemmer-Rachamimov4, Vijaya Ramesh2
1
Neurology, 2Center for Genomic Medicine, 3Biostatistics Center, 4Pathology, Massachusetts General Hospital, Boston, United States

Platform Talk: Management of NF2-Associated Vestibular Schwannomas in Children and Young Adults:
Influence of Surgery on Tumor Volume and Growth Rate
Monday, 5 November, 10:40 – 11:00

Isabel Gugel, Centre of Neurofibromatosis and Rare Diseases; Department of Neurosurgery, University Hospital Tübingen, Tübingen

Background: The occurrence of bilateral vestibular schwannomas (VS) in patients with Neurofibromatosis Type 2 (NF2) is a typical hallmark of the disease
and their growth control by microsurgery or chemotherapy is one of the main issues in treatment concept. In the current study, we evaluated tumor volume
and growth rate of VS before and after hearing preserving surgery (decompression of the internal auditory canal (IAC) with various degrees of tumor reduction
according to acoustic evoked potential (AEP) stability) in NF2 patients under the age of 25.

Methods: In thirty-six NF2 patients (72 tumors) tumor volumetry using contrast T1-weighted MR images with thin slices (< 3 mm) was performed using the
BrainLab Software. Growth rate was calculated using the gradient between a minimum of two MRI points. Significance was tested with two independent sample
t-test and paired sample t-test by the Statistical Package for Social Services (SPSS Version 24) SPSS software.

Results: Data from 28 patients and 47 tumors before and after surgery was included. Two patients were operated twice on one side. 8 patients were excluded
from the analysis because of no treatment, chemotherapy or the presence of collision tumors. The mean follow-up was 37.64 months (range 12-167 months)
per patient. A mean growth rate of 0.693 cm3/year (+/-1.296 cm3/year) before and 0.227 cm3/year (+/-0.416 cm3/year) after surgery with a statistical
significance (p = 0.013) was observed with a total reduction by 32.756 % of the output value. Similarly, tumor volume before (mean: 3.338 cm3 +/- 4.972 cm3)
and after (1.658 cm3 +/- 3.817 cm3) after surgery was significantly lower (p < 0.001), with a total reduction of volume down of 50.33 % by surgery.

Conclusions: Decompression of IAC and hearing preservation by partial tumor resection has no stimulatory influence on tumor growth rate but slows tumor
growth. Thus, timing of surgery is important and should precede medical treatment with growth inhibitors like bevacizumab. The exact effect of surgery on the
extent of hearing preservation and the duration of hearing stability post-surgery is currently evaluated.

Full List of Authors: Isabel Gugel*1, 2, Marcos Tatagiba1, 2, Julian Zipfel1, 2, Victor-Felix Mautner1, 3, Martin Schuhmann1, 2, 4
1
Centre of Neurofibromatosis and Rare Diseases, 2Department of Neurosurgery, University Hospital Tübingen, Tübingen, 3Department of Neurology, University Medical Center Hamburg-
Eppendorf, Hamburg, 4Division of Pediatric Neurosurgery , University Hospital Tübingen, Tübingen, Germany

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 45
 PLENARY SESSION: PAIN AND PRURITIS IN NEUROFIBROMATOSIS
Chairs: Allan Belzberg, MD, Johns Hopkins University, US; Rosalie Ferner, MD, SGuy’s and St. Thomas’ NHS Foundation Trust and King’s College, UK;
Scott Plotkin, MD, PhD, Massachusetts General Hospital, US

Pathophysiology of Itch
Monday, 5 November, 11:20 – 11:45

Laurent Misery, MD, Universite de Bretagne Occidentale

Although poorly known until now, itch is very frequent in patients with NF. It is not restricted to the tumours but frequently associated with tumours. Its
pathophysiology is poorly understood in these conditions but there is probably a role of the inhibitory effect of neurofibromin. Peripheral and central sensitizations
to itch are also involved.

Therapeutic Management of Neuropathic Pain

Monday, 5 November, 11:45 – 12:05

Nadine Attal, MD, PhD, INSERM U 987 and CETD, Boulogne Billancourt, France

Neurofibromatosis is often associated with neuropathic pain. Such pain is considered as extremely difficult to treat. Recent guidelines for pharmacological
treatment recommend tricyclic antidepressants (particularly amitriptyline), serotonin–norepinephrine reuptake inhibitors (particularly duloxetine), pregabalin
and gabapentin as first line, while second line treatments include tramadol, and, for localized peripheral neuropathic pain, lidocaine plasters and capsaicin high
concentration patches. Third line treatments include strong opioids (with rigorous monitoring) and botulinum toxin A (for peripheral neuropathic pain in specialist
settings). Stimulation techniques are increasingly proposed alone or in combination with pharmacotherapy, because of a generally better side effect profile; they
include transcutaneous electrical nerve stimulation and noninvasive brain neurostimulation techniques particularly repetitive transcranial magnetic stimulation.
Invasive techniques such as spinal cord stimulation are proposed for refractory cases. Therapeutic perspectives include the development of compounds acting
on new targets and the implementation of an invididualized therapeutic approach.

Biology of Neuropathic Pain and Potential Links to NF

Monday, 5 November, 12:05 – 12:30

Michael Caterina, MD, PhD, Johns Hopkins Medical Center

Neuropathic pain is a common pathological condition that results from injury or other insults to the nervous system, such as trauma, diabetes mellitus, viral
infection, chemotherapeutic agents, or neoplastic diseases such as peripheral nerve sheath tumors. Neuropathic pain often presents with a characteristic set of
symptoms and clinical signs distinct from those associated with purely inflammatory pain. Accordingly, the pathophysiological mechanisms and the therapies
that have proven most effective at treating these distinct conditions differ from one another. Pain is a common but poorly understood and inadequately treated
symptom of peripheral nerve sheath tumors such as those caused by neurofibromatosis and schwannomatosis. In this presentation, I will discuss some of the
basic biological mechanisms underlying neuropathic pain and how these mechanisms might relate to pain in peripheral nerve sheath tumors.

46 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Molecular Mechanisms Underlying Schwannomatosis Pain
Monday, 5 November, 12:30 – 12:50

Larry Sherman, PhD, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton

Background: Schwannomatosis patients present with intractable pain that can occur in the absence of a detectable mass and which is not always relieved by
tumor resection. A recent study suggested that although most schwannomatosis patients experience some degree of pain, patients with SMARCB1 mutations
have less severe pain compared to patients with LZTR1 mutations (Medicine (Baltimore). 2018; 97(5):e9717).

Methods: We found that inducible conditional disruption of the Smarcb1 gene in Schwann cells does not lead to changes in peripheral nerve morphology or Schwann
cell proliferation or cell cycle-related gene expression. However, mice with targeted Smarcb1 disruption in Schwann cells demonstrate increased pain sensitivity.

Results: Dorsal root ganglion (DRG) and trigeminal ganglion neurons from mice with Schwann cell-targeted disruption of Smarcb1 express elevated levels of the
TRPV1, a non-selective cation channel that can be activated by a number of noxious stimuli including capsaicin. We also find that TRPA1, an ion channel that
acts as a sensor for environmental irritants, and the calcitonin gene related peptide (CGRP), which has been implicated in pain signaling, are elevated in sensory
neurons of mice with Schwann cell-targeted Smarcb1 mutations. Wild type DRG cells grown in Smarcb1-null Schwann cell conditioned media demonstrated
elevated cobalt uptake, a marker of TRPV1 activity, compared to cells grown with wild type Schwann cell conditioned media, and DRG cultures treated with
Smarcb1-null Schwann cell conditioned media or conditioned media from schwannoma cells derived from schwannomatosis patients expressed elevated levels
of TRPV1, TRPA1 and CGRP as indicated by immunocytochemistry. Proteomic, DNA microarray and chromatin immunoprecipitation analyses identified several
proteins that are elevated in Smarcb1 mutant Schwann cells and whose transcription is directly regulated by Smarcb1. Several of these proteins directly influence
the expression of TRPV1 in sensory neurons.

Conclusions: Collectively, these data indicate that loss of Smarcb1 in Schwann cells leads to the secretion of factors that induce the expression of pain
mediators in sensory neurons, and suggest a mechanism for schwannomatosis pain in patients bearing SMARCB1 mutations. We are testing if the effects of
factors derived from Lztr1 mutant Schwann cells are distinct from those of Schwann cells with Smarcb1 mutations to determine if loss of either gene results in
distinct mechanisms of pain signaling.

Full List of Authors: Larry Sherman*1, Steven Matsumoto1, Fatima Banine1, Cristina Fernandez-Valle2
1
Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, 2Biomedical Sciences, College of Medicine, University of Central Florida, Orlando, United States

FONDATION ARC POUR LA RECHERCHE SUR LE CANCER KEYNOTE LECTURE:

Targeted Therapy in Melanoma
Monday, 5 November, 13:30 – 14:30

Caroline Robert, MD, PhD, Gustave Roussy Cancer Center, Paris

Melanoma therapy was revolutionized by two strategies over the last recent years: targeted therapy for BRAF-mutant melanoma, and immunotherapy relying on
checkpoint inhibitors, regardless of BRAF mutation.

Targeted therapy combining anti-BRAF + anti-MEK achieves the highest response rate – around 70% – in patients with metastatic BRAF-mutant melanoma,
however, secondary resistances are frequent. After one year of treatment, half of the patients show signs of relapse.

A vast variety of resistance mechanisms have been identified and several strategies of combination or sequential treatments are evaluated in order to delay the
occurrence of resistance.

These targeted agents are now evaluated in earlier disease stages in the adjuvant and neoadjuvant settings with very promising results.

They are associated with a large spectrum of adverse events, new to the clinicians, but that are usually tolerable and easy to manage.

The complex controlling the initiation of protein translation, eIF4F, is a nexus of resistance that is involved in most of the resistance mechanisms. It can be
targeted by eIF4A inhibitors that are in early clinical development.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 47
 PLATFORM PRESENTATIONS – BASIC SCIENCE
Chairs: Yuan Zhu, PhD, Children’s National Medical Center, US; Conxi Lazaro, PhD, Institut Catala d’Oncologia, Spain

Platform Talk: Transposon Mutagenesis and CRISPR/Cas9 Screening Reveal Pathways Driving Malignant
Peripheral Nerve Sheath Tumor Development and Maintenance
Monday, 5 November, 14:30 – 14:50

German L. Velez Reyes, PhD, Medical School, University of Minnesota, Minneapolis, United States

Background: Malignant Peripheral Nerve Sheath tumors (MPNSTs) are highly aggressive, soft tissue, Schwann cell sarcomas. Half of these tumors occur
sporadically, while the rest occur in the context of Neurofibromatosis Type 1 Syndrome (NF1), usually developing from pre-existing plexiform neurofibromas. NF1
is characterized by loss-of-function mutations in the gene encoding neurofibromin, a negative regulator of the Ras pathway.

Methods: To better understand genetic factors that give rise to MPNSTs, we performed a Sleeping Beauty (SB) transposon screen in mice. The results implicated
Wnt/b-catenin, PI3K-AKT-mTOR, and other pathways. We next sought to validate SB screen gene hits in a human cellular model using CRISPR/Cas9 technology
as a tool to induce loss-of-function mutations in tumor suppressor gene (TSG) and oncogene candidates. A total 103 genes were independently targeted with
multiple single guide RNAs (sgRNA) in immortalized human Schwann and MPNST cell lines and effects on transformation assessed. Transformation was
assessed by anchorage-independent colony formation in soft agar, transwell migration, and xenograft tumor formation in NRG mice.

Results: In these assays, more than 30 genes scored as TSG candidates. Our results revealed a role for the Wnt/b-catenin, Hippo/Yap, RhoA, and growth factor
receptor signaling pathways in human neurofibroma and MPNST development and maintenance. Specifically, we found that TAOK1 and GDI2 loss of function
lead to very potent transformation of immortalized human Schwann cells and tumor formation in a xenograft model. TAOK1, deleted in up to 8% of MPNSTs,
encodes a Ser/Thr kinase that negatively regulates the Hippo/Yap pathway, which has been implicated in the genesis and progression of MPNST. We have
found that expression of TAOK1 is also reduced in human neurofibromas, MPNSTs, and MPNST cell lines. On the other hand, GDI2 is a regulator of the RhoA
pathway, which has also been found to act as a suppressor of lung cancer metastasis. We have found reduced GDI2 copy number, expression, and promoter
hypermethylation in MPNSTs.

Conclusions: Our screening methods have found novel candidate cancer genes involved in the development of human MPNSTs. These have generated
hypothesis-driven pre-clinical studies that we are pursuing at the moment. We are devising methods to target the pathways implicated by our studies and that are
represented in human MPNSTs.

Full List of Authors: German L. Velez Reyes*1, Nicholas Koes2, Gabriel Kaufmann2, Mariah Berner2, David Largaespada3
1
Medical School, 2College of Biological Sciences, 3Medical School, Department of Pediatrics, University of Minnesota, Minneapolis, United States

Disclosure of Interest: G. Velez Reyes: None Declared, N. Koes: None Declared, G. Kaufmann: None Declared, M. Berner: None Declared, D. Largaespada has a conflict with: DL is
the co-founder and co-owner of several biotechnology companies, specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc. (recently acquired by Immunsoft, Inc.), and
B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc., and Genentech, Inc. is funding some of his research. The business of all these companies is unrelated to the contents of
this abstract. Others authors have no conflict of interest to disclose.

48 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Early Development and Autism Risk in Neurofibromatosis Type 1
Monday, 5 November, 14:50 – 15:10

Anna Kolesnik, Centre for Brain and Cognitive Development, Birkbeck University, London

Background: Diagnosis of Neurofibromatosis Type 1 (NF1) occurs based on distinct physical features, although up to 80% children experience cognitive and
behavioral difficulties. No studies to date provide descriptions of early development, nor rotes to social impairment. Prevalence for NF1 and Autism Spectrum
Disorder (ASD) reaches 25% in population-based samples (Garg et al.2013;Garg et al.2013) vs. 1.5% in the general population (Baird et al.2006), which is not
explained by learning difficulties. Identifying early electrophysiological and behavioral indices of cognitive function may help predict later ASD symptoms in NF1
and help construct individualized intervention protocols.

Methods: We present the first report from a prospective study of 10-month old infants with NF1 (n=10,(Kolesnik et al.2017)), compared to a cohort of
infants with familial risk of ASD (BASIS; separated by outcome at age 3; ASD (n=34), atypical development(n=44), or typical development (n=89)) as well
as low-risk controls(n=75). Behavioral domains include observer examinations and parent reports of cognitive and adaptive function. We further discuss
eye tracking quality metrics: lost data, accuracy and precision, from the infants with NF1 (5-14m), infants at risk of ASD and infants without family history of
neurodevelopmental disorders (5-14m).

Results: At 10 months, infants with NF1 show striking impairments in motor function relative to low-risk infants, a pattern similar to infants with later ASD
(Mullen Scales of Early Learning, Fig1). Both NF1 and HR-ASD groups showed communication delays, although social skills were not notably impaired in NF1.
Initial assessment of eye-tracking data suggests high retention rate and adequate accuracy and precision over the gaze-point from infants with NF1. Consistently,
our provisional Index of Head Motion reveals increased movement of the head in infants with NF1 relative to the typically developing controls throughout the eye-
tracking battery.

Conclusions: There are notable developmental difficulties in

10-month-old infants with NF1, which are similar to infants with later
ASD diagnosis. These findings highlight the importance of motor
functioning delay over social skills for future animal models and early
treatment protocols. The next step is assessing whether these delays
persist at 14 months, which will be presented and discussed. We will
further consider the impact of motor delay on neurocognition, based
on quality metrics and head-motion data from our eye-tracking battery.

Full List of Authors: Anna Kolesnik*1 on behalf of Early DEvelopment in

Neurofibromatosis Type 1 (EDEN); The British Autism Study of Infant Siblings
(BASIS), Emily Jane Harrison Jones1 on behalf of Early DEvelopment in
Neurofibromatosis Type 1 (EDEN); The British Autism Study of Infant Siblings
(BASIS), Luke Mason1 on behalf of The British Autism Study of Infant Siblings (BASIS), Shruti Garg2 on behalf of Early DEvelopment in Neurofibromatosis Type 1 (EDEN), Jonathan
Green2 on behalf of Early DEvelopment in Neurofibromatosis Type 1 (EDEN), Tony Charman3 on behalf of The British Autism Study of Infant Siblings (BASIS), Mark Henry Jonhson4 on
behalf of Early DEvelopment in Neurofibromatosis Type 1 (EDEN); The British Autism Study of Infant Siblings (BASIS) and Early DEvelopment in Neurofibromatosis Type 1 (EDEN)
1
Centre for Brain and Cognitive Development, Birkbeck University, London, 2Division of Neuroscience and Experimental Psychology, University of Manchester, Manchester, 3Institute of
Psychiatry, Psychology & Neuroscience, King's College, London, 4Psychology, University of Cambridge, Cambridge, United Kingdom

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 49
 Platform Talk: Antisense Directed Gene Therapy to Treat Intragenic Neurofibromatosis Type I (NF1) Mutations
Monday, 5 November, 15:10 – 15:30

Michael Daniel, Genetics, University of Alabama at Birmingham, Birmingham, United States

Background: Antisense oligonucleotides offer an approach to skipping an exon in which an NF1 mutation is located. Antisense directed gene therapy for exon
skipping has been used successfully to treat various genetic disorders, including Duchenne muscular dystrophy (DMD). In an effort to develop an exon skipping
approach to NF1, we first sought to determine which NF1 exons could be skipped and still maintain neurofibromin function.

Methods: We mined and integrated available data of reported patient mutations resulting in individual exon or multi-exon deletions with computational predictions
of exon skipping effects on the structural and physicochemical features of the mutant protein. Sources for patient information are the Leiden Open Variation
Database, Human Gene Mutation Database, and various online publications. This information was combined with available structural data from the Protein
Database. Changes in features were assessed with respect to their potential to impair neurofibromin functionality.

To determine if skipping of an exon still results in a functional NF1 protein, we generated mNf1 cDNAs that have each exon of interest deleted. These are tested
in vitro in NF1 null HEK293 cells for restoration of neurofibromin function. Each cDNA is transiently transfected into cells and then evaluated for NF1 and p-ERK
protein levels via Western blot as well as levels of GTP-bound Ras, and finally for ELK1 transcriptional activity via a luciferase readout.

Results: We used bioinformatics to prioritize NF1 exons for skipping based on known skips in unaffected individuals, lack of skips reported in NF1 individuals,
and exons that are not within critical NF1 functional domains (such as the GRD); prioritized exons include 12, 17, 18/19, 20, 40, 46, and 51. We created
representative cDNAs for each exon and evaluated NF1 levels and Ras-activity. Deletion of certain exons (e.g. 18/19) abolishes NF1 activity while other deletions
(e.g. exons 40 and 46) retain function in the assay.

Conclusions: Our goal was to assess if antisense directed gene therapy could be used to treat NF1 mutations. To this end, we determined which NF1 exons
could be skipped while still maintaining NF1 function.

Full List of Authors: Michael Daniel*1, Hui Liu1, Robert Kesterson1, Bruce Korf1, Andre Leier1, Deeann Wallis1
1
Genetics, University of Alabama at Birmingham, Birmingham, United States

Feasibility of Gene Replacement Therapy in NF1-Related MPNST

Monday, 5 November, 15:30 – 15:50

Verena Staedtke, MD, PhD, Neurology, Johns Hopkins University, Baltimore, United States

Background: Neurofibromatosis Type 1 (NF-1) is a RASopathy that represents a major risk for the development of the malignant peripheral nerve sheath tumor
(MPNST), in which biallelic-inactivating NF1 mutations in Schwann cells result in tumor development due to Ras hyperactivation. MPNSTs are very difficult
to treat and current therapies have shown little long-term benefit. In in the treatment of NF-1 related MPNSTs, gene therapy remains a largely unexplored area
despite the known uniform monogenic alteration as underlying cause of tumor formation and the fact that clinical gene therapy has been increasingly successful
owing to improved gene delivery technologies. Among these, adeno-associated viruses (AAVs)-based delivery vectors have emerged as safe and effective.
However, AAV’s limited packaging capability prohibits the use of full length NF1 molecule owing to its very large size. Hereby, we initiated the development of an
AAV-based treatment using the NF1-GAP-related domain (NF1-GRD) to deactivate Ras activity in MPNSTs and in pre-cancerous cells in NF1 patients.

Methods: NF1-GRD domain was cloned in viral expression vectors and tested for inhibition of Ras activity and growth in MPNST cells. Thirteen AAV serotypes
expressing EGFP were produced and tested for transduction in MPNST and human Schwann cells. NF1-GRD was further optimized for anti-Ras activity by fusing
with a membrane-attaching motif.

Results: We first demonstrated the ability of NF1-GRD in suppressing the Ras activity and cell growth in MPNST cells using a lentivirus vector. Following, we
systematically assessed 13 AAV subtypes for their capacity of transducing three human NF-1 derived MPNST cell lines and human Schwann cells. Among 13
different strains tested, 5 AAV serotypes appeared particularly promising, with very efficient transduction rates in MPNST and Schwann cells. Finally, to further
optimize the therapeutic efficacy of NF1-GRD, we created a novel NF1-GRD construct with cell membrane-targeting resulting in superior and specific inhibition of
MPNST cells.

Conclusions: This approach has the potential to add a new dimension to the treatment of NF1-related MPNSTs and sets the foundation of engineering a novel
AAV vector with higher transducing efficacy in MPNST and Schwann cells. In the long-term, the scope of such a therapy could ultimately be extended to a
preventative application for NF1 haploid individuals who are at high risk for a higher tumor burden/cancer or NF-1 patients in general.

Full List of Authors: Verena Staedtke*1, Renyuan Bai2

1
Neurology, 2Neurosurgery, Johns Hopkins University, Baltimore, United States

50 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Preclinical Assessment of MEK1/2 Inhibitors for Neurofibromatosis Type 2-Associated
Schwannomas Reveals Differences in Efficacy and Drug Resistance Development
Monday, 5 November, 15:50 – 16:10

Cristina Fernandez-Valle, PhD, University of Central Florida, Orlando

Background: Neurofibromatosis type 2 (NF2) is a genetic tumor disorder caused by loss of function of the NF2/merlin tumor suppressor gene. A hallmark of NF2
is formation of bilateral vestibular schwannomas (VS) for which no FDA-approved drug is presently available. Because merlin modulates activity of the Ras/Raf/
MEK/ERK pathway, we investigated repurposing drugs targeting MEK1/2 to treat NF2-associated schwannomas.

Methods: Mouse and human merlin-deficient Schwann cell (MD-MSC/HSC) lines were screened against six MEK1/2 inhibitors. Efficacious drugs were tested
in orthotopic allograft and NF2 transgenic mouse models. Proteome and pathway analyses were conducted. Drug efficacy was examined in primary human VS
cells with NF2 mutations and correlated with differential DNA methylation patterns.

Results: Trametinib, PD0325901, and cobimetinib were the most potent in reducing MD-MSC/HSC viability. Each decreased pERK1/2 and cyclin D1, increased
p27, and induced caspase 3 cleavage in MD-MSCs. Proteome analysis was consistent with cell cycle arrest and activation of pro-apoptotic pathways in
trametinib-treated MD-MSCs. The three inhibitors slowed the growth of MD-MSC allografts compared to controls. However, decreased pERK1/2, cyclin D1,
and Ki67 levels were observed in PD0325901 and cobimetinib, but not trametinib treated grafts. Eight weeks of trametinib treatment reduced tumor burden
and average tumor size compared to controls in the NF2 transgenic mouse model; tumors did not exhibit reduced pERK1/2 staining compared to control. Both
trametinib and PD0325901 modestly reduced viability of several primary human VS cells with NF2 mutations. DNA methylation analysis of PD0325901-resistant
versus -susceptible VS identified differentially methylated regions in genes that could contribute to drug-resistance.

Conclusions: This comprehensive pre-clinical study demonstrates efficacy differences and possible emergence of drug resistance among MEK inhibitors in
schwannoma models and supports further investigation of MEK inhibitors alone and in combination with other targeted drugs as treatments for NF2-associated
schwannomas.

Full List of Authors: Cristina Fernandez-Valle*1, Marisa Fuse1, Christine Dinh2, Jeremie Vitte3, Joanna Kirkpatrick4, Thomas Mindos4, Stephani Campion5, Konstantin Brnjos1, Maria Clara
Franco6, Jie Huang7, Juan Young2, Alejandra Petrilli1, Denise Yan2, Rahul Mittal2, Rulong Shen8, Fred Telischi2, Long-Sheng Chang7, Helen Morrison4, Marco Giovannini3, Xue-Zhong Liu2
1
University of Central Florida, Orlando, 2University of Miami Miller School of Medicine, Miami, 3David Geffen School of Medicine, UCLA, Los Angeles, United States, 4Leibniz Institute on
Aging, Jena, Germany, 5Orlando Health , Orlando, 6Oregon State University, Corvallis, 7Nationwide Children's Hospital, 8Ohio State University, Columbus, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 51
 Platform Talk: Novel Drug Discovery for NF2-Deficient Meningiomas: Brigatinib Causes Tumor Shrinkage in
NF2-Deficient Meningiomas
Monday, 5 November, 16:10 – 16:30

Long-Sheng Chang, PhD, Center for Childhood Cancer & Blood Diseases/Department of Pediatrics, The Res Inst at Nationwide Children’s Hosp/The Ohio
State Univ College of Medicine, Columbus, OH

Background: Originating from meningothelial cells of the arachnoid layer lining the brain, meningiomas cause significant morbidities by compressing adjacent
brain tissues, cranial nerves, and blood vessels. Meningiomas can occur spontaneously or are frequently found in NF2 patients carrying NF2 mutations. As an
FDA-approved drug is presently not available for NF2-deficient meningioma, we aimed to identify novel targeted drugs that delay tumor progression or cause
tumor shrinkage.

Methods: A panel of isogenic NF2- and NF2+ arachnoidal cells and NF2- meningioma cells was used to screen ~2000 compounds of diverse mechanisms of
action. The anti-tumor activity of the selected drug combination, pharmacokinetic (PK) analysis, PathScan RTK Signaling Antibody Arrays, and Western blots
were performed.

Results: From the single agent screen, 45 potent compounds were selected for further combination screening, and 33 drug combinations, including the
combination of brigatinib, an inhibitor of multi-RTKs including ALK, and MK2206, an AKT inhibitor, were identified that exhibited a greater synergistic growth
inhibition in NF2- cells versus NF2+ cells. As a single agent, brigatinib effectively blocked tumor growth and reduced tumor size in the intracranial NF2-deficient
Ben-Men-1-LucB meningioma model, while MK2206 only modestly suppressed tumor growth. Combined treatment with brigatinib and MK2206 further reduced
tumor size. Upon cessation of treatment, the tumors treated with brigatinib or the brigatinib/MK2206 combination regrew. Importantly, these regrown tumors
were responsive to these drugs when retreated again. PK analysis revealed that both brigatinib and MK2206 crossed the blood-brain barrier and accumulated
in tumor-containing brain tissues. Intriguingly, Ben-Men-1 cells did not express ALK but expressed several phospho-RTKs (p-RTKs) with strong expression of
p-ErbB2, ErbB3, FGFR1, TrkA, and VEGFR2. Brigatinib treatment reduced the levels of most of these p-RTKs with the most significant reduction in EGFR, ErbB2,
ErbB3, VEGFR2, several Eph receptor members, as well as FAK. In addition, brigatinib treatment attenuated the downstream signals of these kinases, including
p-AKT and p-ERKs.

Conclusions: The anti-tumor effects of brigatinib in NF2-deficient meningiomas are mediated through inhibition of multiple growth-promoting RTKs but not ALK.
Brigatinib and its combination with an AKT inhibitor should be further evaluated in patients with NF2-deficient meningiomas

Full List of Authors: Long-Sheng Chang*1, Sarah S Burns1, Janet L Oblinger1, Marc Ferrer2, Jie Huang1, Ming Poi3, Vijaya Ramesh4, On behalf of the Synodos for NF2 Consortium5 and
Full list of Investigators of the Children's Tumor Foundation (CTF)-supported Synodos for NF2 Consortium: https://www.synapse.org/#!Synapse:syn2343195/wiki/62126
1
Center for Childhood Cancer & Blood Diseases/Department of Pediatrics, The Res Inst at Nationwide Children's Hosp/The Ohio State Univ College of Medicine, Columbus, OH, 2National
Center for Advancing Translational Sciences, NIH, Bethesda, MD, 3Division of Pharmacy Practice & Science , The Ohio State Univ College of Pharmacy, Columbus, OH, 4Center for
Genomic Medicine and Department of Neurology, Massachusetts General Hospital, Boston, MA, 5The CTF-supported Synodos for NF2 Consortium, NY, United States

Support: the CTF and US Department of Defense

52 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
PLATFORM PRESENTATIONS – CLINICAL SCIENCE
Chairs: David Viskochil, MD, PhD, University of Utah, US; Patrick Combemale, MD, Cancer Research Center of Lyon, France

Platform Talk: NF105: A Neurofibromatosis Clinical Trials Consortium (NFCTC) Phase II Study of Cabozantinib
(XL184) for Neurofibromatosis Type 1 (NF1) Associated Plexiform Neurofibromas
Monday, 5 November, 14:30 – 14:50

Michael Fisher, MD, Children’s Hospital of Philadelphia, Philadelphia, PA, United States

Background: Pre-clinical mouse models of plexiform neurofibroma (PN) demonstrated the efficacy of altering the microenvironment with Cabozantinib, a multi-
targeted tyrosine kinase inhibitor of KIT, MET, VEGF, and AXL. Here we report the activity of Cabozantinib in adolescents and adults with NF1-associated PN.

Methods: The NFCTC conducted a multi-institutional phase II study (NCT02101736) of Cabozantinib in subjects ≥16 years with NF1 and clinically significant PN
defined as potentially life-threatening, impinging on vital structures or significantly impairing quality of life. The primary study aim was response determined by
volumetric MRI analysis. Cabozantinib was administered once daily on a continuous dosing schedule (1 cycle=28 days) for up to 24 cycles. Patients started at
an initial 40 mg oral daily dose with a dose escalation to a target dose of 60 mg after 2 cycles. Dose reductions for toxicity were allowed to a minimum dose of
20 mg/day. Success was defined as ≥ 25% of subjects achieving and maintaining a partial response (PR) (≥ 20% decrease in PN volume from baseline) after
12 cycles without significant toxicity.

Results: 23 subjects enrolled (mean age 23 years); 19 were deemed evaluable, defined as completion of at least 1 cycle of therapy and having had at least one
follow-up PN imaging. 4 subjects were not evaluable due to failure to meet eligibility (n=1), subject withdrawal (n=1), lost to follow-up during first cycle (n=1)
or enrollment error removed prior to receiving study drug (n=1). The median baseline PN volume was 556.6 cm3 (range 56.8-2954.0 cm3). Eight patients (42%)
met criteria for PR by the 12th cycle. No subject had PN progression on treatment. The most common toxicities were CTCAEv4 grade 1-2 diarrhea, constipation,
nausea, vomiting, asymptomatic hypothyroidism, fatigue, headaches, palmar-plantar erythrodysesthesia (PPE) and leukopenia. Three subjects experiencing
grade 3 serious adverse events (nausea and vomiting, PPE, skin infection) remained on study after dose reductions.

Conclusions: Cabozantinib is only the second class of agents to demonstrate substantial clinical activity for PN. It was well tolerated in patients ≥16 years old.
This trial has been expanded to include a cohort of children aged 3 to 15 years.

Full List of Authors: Chie-Schin Shih1, Jaishri Blakeley2, Wade Clapp1, Pam Wolters3, Eva Dombi3, Gary Cutter4, Nicole Ullrich5, Jeffrey Allen6, Roger Packer7, Stewart Goldman8, David
Gutmann9, Scott Plotkin10, Tena Rosser11, Kent Robertson1, Brigitte Widemann3, Bruce Korf4, Michael Fisher*12
1
Indiana University, Indianapolis, IN, 2Johns Hopkins Medical Center, Baltimore, MD, 3National Cancer Institute-Pediatric Oncology Branch, Bethesda, MD, 4University of Alabama-
Birmingham, Birmingham, AL, 5Boston Children’s Hospital, Boston, MA, 6New York University, New York, NY, 7Children’s National Medical Center, Washington, DC, 8Lurie Children's
Hospital, Chicago, IL, 9Washington University, St. Louis, MO, 10Massachusetts General Hospital, Boston, MA, 11Children’s Hospital of Los Angeles, Los Angeles, CA, 12Children's Hospital
of Philadelphia, Philadelphia, PA, United States

Supported by DOD Award W81XWH-12-1-0155 and Exelixis

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 53
 Platform Talk: Clinical Presentation and 5 Year Survival of 83 Individuals with NF1 Associated Malignant
Peripheral Nerve Sheath Tumour in a National Neurofibromatosis Centre
Monday, 5 November, 14:50 – 15:10

Rosalie Ferner, MD, Neurofibromatosis Centre, Department of Neurology, Guy’s and St Thomas’ NHS Foundation Trust

Background: Neurofibromatosis 1 (NF1) associated malignant peripheral nerve sheath tumour (MPNST) is difficult to diagnose and has been reported as having
a poor prognosis. There is a 15.8% lifetime risk of developing NF1-MPNST. The aim of this study was to contribute to understanding the natural history of NF1-
MPNST by reviewing the clinical manifestations and 5 year survival in a large cohort of patients at a national institution.

Methods: We conducted a retrospective review of patients with NF1 and MPNST under the care of our national neurofibromatosis service from 2009-2018.
Demographics, presenting symptoms, site, size and grade of tumour were assessed. Presence of neuropathy, other malignancy and family history of MPNST
were documented. Follow-up and 5 year survival was determined.

Results: 83 patients had MPNST out of 1596 (4%) with NF1, 47 males, 36 females. Age range at diagnosis was 10-81 years, median 31. Clinically segmental
NF1 was present in 4 patients. 82% had pain; 33% had neurological deficit/breathing problems; 61% had hard texture of palpable tumours and 57% noted rapid
growth of a visible tumour. 7 individuals had clinical and neurophysiological length dependent, axonal neuropathy; 10% had a family history of MPNST; 27% had
another malignancy. Lower limb (25%), abdomen and pelvis (21%) were the commonest sites for MPNST. Size of tumour range was 2-24 cm, median 9 cm.
69% of MPNST were high grade tumours; 5 patients had a previous history of MPNST and presented with new primary tumours. Follow-up range was 0.16-30.1
years, median 4.1 years. 5 year survival was 59% with no significant difference between males and females. 47% of survivors were asymptomatic at last review.

Conclusions: We have presented a comprehensive review of the phenotype and survival of a large cohort of individuals with NF1-associated MPNST. The
majority of MPNST were high grade and the 5 year survival was 59%. Survival has improved since earlier studies and is consistent with 5 year survival rates for
NF1 and Cancer reported in a recent Finnish study. Potentially this reflects earlier diagnosis and/or improved management in cohesive national centres.

Full List of Authors: Rosalie Ferner*1, Roberto Tirabosco2, Una Sheerin3, Andrew Hayes4, Mary Thomas3, John Golding5
1
Neurofibromatosis Centre, Department of Neurology, Guy's and St Thomas' NHS Foundation Trust, 2Histopathology, Royal National Orthopaedic Hospital, 3Neurofibromatosis Centre,
Department of Neurology, Guy's and St. Thomas' NHS Foundation Trust, 4Sarcoma and Melanoma Unit, Royal Marsden Hospital, 5Psychology, University of Westminster, London, United
Kingdom

54 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Risk of Subsequent Neoplasms (SNs) After a Primary Tumor in Patients with NF1
Monday, 5 November, 15:10 – 15:30

Michael Fisher, MD, Pediatrics, Children's Hospital of Philadelphia, Philadelphia, United States

Background: Concern about SNs has resulted in limited use of radiation (RT) in NF1 patients, despite the fact that RT results in superior tumor control (vs.
chemotherapy [CT]) for low grade glioma (LGG) and provides superior nerve-sparing (vs. surgery alone) for malignant peripheral nerve sheath tumors (MPNSTs).
Alkylator CT is used sparingly in NF1 patients, due to concerns about increased risk of therapy-related leukemia, limiting treatment options for LGG. Gaps in knowledge
exist regarding: A) SN risk in childhood cancer survivors with and without NF1; B) SN risk in NF1 patients with a primary tumor treated with RT or alkylator CT.

Methods: Cumulative incidence of SNs was calculated treating death as competing risk. Proportional subdistribution hazards multivariable regression analysis
was used to identify risk factors.

Results: (A) We used the Childhood Cancer Survivor Study cohort of 5+y survivors with NF1 (n=176) and without NF1 (n=24,181). Excluding basal cell
carcinoma from the SNs, a significantly higher 20y cumulative incidence of SNs was observed in the NF1 cohort (10% vs. 4.1%, p=0.0003). Multivariable
analysis (adjusting for primary tumor type, age at diagnosis of primary tumor, alkylator CT and RT) found the NF1 cohort at a 3.1-fold higher risk of SN (95%CI,
1.8-4.9, p<0.0001) than the non-NF1 cohort. (B) In order to capture events from primary tumor diagnosis to sentinel event (and overcome the limitations of the
CCSS 5y survivor cohort), we constructed an independent retrospective cohort of 169 NF1 patients treated at 2 centers with dedicated NF clinics (UAB, CHOP).
Mean age at diagnosis of primary tumor was 5.0±5.0y; 71% had a primary brain tumor; treatment exposure included: alkylators CT (63.3%), RT (14.2%).
Twenty-eight SNs (glioma: n=14; MPNST: n=8; other: n=6) were observed at a median of 8.6y (range 0-18.4) from primary tumor dx. Multivariable analysis
(adjusted for age at diagnosis and primary tumor) found that exposure to RT was associated with a 2.3-fold increased risk of SNs (95%CI, 0.9-5.2, p=0.04). In
contrast, exposure to alkylator CT was not associated with increased risk of SNs (HR=1.1, 95%CI 0.4-2.9, p=0.9).

Conclusions: NF1 patients with a primary tumor were at a significantly higher risk of developing SNs vs. non-NF1 cancer patients. Among NF1 patients, exposure
to RT, but not alkylator CT was associated with increased SN risk. These findings provide evidence for refining management of primary tumors in NF1 patients.

Full List of Authors: Smita Bhatia1, Yanjun Chen1, F. Lennie Wong2, Lindsey Hageman1, Kandice Smith1, Gregory Friedman1, Joseph Neglia3, Michael Arnold4, Lucie Turcotte3, Peter De
Blank5, Gregory Armstrong6, Leslie Robison7, Wade Clapp8, Kevin Shannon9, Jean Nakamura10, Michael Fisher*11
1
Pediatrics, University of Alabama at Birmingham, Birmingham, 2Population Sciences, City of Hope, Duarte, 3Pediatrics, University of Minnesota, Minneapolis, 4Pathology and Lab Medicine,
Nationwide Children's Hospital, Columbus, 5Pediatrics, Cincinnati Children's Hospital, Cincinnati, 6Pediatrics, 7Epidemiology and Cancer Control, St. Jude Children's Research Hospital,
Memphis, 8Pediatrics, Indiana University, Indianapolis, 9Pediatrics, 10Radiation Oncology, UCSF, San Francisco, 11Pediatrics, Children's Hospital of Philadelphia, Philadelphia, United States

Platform Talk: Dorsal Root Ganglia Volume Differentiates Schwannomatosis and Neurofibromatosis Type 2
Monday, 5 November, 15:30 – 15:50

Tim Godel, Heidelberg University Hospital, Heidelberg, Germany

Background: Schwannomatosis and neurofibromatosis type 2 (NF2) are hereditary tumor syndromes that are primarily characterized by the development of
peripheral nerve sheath tumors. Although there is considerable pathomorphological overlap, genetic origins and clinical symptoms vary greatly. Dorsal root ganglia
(DRG) have been shown to play a key pathophysiological role in the development of peripheral neuropathies and pain syndromes. Since they are also described as
a vulnerable site in the NF2 murine model, we here investigated DRG volume of schwannomatosis and NF2 patients in comparison to healthy controls.

Methods: In this prospective multicenter study, 16 schwannomatosis, 14 NF2 patients and 26 healthy controls were examined by MR-Neurography at 3 Tesla.
A 3D T2-weighted sequence was used for imaging the lumbosacral plexus, covering DRG L3 to S2. DRG volume for each patient and level was calculated and
tested for statistical significance using one-way ANOVA with Bonferroni correction. P-values of < 0.05 were considered significant. ROC curves of DRG volume
of NF2 and schwannomatosis patients were used to assess diagnostic accuracy.

Results: Compared to schwannomatosis patients, DRGs of NF2 patients were enlarged by 268% for L3, 233% for L4, 257% for L5, 285% for S1, and 207%
for S2 (p<0.001 for all levels). Compared to healthy controls, schwannomatosis patients showed no difference in DRG volume for all levels (p>0.05). ROC
analysis of DRG volume (L3 to S2) as a diagnostic marker for discrimination of NF2 from schwannomatosis showed an AUC of 0.90. Further, schwannoma
formations of the lumbosacral plexus involving or originating from a DRG were found in 5/14 NF2. In contrast, while schwannomas of the lumbosacral plexus
were also found in 7 Schwannomatosis patients, all of theses were located at least >2mm either proximal or distal to the DRG.

Conclusions: Pathological alterations in DRG structure may play a crucial role in the development of areflexia and sensory loss in NF2. Moreover, our ROC
analysis indicated that DRG hypertrophy might serve as a highly accurate and easily investigated diagnostic marker in the discrimination of the two tumor
syndromes. This observation could be useful, if vestibular schwannomas are not present at the time of initial diagnosis or NF2 mutations in blood-DNA are
missing. It remains to be solved whether DRG hypertrophy in human NF2 occurs prior to development of vestibular schwannoma.

Full List of Authors: Tim Godel*1, Philipp Bäumer1, Said Farschtschi2, Mirko Pham3, Daniel Schwarz1, Moritz Kronlage1, Isabel Gugel4, Sabine Heiland1, Martin Bendszus1, Victor-Felix Mautner2
1
Heidelberg University Hospital, Heidelberg, 2UKE, Hamburg, 3Würzburg University Hospital, Würzburg, 4University Hospital Tübingen, Tübingen, Germany

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 55
 Platform Talk: Use of an NF2 Genetic Severity Score to Incorporate Genotype into Routine Patient Care
Monday, 5 November, 15:50 – 16:10

Dorothy Halliday, Oxford Centre for Genomic Medicine, Nuffield Orthopaedic Centre; NF2 Unit, Neurosciences, Oxford University Hospital NHS
FoundationTrust, Oxford, United Kingdom

Background: Genotype information in NF2 is important diagnostically to delineate NF2, mosaic NF2 and schwannomatosis, as the three conditions have
phenotypic overlap. Published work has delineated clear differences in phenotype between severe truncating NF2 mutations and milder missense mutations.
Truncating mutations however at the beginning and end of the NF2 gene cause a less severe phenotype than mutations more centrally. In addition, the high
rate of mosaicism in de-novo disease is important; as a patient mosaic for a truncating mutation may have a more severe phenotype than a patient with a
constitutional missense mutation.

The validated UK genetic severity score draws together the published data on genotype phenotype correlation, to create a simple 1-3 score where 1 represents
patients with Mosaic NF2 with no mutation detected in blood; and 2A-mild, 2B-moderate and 3-severe represents those with different types of NF2 mutation
(constitutional or mosaic) detected in blood. The score was devised to allow categorisation of patients into likely phenotypic severity groups, based on their genotype.

Methods: Using the UK NF2 severity score we assigned all 154 patients cared for by the Oxford and South-West of England NF2 team a genetic severity. We
have used this score in natural history and interventional service evaluations, approved by the Oxford University Hospitals NHS trust.

Results: In a retrospective review, we demonstrated a clear trend towards later age of deafness for patients in group 1 compared to group 3. We reviewed
ophthalmological findings and showed that various eye findings including development of optic atrophy, retinal abnormalities and cataract were related to the
genetic severity and that group 3 patients developed visual impairment at a younger age. We showed that psychological measures of distress in patients with
NF2 were increased, and that the level of distress was not related to severity of disease. We have shown that NF2 patients with impaired balance respond well to
physiotherapy intervention and that the level of response was not related to genetic severity.

Conclusions: By incorporating genotype information routinely into natural history studies we aim to further develop prognostic data on NF2 based on genotype,
to better facilitate patient management decisions.

Full List of Authors: Dorothy Halliday*1, 2, Beatrice Emmanouil2, Sam MacKeith3, Katherine Browne2, Sally Painter4, Rachel Woolrich5, Louise Dalton5, Allyson Parry2, 6
1
Oxford Centre for Genomic Medicine, Nuffield Orthopaedic Centre, 2NF2 Unit, Neurosciences, 3ENT Department, 4Department of Ophthalmology, 5Department of Psychological Medicine,
6
Department of Neurology, Oxford University Hospital NHS FoundationTrust, Oxford, United Kingdom

56 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Smile Reanimation Using the Lengthening Temporalis Myoplasty in Patients with Facial Paralysis
Monday, 5 November, 16:10 – 16:30

Andre Panossian, MD, Beverly Hills, United States

Background: In patients with neurofibromatosis (NF1 or NF2), dysfunction of the facial nerve may happen directly by tumor involvement or secondary to
resection of facial tumors. The lengthening temporalis myoplasty (LTM) is a relatively new technique for dynamic lip reanimation. It utilizes cranial nerve V and
is, therefore, relatively spared during resection of facial tumors. In this setting, the temporalis is used in an orthograde fashion as opposed to a turnover flap to
attach to the oral commissure for smile activation. There are several potential advantages that LTM may have over other methods.

Methods: A single-surgeon retrospective review was performed between 2012 and 2017, examining patients undergoing smile reanimation using LTM. Patient
demographics, diagnosis, operative time, smile excursion, laterality, complications, and length of stay were recorded. Photos and video were obtained for all
patients both pre- and postoperatively.

Results: A total of 29 patients underwent smile reanimation between 2012 and 2017 using LTM. Twenty-two patients underwent unilateral and 7 patients
underwent bilateral procedures. Only 2 patients had a diagnosis of NF1 and one with NF2. The remainder had other diagnoses resulting in facial nerve palsy.
Average operative times for unilateral LTM was 248 minutes, and bilateral LTM was 418 minutes. Smile excursion averaged 8.9 mm (range 5-17 mm). There
were no failures but approximately 15 patients (52%) required at least one revision to re-establish movement due to tendon avulsion and/or adhesions. An
additional 3 patients demonstrated temporal hollowing, requiring fat injection.

Conclusions: Lengthening temporalis myoplasty produces decreased cheek bulk, shorter operative times, ease of on-table adjustments and revisions, and
outpatient surgery over other techniques. However, revision rates for LTM are still high, and the procedure requires a nasolabial fold incision. Patients who
have undergone LTM also lack spontaneity due to its function as a masticatory muscle. In some cases, patients may require extended therapy postoperatively.
However, the incorporation of LTM in patients undergoing extensive facial neurofibroma resection may help achieve more aggressive debulking while preserving
the ability to animate the lower face.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 57
 PARALLEL SESSION: BIOLOGY OF NF2 AND SCHWANNOMATOSIS
Chairs: Helen Morrison, PhD, Leibniz Institute on Aging, Germany; Andrea McClatchey, PhD, Harvard University, US; Jeremie Vitte, PhD, UCLA, US

Late Inactivation of Smarcb1 and Additional Nf2 Loss Are Necessary for Schwannoma Development in
Schwannomatosis Patients
Monday, 5 November, 16:45 – 17:10

Jeremie Vitte, PhD, Department of Head and Neck Surgery, University of California, Los Angeles, United States

Germline mutations of the SMARCB1 gene are found in schwannomatosis patients developing benign tumors of cranial and peripheral nerves in their adulthood.
SMARCB1 germline mutations are also the hallmark of rhabdoid tumors (RTs), malignant pediatric tumors mostly developing in brain and kidney, and leading to
a familial cancer predisposition syndrome. The mechanisms by which SMARCB1 germline mutations predispose to RTs versus schwannomas are still unknown.
To understand the origin of these two types of SMARCB1-associated tumors, we generated different tissue- and developmental stage-specific conditional
knockout mice carrying Smarcb1 and/or Nf2 deletion. We demonstrated the existence of a time window in neural crest cells where early loss of Smarcb1 was
necessary to initiate tumorigenesis in the cranial nerves and meninges with typical histological features and molecular profiles of human rhabdoid tumors.
Importantly, the induction of Smarcb1 loss alone at later developmental stages in the Schwann cell lineage was not tumorigenic and additional biallelic Nf2 gene
inactivation was necessary to initiate schwannoma development, thus generating the first mouse model developing schwannomas with the same underlying
gene mutations found in schwannomatosis patients.

Correlation of the Tumor-Suppressive Function and the Confirmation of NF2

Monday, 5 November, 17:10 – 17:35

Tina Izard, PhD, Scripps Institute, US

We recently determined the crystal structure of lipid-bound neurofibromin 2, the protein responsible for neurofibromatosis type 2 (NF2). Since neurofibromin 2 is a
member of the ezrin, radixin, moesin family of cytoskeletal proteins, it seemed likely that neurofibromin 2 would function similarly to ERM proteins. Our new structural
and functional data demonstrated the correlation between the tumor-suppressive function of neurofibromin 2 and its conformation. We show that membrane
attachment is necessary for inhibiting cell proliferation and what lessons can be learned from the various neurofibromin 2 structures with respect to NF2.

Parallels Between Nerve Regeneration and Tumour Formation in the Peripheral Nervous System
Monday, 5 November, 17:35 – 18:00

Alison C Lloyd, PhD, MRC Laboratory for Molecular Cell Biology, University College London, Gower Street, London, WC1E 6BT

Peripheral nerves have remarkable regenerative properties and there has been great progress in understanding the multicellular processes involved. Underlying
this regenerative capacity is the regenerative nature of the major glial cell type of the PNS, the Schwann cell. Mature, adult Schwann cells dedifferentiate to a
progenitor-like state following an injury and these cells have multiple roles in orchestrating the multicellular response required to regenerate a nerve. Tumours
that arise in the PNS, in both NF1 and NF2, arise mostly from Schwann cells, which appear to be mimicking many of the roles of these progenitor-like Schwann
cells in the injured state. However, whereas an injury usually resolves, these tumours resemble an unrepaired wound and therefore increasing our understanding
of the resolution of the injury response is an important approach for the development of novel therapeutic approaches for the treatment of these tumours. In this
talk, I will discuss our current understanding of how Schwann cells orchestrate the regeneration of peripheral nerves and the implications of these findings for the
biology of NF1 and NF2.

References:

Cattin, A.L., and Lloyd, A.C. (2016). The multicellular complexity of peripheral nerve regeneration. Current Opinion in Neurobiology 39, 38-46.

Cattin, A.L., Burden, J.J., Van Emmenis, L., Mackenzie, F.E., Hoving, J.J., Garcia Calavia, N., Guo, Y., McLaughlin, M., Rosenberg, L.H., et al. Lloyd A.C. (2015). Macrophage-Induced
Blood Vessels Guide Schwann Cell-Mediated Regeneration of Peripheral Nerves. Cell 162, 1127-1139.

Napoli, I., Noon, L.A., Ribeiro, S., Kerai, A.P., Parrinello, S., Rosenberg, L.H., Collins, M.J., Harrisingh, M.C., White, I.J., et al. and Lloyd A.C. (2012). A central role for the ERK-signaling
pathway in controlling Schwann cell plasticity and peripheral nerve regeneration in vivo. Neuron 73, 729-742.

58 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: LZTR1 is a MAM Protein Regulating Mitochondrial Calcium and Respiration in Schwann Cells
Monday, 5 November, 18:00 – 18:20

Asha Kiran Akula, Morrison Group, Leibniz Institute on Aging-Fritz Lipmann Institute, Jena, Germany

Background: Schwannomatosis is the third major form of neurofibromatosis family of genetically inherited peripheral neuropathies that appear with age.
They can be caused by chronic axonal damage resulting in axonopathies characterized by neuronal degeneration. The origin of these axonopathies in NF1,
NF2 and schwannomatosis is not understood.Schwannomatosis is mainly characterized by chronic neuropathic pain and the formation of multiple peripheral
schwannomas. Recently, LZTR1 gene was identified as a new candidate gene for schwannomatosis (Piotrowski et.al., and Hutter et.al., 2014). The LZTR1
gene encodes a member of the BTB-kelch protein superfamily called Leucine-zipper-like transcriptional regulator 1 (LZTR1). It is not understood whether LZTR1
germline mutations in humans are causative for both chronic pain and schwannoma development, or whether additional somatic mutations (e.g. in the NF2 gene
encoding for the tumour suppressor protein merlin) are necessary for tumour formation. Nothing is known about the expression, subcellular localization and
function of LZTR1 except that it was suggested to be localized to the Golgi network, where it might stabilize the Golgi complex (Nacak et al., 2005). In the current
study, we aim to delineate the molecular mechanisms underlying neuropathic pain associated with Schwannomatosis.

Methods: Subcellular Fractionation, flow cytometry, Immunofluorescence, mouse and human Schwann cell culture, protein biochemistry, lentiviral knock-down,
electron microscopy, Seahorse metabolic assay

Results: LZTR1 is specifically found to be expressed in the unmyelinated fibres/c-fibres of the peripheral nervous system, Schwann cells and DRG neurons
(dorsal root ganglion). Using subcellular fractionation of HeLa cell lysates expressing recombinant myc-tagged LZTR1, we found that LZTR1 is a membrane
protein localized to mitochondria and mitochondrial associated ER membrane (MAM). Not only that,knock-down of LZTR1 in Schwanncells showed
mitochondrial dysfunction with an augmented mitochondrial Ca2+ levels and respiration; and decreased reactive oxygen species production.

Conclusions: data suggest that LZTR1 is a protein associated with MAMS possibly regulating calcium homeostasis which could indirectly contribute to
neuropathic pain states by modulating voltage-activated Ca2+ channels and pumps present in the endoplasmic reticulum and mitochondria.

Full List of Authors: Asha Kiran Akula*1 and Helen Morrison

1
Morrison Group, Leibniz Institute on Aging-Fritz Lipmann Institute, jena, Germany

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 59
 Merlin/ERM Proteins Regulate Growth Factor-Induced Macropinocytosis and Receptor Recycling by
Organizing the Plasma Membrane:Cytoskeleton Interface
Monday, 5 November, 18:20 – 18:40

Christine C. Chiasson-Mackenzie, PhD, Department of Pathology, Massachusetts General Hospital/ Harvard Medical School; Center for Cancer
Research, Massachusetts General Hospital, Boston, United States

Background: The architectural and biochemical features of the plasma membrane are governed by its association with the underlying cortical cytoskeleton. The
neurofibromatosis type 2 (NF2) tumor suppressor, merlin, and closely related membrane:cytoskeleton linking protein ezrin, organize the membrane:cytoskeleton
interface, a critical cellular compartment that both regulates and is regulated by growth factor receptors. An example of this poorly understood interrelationship is
macropinocytosis, a process of nutrient uptake and membrane remodeling that can both be triggered by growth factors and manage receptor availability. A newly
appreciated role for macropinocytosis in tumorigenesis and evidence that cells use macropinocytosis for uptake of therapeutics has sparked interest in exploiting
macropinocytosis for drug delivery. Because NF2 mutations are associated with the development of benign schwannomas and meningiomas that reflect the
persistance of normal signaling rather than aberrant activation of oncogenic signaling, drug delivery mechanisms that favor the targeting of NF2-mutant cells are
desperately needed.

Methods: We used high-powered confocal imaging together with cell biology and biochemical approaches to investigate the relationship between ErbB signaling
and membrane:cytoskelton organization in NF2-mutant cells. We utilized extracellular vesicle uptake as a proof of principle of the therapeutic potential of
exploiting macropinocytosis in NF2-mutant cells.

Results: We show that merlin-deficiency primes the membrane:cytoskeleton interface for growth factor-induced macropinocytosis via a mechanism involving increased
cortical ezrin, altered actomyosin and stabilized cholesterol-rich membranes. These changes profoundly alter EGFR trafficking in merlin-deficient cells, favoring
increased membrane levels of its heterodimerization partner ErbB2 and increased recycling. Our work suggests that merlin-deficient cells exploit macropinocytosis
for receptor recycling. Finally, we provide evidence that the macropinocytic proficiency of NF2-deficient cells can be employed for therapeutic uptake.

Conclusions: This work provides new insight into fundamental mechanisms of macropinocytotic uptake and processing, and suggests new ways to interfere
with or exploit macropinocytosis in NF2-mutant and other tumors. Our discovery that NF2-deficiency facilitates growth factor-induced macropinocytosis and
therapeutic uptake in Schwann and meningioma cells sets the stage for advancing this therapeutic strategy.

Full List of Authors: Christine C. Chiasson-Mackenzie*1, 2, Zachary Morris1, 2, Ching-Hui Liu2, Thijs Koorman2, William Bradford2, Andrea McClatchey1, 2
1
Department of Pathology, Massachusetts General Hospital/ Harvard Medical School, 2Center for Cancer Research, Massachusetts General Hospital, Boston, United States

60 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Platform Talk: Risk of Contralateral Breast Cancer and Survival in Neurofibromatosis Type 1:
A Five Country Cohort Study

Gareth Evans, MD, Genomic Medicine, University of Manchester, Manchester, United Kingdom

Background: Neurofibromatosis type 1 (NF1) is an autosomal dominant condition caused by mutation/deletion of the NF1 gene. Although nervous system
tumours have long been associated with NF1 recent evidence has demonstrated an increased risk of breast cancer.

Methods: We obtained details of NF1 patients with breast cancer

from five population based cohorts in Europe. Risk of contralateral
breast cancer and death were assessed by Kaplan-Meier analysis
using a delayed entry method for those who were ascertained by
the NF center after breast cancer diagnosis. A control group of 335
women screened for familial breast cancer was used to assess
relative survival.

Results: 143 women with breast cancer diagnosed at a median

age of 47.2 years (range 26.9-84.4 years). There were 1208 years
of clinical follow up from breast cancer diagnosis (mean=8.45
years). Thirteen women had developed contralateral breast cancer
at a median age of 53.7 years (range=28.5-81.7) with a rate of
11.3 per 1000 years and cumulative risk to 20 years was 29.7%.
Five and 10-year all cause survival in unscreened NF1 women was
72% (95%CI=63.0-82.2) and 58.7% (95%CI=48.5-71.0) and
breast cancer specific 10-year survival was 73.5% (95% CI 62-
82%) compared to the control familial population of 89.5% (95% CI
84.4-93%). There were no breast cancer deaths in the 27 women in
Padova who had undergone breast screening. NF1 invasive breast
cancers were less likely to be stage 1 (22% vs 61%) p<0.0001. There was also a higher rate of HER2+ cancers at 24.2% vs 11% p=0.02. 30-year all cause
survival was only 18.2% (95%CI=5.97-55.6%).

Conclusions: NF1 women have a substantial contralateral breast cancer incidence and poor survival. It is likely that breast screening starting in the early 30’s
will improve survival.

Full List of Authors: Gareth Evans*1, Roope Kallionpaa2, Maurizio Clementi3, Pierre Wolkenstein4, Oidad Zehou5, Victor Mautner6, Lauren Lewis1, Juha Peltonen2
1
Genomic Medicine, University of Manchester, Manchester, United Kingdom, 2Cell Biology, Turku University Hospital, Turku, Finland, 3Clinical Genetics Unit, University of Padova,
Padova, Italy, 4Oncodermatologue, Service de Dermatologie, Hôpital Henri Mondor, Paris, France, 5Oncodermatologue, Service de Dermatologie, Hôpital Henri Mondor, Paris, United
Kingdom, 6Neurofibromatose, Ambulanz–Hamburg, Hamburg, Germany

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 61
62 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
POSTERS: Basic Science
Poster Presentation (odd numbers)
SATURDAY, 3 NOVEMBER 2018 (18:00 – 19:30) (Location – Room 101)
LAST FIRST POSTER TITLE
Abramowicz Anna 1 Spectrum of Mutations in the NF1 Gene in Polish Patients with Neurofibromatosis Type I and
Neurofibromatosis-Noonan Syndrome
Adrien Pécriaux 3 Genetic Variants of the Androgen Receptor and Phenotype of Neurofibromatosis Type 1
Allaway Robert 5 Drug-Target Explorer: An Interactive Tool for Examining Chemical-Biological Interactions
Allaway Robert 7 Identifying Resources and Gaps in Preclinical to Clinical Translation in NF1
Alveshere Andrea J. 9 The NF1 Gene in Early Modern Humans
Arima Yoshimi 11 Establishment of Neurofibroma Cells and Dedifferentiated Fat (DFAT) Cells from Neurofibromas of NF1 Patients
Ayter Sükriye 13 EVI2B Gene Assessment in NF1 Tumors and Leukemia’s
Benedetti Helene 15 LARP6, a RNA-Binding Protein Involved in Translation and Stability of Collagen mRNA, Is a New Partner of
Neurofibromin (Nf1): Molecular Studies and Functional Implications
Ben-Shachar Shay 17 Functional Predictions and Characterization of Neurofibromin Architecture
Bonsang Benjamin 19 Expression of VEGF and VEGFR Isoforms in Nerve Tumors of Neurofibromatosis Type 1:
A New Oncogenesis Pathway Relying on an Autocrine/Paracrine Activation Loop in Tumor Cells?
Bornhorst Miriam 21 Identification of a Therapeutic Time Window That Improves Vision in an NF1-Deficient Optic Pathway
Glioma Mouse Model
Borrie Sarah C. 23 ASD-like Social Behavior Deficits in Mouse Models for Neurofibromatosis Type 1 and Legius Syndrome
Bouley Stephanie 25 Targeting NF1-Dysregulated Tumors Using a Synthetic Lethal Approach
Brekelmans Carlijn 27 Characteristics of NF1-Inactive Periosteal Derived Cells
Brosseau Jean-Philippe 29 NF1 Heterozygosity Fosters De Novo Tumorigenesis But Impairs Malignant Transformation
Carrió Meritxell 31 Reprogramming Vestibular Schwanoma Cells: Generation of Patient-Specific NF2(+/-) and (-/-) Induced
Pluripotent Stem Cells (iPSC)
Castellanos Elisabeth 33 Mutational Spectrum by Phenotype: Patients with Clinical Suspicion of RASopathy and Children with
Multiple Café-Au-Lait Macules Tested with a Custom NGS Panel
Chaney Katherine 35 A New Mouse Model of Neurofibroma, Atypical Neurofibroma, and Peripheral Nerve Sheath Tumors
Chang Long-Sheng 37 Identification of Silvestrol-Related Rocaglates with Better Bioavailability and High Potency Against
Malignant Peripheral Nerve Sheath Tumors
Chen Hongsai 39 Differential NF2 Gene Status in Sporadic Vestibular Schwannomas and its Prognostic Impact on Tumour
Growth Patterns
Chen Zhiguo 41 Developmental Origin and Spatiotemporal NF1 Loss of Heterozygosity Give Rise to Different Types of
Cutaneous Neurofibroma
Chiara Federica M. 43 Mechanotransduction and NF1 Loss Partner in Crime: A “Mature” Story
Choi Kwangmin 45 Genomic Variants Enriched in Patients with Plexiform Neurofibroma
Danna Jeremy 47 Are Reading and Motor Skills Related to Sequential Learning and Motor Adaption in Children with
Neurofibromatosis Type 1?
Dodd Rebecca 49 Combining Cre/loxP and CRISPR/Cas9 Technology for In Vivo Studies of Myeloid Cell Function in MPNST
Doucet Emile 51 Targeting the Serotonin the 5-HT6 Receptor/mTOR Complex to Reverse Cognitive Deficits in
Neurofibromatosis Type 1
Drelich Lauranne 53 Characterization of Vestibular Schwannomas in NF2 Using a MALDI Mass Spectrometry Imaging
Combined with Microproteomic Approach
Fenckova Michaela 55 Increase of Inward Cationic Currents (Ih) with Lamotrigine Partially Corrects Deficits in Habituation
Learning in a Drosophila Model of NF1
Foray Nicolas 57 Radioprotective Effect of the Combination of Statins and Bisphosphonates in Cells Derived From
Neurofibromatosis Type 1 (NF1)
Foray Nicolas 59 Radiobiological Characterization of Neurofibromatosis Type I: Impact of Neurofibromin Protein on the
ATM-Dependent DNA Damage Repair and Signaling Pathway. A Novel Mechanistic Model.
Franco Maria C. 61 Tyrosine Nitration Regulates Neurofibromatosis Type 2-Associated Schwannoma Energy Metabolism and
Cell Survival
Gel Bernat 63 Genomic Structural Alterations as a Driving Force in MPNST Development
Godec Abigail 65 PARP Inhibition in Combination with Cytotoxic Chemotherapy as a Therapeutic Option for MPNSTs

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 63
 POSTERS: Basic Science
Poster Presentation (odd numbers)
SATURDAY, 3 NOVEMBER 2018 (18:00 – 19:30)
LAST FIRST POSTER TITLE
Godec Abigail 67 Whole Exome Sequencing Leads to the Identification of TRIM23 as a Potential Suppressor of Metastasis in
NF1-MPNST
González Muñoz Teresa 69 Defining the Role of Endoglin in Malignant Peripheral Nerve Sheath Tumor Progression
Gosline Sara 71 Predicting In Vitro Drug Efficacy in Neurofibromatosis via Integration of Protein and Drug-Target Networks
Gosline Sara 73 A Pan-NF Knowledge Portal
Guo Chengrui 75 Identification of Key Candidate Genes and Pathways in NF1 by Integrated Bioinformatical Analysis
Hanemann Oliver 77 Phase 0 Trial Investigating the Intra-Tumoural Concentration and Activity of Sorafenib in Neurofibromatosis Type II
Hawley Eric 79 Genetic Ablation of Pak1 Extends Lifespan and Reduces the Size of Tumor Bearing Tissue in a Genetically
Engineered Mouse Model of Neurofibromatosis Type 2 (NF2)
Helbing Dario-Lucas 81 The Tumor Suppressor Protein Merlin Mediates Macrophage Activation and Myelin Phagocytosis
Hennigan Robert 83 Merlin Regulates p53 and YAP via Direct Binding to the Tumor Suppressor ASPP2
Hirbe Angela 85 Development of a Preclinical NF1-MPNST Platform Suitable for Precision Oncology Drug Discovery and Evaluation
Isakson Sara H. 87 NF1 Minipigs Exhibit Spontaneous Loss of Heterozygosity and Attain Clinically Relevant Plasma
Concentrations of MEK Inhibitors
Jung Marie Juliane 89 Muscle Stem Cell Intrinsic Role of Merlin in NF2 Associated Muscle Atrophy
Keller Bryant J. 91 Targeting the Hyaluronan-Rich Peripheral Nerve Sheath Tumor Microenvironment to Improve Drug Efficacy
and Delivery
Ki Dong Hyuk 93 Synergism between Topo I and mTOR inhibitors in Malignant Peripheral Nerve Sheath Tumors Affecting
NF1 Patients
Kissil Joseph 95 Genome-Wide Mapping Reveals New Nuclear Functions for YAP in Schwann Cells
Kohlmeyer Jordan 97 Targeting a Novel RABL6A-RB1 Pathway Suppresses MPNST Pathogenesis
La Rosa Salvatore 99 Delivering on the Vision of Bench to Bedside: A Funding Community Effort to Develop Effective Therapies
for Neurofibromatosis Type 1 Tumors
Lalancette Eve 101 Habituation in Children with Neurofibromatosis Type 1 Measured by Electroencephalography
Lambert Laura 103 Characterization of a Novel Patient Specific Neurofibromatosis Type I (NF1) Rat Model
Lambert Laura 105 Generation of Patient-Specific Neurofibromatosis Type I (NF1) Rodent Models
Laraba Liyam J. 107 The Role of Hippo Signalling in Merlin Null Schwannomas and Meningiomas
Larsson Alex 109 Modeling an MPNST-Like State in Immortalized Human Schwann Cells for High-Throughput Drug Screening
Lemberg Kathryn 111 JHU395, A Nervous Tissue Penetrant Glutamine Antagonist, Restricts Growth of Malignant Peripheral
Nerve Sheath Tumor Through Inhibition of Biosynthesis
Mansouri Sheila 113 Schwannomatosis Schwannomas Harbor Distinct DNA Methylation Profiles
Maziero Stéphanie 115 Memory Systems and Neurofibromatosis Type 1 (NF1): What is Impaired, What is Spared?
Mazuelas Helena 117 Understanding the Role of NF1 in Schwann Cell Differentiation and Tumor Formation through the Analysis
of an In Vitro iPSC-Based Differentiation Model
Michael George 119 Processing of Visual and Vocal Emotions in Children and Adults with Neurofibromatosis Type 1, and the
Way They Combine with Attention to Determine Agressive Behavior
Miller David T. 121 Genomics of MPNST (GeM) Consortium
Mills Kristen 123 Linking the Mechanosensing Ability of NF1+/-, NF2+/-, and Healthy Fibroblasts to Their Force-Generating
Capabilities
Mullin Rebecca 125 Reliability of Functional Outcome Measures in Adults with Neurofibromatosis Type 2
Mund Julie A. 127 Identification of Rac1 by an shRNA Library Screen and Genetic Proof of Concept That Disruption of Rac1
Prevents Plexiform Neurofibroma Formation in a Mouse Model of Neurofibromatosis Type 1
Napolitano Filomena 129 Potential Diagnostic and Prognostic Biomarkers for Neurofibromatosis Type 1 and Related Cancers:
Preliminary Results of Expression Profiling of Serum Circulating miRNAs in an Italian Study Cohort.
Nemmi Federico 131 MRI Multimodal Multivariate Signature of NF1: A Multicentric Study
Nikrad Julia A. 133 A CRISPR/Cas9 Based Synthetic Lethality Screen in NF1-deficient Human Schwann Cells
Ostrow Kimberly L. 135 The Secretomes of Painful Versus Nonpainful Human Schwannomatosis Tumor Cells Differentially
Influence Sensory Neuron Gene Expression and Sensitivity

64 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
POSTERS: Basic Science
Poster Presentation (odd numbers)
SATURDAY, 3 NOVEMBER 2018 (18:00 – 19:30)
LAST FIRST POSTER TITLE
Parfait Béatrice 137 Targeted Next-Generation Sequencing for Differential Diagnosis of Neurofibromatosis Type 2,
Schwannomatosis, and Meningiomatosis
Pemov Alexander 139 Estimation of Frequency of Pathogenic Variation in NF1, SPRED1 and MAPK1/3 in Unaffected Populations
Using Public Genomic Databases
Piotrowski Arkadiusz 141 Deep Targeted Resequencing of the LZTR1, SMARCB1 and NF2 Genomic Loci for Non-Coding or Mosaic
Mutations in Schwannomatosis
Pratilas Christine 143 Establishment of a Comprehensive Clinically-Annotated Nerve Sheath Tumor Bank from Patients with
Neurofibromatosis Type 1
Qin Wenjing 145 Tyk2 Promotes Malignant Peripheral Nerve Sheath Tumor Progression through Inhibition of Cell Death
Radomska Katarzyna 147 Injury Signals Promote the Development of Cutaneous Neurofibroma in Mouse Model of NF1
Rasola Andrea 149 TRAPping the Metabolic Adaptations of NF1-Associated Tumors
Rocha Melissa 151 Development of an Antisense-Mediated Exon Skipping Therapeutic Strategy for Neurofibromatosis Type 1
Scheffzek Klaus 153 Restoring Functional Neurofibromin by Protein Transduction
Shao Diane D. 155 Tumor Evolution Revealed by Deep Sequencing of Primary and Locally Recurrent NF1-Associated Malignant
Peripheral Nerve Sheath Tumors (MPNST)
Sharafi Parisa 157 The Evaluation of Telomer Length and Telomerase Activity in NF1 Tumors
Sokolowski Mark 159 CRISPR Based Engineering of NF1-Associated MPNST-Like Cells
Soragni Alice 161 Organoid Models of Cutaneous Neurofibromas for High-Throughput Drug Screenings
Staedtke Verena 163 Reducing Mortality from C. Novyi-NT-Induced Cytokine Release Syndrome via Bio-Engineering of a Novel
ANP-Expressing Strain
Stepanova Dina S. 165 Geranylation Inhibitors Show Selective Toxicity on Nf2-Null Cells, Which is Driven by Activated Rac1
Translocation to the Nucleus
Sun Daochun 167 Cancer Stem Cells as Targets for MPNST Tumorigenesis and Relapse
Temiz Nuri A. 169 Identification of Candidate Synthetic Lethal Genetic Interactions with NF1 Loss Using Cancer Dependency Maps
Teranishi Yu 171 Prediction of Functional Prognosis in Japanese NF2 Patients by Phenotype-Genotype Correlation
Vallee Beatrice 173 Development of Small Molecule Inhibitors of LIM Kinases, New Therapeutic Targets to Treat
Neurofibromatosis Type I
Vasiljevski Emily R. 175 Evaluating Modified Diets and Dietary Supplement Therapies for Reducing Muscle Lipid and Improving
Muscle Function in Neurofibromatosis Type 1 (NF1)
Verma Sharad 177 NF1 Funding: A State of the Field
Verma Sharad 179 An Efficient Operational Model for Funding NF1 Research
Wallis Deeann 181 Comprehensive Spectrum of NF1 Binding Partners
Wallis Deeann 183 Analysis of Patient-Specific NF1 Variants Leads to Functional Insights
Wallis Deeann 185 Multi-omics Profiling for Biomarker Discovery in Neurofibromin (NF1) Deficient Cells
Wang Jiawan 187 Adaptive Signaling Response to MEK Inhibition in NF1-Associated Malignant Peripheral Nerve Sheath Tumor
Wassef Michel 189 EZH2 is Unable to Regulate Transcription in the Absence of SUZ12 or EED
Watson Adrienne 191 Genetically Engineered Minipigs Model Major Clinical Features of Human Neurofibromatosis Type 1
Weimer Jill M. 193 An Innovative Pig Model of Neurofibromatosis Type 1 (NF1) That Mimics the Human Disease
Williams Kyle 195 Pharmacologic Vulnerabilities of NF1-Associated Tumors and Malignancies: A Small Molecule Screen
Identifies Compounds Capable of Selectively Killing NF1 Deficient Human Schwann Cells and Controlling
Tumor Growth In Vivo
Wimmer Katharina 197 Interruption of the AG-Exclusion Zone is the Major Mechanism of 90 NF1 3′ Splice Site Mutations Outside
the Canonical AG Di-Nucleotides: Lessons to Learn for Intronic Variants of Unknown Significance
Wu Natalie 199 Hippo Signaling-Mediated Schwann Cell Reprogramming Drives Malignant Peripheral Nerve Sheath Tumorigenesis
Wu Jianqiang 201 Runx1 and Runx3 Cooperatively Repress Pmp22 to Drive Neurofibromagenesis
Zhang Xiyuan 203 Novel Function of Enhancer of Zeste Homolog 2 (EZH2) in Malignant Peripheral Nerve Sheath Tumor (MPNST)
Zhang Xiaochun 205 Beta-III-Spectrin Promotes Tumor Cell Survival in NF1-MPNST

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 65
 Spectrum of Mutations in the NF1 Gene in Polish Patients with Neurofibromatosis Type I and
Neurofibromatosis-Noonan Syndrome

Anna Abramowicz, Department of Medical Genetics, Institute of Mother and Child, Warsaw, Poland

Background: Neurofibromatosis type I (NF1) and neurofibromatosis-Noonan syndrome (NFNS) are included to the RASopathies. Both are caused by the
presence of a pathogenic variant in the NF1 gene, are inherited in an autosomal dominant manner and display high phenotypic variability.

The aim of our study was to assess the spectrum of mutations identified in NF1/NFNS patients and an attempt to assess the genotype-phenotype correlation.

Methods: One hundred fifteen (115) and 23 NF1/NFNS patients, respectively, were included in the analysis. The identification of point mutations was based
on next generation sequencing (60), Sanger sequencing on RNA / cDNA (42) or DNA (35) template and MLPA technique (88). Potentially pathogenic variants
were tested in-silico using prediction algorithms and available databases (HGMD, LOVD) and were analysed in families.

Results: The analysis of the NF1 gene allowed to identify a pathogenic / potentially pathogenic variant in 92% and 95% of NF1 and NFNS patients, respectively.
The analysis of genotype-phenotype correlation showed that the missense mutations (including p.Arg1809Cys and p.Leu1196Phe variants) occurred more
frequently in NFNS than NF1 patients (9/23; 39% vs. 21/98; 21%). In NF1 patients, most of the identified variants were loss-of-function mutations (77, 79%),
including the splicing ones (30, 31%). The distribution of mutations depended on the applied technique. In case of cDNA sequencing, splicing mutations were
quite common (30%) including 8 (23%) mutations identified in exons as the missense or nonsense variants. In contrast, the NGS-based analysis revealed the
presence of missense and nonsense variants in 17 (34%) and 9 (18%) patients, and 18 of them might affect splicing as shown by in-silico analysis.

Large deletions encompassing several exons or the entire NF1 gene were identified in 9 (10%) probands. One (deletion of exons 39-42) was identified by RNA/
cDNA analysis in a patient with no splicing mutation and the MLPA revealed only the presence of exons 40-42 deletion.

Conclusions: The spectrum of mutations identified in the NF1 and NFNS patients was different suggesting that the specific variants were associated with
dysmorphy and other features typical of Noonan syndrome. The NGS analysis of the NF1 gene should include at least in-silico analysis for splicing alterations.

Full List of Authors: Anna Abramowicz*1, Marta Długoszewska1, Emilia Debek1, Aleksandra Landowska1, Olga Kordowska1, Sylwia Rzonca1, Katarzyna Bilska2, Marek Karwacki2,
Anna Kutkowska- Kazmierczak1, Jennifer Castaneda1, Natalia Braun-Walicka1, Damian Bieszczad3, Tatiana Chilarska4, Jacek Pilch5, Małgorzata Piotrowicz4, Renata Posmyk6, Robert
Smigiel7, Katarzyna Wojciechowska8, Ewa Obersztyn1, Anna Raciborska9, Jerzy Bal1, Monika Gos1
1
Department of Medical Genetics, 2Neurofibromatosis Counselling Unit, Institute of Mother and Child, Warsaw, 3Clinic of Pediatrics, Hematology and Oncology, University Clinical
Centre, Gdansk, 4Department of Genetics, Mother’s Memorial Hospital Research Institute, Lodz, 5Department of Pediatric Neurology, Medical University of Silesia, Katowice,
6
Department of Clinical Genetics, Podlaskie Medical Center, Bialystok, 7Department of Paediatrics and Rare Disorders, Wroclaw Medical University, Wroclaw, 8Department of
Paediatric Haematology and Oncology and Transplantology, Medical University, Lublin, Clinic of Oncological Surgery of Children and Adolescents, Institute of Mother and Child,
Warsaw, Poland

Acknowledgment: The work was supported from the National Science Centre (2013/09/B/NZ2/03164) and Institute of Mother and Child intramural grants.

66 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Genetic Variants of the Androgen Receptor and Phenotype of Neurofibromatosis Type 1

Pécriaux Adrien, Inserm U955, Hôpitaux Universitaires Henri Mondor, Créteil, France

Background: Neurofibromatosis type 1 (NF1) is a genetic disorder with variable

expressivity. Among young adults with NF1, men tend to develop more subcutaneous
neurofibromas (SCNF) than women, suggesting an influence of androgens on this
phenotype. This is not the case for cutaneous neurofibromas (CNF). In addition,
patients with SCNF are at higher risk for presence of internal neurofibromas and for
mortality. Previous studies had shown that genetic polymorphisms in the Androgen
Receptor (AR) gene on chromosome X can modulate the risk of androgen-sensitive
diseases in men. The most studied is a coding CAG repeat in exon 1, for which the
presence of a short allele (<20 repeats) has been associated with an increased risk of
prostate cancer, whereas the presence of a long allele (> 24) may confer an increased
risk of testicular cancer. We studied the influence of two common polymorphisms of
the AR gene (including the CAG repeat) on NF1 phenotype in men with NF1.

Methods: In 190 men with NF1, the size of CAG and GGN repeats in AR exon 1 was
assessed by fragment analysis. NF1 patients were classified according to their age and
the presence of at least two SCNF or CNF at examination.

Results: 58 men aged 18 to 30 years with at least 2 SCNF (cases) were compared
to 85 men over 30 without SCNF (controls). There was a significant difference in the
distribution of CAG repeats between groups, with an excess of short alleles (<20
repeats) in controls (28% vs. 9% in cases) and an excess of long alleles (> 24) in
cases (26% vs. 9% in controls) (p=0.0004). The presence of a short allele appears to protect against the occurrence of SCNF (OR=0.24, 95% CI 0.09-0.67),
while the presence of a long allele is associated with an increased risk (OR=3.36, 95% CI 1.32-8.56). No difference in the distribution of GGN repeats was
observed between the two groups.

Concerning CNF, no difference in the CAG or GGN repeat size was found between men aged 18-30 and at least 2 CNF and men over 30 without CNF.

Conclusions: Our study shows an association between the number of CAG repeats in exon 1 of the AR gene and the risk of early-onset SCNF in men with NF1. The
mechanism responsible for this association remains to be clarified. This result points AR as a potential modifier gene in NF1.

Full List of Authors: Pécriaux Adrien*1, Nicolas Ortonne1, 2, 3, Rakia Bhouri1, 3, 4, Salah Ferkal3, 5, Emilie Sbidian3, 5, Tu-anh Duong3, 5, Jean-Christophe Moreno3, 5, Romain Bosc6, Oana
Hermeziu3, 6, Caroline Barau1, 7, Pierre Wolkenstein3, 5, Benoît Funalot1, 3, 4
1
Inserm U955, 2Département de Pathologie, 3Centre de Référence des Neurofibromatoses, 4Département de Génétique, 5Service de Dermatologie, 6Service de Chirurgie Plastique,
Reconstructrice et Esthétique, 7Plateforme de Ressources Biologiques, Hôpitaux Universitaires Henri Mondor, Créteil, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 67
 Drug-Target Explorer: An Interactive Tool for Examining Chemical-Biological Interactions

Robert Allaway, Computational Oncology, Sage Bionetworks, Seattle

Background: Phenotypic high-throughput screens are often utilized in modern drug discovery pipelines. Such screens are conducted with an array of
molecules, ultimately measuring a biological change within a disease model. For example, a screen might test drugs in NF1 wild-type and mutant cell lines
to find molecules that are selectively toxic to NF1-/- cells. While these screens can yield valuable hits, they also present challenges including identifying the
target(s) that mediate the effect seen in the screen, characterizing ‘hits’ with a polypharmacologic target profile, and contextualizing screen data within the large
space of drugs and screening models. To address these challenges, we developed an application that enables exploration of the chemical-biological interaction
space.

Methods: Data from ChEMBL, PubChem, DrugBank, Klaeger et. al. 2017, and the Drug-Gene Interaction Database (DGIdb) for over 280,000 small molecules
were downloaded. Drug-target interaction data points were quantified for each interaction. In addition, when quantitative data were obtained, summarized
potency metrics were calculated. Each molecule was annotated with a name and chemical structure, and every target was annotated with protein & gene
identifiers. To enable exploration of this database, an interactive web interface was developed using the R Shiny platform and cheminformatics R packages.

Results: The app enables the end user with a specific query molecule to search a database of experimentally-derived drug-target interactions. The database can
be queried using drug names or structures. A chemical similarity parameter allows the user to expand their search to other structurally related molecules. The
structure of the query molecule is compared to every database molecule, and similar molecules and targets are presented in interactive tabular and network-
based forms for in-depth exploration. The app also performs enrichment analysis on the target lists and allows the user to evaluate structure-activity relationships
for drug response data. Finally, if a user has a target of interest, they can search for molecules that bind that target and explore the resulting data interactively.

Conclusions: The Drug Target Explorer is a multifunctional platform for exploring chemical space as it relates to biological targets, and may be useful at several
steps along the drug development pipeline including target discovery, structure-activity relationship, and lead compound identification studies.

Full List of Authors: Robert Allaway*1, Salvatore La Rosa2, Justin Guinney1, Sara Gosline1
1
Computational Oncology, Sage Bionetworks, Seattle, 2Children's Tumor Foundation, New York, United States

Identifying Resources and Gaps in Preclinical to Clinical Translation in NF1

Robert Allaway, Computational Oncology, Sage Bionetworks, Seattle

Background: Neurofibromatosis type I (NF1) is an autosomal dominant genetic condition characterized by peripheral nervous system tumors, including plexiform
neurofibroma (pNF) that can undergo malignant transformation (MPNST). Other than early results with selumetinib, the identification of effective treatments for
these tumors has remained elusive. Recent studies using in vitro and in vivo models of NF1-related tumors have identified several candidates, but it is unclear
which molecule-model combinations have been tested, which models are highly- or under-utilized, and which models align with clinical outcomes.

Methods: To identify resources and clarify gaps, we compiled a dataset of clinically tested drugs in NF1-related tumors and an array of in vitro and in vivo
models currently available with the goal of identifying gaps in preclinical pipelines. These data include the results of NF1 based clinical trials from ClinicalTrials.
gov, model data curated from PubMed, and drug screening data across a variety of models of NF1-related tumors. Additional curation was performed to define
and standardize the nomenclature and features of available models.

Results: Curation of clinical trial data revealed that 57 pharmacologic clinical trials for NF1 related tumors have been or are being conducted. 27 are ongoing,
10 are completed and did not meet efficacy endpoints (10/15), and 15 are completed with no results published. Most interventions have been tested only
once. Curation of NF1 models identified 45 in vivo animal models (5 species), and 87 cell lines (3 species), but few of these appear to be readily acquirable.
While several in vitro models of MPNST exist, there are fewer readily available in vitro models of pNF, OPG and other NF1 related tumors. Finally, it is unclear
which model features are required to successfully predict clinical outcomes.

Conclusions: There are several research tools available for NF1 research predominantly focused on pNF and MPNST. However, there are challenges in
translating preclinical discoveries to clinical trials. There are opportunities to fill these gaps by developing and systematically assessing in vitro and in vivo
tumor models for their predictive value for clinical performance. Freely disseminating tumor models would further fill gaps in the NF1 drug development
pipeline. There is also a need to publish unreported preclinical and clinical trial results, and use these data to assess the performance of preclinical models in
predicting clinical outcomes.

Robert J. Allaway*1, Sharad Verma2, Wade Clapp3, Nancy Ratner4, Justin Guinney1, Salvatore La Rosa5, Annette Bakker5, Sara Gosline1, Jaishri Blakeley2
1
Computational Oncology, Sage Bionetworks, Seattle, WA, 2Neurofibromatosis Therapeutic Acceleration Program, Baltimore, MD, 3Indiana University, Indianapolis, IN, 4Cincinnati
Children’s Hospital Medical Center, Cancer and Blood Diseases Institute, Cincinnati, OH, 5Children's Tumor Foundation, New York, NY, United States

68 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
The NF1 Gene in Early Modern Humans

Andrea J. Alveshere, Department of Sociology and Anthropology, Western Illinois University, Macomb, IL

Background: Increasing data suggest that many unique characteristics of modern humans reflect the presence and impact of the ubiquitous eukaryotic
Supergene, NF1, originally identified from instances of gene mutations accounting for Von Recklinghausen disease or the NF1 syndrome.

Methods: From genetic and anthropological vantage points, keying off a publication of a Cro-Magnon person with an apparent NF1 mutation, plus our
general respect for NF1 mutant phenotypic elements involving adverse effects on social skills, learning and musicality and, as well, positive effects including
minimization of obesity, diabetes mellitus, alcoholism and opiate addiction, we have documented the persistent and expanding presence of NF1 alleles in
multiple lineages of ancient and modern humans.

Results: In this brief presentation we document multiple bases for pursuing a unique role of NF1 as a genetic mechanism for “finessing” multiple elements of
the modern human phenotype. As demonstrated here, expansion of the cranial vault associated with NF1 variants is an example.

Conclusions: As we learn more about how wildtype NF1 enhances modern humans we will both improve the clinical management of those bearing the mutant
alleles and improve the health and well-being of the general population as well.

Full List of Authors: Andrea J. Alveshere*1, Vincent M. Riccardi2, Pierre Wolkenstein3, Luiz Oswaldo C Rodruigues4, Marco Giovannini5
1
Department of Sociology and Anthropology, Western Illinois University, Macomb, IL, 2The Neurofibromatosis Institute, La Crescenta, CA, United States, 3Department of
Dermatology, Henri-Mondor Hospital, APHP, Créteil, France, 4Clínico do Centro de Referência em Neurofibromatoses, Hospital das Clínicas da Universidade Federal de Minas Gerais,
Belo Horizonte, MG, Brazil, 5Department of Head and Neck Surgery, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA, United States

Establishment of Neurofibroma Cells and Dedifferentiated Fat (DFAT) Cells from Neurofibromas of NF1 Patients

Yoshimi Arima, Division of Gene Regulation, Institute for Advanced Medical Research

Background: Neurofibroma is a critical symptom in NF1 patients, and surgical resection is the standard and only available treatment for neurofibromas. To
develop therapeutic strategies for NF1-associated neurofibromas, we are attempting to develop in vitro models that recapitulate the pathological and clinical
properties of neurofibromas.

Methods: Neurofibroma cells and dedifferentiated fat (DFAT) cells were established from NF1 patient tumors. Tissue samples were obtained during tumor
resection surgery at our hospital from patients who met the NIH clinical diagnostic criteria for NF1. Whole-blood specimens were also obtained for gene
analysis. All patients provided written informed consent. The institutional review board of our university approved this aspect of the study. Tumor tissues were
dissociated in DMEM containing collagenase. The neurofibroma cells at the bottom of the tube and the floating stromal adipocytes were collected separately after
centrifugation. To establish DFAT cells, the stromal adipocytes were placed in a culture flask filled with 20% FBS-DMEM, and then the flask was inverted and
incubated at 37 °C in a humidified atmosphere of 5% CO2. The stromal adipocytes floated up through the medium and adhered to the ceiling of the flask. After 1
week, the cells were firmly attached to the ceiling and had dedifferentiated. Those DFAT cells can be passaged as well as the neurofibroma cells.

Results: We established neurofibroma cells and DFAT cells from NF1-associated neurofibromas. We performed flow cytometry analysis and found that those
cells derived from NF1 patients were expressed SOX10, S100, and CD90, all of which are expressed in Schwann cells. We identified the NF1 mutations in the
patients by next-generation sequencing. Peripheral blood specimens from patients 1 and 2 were positive for c.1466A>G, p.Tyr489Cys and c.3213_3214delAA,
p.Ser1072Hisfs*16 mutations of NF1, respectively. We also identified NF1 mutations in the cells that we had established from tumors. In the tumor specimen of
patient 1, we identified an additional somatic mutation, c.6772C>T, p.Arg2258X of NF1 gene.

Conclusions: We established neurofibroma cells and DFAT cells from neurofibromas of NF1 patients. DFAT cells exhibit multipotent ability of differentiation into a
variety of cell types. These cells may prove useful for studies of the pathophysiology of NF1-associated neurofibromas as well as cell-based drug screening for
facilitating the development of new treatments.

Full List of Authors: Yoshimi Arima*1, Hiroyuki Nobusue1, Shigeki Sakai2, Kazuo Kishi2, Toshiki Takenouchi3, Kenjiro Kosaki4, Hideyuki Saya1
1
Division of Gene Regulation, Institute for Advanced Medical Research, 2Department of Plastic and Reconstructive Surgery, 3Department of Pediatrics, 4Center for Medical Genetics,
Keio University School of Medicine, Tokyo, Japan

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 69
 EVI2B Gene Assessment in NF1 Tumors and Leukemia’s

Sükriye Ayter, Medical Biology and Genetics, TOBB University of Economics and Technology, Faculty of Medicine

Background: Neurofibromatosis Type 1 (NF1) is an autosomal dominant disease affecting the development and growth control of a variety of tissues. NF1
represents a major risk factor for development of malignancies, particularly malignant peripheral nerve sheath tumors, optic gliomas, and leukemia. The NF1
gene encodes a transcript estimated at 11 to 13 kb containing 61 exons. Three genes, EVI2A, EVI2B, and OMGP, are embedded within intron 27b of the NF1
gene. These genes are transcribed in the direction opposite that of the NF1 gene. Little is known about the function of these genes. Both EVI2A and EVI2B
encode putative transmembrane proteins. The mouse homologs (Evi-2a and Evi-2b; ecotropic viral integration sites) are associated with viral insertions
involved in leukemia in mice and its relation to NF1 symptoms is unknown. It has been already identified that Evi2b as a direct target gene of C/EBPa, a
transcription factor critical for myeloid differentiation. It is possible that these genes are related to the leukemia observed in NF1 patients, although there are
no data confirming this association. Expression of this gene is altered by viral integration and this altered expression may predispose cells to myeloid disease.
These genes might act as a modifier in the NF1 phenotypic variations. Therefore we investigated EVI2A and EVI2B gene in NF1 tumors and leukemia.

Methods: We analyzed 10 NF tumors, 20 leukemia and 3 NF1- leukemia by PCR based technics. DNA samples were sequenced to detect variations in each
exon. The pathological status of tumor tissues was confirmed by routine pathological examination. Standard immunohistological procedure was performed for
EVI2B protein and S100 in tumor samples to prove the existence of Schwann cells.

Results: We observed viral integration for EVI2B gene in 15 out of 20 leukemia and in all 3 NF1-leukemia patients. We did not observe any integration in NF1
patients. Moreover we do not detected any integration in EVI2A gene in none of the patients. The immunofluorescence staining results demonstrated intact
protein localization in cell membranes of NF1 tumors.

Conclusions: No integration in EVI2A gene were detected in none of NF1, leukemia’s and NF1-leukemia’s patients. The integrations in EVI2B were discovered
only in leukemia’s and NF1-leukemia’s but not in NF1 patients. This shows that EVI2B is a candidate effector gene for leukemia’s but not for NF1 disease. This
study was supported by TÜBİTAK Project No: 107S96.

Full List of Authors: Sükriye Ayter*1, Parisa Sharafi1, Banu Anlar2, Ali Varan3, Mualla Cetin4
1
Medical Biology and Genetics, TOBB University of Economics and Technology, Faculty of Medicine, 2Pediatric Neurology, 3Pediatric Oncology, 4Pediatric Hematology, Hacettepe
University, Ankara, Turkey

70 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
LARP6, a RNA-Binding Protein Involved in Translation and Stability of Collagen mRNA, Is a New Partner of
Neurofibromin (Nf1): Molecular Studies and Functional Implications

Helene Benedetti, CBM, CNRS, Orleans, France

Background: To improve our understanding of the molecular functions of Nf1, we have performed a two-hybrid screen on a human brain cDNA library using
the SecPH domain of Nf1 as a bait. LARP6 was identified as an interactor. LARP6 is a RNA-binding protein involved in posttranscriptional regulation of type I
collagen mRNA. In specific conditions, type I collagen synthesis can be upregulated several hundred fold. This process is predominantly achieved by LARP6
which stabilizes and stimulates translation of mRNA. For this purpose, LARP6 binds a secondary stem-loop structure (termed 5’ SL) in the 5’ untranslated
region (5’UTR) of mRNA.

This interaction is of high interest since type I collagen is overproduced in neurofibroma and by Nf1 deficient osteoblasts resulting in non-mineralized
collagenous matrix and bone fragility.

Methods: Interactions between Nf1, LARP6 and domains of these proteins were tested by co-immunoprecipitations and two-hybrid technics. SecPH, LARP6
and its Nt-LAM-RRM domain were produced and purified and gel mobility shift assays were conducted to visualize the formation of complexes with collagen
mRNA 5’SL.

Primary human lung fibroblasts were transfected by SecPH and collagen synthesis was monitored by western-blot.

Results: Nf1 and LARP6 interact independently of the presence of collagen mRNA. This interaction implies the Nt-LAM domain of LARP6 and the SecPH
domain of Nf1

Two complexes, C1 and C2 are formed between LARP6 or its Nt-LAM-RRM domain and 5’SL. SecPH was shown to abolish their formation and to induce the
formation of a new complex, Cx. We are currently working on Cx to know if it contains SecPH.

Synthesis of type I collagen by human lung fibroblasts seems to be affected by the overproduction of SecPH. These results have to be confirmed by silencing
of Nf1 in these cells and overproduction of a SecPH mutant affected in its interaction with LARP6.

Conclusions: Nf1 interacts with LARP6 and its SecPH domain affects the complex formed between LARP6 and the 5’ SL region of type I collagen mRNA. In
addition, SecPH seems to affect type I collagen synthesis by human lung fibroblasts. We now want to go further in the experiments to demonstrate that these
two results are linked and to unravel the precise molecular mecanism involved. This work will tell us if LARP6 can constitute a new therapeutic target for
phenotypes associated with excessive collagen production in NF1.

Full List of Authors: Helene Benedetti*1, Michel Doudeau1, Fabienne Godin1, Beatrice Vallee1, Aurelie Cosson1, Mohammed Bergoug1
1
CBM, CNRS, Orleans, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 71
 Functional Predictions and Characterization of Neurofibromin Architecture

Shay Ben-Shachar, Sackler Faculty of Medicine, Tel Aviv University; Gilbert Israeli Neurofibromatosis Center, Tel-Aviv Medical Center, Tel Aviv, Israel

Background: Neurofibromatosis type I (NF1) is caused by heterozygous loss-of-function variants in the NF1 gene encoding neurofibromin which serves as a
tumor suppressor that inhibits RAS signaling and regulates cell proliferation and differentiation. While, the only well-established functional domain in the NF1
protein is the GAP-related domain (GRD), most of the identified non-truncating disease-causing variants are located outside of this domain, supporting the
existence of other important disease-associated domains. Identifying these domains may reveal novel functions of NF1

Methods: By implementing inferential statistics combined with machine-learning methods, we developed a novel NF1-specific functional prediction model that
focuses on nonsynonymous single nucleotide variants (SNVs). The model enables annotating all possible NF1 nonsynonymous variants, thus mapping the
range of pathogenic non-truncating variants at the codon level across the NF1 gene.

Results: The generated model demonstrates high absolute prediction value for missense and splice-site variations (area under the ROC curve of 0.94)
outperforming 14 other established models.

By reviewing the entire dataset of nonsynonymous variants, two novel domains (Armadillo type fold 1 and 2) were identified as being associated with
pathogenicity (OR 1.86; CI 1.04 to 3.34 and OR 2.08; CI 1.08 to 4.04, respectively; P<0.05). Specific exons and codons associated with increased
pathogenicity were also detected along the gene inside and outside the GRD domain.

Conclusions: Exons and codons with higher and lower likelihood to be associated with NF1 were detected using a novel model, enabling better prediction of
pathogenicity for variants in NF1 gene, as well as revealing novel NF1-associated domains in addition to the GRD.

Full List of Authors: Shay Ben-Shachar*1, 2, D. Gareth Evans3, Felix Bokstein1, 2, Deeann Wallis4, Ofer Isakov1, 5
1
Sackler Faculty of Medicine, Tel Aviv University, 2Gilbert Israeli Neurofibromatosis Center, Tel-Aviv Medical Center, Tel Aviv, Israel, 3Manchester Centre for Genomic Medicine, Central
Manchester University Hospitals NHS Foundation Trust, Manchester, United Kingdom, 4Department of Genetics, University of Alabama at Birmingham, Birmingham, AL, United
States, 5Department of Internal Medicine T, Tel-Aviv Medical Center, Tel Aviv, Israel

72 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Expression of VEGF and VEGFR Isoforms in Nerve Tumors of Neurofibromatosis Type 1: A New Oncogenesis
Pathway Relying on an Autocrine/Paracrine Activation Loop in Tumor Cells?

Benjamin Bonsang, Department of Pathology and INSERM U955 team 9, Hôpital Henri Mondor, APHP, Créteil; Department of Pathology, Institut Curie

Background: Malignant peripheral nerve sheath tumors (MPNST) are highly aggressive sarcomas which most often arise from deep neurofibroma (NF) and
occur in up to 10% of neurofibromatosis type 1 (NF1) patients. It has been suggested that angiogenesis induced by VEGF (vascular endothelial growth factor)
through its receptor VEGFR could play a key role in the development of theses tumors.

Methods: We have studied the expression of the different isoforms of VEGF and VEGF receptor (VEGFR) using immunochemistry (IHC, VEGFA and VEGFR1) and
real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) (VEGF A/B/C/D and VEGFR1/2/3) in tumor cells (Schwann cells or fibroblastic
cells), in a series of 13 MPNST and 26 deep NF (17 plexiform NF and 19 diffuse NF) from NF1 patients. Micro-vascular density (number of capillaries in 20 fields
X400) and endothelial cells activation (percentage of pAKT+ and pERK+ endothelial cells in 10 capillaries) were quantified in tissue sections.

Results: IHC study showed higher positive staining for VEGFR1 in MPNST and plexiform NF in comparison with diffuse NF (p=0.05) and for VEGFA in MPNST only
(p=0.01). qRT-PCR showed overexpression of VEGF and VEGFR isoforms transcripts in MPNST, especially VEGFC (p=0.03) and VEGFR2 (p<0.001). Micro-
vascular density was significantly higher in NF than in MPNST (p=0.0025), with no difference regarding the expression of endothelial cells activation markers.

Conclusions: As already reported in cell lines, tumor cells of MPNST co-express VEGF and VEGFR isoforms. The expression appears quantitatively more
important than in deep NF but is paradoxically associated to a reduced micro-vascular density. This may suggest that the VEGF VEGFR pathway could play an
oncogenic role, rather by autocrine/paracrine activation of tumor cells than by stimulation of angiogenesis in the tumor environment.

Full List of Authors: Benjamin Bonsang*1, 2, Ivan Bièche3, 4, Eric Pasmant3, 5, Ingrid Laurendeau3, 5, Laurent Maksimovic1, Nadine Martin1, Pierre Wolkenstein6, Nicolas Ortonne1
1
Department of Pathology and INSERM U955 team 9, Hôpital Henri Mondor, APHP, Créteil, 2Department of Pathology, Institut Curie, 3EA7331, Faculty of Pharmacy of Paris , Paris
Descartes University, 4Department of Genetics, Institut Curie, 5Department of Genetics and Molecular Biology, Hôpital Cochin, APHP, Paris, 6Department of Dermatology, Hôpital
Henri Mondor, APHP, Créteil, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 73
 Identification of a Therapeutic Time Window That Improves Vision in an NF1-Deficient Optic Pathway Glioma
Mouse Model

Miriam Bornhorst, Children's National Health System

Background: Optic pathway gliomas (OPGs) are glial tumors that arise early during development in 15-20% of patients with Neurofibromatosis Type I (NF1).
While chemotherapy is often used for patients who have growing tumors and/or vision loss associated with the tumor, a minority of patients have improvement
in their vision following treatment. Thus, a new approach to treatment is needed to improve outcomes in this patient population.

Methods: In our laboratory, we have a Nf1-deficient conditional knockout mouse model that develops optic pathway gliomas around 60 days of age (P60). These
mice have associated axon degeneration, loss of retinal ganglion cells, and evidence of vision loss on behavioral testing, thus making this a good model to use
in order to identify the optimal therapeutic time window that improves or prevents vision loss. In this study, we treated NF1-deficient mice with a Mek-inhibitor
during different developmental time windows and measured response to treatment in terms of nerve cellularity, axon integrity, and retinal ganglion cell loss.

Results: We found that response to treatment is dependent on when the therapy is initiated during OPG development, and the best response occurred when
therapy was initiated early during this process.

Conclusions: This study suggests that the timing of therapy initiation is critical for optimal outcomes in NF1-associated OPGs.

Full List of Authors: Miriam Bornhorst*1, Francisco Nadal-Nicolas2, Emmanuelle Jecrois1, Daphine Kwesiga3, Wang Zheng1, Steven Stasheff1, 2, Austin Friend1, Patrick Mateas1, Yuan Zhu1
1
Children's National Health System, 2National Institute of Health, Washington, 3University of Maryland, Baltimore, United States

ASD-like Social Behavior Deficits in Mouse Models for Neurofibromatosis Type 1 and Legius Syndrome

Sarah C. Borrie, Department of Human Genetics, KU Leuven, Leuven, Belgium

Background: Neurofibromatosis type 1 and Legius syndrome are two related RASopathy disorders, stemming from mutations in the RAS–MAPK pathway,
in the NF1 and SPRED1 genes respectively. Neurofibromin is a RAS-GAP protein, negatively regulating activation of RAS, and SPRED1 interacts with
neurofibromin. Both disorders have similarities, including neurological problems that span cognitive deficits and increased incidence/reports of autism
spectrum disorder (ASD), however the presentation is very different. Mouse models for these disorders, Nf1+/- and Spred1-/- mice, exhibit cognitive deficits
consistent with human phenotypes, but little is known about the ability of these models to recapitulate the ASD-like symptoms present in patients, and what the
molecular mechanisms underlying such phenotypes are.

Methods: Here we examined social behaviours in Spred1-/- and Nf1+/- mice, to ask if social deficits are observed in these mouse models, and whether any
observed deficits depend upon RAS-MAPK signalling.

Results: Spred1-/- mice displayed abnormal social behavior in the automated tube test compared to wildtype controls, and also impairments in nesting
behaviour. Studies in Nf1+/- mice also demonstrated social deficits in the tube test. Social deficits in Spred1-/- adult mice could be reversed in by acutely
inhibiting the RAS-MAPK pathway with MEK inhibitors. Expression of Spred1 mRNA was seen in both inhibitory and excitatory neurons in the mouse brain,
similar to the known localisation of Nf1 in these cell types, indicating that further investigation of the relative contribution of these cell types to social behaviour
in mouse models for RASopathy disorders is warranted.

Conclusions: These findings indicate that social deficits relevant to ASD can be modelled in Spred1-/- and Nf1+/- mice, and suggest that RAS–MAPK pathway
over-activation can underpin these deficits.

Full List of Authors: Sarah C. Borrie*1, Ellen Plassschaert1, Ype Elgersma2, Steven Kushner2, Akihiko Yoshimura3, Eric Legius1, Hilde Brems1
1
Department of Human Genetics, KU Leuven, Leuven, Belgium, 2Eramus Medical Center, Rotterdam, Netherlands, 3Department of Microbiology and Immunology, Keio University
School of Medicine, Toyko, Japan

Funding: Children’s Tumor Foundation Young Investigator Award (SCB), FWO postdoctoral fellowship (HB), KU Leuven Opening the Future

74 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Targeting NF1-Dysregulated Tumors Using a Synthetic Lethal Approach

Stephanie Bouley, Molecular and Systems Biology, Geisel School of Medicine at Dartmouth

Neurofibromatosis type 1, a genetic disorder caused by loss of the gene neurofibromin 1 (NF1), predisposes patients to the development of aggressive,
NF1-deficient neurological tumors, including malignant peripheral nerve sheath tumors (MPNSTs), highly resistant sarcomas with a 5-year survival of 16-38%.
While there are clinical trials investigating new therapies for NF1-deficient MPNSTs, such as MEK inhibitor therapy, there still exists a critical need for novel
therapeutics. Our group addressed this clinical unmet need by developing a yeast-based screening platform to identify compounds that are synthetic lethal
with NF1 loss. Yeast lacking the NF1 homolog IRA2 (NF1 mutant) were screened with drug-like compounds representative of a >300,000 compound library.
Compounds were considered a hit if they inhibited the growth of NF1 mutant cells at concentrations that had no effect on wild-type control strains. Our first
screen, carried out in collaboration with Dr. Nancy Ratner at the University of Cincinnati Drug Discovery Center, also included testing compounds in NF1 wild-
type and mutant MPNST cell lines, and resulted in the identification of our first lead compound UC1. Using a high-copy suppressor screen, we identified CDK9
as UC1’s target. In a second screen carried out in our yeast platform, we identified several other lead compounds, including Y100 and Y102. Y102 treatment
resulted in the accumulation of reactive oxygen species and the autophagy marker p62, as well as an altered mitochondrial phenotype in NF1-dysregulated
cells. Treatment with Y102 also resulted in accumulation of the mitophagy marker BNIP3L/Nix, and increased localization of lysosomes to the perinuclear
region of the cell. Together our findings suggest that Y102 prevents lysosomal-directed mitochondrial clearance. Currently, we are implementing proteomic
strategies to identify cellular targets of Y102 as well as Y100. I hypothesize that NF1-deficient cancers can be selectively targeted with the proposed mitophagy
modulator Y102 and that modulation of autophagy/mitophagy is an effective therapeutic strategy in cancers with NF1 loss. We expect that this work will result
in the identification of new targets and therapeutic leads for aggressive neurological cancers driven by NF1 loss like MPNSTs.

Full List of Authors: Stephanie Bouley*1, Robert Allaway1, Matthew Wood1, Andrew Grassetti1, 2, Helen Hou1, Scott Gerber1, 2, Jimmy Wu3, William Seibel4, Nancy Ratner4, Konstantin
Dragnev5, 6, 7, Yolanda Sanchez1, 5, 7
1
Molecular and Systems Biology, 2Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth, 3Chemistry, Dartmouth College, Hanover, 4Cincinnati Children’s Hospital,
University of Cincinnati, Cincinnati, 5Medicine, 6Hematology and Oncology, Geisel School of Medicine at Dartmouth, Hanover, 7Norris Cotton Cancer Center, Lebanon, United States

Characteristics of NF1-Inactive Periosteal Derived Cells

Carlijn Brekelmans, Department for Human Genetics, KULeuven

Background: The periosteum plays an important role in osteogenesis, both during development and fracture repair. Periosteal derived cells (PDCs) were
shown to be an important source of mesenchymal stem cells (MSCs) during this process. MSCs are primarily characterized by their tri-lineage differentiation
potential, to the adipogenic, chondrogenic and osteogenic lineage. Recently we found a somatic NF1 mutation in cells cultured from the pseudarthrosis site
from all resections taken during the first intervention and in 60% of later surgeries removing pseudarthrosis tissue. The same somatic mutation was found in
PDCs from periosteum proximal and distal to the pseudarthrosis site in three extensively sampled individuals (2 tibiae and 1 radius). Therefore we tested the
functional effect of NF1 inactivation on PDCs, since we suspect them to be involved in the development of the NF1-related pseudarthrosis.

Methods: Tri-lineage differentiation was induced to the adipogenic, chondrogenic and osteogenic lineage. Differentiation was quantified using staining and
qPCR. Also, VEGF secretion in the cell culture medium was determined.

Results: A pilot experiment uncovered diminished osteogenic differentiation in the NF1-inactive PDCs. Chondrogenic and adipogenic differentiation were also
reduced in these cells. Strikingly though, PPARG expression was increased in undifferentiated NF1-inactive PDCs and small lipid droplets were observed in the
cytoplasm of these cells. Furthermore, VEGF expression was increased in the NF1-inactive PDCs compared to controls.

Conclusions: NF1 inactivation diminishes the MSC characteristics of PDCs. Tri-lineage differentiation was reduced in these cells, as was VEGF secretion. The
extent of the effect of NF1 inactivation on PDCs is being investigated in ongoing functional characterizations. Of special interest to us is the unexpected finding
of lipid droplets and increased PPARG expression in undifferentiated NF1-inactive PDCs.

Full List of Authors: Carlijn Brekelmans*1, Silke Hollants2, Caroline De Groote2, Marina Maréchal3, Frank P Luyten3, Johan Lammens4, Eric Legius1, Hilde Brems1
1
Department for Human Genetics, KULeuven, 2Department for Human Genetics, UZ Leuven, 3Prometheus, Division of Skeletal Tissue Engineering, KULeuven, 4Department of
Orthopedic Surgery, UZ Leuven, Leuven, Belgium

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 75
 NF1 Heterozygosity Fosters De Novo Tumorigenesis But Impairs Malignant Transformation

Jean-Philippe Brosseau, Dermatology, University of Texas Southwestern Medical Center, Dallas, United States

Background: NF1 biallelic inactivation in Schwann cell lineage and melanocytes trigger the development of a characteristic set of benign tumors and
hamartoma (neurofibromas, café-au-lait-macules) present in >99% of NF1 patient by the age of 20. Strikingly, progression to malignant skin cancer was
rarely if ever reported. However, recent sequencing efforts revealed that the NF1 gene is frequently mutated in malignant tumors such as melanomas and
squamous cell carcinomas. This is puzzling because NF1 patients are not typically associated with these malignancies even though they already have a
predisposing NF1 mutation in all cells. We reasoned that individual from the general population are actually more prone to certain malignant tumors than NF1
patient and hypothesized that the NF1+/- cells from the microenvironment is impairing certain malignant transformation in NF1 patients.

Methods: We monitored the kinetic of benign tumor onset and benign to malignant conversion rate in mice with Nf1+/+ and Nf1+/- background in two mouse
models. In addition, we performed retrospective analysis of published clinical data of benign and malignant tumor commonly associated with NF1 patients as
well as those with high frequency mutation for the NF1 gene in the general population.

Results: Both the Nf1+/- and Nf1+/+ mice develop papilloma (benign) but the later develops it slower and further progress to squamous cell carcinoma
(malignant) in the two steps carcinogenesis model. Similarly, PLP-CreERT2; Nf1f/- mice develop plexiform neurofibroma (benign) faster than their PLP-CreERT2;
Nf1f/f littermates and only the later spontaneously develop MPNST more robustly. This is corroborated by the clinical data where NF1 patients develop much
more benign tumors than advanced malignant/metastatic tumors; and non-NF1 patient rarely develop the benign NF1-related tumor sporadically but do
spontaneously gets NF1 inactivation in malignant tumor not frequently observed in NF1 patient.

Conclusions: NF1 heterozygosity has an antagonistic role in tumor initiation and malignant transformation. It suggests that NF1 patients would develop cancer
much more frequently if it was not due to the protective effect of the NF1 heterozygosity on cancer progression. It predicts that identifying the cells and factors
responsible for the malignant transformation impairment could potentially be translated into a general anti-cancer strategy for NF1 patients and beyond. JPB
and CPL are CTF YIA. This work was also supported by the NCI and the US DoD.

Full List of Authors: Jean-Philippe Brosseau*1, Chung-Ping Liao1, Amisha Patel1, Lu Le1, 2, 3
1
Dermatology, 2Hamon Center for Regenerative Science and Medicine, 3Simmons Comprehensive Cancer Center, University of Texas Southwestern Medical Center, Dallas, United States

76 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Reprogramming Vestibular Schwanoma Cells: Generation of Patient-Specific NF2(+/-) and (-/-) Induced
Pluripotent Stem Cells (iPSC)

Meritxell Carrió, Hereditary Cancer Group, The Institute for Health Science Research Germans Trias i Pujol (IGTP - PMPPC), Badalona (Barcelona)

Background: The development of bilateral Vestibular Schwanomas (VS) in NF2 patients can lead to total hearing loss, representing one of the most aggressive
manifestations of this disease. There is the necessity to understand the formation of these tumors in order to develop new therapies to prevent their formation
and growth. In our laboratory, we have expertise in culturing Schwann cells from both NF1 and NF2-related tumors. However, primary cultures are perishable
and can only be used for a limited number of experiments. To overcome this problem, we have recently generated NF1(+/-) and NF1(-/-) induced pluripotent
stem cells (iPSC) from plexiform neurofibromas and differentiated them up to Schwann cells. The objective of this project was to establish a similar iPSC-
based model for VS of NF2 patients.

Methods: Tumor samples were kindly provided by NF2 patients from the Spanish Reference Centre (CSUR) on Phakomatoses HUGTiP-ICO. We selected three
VS from three independent NF2 patients, one of them being a mosaic, with known constitutive and somatic mutations. Tumors were dissociated and plated in
Schwann Cell Media. Reprogramming was performed using Sendai virus containing the 4 Yamanaka factors. iPSC clones were genotyped for NF2 germline
and somatic mutations. Selected iPSC clones were further characterized for pluripotency and differentiation potential.

Results: We obtained iPSC clones from the three VS reprogrammed with a different efficiency. We obtained NF2(+/-) iPSC clones from two tumors. From
the VS excised from a mosaic patient, we obtained NF2(-/-) and NF2(+/+). NF2(+/-) and (+/+) iPSC clones exhibited the typical morphology and growth
pattern of iPSC colonies. However, all NF2(-/-) iPSC colonies exhibited an aberrant growth pattern and morphology, difficulting their culture and expansion.
These results are consistent with recent findings pointing to NF2 as a growth restricting gene in pluripotent stem cells. Nonetheless, some of these clones have
been characterized for pluripotency and differentiation potential.

Conclusions: As far as we know, this is the first time NF2(-/-) iPSCs have been obtained directly from VS of NF2 patients. This resource constitutes a non-
perishable source of cells for disease modelling and for assaying new therapeutic strategies for VS.

Full List of Authors: Meritxell Carrió*1, Bernd Kuebler2, Inma Rosas1, Álex Negro1, Núria Catasús1, Conxi Lázaro3, Begoña Aran2, Anna Veiga2, Ángel Raya4, Ignacio Blanco5, Elisabeth
Castellanos1, Eduard Serra1
1
Hereditary Cancer Group, The Institute for Health Science Research Germans Trias i Pujol (IGTP - PMPPC), Badalona (Barcelona), 2Stem Cell Bank of Barcelona (BLCB), Center of
Regenerative Medicine in Barcelona (CMRB), 3Hereditary Cancer Program, Catalan Institute of Oncology (ICO), 4Center of Regenerative Medicine in Barcelona (CMRB), Hospitalet de
Llobregat (Barcelona), 5Medical Genetics and Genetic Counseling Program, Germans Trias i Pujol Hospital (HUGTiP), Badalona (Barcelona), Spain

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 77
 Mutational Spectrum by Phenotype: Patients with Clinical Suspicion of RASopathy and Children with
Multiple Café-Au-Lait Macules Tested with a Custom NGS Panel

Elisabeth Castellanos, Hereditary Cancer Group, Program of Predictive and Personalized Medicine of Cancer (PMPPC), Germans Trias & Pujol
Research Institute (IGTP), CIBERONC

Background: The RASopathies are a clinically defined group of hereditary syndromes caused by germline mutations in genes that encode components of the
RAS/MAPK pathway. These disorders include Neurofibromatosis type 1 (NF1), Legius syndrome and Noonan syndrome, among others, and are caused by
mutations in more than 20 genes. Many of these conditions exhibit numerous overlapping phenotypic features, especially during childhood, making clinical
diagnosis difficult. The sequencing capacity provided by NGS-based panels allows the genetic testing of all RASopathy genes in a single assay facilitating a
genetic diagnostics of patients with inconclusive clinical presentations, to provide a timely appropriate management for each disorder.

Methods: I2HCP v2.2 custom panel was validated for its use in the genetic testing of RASopathies using a set of 29 samples with known mutations in
RASopathy genes. Once validated, I2HCP v2.2 was used to analyze 48 children with a clinical suspicion of a RASopathy. In addition, 94 unrelated children with
multiple café-au-lait macules (CALMs) with or without other symptoms non-related to NF1 were genetically tested and also 87 patients fulfilling NIH clinical
criteria for NF1. NGS-based test was performed from gDNA. NF1 and SPRED1 gene analysis were complemented by MLPA and cDNA analysis.

Results: The sensitivity and specificity of I2HCP v2.2 in RASopathy gene testing was greater than 99%. We identified pathogenic variants in 22 out of 48
patients with clinical suspicion of RASopathy, with mutations in the NF1 gene accounting for 10% of the cases in this group. Furthermore, 48% of children
with only CALMs carried an NF1 pathogenic variant and 4% a SPRED1 pathogenic mutation. When in addition to CALMs other clinical manifestations were
considered, mutations in the NF1 gene represented from 43% to 80% of the individuals, depending on each clinically defined group. Finally, the NF1 mutational
spectrum in children with clinical suspicion of NF1 was different from the one in patients fulfilling NF1 NIH clinical criteria.

Conclusions: The I2HPC NGS-panel has been validated for its use in the genetic testing of patients with clinical suspicion of a RASopathy. In 10% of these
patients we identified a pathogenic mutation in the NF1 gene. The I2HPC genetic testing strategy could also be applied to children with only CALMs as a clinical
manifestation of NF1. At least in 50% of these children a mutation in the NF1 or SPRED1 genes could be identified.

Full List of Authors: Elisabeth Castellanos*1, Imma Rosas2, Alejandro Negro2, Bernat Gel1, Andreu Alibés3, Hilde Brems4, Graciela Pi5, Neus Baena6, Mercè Pineda7, Guillem Pintos8, 9,
Hector Salvador10, Conxi Lazaro11, Ignacio Blanco12, Lluïsa Vilageliu13, Daniel Grinberg13, Eric Legius4, Eduard Serra1
1
Hereditary Cancer Group, Program of Predictive and Personalized Medicine of Cancer (PMPPC), Germans Trias & Pujol Research Institute (IGTP), CIBERONC, 2Hereditary Cancer
Group, Program of Predictive and Personalized Medicine of Cancer (PMPPC), 3Program of Predictive and Personalized Medicine of Cancer (PMPPC), Germans Trias I Pujol Research
Institute (IGTP), Badalona, Spain, 4Laboratory for Neurofibromatosis Research, Department of Human Genetics, KU Leuven, Leuven, Belgium, 5Neuropaediatrics Unit, La Ribera
Hospital, Valencia, 6Genetics Laboratory, UDIAT-CD, Parc Tauli Sabadell Hospital Universitari, Sabadell, 7Neuropaediatrics Unit, Hospital Sant Joan de Déu, Esplugues, 8Teaching Unit
of Germans Trias i Pujol, Universitat Autònoma de Barcelona, Badalona, Spain, 9Centre for Rare Diseases, University Hospital Vall d'Hebron, BARCELONA, 10Paediatrics oncology
Unit, Hospital Sant Joan de Déu, Esplugues, 11Hereditary Cancer Program, Catalan Institute of Oncology (ICO-IDIBELL), CIBERONC, L'Hospitalet de Llobregat, 12Clinical Genetics and
Genetic Counseling Program, Germans Trias i Pujol Hospital, Badalona, 13Department of Genetics, Microbiology and Statistics, Facultat de Biologia, Universitat de Barcelona (UB),
IBUB, IRSJD, CIBERER, Barcelona, Spain

78 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
A New Mouse Model of Neurofibroma, Atypical Neurofibroma, and Peripheral Nerve Sheath Tumors

Katherine Chaney, Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio

Background: Plexiform neurofibromas (PN) are histologically benign peripheral nerve sheath tumors that occur in 25-50% of patients with the autosomal
dominant tumor predisposition syndrome neurofibromatosis type 1 (NF1). PN can cause substantial morbidity and are believed to serve as precursor lesions
for atypical neurofibromas (ANF). Many ANF show CDKN2A (Ink/Arf in mouse) loss, and may histologically show atypical neurofibromatous neoplasms of
uncertain biologic potential (ANNUBP) characteristics. ANF themselves may be precursor lesions to malignant peripheral nerve sheath tumors (MPNST), a
highly aggressive soft tissue sarcoma with no effective medical therapy.

Methods: In an attempt to develop an accurate model of neurofibroma to ANF to GEM-PNST we generated a mouse model based on Joseph NM, et al., Cancer
Cell. 2008, in which 26% of genetically engineered mice developed (GEM)-PNST (Nf1+/-;Ink/Arf-/-), and our previously described neurofibroma model
(DhhCre;Nf1fl/fl) Wu J., et al., Cancer Cell. 2008. We generated cohorts of DhhCre; Nf1fl/fl; Ink4a/Arf+/- and DhhCre; Nf1fl/fl; Ink4a/Arf-/- mice.

Results: DhhCre; Nf1fl/fl; Ink/Arf+/- and DhhCre; Nf1fl/fl; Ink/Arf -/- mice showed poor survival versus Ink4a/Arf controls. Each cohort developed superficial
GEM-PNST, which histologically were Grade 2/3 GEM-PNST (40%). Histology showed areas of GEM-PNST (superficial) arising from ANF. Each cohort also
developed paraspinal tumors, half of which showed neurofibroma histology, and half ANF histology. While most DhhCre; Nf1fl/fl; Ink/Arf -/- mice died rapidly
and of unrelated disease, half of DhhCre; Nf1fl/fl; Ink/Arf+/- mice died of NF-tumor-related pathologies.

Conclusions: We have developed a new model of progression from neurofibroma to GEM-ANF to GEM-PNST. DhhCre; Nf1fl/fl; Ink/Arf+/- mice may be useful
as a preclinical therapeutic model of neurofibroma transition to ANF and GEM-PNST.

Full List of Authors: Katherine Chaney*1, Ami Patel1, Sara Szabo2, Eva Dombi3, David Largaespada4, Nancy Ratner1
1
Experimental Hematology and Cancer Biology, 2Pediatrics and Pediatric Pathology, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, 3Center for Cancer Research,
National Cancer Institute , Bethesda, Maryland, 4Pediatrics and Genetics , University of Minnesota , Minneapolis, Minnesota, United States

Supported by 1R01-NS086219 to D.A.L. and N.R.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 79
 Identification of Silvestrol-Related Rocaglates with Better Bioavailability and High Potency Against Malignant
Peripheral Nerve Sheath Tumors

Long-Sheng Chang, Center for Childhood Cancer & Blood Diseases/Department of Pediatrics, The Research Institute at Nationwide Children's Hospital/
The Ohio State University College of Medicine, Columbus, OH

Background: Malignant peripheral nerve sheath tumors (MPNSTs) frequently overexpress eIF4F components, and the eIF4A inhibitor silvestrol effectively
suppresses MPNST growth. However, silvestrol has suboptimal drug-like properties, including a bulky structure and poor oral bioavailability. Our objectives are
to identify potent silvestrol-related rocaglates and to determine their bioavailability, anti-tumor effects, and mechanisms of action.

Methods: NF1+/+ STS26T and NF1-/- ST8814 MPNST and 697 and silvestrol-resistant 697-R leukemic cells were treated with various concentrations of each
rocaglate. Cell proliferation assays, flow cytometry, Western blots, and pharmacokinetic (PK) analysis were performed. A quantifiable, orthotopic NF1-deficient
MPNST mouse model and immunohistochemistry were conducted to assess antitumor effects.

Results: Among 10 silvestrol-related rocaglates lacking the dioxanyl ring examined, didesmethylrocaglamide (DDR) and rocaglamide (ROC) had potent growth-
inhibitory activity comparable to silvestrol in MPNST cells. Structure-activity relationship analysis revealed that the dioxanyl ring in silvestrol was dispensable
while the C-8b hydroxyl group was essential for cytotoxicity. DDR and ROC arrested MPNST cells at G2/M and significantly increased the sub-G1 fraction.
Accordingly, these rocaglamides induced cleavage of caspases 3 and 7 and poly(ADP-ribose) polymerase, while decreasing total protein levels of these
apoptotic markers, consistent with translation inhibition. Additionally, DDR and ROC reduced the levels of mitogenic kinases AKT and ERK1/2. Unlike silvestrol,
DDR and ROC inhibited proliferation of silvestrol-resistant 697-R leukemic cells, which over-express the MDR1 multidrug transporter, at IC50 values similar
to silvestrol-sensitive 697 cells, suggesting that these rocaglamides may be more bioavailable. PK analysis confirmed that ROC had 50% oral bioavailability.
Importantly, ROC, when administered intraperitoneally or orally, potently suppressed the growth of luciferase-expressing NF1-/- ST8814-Luc MPNST xenografts
with no overt toxicity. Treated tumors had abundant phospho-histone H3 labeling and more cleaved caspase 3-positive cells, consistent with G2/M arrest and
indicative of increased apoptosis, respectively.

Conclusions: The more favorable drug-like properties and potent anti-tumor effects suggest that ROC and DDR have potential to become viable MPNST treatments.

Full List of Authors: Long-Sheng Chang*1, Janet L Oblinger1, Sarah S Burns1, Jie Huang1, Larry Anderson2, Rulong Shen3, Li Pan4, Yulin Ren4, Barry R O'Keefe5, A Douglas
Kinghorn4, Jerry M Collins2
1
Center for Childhood Cancer & Blood Diseases/Department of Pediatrics, The Research Institute at Nationwide Children's Hospital/The Ohio State University College of Medicine,
Columbus, OH, 2Division of Cancer Treatment and Diagnosis, National Cancer Institute, NIH, Bethesda, MD, 3Department of Pathology, The Ohio State University College of
Medicine, 4Division of Medicinal Chemistry and Pharmacognosy, The Ohio State University College of Pharmacy, Columbus, OH, 5Division of Cancer Treatment and Diagnosis and
Molecular Targets Program, Center for Cancer Research, National Cancer Institute, NIH, Bethesda, MD, United States

Support: CancerFree Kids, the US Department of Defense, and National Cancer Institute, NIH

80 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Differential NF2 Gene Status in Sporadic Vestibular Schwannomas and its Prognostic Impact on Tumour
Growth Patterns

Hongsai Chen, Department of Otolaryngology Head & Neck Surgery, Ear Institute, School of Medicine, Shanghai Jiao, Shanghai, Cocos (Keeling) Islands

Background: The great majority of sporadic vestibular schwannomas (VSs) are due to the inactivation of the NF2
gene. Some tumours remain stable for years, whereas others grow relatively fast. However, the biological background of these phenotypical heterogeneities is
largely unknown.

Methods: A total of 282 patients with sporadic VSs were included in a multi-center study.Bidirectional sequencing was conducted to detect microlesions
in the NF2 gene. To identify large/exonic deletions, we used a commercial MLPA kit for the analysis. Immunoblot analysis, Immunohistochemistry and
Immunofluorescence were conducted to evaluate the expression levels of merlin proteins.

Results: We found age-dependent differences in the clinical parameters of sporadic VSs. Young patients were characterized by progressive tumour behaviours,
including earlier onset of initial symptoms, shorter symptom duration and larger tumour size. An increased rate of “two-hits” of both NF2 alleles, usually by
mutation and allelic loss, was observed in young cases compared to older, and this correlated with the loss of protein and mRNA expression. Following the
loss of NF2 expression, schwannoma cultures demonstrated increased proliferation rates.

Conclusions: We have identifed a correlation between the NF2 status and the growth patterns of sporadic VSs. The treatment decision-making, microsurgery
or “wait and scan” strategy, should be carried out according to the tumour’s genetic background.

Full List of Authors: Hongsai Chen*1, zhaoyan wang2, hao wu2

1
Department of Otolaryngology Head & Neck Surgery, Ear Institute, School of Medicine, Shanghai Jiao, Shanghai, Cocos (Keeling) Islands, 2Department of Otolaryngology Head &
Neck Surgery, Ear Institute, School of Medicine, Shanghai Jiao, Shanghai, China

Developmental Origin and Spatiotemporal NF1 Loss of Heterozygosity Give Rise to Different Types of
Cutaneous Neurofibroma

Zhiguo Chen, Dermatology, UT Southwestern Medical Center, Dallas, TX, United States

Background: Dermal or cutaneous neurofibromas (cNF), a Schwann cell tumor of the peripheral nerves in the skin, affect more than 95 percent of adults
with Neurofibromatosis Type 1 (NF1) and are a major source of emotional, physical and social distress as NF1 patients can have thousands of these tumors
covering most of their skin. Thus, patients with NF1 often identify these tumors as their greatest burden. To date, there is no available medical treatment for
cNF, no known way to prevent them from developing. The major barriers that impede progress in this field are the lack of accurate models of these common
cNF tumors for drug evaluation and a limited understanding of their pathogenesis as well as the identity of specific cell of origin that directly gives rise to cNF.
In this regard, mouse models are important tool for elucidating the molecular mechanism and preclinical drug screening. Ironically, none of the numerous NF1
mouse models developed so far recapitulate cNF.

Methods: We take advantage of genetic labeling for cell lineage tracing to identify mouse neural crest Cre lines that are expressed in the sub-population of
Schwann cell lineage that give rise to cNF when NF1 is deleted.

Results: We discovered that a Homeobox B transcription factor serve as the lineage marker to trace the developmental origin of cNF neoplastic cells that is
completely recapitulates human neurofibromatosis, generating a novel mouse model that spontaneously develops both cutaneous and plexiform neurofibroma.
In addition, we discovered that the modulation of the Hippo pathway acts as a modifier to promote neurofibromagenesis, suggesting that dampen the Hippo
pathway may serve as part of the comprehensive treatment approach for neurofibroma.

Conclusions: This study provides insights into the developmental origin of cNF, the most common tumor in NF1, and generates the first mouse model
that faithfully recapitulates both human cutaneous and plexiform neurofibroma. This novel mouse model has begun to yield vital clues to neurofibroma
pathogenesis and now opens the doors for deciphering the evolution of cNF to identify potential effective therapies.

Full List of Authors: Zhiguo Chen*1, Juan Mo1, Jean-Philippe Brosseau1, Tracey Shipman1, Yong Wang1, Chung-Ping Liao1, Jonathan Cooper1, Lu Le1, 2, 3, 4
1
Dermatology, 2Neurofibromatosis Clinic, 3Simmons Comprehensive Cancer Center, 4Hamon Center for Regenerative Science and Medicine, UT Southwestern Medical Center, Dallas,
TX, United States

Funding: J.P.B and C.P.L are recipients of the YIA from Children’s Tumor Foundation. This work was supported by funding from the National Cancer Institute, the US Department of
Defense, the Giorgio Foundation, the Neurofibromatosis Therapeutic Acceleration Program and the NF1 Research Consortium Fund.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 81
 Mechanotransduction and NF1 Loss Partner in Crime: A “Mature” Story

Federica M. Chiara, Cardiological, Thoracic and Vascular Sciences, University, Padova, Italy

Background: In several tumor models, the potent activation of Ras/Raf/ERK signaling by a mutated Ras oncogene drives the initiating steps of cell
transformation. In NF1, neurofibromin loss leads to a modest decrease of RAS-GTPase activity, which in turns leads to increased Ras activation. Thus, Ras
activation alone does not trigger transformation of NF1-/- Schwann cells (SCs) (1), but requires further molecular hits to drive neoplastic progression (2). Mouse
models of NF1 revealed that Ras-mediated transformation of SCs probably relies on a step-wise process that integrates circuits of amplification signals from
the local niche, whose the major component is the extracellular matrix (ECM), whose mechanical and biochemical properties provide environmental cues
influencing cell behavior (3). We have proposed the hypothesis that the ECM contributes to neurofibroma development and progress towards malignancy. Our
hypothesis is sustained by the following observations i) in pancreatic and breast cancer progression models, loss of tensional homeostasis helps Ras oncogenic
transformation (4); ii) Neurofibromas are composed of a fibrotic milieu; iii) our previous data show that primary NF1-/- SCs formed 3D colonies in stiff Matrigel
and showed FAK/Src axis hyperactivation.

Methods: We settled a new 3D cellular model in vitro of soft and stiff matrix that we investigated by ìmmunofluorescence and biochemical means.

Results: In the present report, we confirm our preliminary data in vitro in a 3D experimental model using immortalized SCs from plexiform neurofibroma. In the
effort to elucidate how these cells transduce mechanical signals, we found a FAK-dependent activation of AKT, which disrupts the degradation complex and
activates the oncogenic transcription factor beta-catenin. RHO/YAP transduction axes, in a collagen rich, stiff matrix further potentiate this process. Further,
SCs build up a platelet growth factor autocrine loop and as carcinoma cells, they express a collagen linker enzyme—lysyl oxidase (5)— and secrete increased
amounts of metalloproteinases MMP-2 and 3 and IL-8.

Conclusions: These findings point towards an important role of the fibrotic process not only in sustaining, but also in triggering neurofibroma progression.

1. Cichowski K and Jacks T, Cell; 104:59, 2001

2. Harrisingh MC and Lloyd AC, Cell Cycle; 3:1255, 2004
3. Aragona M et al, Cell,
4. Hanane Laklai et al, Nature Medicine, 22, 5, 2016
5. Erez, N. et al Nature, 2015 Jun 4;522(7554):41-2.

Full List of Authors: Federica M. Chiara*1, Kristen Mills2, Anna Stocco3, Andrea Errico1, Robert Altman2, Xiangyu Gong2, Jonathan Kulwatno2
1
Cardiological, Thoracic and Vascular Sciences, University, Padova, Italy, 2Department of Mechanical, Aerospace, and Nuclear Engineering Center for Biotechnology and
Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NY 12180, United States, 3Biomedical Sciences, University, Padova, Italy

82 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Genomic Variants Enriched in Patients with Plexiform Neurofibroma

Kwangmin Choi, Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, Cincinnati

Background: Bi-allelic NF1 gene mutations characterize benign plexiform neurofibroma (PNF) Schwann cells (SC). Although no other genomic changes have
been identified to explain PNF tumorigenesis, clinical studies imply that formation of PNF may require additional mutations.

Methods: We performed the Whole-Exome Sequencing (WES) using sorted SCs and their matched fibroblasts (FBs) from 8 NF1 patients. Rare non-
synonymous germline and somatic variants were predicted by GATK, MuTect1, VarScan2, Strelka, and SomaticSniper. We also re-analyzed independent dermal
neurofibroma (DNF) and PNF sequencing data sets to verify our findings. We measured genomic instability of PNF tumors and DNF SC in tissue sections and
cultured cells, and used cell culture to define how loss of the DNA repair protein ATM affects SC and SC precursors.

Results: We identified possibly pathogenic variants in 47 genes, in addition to the known driver mutations in NF1. None of them have previously been evaluated
in the context of the NF1 tumorigenesis. Somatic variants in SC were present at low read number, in few tumors, and were largely missense alterations of
unknown significance of uncertain relevance to tumorigenesis. Importantly, we identified frequent, probably damaging germline variants in 16 genes, including
ATM, many of which showed decreased gene expression. Phenotypes relevant to the DNA repair-related gene ATM were present at elevated levels in a subset
of neurofibromas, and of neurofibroma SC. In mice, the Atm gene loss promoted SC precursor self-renewal, and increased tumor formation in vivo. Thus,
genetic alteration of ATM may modify NF1 pathogenesis driven by NF1 mutations.

Conclusions: Recent studies identify ATM germline risk alleles for many sarcoma types. Our finding of increased mutational burden in ATM-mutant SC, not
fibroblasts, suggests that, in SC with biallelic NF1 mutations and elevated RAS-GTP, the effects of ATM variants are enhanced. The DNA comet assay also
suggests that DNA damage repair capacity of SC from some DNF is reduced. Genes, including ATM, showing common germline variants in neurofibroma are
potential modifiers of tumor formation and/or transformation to malignancy, and may contribute to clinical variability. The DNA comet assay also suggests that
DNA damage repair capacity of SC from some DNF is reduced.

Full List of Authors: Kwangmin Choi*1, Yanan Yu1, Jianqiang Wu1, Paul Andreassen2, Philip Dexheimer3, Mehdi Keddache4, Jose Cancelas1, Hilde Brems5, Robert Spinner6, Eric
Legius7, Kristine Vogel8, Nancy Ratner1
1
Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, Cincinnati, 2Experimental Hematology and Cancer Biology, Cincinnati Children's
Hospital Medical Center, Cinciinati, 3Biomedical Informatics, 4Human Genetics, Cincinnati Children's Hospital Medical Center, Cincinnati, United States, 5Center for Human Genetics,
University Hospitals Leuven and Department of Human Genetics, The Katholieke Universiteit Leuven, Leuven, Belgium, 6Neurological Surgery, Mayo Clinic, Rochester, United States,
7
Center for Human Genetics, University Hospitals Leuven and Department of Human Genetics, KU Leuven, Leuven, Belgium, 8Cell systems and Anatomy, UT Health San Antonio,
San Antonio, United States

This work was funded by NIH-R37-NS083580 (NINDS Javits Neuroscience Investigator Award to NR) and an Innovation Award from Cincinnati Children’s Hospital (NR)

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 83
 Are Reading and Motor Skills Related to Sequential Learning and Motor Adaption in Children with
Neurofibromatosis Type 1?

Jeremy Danna, Aix Marseille Univ, CNRS, LNC, Marseille

Background: About half of children with Neurofibromatosis type 1 (NF1) have learning disabilities (Cutting & Denckla, 2003), including deficits in reading and/
or motor learning. In the case of neurodevelopmental disorders (APA, 2013), reading and motor difficulties have been related to a deficit in the procedural
learning system (model of Nicolson & Fawcett, 2007), within which two learning processes have been distinguished: sequential learning and motor adaptation
(model of Doyon and Benalli, 2005). The present study aims to better characterize learning disabilities in reading and motor skills from the rationale provided
by these models.

Methods: Twenty-three children with NF1 (10;2 ± 1;6 years, 16 girls) participated in the study. Parents and children gave their informed consent prior to
the experiment, approved by the local ethics committee. Reading and motor skills were respectively evaluated with the Alouette test (Lefavrais, 2005) and
the French M-ABC motor test (Soppelsa & Albaret, 2004). Two experimental tasks were conducted, one requiring a motor adaption, namely writing trigrams
in conventional (from left to right) versus unconventional directions, and one requiring a sequential learning, namely learning a new ideogram following the
correct sequence of strokes.

Results: Motor adaptation. A cluster analysis completed with an ANOVA revealed two categories of children: a first cluster including 12 children whose
performance was not different between the two writing directions, and a second cluster including 11 children in whom this difference was significant (p <
0.001). The M-ABC score was lower for the cluster 2 than for the cluster 1. The Alouette scores did not differ between the clusters.

Sequential learning. The analyses revealed two categories of children: a first cluster including 13 children in whom the learning phase was correct, and
a second cluster including 10 children in whom the sequential errors were greater. The difference between the clusters was not significant for the scores
obtained in the Alouette and M-ABC tests.

Conclusions: A link between the performance in the adaption task and motor skills was evidenced in NF1 children. No link was observed between reading
skills and performance in both tasks. This finding runs counter the reference to the model of Nicolson & Fawcett (2007) concerning learning disabilities in
reading and motor skills in children with NF1.

Full List of Authors: Jeremy Danna*1, Marianne Jover2, Stephanie Ducrot3, Jean-Luc Velay1, Jean-Michel Albaret4, Frederique Audic5, Yves Chaix6
1
Aix Marseille Univ, CNRS, LNC, Marseille, 2Aix Marseille Univ, PSYCLE, 3Aix Marseille Univ, CNRS, LPL, Aix-en-Provence, 4ToNIC, Toulouse NeuroImaging Center, Université de
Toulouse, Inserm, UPS, Toulouse, 5Service de Neurologie Pédiatrique, CHU Timone-Enfants, Marseille, 6ToNIC, Toulouse NeuroImaging Center, Université de Toulouse, Inserm, UPS
& Centre Hospitalier Universitaire de Toulouse-Purpan, Toulouse, France

Disclosure of Interest: J. Danna has a conflict with: None. This work has benefited from support from the French Government, managed by the French National Agency for Research
(ANR), under the project title DYSTAC-MAP (ANR-13-APPR-0010)., M. Jover: None Declared, S. Ducrot: None Declared, J.-L. Velay: None Declared, J.-M. Albaret: None Declared,
F. Audic: None Declared, Y. Chaix: None Declared

84 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Combining Cre/loxP and CRISPR/Cas9 Technology for In Vivo Studies of Myeloid Cell Function in MPNST

Rebecca Dodd, Internal Medicine, University of Iowa, Iowa City, United States

Background: NF1 haploinsufficiency is a hallmark of Neurofibromatosis Type 1. NF1 +/- myeloid cells are key players in neurofibroma development, but
the mechanisms driving myeloid cell function in malignant peripheral nerve sheath tumors (MPNSTs) are poorly understood. Using mouse models, we have
shown that NF1 +/- bone marrow can accelerate MPNST initiation, which is characterized by elevated levels of tumor-infiltrating mast cells and CD11b+
myeloid cells. Importantly, mast cells are also enriched in human NF1-associated MPNSTs, suggesting they may be attractive therapeutic targets. While our
recent findings underscore the importance of NF1 +/- myeloid cells in MPNST biology, further work is needed to understand the significance and molecular
mechanisms of these immune cells.

Methods: To further define the population of myeloid cells influencing MPNST initiation, we have developed unique mouse models that combine CRISPR/
Cas9 tools with Cre-loxP-driven, lineage-specific cell depletion. We recently established technology to generate CRISPR-induced MPNSTs in wild type mice
by injecting the sciatic nerve with Adenovirus expressing Cas9 and guide RNAs to NF1 and p53. We are combining this CRISPR/Cas9 somatic tumorigenesis
approach with Cre-loxP control of immune cell populations to determine the role of myeloid and mast cells in MPNST formation. Using lineage-specific Cre
drivers, we are generating MPNSTs in mice where myeloid or mast cells are conditionally or temporally depleted.

Results: Using mice with Cre-inducible diphtheria toxin (LSL-DTA) crossed to a mast cell-specific promoter (Mcpt5-Cre), we have successfully depleted mast
cells in multiple tissues, including CRISPR/Cas9-generated MPNSTs. Deletion of mast cells was sufficient to slow MPNST initiation in this mouse model. Flow
cytometry analysis of MPNSTs from Mcpt5-Cre; LSL-DTA mice showed that mast cell deletion alters the immune cell infiltrate of the tumors.

Conclusions: By combining CRISPR/Cas9 somatic tumorigenesis technology with sophisticated Cre-loxP control of the tumor microenvironment, we are
able to differentially manipulate tumor cells and the surrounding stroma to investigate mechanisms of myeloid and mast cell function in MPNSTs. These
experiments have identified a fundamental role for mast cells in MPNST initiation, and current studies are investigating the mechanisms influencing mast cell
and monocyte function in NF1-associated MPNSTs.

Full List of Authors: Amanda Scherer1, Vickie Knepper-Adrian1, Victoria Stephens1, Emily Laverty1, Rebecca Dodd*1
1
Internal Medicine, University of Iowa, Iowa City, United States

Funding Support: Department of Defense NF1 New Investigator Award

Targeting the Serotonin the 5-HT6 Receptor/mTOR Complex to Reverse Cognitive Deficits in
Neurofibromatosis Type 1

Emile Doucet, Neurobiologie, Institut de Génomique Fonctionnelle, CNRS, INSERM, Université de Montpellier, Montpellier

Background: Among the 14 serotonin receptor subtypes, the 5-HT6 receptor has emerged as a promising target for the treatment of cognitive impairment
associated with several neuropsychiatric disorders, including Alzheimer’s disease and schizophrenia: 5-HT6 receptor blockade consistently enhances mnemonic
performance in a broad range of procedures in rodents and there is preliminary evidence for pro-cognitive properties of 5-HT6 receptor antagonists in human.
Using a proteomics approach, we showed that 5-HT6 receptors recruit several proteins of the mTOR pathway, including mTOR itself and Neurofibromin.
Further biochemical and behavioural studies showed that the interaction of Neurofibromin with 5-HT6 receptor increases agonist-independent receptor-operated
signalling and that mTOR activation, under the control of 5-HT6 receptor, underlies cognitive deficits in two rat neuro-developmental models of schizophrenia.
The present study was undertaken to explore the role of 5-HT6 receptor in the increased mTOR activation in brain of Nf1+/− mice, a preclinical model of
neurofibromatosis type 1, and whether blocking 5-HT6 receptor-operated mTOR signaling rescues social discrimination deficits in these mice.

Methods: mTOR signalling was assessed by Western blotting using antibodies against phosphorylated (activated) mTOR and p70 S6 kinase. Social
discrimination in wild type and Nf1+/− mice was assessed using the three-chamber Crawley’s test.

Results: Administration of the 5-HT6 receptor antagonist SB258585, but not the mTOR inhibitor rapamycin, to Nf1+/- mice abolished the long-term social discrimination
deficit observed in these mice. Neither SB258585 nor rapamycin affected social discrimination in wild-type mice. In addition, the increase in the phosphorylation
level of mTOR and its substrate p70S6 kinase, in prefrontal cortex of Nf1+/- mice, compared with wild type mice, was abolished by SB258585 administration.

Conclusions: Collectively, these findings indicate that 5-HT6 receptor blockade rescues the deficit in long-term social discrimination observed in a preclinical
model of Neurofibromatosis type 1, and that the associated reduction in brain mTOR signalling does not account for the beneficial effect of 5-HT6 receptor
inhibition upon social discrimination. They also suggest that repositioning 5-HT6 receptor antagonists that are currently in phases 2 and 3 of clinical trials in
dementia would be a promising strategy to alleviate cognitive deficits in patients with neurofibromatosis type 1.

Full List of Authors: Emile Doucet*1, Séverine Chaumont-Dubel1, Séverine Morisset-Lopez2, Joël Bockaert1, Carine Bécamel1, Philippe Marin1
1
Neurobiologie, Institut de Génomique Fonctionnelle, CNRS, INSERM, Université de Montpellier, Montpellier, 2Centre de Biophysique Moléculaire, CNRS, Université d'Orléans, Orléans, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 85
 Characterization of Vestibular Schwannomas in NF2 Using a MALDI Mass Spectrometry Imaging Combined
with Microproteomic Approach

Lauranne Drelich, U1192 - Protéomique Réponse Inflammatoire Spéctrométrie de Masse - PRISM, Inserm, Villeneuve d’Ascq

Background: Among the pathways identified controlling schwannoma growth in in-vitro and in-vivo models, none as demonstrated sufficient clinical relevance
to validate its clinical use. Therefor new potential therapeutic targets are needed. The aim of our study was to identify the protein expression and networks
involved in NF2 vestibular schwannomas using an unbiased, large scale proteomic approach.

Methods: An in-vitro approach was used to identify the intracellular proteins and secreted proteins from matched primary cultures of mouse Schwann cells
harboring either a NF2 competent or NF2-/- status. Then paraffin embedded tissue sections from NF2 schwannomas and non-NF2 vestibular nerves (Tumor
Bank protocol CSTMT234) were digested in-situ using FASP-trypsin Digestion.Molecular images are generated from thin tissue section in order to determine
the spatial localization of digested peptides. Thanks to unsupervised statistical analysis, we then generated hierarchical clustering of homogeneous molecular
regions. According to these regions, microproteomic gave access to large scale protein identification and relative quantification. Statistical analysis was then
carried out using ANOVA (threshold p value 0.05) and presented as heatmap and Venn diagram.

Results: A first selection of ten schwannomas operated on NF2 patients and two control non-NF2 vestibular nerves were selected. At the present time,
microproteomic analysis were realized on 2/10 tumor tissues. The results on the 8 other tumors are expected for the final presentation. A total of 971 were
identified. 626 proteins are common between the two patient and two panel of proteins are specifically overexpressed in each patient. One panel is associated
with cellular division process whereas the other one correspond to enzyme regulation. A total of 355 intracellular proteins were overexpressed in the NF2-
/- Schwann cells compared to WT cells. DAVID ontology analysis highlighted an enrichment of proteins involved in ribosomes and mRNA translation (p value
9.3E-24) which might be the consequence of mTORc activation in these NF2-/- Schwann cells. A comparison between proteins overexpressed in tumoral
tissue and NF2-/- cell line is in process.

Conclusions: This first report demonstrates the possibilities raised by MALDI MSI and combined micro-proteomic approach to identify numerous modifications
of protein expression secondary to the loss of the tumor suppressor gene NF2 even from paraffin embedded tissues.

Full List of Authors: Lauranne Drelich*1, Marie Duhamel1, Nicolas Bonne1, 2, Dominique Lallemand3, Michel Salzet1, Isabelle Fournier1
1
U1192 - Protéomique Réponse Inflammatoire Spéctrométrie de Masse - PRISM, Inserm, Villeneuve d'Ascq, 2Otologie et Otoneurologie, CHU Lille, Lille, 3EA7331, Inserm &
Faculté de pharmacie, Université Paris Descartes, Paris, France

86 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Increase of Inward Cationic Currents (Ih) with Lamotrigine Partially Corrects Deficits in Habituation Learning
in a Drosophila Model of NF1

Michaela Fenckova, Department of Human Genetics, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center,
Nijmegen, Netherlands

Background: Neurofibromatosis Type 1 (NF1) is caused by heterozygous mutations in the gene NF1, encoding the Ras GTPase-activating protein
neurofibromin. NF1 Individuals present with high (>50%) prevalence of learning difficulties and autism spectrum disorder (ASD). Research in the NF1 mouse
model has shown that learning deficits are caused by increased activity of Ras signaling through the classic MAPK pathway. Inhibiting Ras activity with
lovastatin reversed these deficits in mice but was not sufficient to significantly improve cognitive impairments in individuals with NF1 in long-term clinical trials.
Recently, it has been shown that attenuation of inward cationic currents (Ih) through the potassium/sodium hyperpolarization-activated cyclic nucleotide-gated
channel 1 (HCN1) contributes to learning deficits in the NF1 mouse model. Increase of Ih current with the HCN channel agonist, lamotrigine restored these
deficits (Omrani et al., 2015, Mol. Psy.). Whether Ih currents and Ras-MAPK are converging or independent pathways downstream of NF1, and if both need to
be corrected to reverse cognitive deficits remains to be investigated.

Methods: We use Drosophila, a powerful model, to investigate the role of neurofibromin (Drosophila Nf1) and its downstream effectors. Drosophila genetics
allows to dissect the functional relationship between Ih currents and Ras-MAPK by genetic interaction experiments and assess cognitive phenotypes and drug
rescue experiments with high efficiency.

Results: We found that decreased Drosophila Nf1 expression, increased Ras-MAPK signaling as well as decreased Ih currents cause deficits in habituation,
a form of learning that is highly relevant to cognitive functioning. Genetic interaction between Drosophila Nf1 and Ih confirmed the involvement of these two
proteins in shared cellular process. Increasing Ih currents with lamotrigine partially restored the habituation deficits of Nf1 knockdown flies, providing a suitable
system for future combinatorial drug testing.

Conclusions: Our research provides a novel approach to identify targets and treatment strategy to alleviate cognitive defects associated with NF1. Using
habituation, a fundamental form of learning and a prerequisite for higher cognitive functioning that can be measured in both patients and flies we increase the
translatability of our research. Our model thus offers an unprecedented throughput and flexibility, ideal for pre-clinical testing.

Full List of Authors: Michaela Fenckova*1, Laura Blok1, Mireia Coll-Tane1, Monique Voet, van der1, Annette Schenck1
1
Department of Human Genetics, Donders Institute for Brain, Cognition and Behaviour, Radboud University Medical Center, Nijmegen, Netherlands

Disclosure of Interest: M. Fenckova: None Declared, L. Blok: None Declared, M. Coll-Tane: None Declared, M. Voet, van der: None Declared, A. Schenck has a conflict with: The
research was funded by European Union’s FP7 large-scale integrated network Gencodys (HEALTH-241995) and DCMN Radboudumc junior research fellowship.

Radioprotective Effect of the Combination of Statins and Bisphosphonates in Cells Derived From
Neurofibromatosis Type 1 (NF1)

Nicolas Foray, Inserm

Background: The major molecular radioprotection approaches are generally based on the reduction of the number of radiation-induced DNA damage via
anti-oxidant drugs. However, an increasing body of evidence suggests that radiosensitivity is rather caused by unrepaired DNA double-strand breaks (DSB)
than induced ones. Furthermore, the number of DSB produced immediately after irradiation does not necessarily condition the number of unrepaired residual
DSB. By analyzing hundreds radiosensitive cell lines, the nucleo-shuttling of the ATM protein has been shown to condition the cellular response to radiation.
Since the combination of bisphosphonates (zoledronate) and statins (pravastatin) (ZOPRA) results in an enhanced diffusion of ATM in the nucleus, the ZOPRA
treatment may lead to a better radioprotection of cells. Here, we investigated such hypothesis in human skin fibroblasts derived from NF1 patients.

Methods: The effect of ZOPRA treatment was analyzed on a collection of 40 cell lines derived from NF1 patients by using immunofluorescence against the
major DSB biomarkers.

Results: The ZOPRA treatment appeared to significantly reduce the number of unrepaired DSB and the level of genomic instability specific to NF1 cells. It
appeared that statins and bisphosphonates contribute in a supra-additive manner to such effect.

Conclusions: The use of statins and bisphosphonates may be an interesting approach to reduce the radiation-induced risk specifically linked to NF1 syndrome.
Further clinical investigations are needed to confirm the hypothesis and ask whether ZOPRA may also contribute to increase the life span of NF1 patients.

Full List of Authors: Mélanie Ferlazzo1, Nicolas Foray*1, Laurene Sonzogni1, Clément Devic1, Adeline Granzotto1, Bernard Guennoc2, Patrick Combemale2
1
Inserm, 2Centre Leon Berard, Lyon, France

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 87
 Radiobiological Characterization of Neurofibromatosis Type I: Impact of Neurofibromin Protein on the ATM-
Dependent DNA Damage Repair and Signaling Pathway. A Novel Mechanistic Model.

Nicolas Foray, Inserm

Background: Neurofibromatosis type 1 (NF1) is a syndrome characterized by high occurrence of benign and malignant brain tumours and is caused by
mutations of neurofibromin. While NF1 was shown to be associated with radiosensitivity, few studies have dealt with the molecular and cellular response of
NF1 cells to ionising radiation.

Methods: Here, we examined the ATM-dependent signaling and repair pathways of the DNA double-strand breaks (DSB), the key-damage induced by ionizing
radiation (medical X-rays), in skin fibroblasts from NF1 patients.

Results: Neurofibromin mutations are associated with an abnormal abundance of neurofibromin molecule in cytoplasm, independently of irradiation. After
irradiation, quiescent NF1 cells showed abnormally low rate of recognized DSB assessed by H2AX immunofluorescence. Irradiated NF1 cells also showed a
delayed nucleo-shuttling of phosphorylated forms of the ATM kinase, caused by a specific binding of ATM to mutated neurofibromin in cytoplasm. Lastly, NF1
fibroblasts showed abnormally high MRE11 nuclease activity suggesting high genomic instability. A similar analysis of 40 cell lines derived from NF1 patients
consolidate these results.

Conclusions: Our results showed that NF1 belongs to the group of syndromes associated with low but significant defect of DSB signalling and delay in
the ATM nucleo-shuttling associated with radiosensitivity and high-cancer proneness. Our findings are therefore consistent with a sequestration of ATM
in cytoplasm due to a more frequent formation of the ATM-neurofibromin complexes. The formation of such complexes prevents the diffusion of ATM into
nucleus after irradiation and impairs DSB repair and signaling. This mechanistic model about NF1 syndrome and raised the potential radiation-induced risks
specifically linked to the NF1 syndrome.

Full List of Authors: Nicolas Foray*1, Mélanie Ferlazzo1, Laurene Sonzogni1, Clément Devic1, Adeline Granzotto1, Bernard Guennoc2, Patrick Combemale2
1
Inserm, 2Centre Leon Berard, Lyon, France

Tyrosine Nitration Regulates Neurofibromatosis Type 2-Associated Schwannoma Energy Metabolism and
Cell Survival

Maria C. Franco, Biochemistry and Biophysics, College of Science, Oregon State University, Corvallis, Oregon

Background: Neurofibromatosis type 2 (NF2) is a genetic disorder of the nervous system caused by inactivation of the merlin tumor suppressor gene. NF2
patients develop bilateral vestibular schwannomas (VS) for which there is no effective drug treatment. Production of the oxidant peroxynitrite occurs in
pathological conditions and leads to tyrosine (Tyr) nitration of proteins. Although Tyr nitration is found in multiple tumor types, its role in tumorigenesis is
unknown. Our goal is to discover nitrated proteins that support VS growth as new potential tumor-specific therapeutic targets, enabling the development of safe
pharmacological strategies for NF2 schwannomas.

Methods: To prevent Tyr nitration, isogenic human wild-type and merlin-deficient Schwann cells (WT- and MD-HSC) were cultured with and without
scavengers and inhibitors of peroxynitrite formation and Tyr nitration. Cell survival and metabolic parameters were assessed using biochemical and molecular
methods, and extracellular flux analysis. Nitrated proteins in VS from NF2 patients were identified by immunoprecipitation with anti-nitroTyr antibodies followed
by mass spectrometry analysis.

Results: VS and MD-HSC showed significantly increased levels of peroxynitrite and Tyr nitration compared to WT-HSC. Notably, prevention of tyrosine nitration
selectively decreased MD-HSC survival. A study of the MD-HSC metabolic phenotype revealed a shift in energy metabolism compared to WT-HSC characterized by
decreased levels and activity of mitochondrial oxidative phosphorylation (complex I to IV), decreased mitochondrial oxygen consumption and membrane potential,
and increased glycolysis and glutamine dependency. Prevention of Tyr nitration reversed the metabolic phenotype of MD-HSC back to that of WT-HSC. We previously
identified nitrated Heat shock protein 90 (Hsp90) as a key inhibitor of mitochondrial activity in tumor cells. Nitrated Hsp90 was present in VS and MD-HSC but not
in WT-HSC, representing an excellent target for drug development. We also identified 5 additional nitrated proteins in VS related to pathways deregulated in NF2.

Conclusions: Nitration induces a metabolic reprogramming and supports survival of human NF2-associated schwannoma cells. Identification of the signaling
pathways regulated by specific nitrated proteins that promote schwannoma growth could provide additional novel targets for the treatment of NF2 and possibly
other tumor types as well.

Full List of Authors: Maria C. Franco*1, Stephani Klingeman Plati2, Jeanine Pestoni1, Oliver Valdivia Camacho1, Marisa Fuse2, Maria Onatunde2, Nicklaus Sparrow2, Cristina
Fernandez-Valle2
1
Biochemistry and Biophysics, College of Science, Oregon State University, Corvallis, Oregon, 2Burnett School of Biomedical Sciences, College of Medicine, University of Central
Florida, Orlando, Florida, United States

88 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Genomic Structural Alterations as a Driving Force in MPNST Development

Bernat Gel, Hereditary Cancer, Germans Trias i Pujol Research Institute (IGTP), Badalona, Spain

Background: Malignant peripheral nerve sheath tumors (MPNST) are soft tissue sarcomas with bad prognosis and lack of curative treatments. NF1 patients
have an 8-13% lifetime risk of developing an MPNST. MPNST may arise from a preexisting benign plexiform neurofibroma (PNF), often after the formation of a
pre-malignant distinct nodule termed atypical neurofibroma (aNF).

Methods: We generated genomic structure (SNP-array), mutational (exome), transcriptomic (RNA-seq, microarray) and epigenomic (DNA methylation) data
from a set of 15 MPNSTs, and also collected available data on MPNST, plexiform and atypical neurofibromas. We performed an integrative bioinformatic
analysis this data to infer the mechanisms of MPNST development.

Results: Regarding the genomic structure, PNFs have no structural

alterations except those affecting chromosome 17q involved in the
somatic inactivation of NF1. In addition to mutation of both NF1
alleles, aNFs present additional recurrent losses of the CDKN2A/B
locus, involving the 9p arm. Contrasting with these nearly-normal
karyotypes, the MPNSTs studied have hyperploid and highly
rearranged genomes with somatic copy number alterations
(SCNAs) affecting most chromosomes. However, MPNST genome
structure is highly stable. In contrast to SCNAs, MPNSTs have a low
number of point mutations, with no clear recurrently affected genes.
Most point mutations appear to be acquired after the genome
reorganization.

This collective data suggest a model for MPNST origin, with a first
progression towards a proliferative cell with reduced senescence
due to the loss of NF1 and CDKN2A/B, followed by a number of
random catastrophic events of genomic alteration and the selection
of a viable stable genomic combination.

Furthermore, SCNA have a profound impact on gene transcription

levels and create regions with an accumulation of over- and
under- expressed genes, transcriptional imbalances (TI). TIs mostly
capture passenger gene expression but allow identification of genes
with SCNA-independent expression regulation. The analysis of
these genes provides insight into the biology of MPNSTs.

Conclusions: Gross genomic structural alterations are a driving force in MPNST biology and their genomic stability suggest a catastrophic event mediated by
loss of senescence capacity as a probable origin.

Full List of Authors: Bernat Gel*1, Ernest Terribas1, Elisabeth Castellanos1, Inma Rosas1, Tapan Mehta2, 3, Peggy Wallace4, Nancy Ratner5, Ignacio Blanco6, Juana Fernández-
Rodríguez7, Conxi Lázaro7, Eduard Serra1
1
Hereditary Cancer, Germans Trias i Pujol Research Institute (IGTP), Badalona, Spain, 2Nutrition Obesity Research Center, 3Department of Health Services Administration, University
of Alabama at Birmingham, Birmingham, 4Department of Molecular Genetics & Microbiology, University of Florida College of Medicine, Gainesville, 5Division of Experimental
Hematology and Cancer Biology, Department of Pediatrics, Cincinnati Children's Hospital Medical Center, University of Cincinnati, Cincinnati, United States, 6Clinical Genetics and
Genetic Counseling Program, Germans Trias i Pujol Hospital, Can Ruti Campus, Badalona, 7Hereditary Cancer Program, Joint Program on Hereditary Cancer, Catalan Institute of
Oncology, IDIBELL campus, L’Hospitalet de Llobregat, Spain

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 89
 PARP Inhibition in Combination with Cytotoxic Chemotherapy as a Therapeutic Option for MPNSTs

Abigail Godec, Washington University in St. Louis, St. Louis, United States

Background: Malignant peripheral nerve sheath tumors (MPNSTs) are aggressive sarcomas that occur at increased frequency in individuals with
neurofibromatosis type 1 (NF1). These tumors have limited treatment options and a high propensity to metastasize leading to a dismal overall survival.
Cytotoxic chemotherapy is basis of first line therapy for metastatic MPNSTs. This alone, however, is very unlikely to result in a durable response. Soft tissue
sarcomas including MPNSTs have been shown to have mutations and copy number aberations in DNA repair genes, suggesting the potential for response to
PARP inhibition. Given these data, we decided to assess the response of MPNST cell lines to PARP inhibition.

Methods: We utilized several different MPNST cell lines, including 2 human-derived NF1-associated (96.2 and MP642), 1 human-derived sporadic (STS 26T)
and murine Nf1/Tp53-mutant NPcis MPNST cells (JW23.3). Proliferation and cell death were evaluated in vitro using the IncuCyte Live Cell Analysis System
(Essen BioScience) in the presence of a PARP inhibitor or drug vehicle. Next, we examined proliferation and cell death in the presence of a PARP inhibitor in
combination with cytotoxic chemotherapy or drug vehicle.

Results: We observed decreased proliferation in all cell lines examined as assessed by measurement of cell confluence over 48 hours. Additionally, we
observed an increase in cell death as measured by fluorescence.

Conclusions: Taken together this data suggests that a PARP inhibitor in combination with cytotoxic chemotherapy is a potential therapy for soft-tissue
sarcomas. Future work is aimed at assessing response to this combination in vivo using both cell lines and patient derived xenografts developed in our lab and
will be presented this fall.

Full List of Authors: Abigail Godec*1, Xue-Liang Zhou1, Xiao Zhang1, Wenjing Qin1, Vanessa Eulo1, Angela Hirbe1
1
Washington University in St. Louis, St. Louis, United States

This work was funded by Dr. Hirbe’s start up funds. The PARP inhibitor was supplied by AstraZenica. Dr. Hirbe is supported by the Francis Collins Award through NTAP.

90 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Whole Exome Sequencing Leads to the Identification of TRIM23 as a Potential Suppressor of Metastasis in
NF1-MPNST

Abigail Godec, Washington University in St. Louis, St. Louis, United States

Background: Malignant peripheral nerve sheath tumors (MPNSTs) are aggressive sarcomas that occur at increased frequency in individuals with
neurofibromatosis type 1 (NF1). These tumors have limited treatment options and a high propensity to metastasize leading to a dismal overall survival. While
previous studies have employed a variety of discovery approaches to discover genes associated with MPNST pathogenesis, little is known about the genetic
events leading to metastasis of these tumors.

Methods: Whole exome sequencing was performed on resection and biopsy materials from two patients with NF1-MPNST. Samples from patient one included
primary tumor, a lung metastasis, and a bone metastasis. Samples from patient two included primary tumor, and two different lung metastases. To further
explore one interesting target, we used shRNA-mediated knock down of Trim23 expression in Nf1/Tp53-mutant NPcis MPNST cells (JW23.3) to explore the
role of Trim23 in MPNST pathogenesis. Subcutaneous and left ventricular (LV) tumor injections were performed to assess the role of Trim23 in growth of
a primary tumor and metastasis, respectively. Injections were performed in C57BL/6 ALBINO immunocompetent mice. All human and animal studies were
performed under active protocols approved by the Institutional Review Board and the Institutional Animal Care and Use Committee, respectively.

Results: First, we identified the NF1 mutations in each patient as well as copy number loss of CDKN2A across all primary and metastatic samples consistent
with these genes serving as drivers of progression. Second, we have identified mutations in TRIM family members enriched in the metastatic lesions. Third, we
found that loss of nuclear TRIM23 expression was associated with poor overall survival in humans (p=0.026). Finally, we identified and pursued a member of
the TRIM family, TRIM23, with both in vitro and in vivo models,
and have demonstrated that decreased expression of Trim23 leads
to increased tumor burden and decreased overall survival in a
mouse model of MPNST growth and metastasis.

Conclusions: Collectively, the ability to track the molecular

evolution of MPNST in two individuals with metastatic disease
offers new insights into the genetic events important for
metastatic progression and metastasis. Additionally we have
identified TRIM23 as a protein that may play a role in MPNST
progression. Current work is aimed at better understanding the
pathway by which TRIM23 affects MPNST progression.

Full List of Authors: Abigail Godec*1, John Chrisinger1, Madhurima

Kaushal1, Reyka Jayasinghe1, Hilary Dietz1, Xiao Zhang1, Angela Hirbe1
1
Washington University in St. Louis, St. Louis, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 91
 Defining the Role of Endoglin in Malignant Peripheral Nerve Sheath Tumor Progression

Teresa González Muñoz, Microenvironment and Metastasis Group, Molecular Oncology Program, Spanish National Cancer Research Centre (CNIO),
Madrid, Spain

Background: Malignant peripheral nerve sheath tumors (MPNSTs) are highly aggressive sarcomas with poor prognosis commonly related to neurofibromatosis
type 1 (NF1). Recent data demonstrate that tumor-microenvironment communication plays a crucial role in the progression of these tumors. While soluble
factors have been described as the main cell-cell communication mechanism in this crosstalk, the role of secreted exosomes (30-100nm extracellular
vesicles) in this scenario is completely unknown. Our previous data characterizing MPNST-derived exosomes highlighted endoglin, a TGF-β co-receptor with
an important role in angiogenesis, as one of the top candidates secreted by MPNST cells. Here, we aim to analyze the role of endoglin in MPNST progression
and to examine its potential use as a new biomarker and therapeutic target for these tumors.

Methods: Exosomes from plasma of NF1 patients in different stages were isolated by ultracentrifugation methods. Endoglin levels were tested in plasma
circulating exosomes by ELISA and in human peripheral nerve sheath tumors by immunohistochemistry. Endoglin knockdown was performed in the MPNST
cell line STS26T and its influence on gene expression and signaling pathways was analyzed by RNA-Seq and validated by qRT-PCR and Western blot.
Moreover, the anti-human and -mouse endoglin antibodies, TRC105 and M1043, respectively, were tested in vivo.

Results: Endoglin levels were significantly increased in plasma-circulating exosomes and in peripheral nerve sheath tumors along the progression of the
disease. Mechanistically, endoglin knockdown resulted in the downregulation of the BMP and MAPK/ERK signaling and the impairment of several pathways
related to pro-angiogenic behavior in MPNST-derived cell lines. Endoglin knockdown also led to the reduction of primary tumor growth and metastasis in vivo.
Finally, our data show that blocking tumor and host endoglin function using anti-human and –mouse endoglin antibodies TRC105 and M1043, respectively,
significantly reduced MPNST tumor growth and metastasis in vivo.

Conclusions: Overall, our data suggest that both tumor and host-derived endoglin play an active role in MPNST progression and support the use of this protein
as a new promising biomarker and a potential target to block the progression of these tumors. Moreover, the analysis of exosomal endoglin levels could serve
as a novel readout of MPNST malignancy and treatment response using liquid biopsy.

Full List of Authors: Teresa González Muñoz*1, Angela Di Giannatale2, Claudia Savini1, Susana García-Silva1, Alberto Benito-Martin3, David Pisapa4, Nancy Ratner5, Kaleb Yohay6,
Charles P. Theuer7, Héctor Peinado1
1
Microenvironment and Metastasis Group, Molecular Oncology Program, Spanish National Cancer Research Centre (CNIO), Madrid, Spain, 2Deparment of Oncohematology,
Bambino Gesù Children's Hospital (IRCCS), Rome, Italy, 3Deparment of Pediatrics, Drukier Institute for Children's Health and Meyer Cancer Center, Weill Cornell Medical College,
4
Deparment of Neurosurgery, Weill Cornell Medicine, New York, 5Deparment of Pediatrics, Cancer and Biology and Neural Tumours Program, Cincinnati Children’s Hospital,
Cincinnati, OH, 6Deparment of Neurology, New York University Langone, New York, 7Tracon Pharmaceuticals, San Diego, United States

92 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Predicting In Vitro Drug Efficacy in Neurofibromatosis via Integration of Protein and Drug-Target Networks

Sara Gosline, Sage Bionetworks, Seattle, United States

Background: The development of diverse cell line models of NF1, NF2, and schwannomatosis-related tumors has expanded the ability to profile and model the
response of these tumors to various drugs. By coupling high throughput drug screening data with molecular data such as gene expression data or somatic
variants, computational approaches can be used to better guide drug development. Common approaches applied to large compendium of cancer cell lines and
drug screens - such as the pharmacogenomic screens of Cancer Cell Line Encyclopedia from the Broad Institute and the Cancer RX Gene dataset from the Sanger
Institute - identify statistically significant correlations between genomic features of a cell line and its resulting response to a compound of interest. However,
these techniques fail in cases where there are not hundreds of cell lines and drugs to compare and are therefore not applicable to rare diseases such as NF.

To overcome the paucity of samples in the modeling of NF cell line data we supplement cell-line-specific gene expression data with quantitative measurements
of drug-targeting experiments and protein-protein interaction data to build a network model of drug efficacy.

Methods: We collected high throughput drug screening experiments and gene expression data from 1883 published cell lines to test the ability to predict in
vitro drug efficacy. We used metaViper to identify master regulators that are differentially active between two sets cell lines and integrated these proteins with
published protein-protein interaction data (from STRING) and drug-target data derived from our previously-published Drug Target Explorer. We then use the
Prize Collecting Steiner Forest (PCSF) algorithm to identify the specific protein-protein interactions that link the master regulators to drugs.

Results: We applied this approach to both cancer cell lines and plexiform neurofibroma cell lines. Preliminary results suggest that this approach can not only
identify previously tested drugs that can give rise to a differential in vitro response but also identify novel, untested drugs.

Conclusions: While still in development, network integration and subsequent reduction provides a flexible way to characterize the diverse changes that occur
across a set of cells using gene expression data. Moving forward we aim to study the robustness of this network approach and apply it to other NF cell line models.

Full List of Authors: Sara Gosline*1, Robert Allaway1, Justin Guinney1

1
Sage Bionetworks, Seattle, United States

A Pan-NF Knowledge Portal

Sara Gosline, Sage Bionetworks, Seattle

Background: Recent developments in neurofibromatosis (NF) research have led to a deluge of data derived from diverse experimental assays including cellular
imaging, drug dosage experiments, clinical trials, patient surveys and high-throughput approaches such as automated drug screens, DNA sequencing, and gene
expression measurements. While some data are being shared via publicly funded data repositories such as the Gene Expression Omnibus (https://www.ncbi.nlm.
nih.gov/geo/), many datasets remain unpublished or hidden in supplemental files of respective publications. The lack of a central, well-curated, and harmonized
data repository for NF data makes it difficult to integrate the results of NF studies across the community, ultimately impeding the progress of NF research.

Here we introduce a pan-NF knowledge portal that includes all work funded by Children’s Tumor Foundation (CTF) and Neurofibromatosis Therapeutic
Acceleration Program (NTAP) for the past three years. To supplement this work we have also collated other published NF datasets and described ongoing
experiments that will be released in the future. This portal represents a milestone in a collaboration between CTF, NTAP and Sage Bionetworks to collect and
curate research funded by the respective organizations.

Methods: Projects supported by CTF and NTAP uploaded results of their experiments by pre-defined milestones to the Synapse collaborative platform (www.
synapse.org). These data were then annotated with a pre-defined metadata dictionary aligned with industry standards to enable facile searching and indexing. This
dictionary was then used to build an elastic search-powered web portal to enable users to query and sort the data, view response summaries, and access data.

Results: This web portal is currently under development and will launch November 1, 2018 at http://nf.synapse.org with the ability to browse ongoing projects
funded across the NF community, as well as data from completed projects. The knowledge portal will be open to all for new submissions for any interested NF
researchers.

Conclusions: As more data becomes publicly available the pan-NF knowledge portal will serve as a central gathering area for scientists and clinicians
interested in studying the disease. Ongoing work includes expanding the knowledge portal to incorporate new types of data as well as building interactive tools
to explore specific data sets.

Full List of Authors: Sara Gosline*1, Robert Allaway1, Salvatore La Rosa2, Sharad Verma3, Justin Guinney1, Jaishri Blakeley3, Annette Bakker2
1
Sage Bionetworks, Seattle, 2Children's Tumor Foundation, New York, 3Neurofibromatosis Therapeutic Acceleration Program, Baltimore, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 93
 Identification of Key Candidate Genes and Pathways in NF1 by Integrated Bioinformatical Analysis

Chengrui Guo, Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of
Medicine, Shanghai, China

Background: Neurofibromatosis type 1 (NF1) is one of the most frequently hereditary diseases in the world. This study aimed to identified some differentially
expressed genes(DEGs) and the involved pathway in NF1 by bioinformatics analysis.

Methods: Four datasets (GSE14038, GSE41747, GSE60082 and GSE1482) of expression microarray were reanalyzed after downloaded from GEO database.
Differentially expressed genes (DEGs) were obtained by the 'limma' data packet in 'R' software. Gene ontology (GO) and Kyoto Encyclopedia of Genes and
Genomes (KEGG) pathway enrichment were analyzed on the DAVID database, and a DEG-associated protein-protein interaction (PPI) network was constructed
using STRING and Cytoscape software.

Results: A total of 147 DEGs (including 87 up and 58 down-regulated genes) were identified. GO enrichment analysis showed that those DEGS were mainly
enriched in 'regulation of adiponectin secretion' and 'regulation of chemokine biosynthetic process'. KEGG pathway enrichment analysis revealed that the DEGs
were mainly enriched in 'Transcriptional misregulation in cancer', 'Toll-like receptor signaling pathway', Based on the PPI network, MFAP5, EFEMP1, ADH1B,
NEGR1, COL3A1, TAC1, COL28A1, SYNPO2 were screened as hub genes.

Conclusions: This bioinformatics analyses identified some key genes which promote the development of NF1. Especially, some key DEGs like MFAP5 and
EFEMP1 could be used as a potential biomarker in diagnosis for NF1.

Full List of Authors: Chengrui Guo*1, Xiaojie Hu1

1
Department of Plastic and Reconstructive Surgery, Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China

94 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Phase 0 Trial Investigating the Intra-Tumoural Concentration and Activity of Sorafenib in Neurofibromatosis
Type II

Oliver Hanemann, Institute of Translational and Stratified Medicine, University of Plymouth, Plymouth

Background: Drug therapies for Neurofibromatosis type 2 (NF2) patients, suffering from schwannomas, meningiomas and ependymomas, are needed. The
aim for this study was to establish the pharmacokinetics (PK) and pharmacodynamics (PD) of the PDGFR/Raf inhibitor Sorafenib (Nexavar) in tissue samples
from NF2 patients (primary endpoint), as well as to test potential blood biomarkers.

Methods: In this phase 0 trial five NF2 patients received 400mg Sorafenib twice daily (Maximum Tolerated Dose, MTD) for 11 days in-between two tumour
biopsies. PK was measured using liquid chromatography-tandem mass spectrometry (LC-MS/MS) and PD by western blotting and immunohistochemistry in
peripheral schwannomas (PS) and blood.

Results: Sorafenib was detected in plasma (3316.9 ng/ml-20792.2 ng/ml), and in peripheral schwannoma (PS) samples (1425.89ng/g-6242.06ng/g) from
all patients. In respect to PD western blotting on PS samples demonstrated non-significant changes of pPDGFR, pAKT, cyclin D1, and cleaved caspase 3 in all
patients; non-significant changes of pERK in four patients and significant reduction in one patient (1/5); non-significant changes of ps6 in three patient (3/5)
and significant reduction in two (2/5). All results from PBMC; pPDGR, pERK/ERK, pAKT/AKT were not significant. Immunohistochemistry results were in line
with above. Side effects were mostly mild to moderate. A rash in one patient was classed as severe.

Conclusions: In this study we confirmed the usefulness of Phase 0 trials for drug development in NF2 patients. The trial was, however, negative for the primary
outcome and patients experienced several side effects.

Full List of Authors: Oliver Hanemann*1, Gareth Evans2, Sylwia Ammoun1, Adam Streeter1, David Hilton3, Christopher Hayword1 and Plymouth University Peninsula Schools of
Medicine and Dentistry
1
Institute of Translational and Stratified Medicine, University of Plymouth, Plymouth, 2Regional Genetic Service, St Mary's Hospital, Manchester, 3Histopathology, University Hospitals
Plymouth NHS Trust, Plymouth, United Kingdom

Genetic Ablation of Pak1 Extends Lifespan and Reduces the Size of Tumor Bearing Tissue in a Genetically
Engineered Mouse Model of Neurofibromatosis Type 2 (NF2)

Eric Hawley, Biochemistry, Indiana University School of Medicine, Indianapolis

Background: Although it has been well established that Neurofibromatosis Type II (NF2) is caused by loss of heterozygosity of the NF2 gene, which encodes
the protein Merlin, relatively little is known about how Merlin functions as a tumor suppressor. One of the few known functions of Merlin is as a negative
regulator of the group A serine/threonine p21 activated kinase, PAK1. PAK1 is a known oncogene which serves as a critical signaling node regulating cell
proliferation, evasion of apoptosis, and DNA damage repair and is commonly amplified in a variety of human malignancies including up to a third of breast
cancers. PAK1 has significantly increased basal activity in Merlin deficient Schwann cells and we hypothesized that Merlin is a critical negative regulator of
PAK1 in Schwann cells and that loss of Merlin leads constitutive activation of PAK1, which in turn drives oncogenic transformation and tumor growth. We
therefore wanted to test whether or not genetic disruption of Pak1 would be protective against the prototypic pathologies observed in our genetically engineered
mouse model of NF2, Nf2flox/flox; Postn-Cre (NF2 GEMM).

Methods: In IACUC approved studies, we crossed our NF2 GEMM animals with Pak1 deficient animals and observed the mice over the period of 15 months to
assess for the development of sensorineural hearing loss, tumor formation, and overall survival. We also assessed the potency and growth inhibitory effects of
a new PAK1 small molecule inhibitor in a murine Nf2 deficient schwannoma cell line.

Results: In the context of our NF2 GEMM, mice genetically deficient in Pak1 but not Pak2 have significantly prolonged overall survival and are significantly
protected from sensorineural hearing loss with a reduction in the size of tumor bearing tissue. The PAK1 small molecule inhibitor NVS-PAK1-1 potently inhibits
PAK1 activation at submicromolar concentrations in vitro and reduces the proliferation of an Nf2 deficient schwannoma cell line.

Conclusions: Genetic deletion of Pak1 but not Pak2 extends the average lifespan and reduces the size spinal dorsal root ganglia in our NF2 GEMM. Total
deletion of Pak1 was well tolerated in Cre-negative animals. The reduced tumor burden and longer lifespan observed in these mice supports the further
development of the PAK1 specific inhibitor NVS-PAK1-1 for preclinical and clinical trials.

Full List of Authors: Eric Hawley*1, Su-Jung Park1, Li Jiang2, Jonathan Chernoff3, D. Wade Clapp2
1
Biochemistry, 2Pediatrics, Indiana University School of Medicine, Indianapolis, 3Fox Chase Cancer Center, Philadelphia, United States

Disclosure of Interest: E. Hawley has a conflict with: Funding for this work was provided by the US Department of Defense through the NFRP., S.-J. Park: None Declared, L. Jiang:
None Declared, J. Chernoff: None Declared, D. W. Clapp: None Declared

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 95
 The Tumor Suppressor Protein Merlin Mediates Macrophage Activation and Myelin Phagocytosis

Dario-Lucas Helbing, Morrison Laboratory

Background: It has been shown that macrophage infiltration into NF2-associated schwannoma is commonly observed and that this inflammatory
microenvironment has an impact on tumor development and progression. Because both Schwann cells and Macrophage acitivities for example cytokine
production and myelin clearance are critical for proper nerve de-and regeneration processes, we wanted to test whether the merlin (Nf2) protein also plays a
role in control of macrophage function in the context of the peripheral nervous system [PNS] function and dysfunction.

Methods: We used the mouse macrophage cell line RAW 264.7 and primary mouse peritoneal macrophages as an in vitro model to study the effect of
macrophage polarization on merlin protein level. Consequently we knocked down merlin to investigate the cytokine profile of merlin deficient macrophages via
qPCR. Furthermore we used sciatic nerve explants as a described model system of ex vivo Wallerian degeneration to address the role of Schwann cell intrinsic
merlin in myelin clearance after a nerve injury.

Results: We show that merlin protein level is downregulated in activated M1 macrophages. After merlin knockdown the proliferation of RAW264.7
macrophages is elevated while we did not observe any significant changes in cytokine production per se. We show that merlin deficient macrophages
react stronger to proinflammatory stimuli in terms of a higher expression of proinflammatory cytokines. In a myelin phagocytosis assay merlin deficient
macrophages displayed elevated engulfment of myelin as well als increased lysosomal abundance suggesting enhanced myelin phagocytosis and processing.
qPCR analysis revealed that merlin deficiency leads to increased mRNA levels of proinflammatory cytokines after myelin stimulation. In agreement with the
observed merlin deficient macrophage phenotype, an ex vivo nerve degeneration assay, carrying a Schwann cell specific merlin knockout also exhibited
enhanced clearance of myelin.

Conclusions: Our results show that merlin can regulate the activation of macrophages and therefore cytokine production in these innate immune cells. We
furthermore found that merlin can regulate phagocytosis of myelin in macrophages and Schwann cells which is of high interest for understanding peripheral
nerve de- and regeneration and maintenance as well as schwannoma development. Our data imply that manipulation of phagocytosis and autophagocytosis
might be a possible therapeutic opportunity for schwannoma which should be tested in future studies.

Full List of Authors: Dario-Lucas Helbing*1, Alexander Schulz1, Michael Reuter1, Joanna Kirkpatrick2, Lars Björn Riecken1, Reinhard Bauer3, Helen Morrison1
1
Morrison Laboratory, 2Core Facility Proteomics, Leibniz Institute on Aging, 3Institute of Molecular Cell Biology, Jena, Germany

Merlin Regulates p53 and YAP via Direct Binding to the Tumor Suppressor ASPP2

Robert Hennigan, Experimental Hematology and Cancer Research, Cincinnati Children's Hospital Medical Center, Cincinnati, United States

Background: Schwannomas are benign peripheral nerve tumors that are characterized by biallelic inactivation of the NF2 tumor suppressor gene, resulting in loss
of function of the NF2 gene product, merlin. In Schwann cells in vitro, merlin mediates contact inhibition of growth, in part by activating the growth suppressive
HIPPO pathway. However, there is no consensus on the mechanism by which merlin acts as a tumor suppressor. Merlin has no intrinsic catalytic activity; its
function is mediated by the proteins with which it interacts. Previously, we used a global proteomic screening technique, proximity biotinylation, to perform
a census of merlin interactions in living Schwann cells. We found that the merlin interactome predominately consists of components of cell junctions; focal
adhesions, adherens junctions and tight junctions, including actin binding proteins and associated signaling molecules. Cell junctions are now seen as signaling
complexes that regulate growth in response to intra- and extracellular mechanical forces

Methods: We designed a powerful set of protein interaction techniques to explore the functional relationships among the proteins defined by our interactome
analysis and to fully define the role that merlin plays in mechanosensing

Results: We identified a set of proteins that bind merlin directly and mediate signaling in response to mechanical forces. This includes ASPP2, a tumor suppressor
that interacts with a range of oncogenic signal transduction molecules; including p53, ras, NF-kB and YAP. We have found that merlin inhibits ASPP2 transcriptional
regulation of both p53 and YAP.

Conclusions: Our data suggest that the merlin-ASPP2 interaction plays a critical role in multiple oncogenic signaling pathways. Loss of merlin in NF2 results in
disruption of ASPP2s ability to regulate a key node in onco-regulatory signaling network, causing the dysregulation of critical signaling systems for p53, YAP and
Ras that contribute to tumor growth and progression.

Full List of Authors: Robert Hennigan*1, Nancy Ratner1, Nancy Ratner1, Nancy Ratner1, Nancy Ratner1
1
Experimental Hematology and Cancer Research, Cincinnati Children's Hospital Medical Center, Cincinnati, United States

96 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Development of a Preclinical NF1-MPNST Platform Suitable for Precision Oncology Drug Discovery and Evaluation

Angela Hirbe, Medical Oncology, Washington University in St. Louis, St. Louis, United States

Background: One of the most common malignancies affecting adults with neurofibromatosis type 1 (NF1) are malignant peripheral nerve sheath tumors
(MPNSTs). 8-13% of individuals with NF1 will develop MPNSTs during young adulthood. There are few therapeutic options, and the vast majority of people
with these cancers will die within 5 years of diagnosis. Despite advances in our understanding of the pathobiology of these tumors and the identification of
seemingly-promising therapeutic targets using a single mouse MPNST model system (NPCis mice and derivative cell lines), no investigational agents have
demonstrated efficacy following translation to human clinical trials. We hypothesize that one of the major reasons that rationally-chosen drugs underperform
in human clinical trials is that the preclinical models used to discover and evaluate promising therapeutic agents do not capture the genetic heterogeneity of
the human cancer. To optimize clinical translation, we aim to develop a collection of MPNST Patient Derived Xenografts (MPNST-PDXs) for the NF community,
which more fully represent the spectrum of genetic heterogeneity seen in the human condition.

Methods: PDX lines were generated from 5 different NF1 patients in each case by implanting a small piece of tumor on the dorsal surface of a nude mouse.
These tumors were then serially passaged in vivo. Whole exome sequencing, RNA sequencing, and copy number analysis was performed on the PDX lines,
and compared to their parental MPNSTs. Blood samples were obtained as germline DNA controls.

Results: To date, we have generated 5 clinically annotated NF1-MPNST PDX lines that we have histologically and genomically characterized. These lines have also
been labeled with GFP-luciferase to allow for the monitoring of tumor burden in live animals in preclinical drug studies. We are in the process of creating 5 more lines.

Conclusions: A collection of histologically and genomically characterized NF1-MPNST PDX lines are now available for pre-clinical testing. Future work with
continue to expand this collection.

Full List of Authors: Angela Hirbe*1, Xiao Zhang1, Abigail Godec1, Tiandao Li2
1
Medical Oncology, 2McDonnel Genome Institute, Washington University in St. Louis, St. Louis, United States

This work was funded by the Francis Collins Scholar Award through NTAP and a pre-clinical grant through the NF Research Initiative (NFRI) at Boston Children's Hospital.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 97
 NF1 Minipigs Exhibit Spontaneous Loss of Heterozygosity and Attain Clinically Relevant Plasma
Concentrations of MEK Inhibitors

Sara H. Isakson, University of Minnesota, Minneapolis

Background: Neurofibromatosis Type I (NF1) is a genetic disease characterized by changes in skin pigmentation and development of nervous system tumors
throughout the body. While symptoms vary widely between individuals, nearly all patients develop benign neurofibromas (NFs) and café au lait macules
(CALMs). Traditional mouse models provide some insight into the pathogenesis of NF1, but do not accurately model the spectrum of disease seen in human
patients. We have developed a porcine model of NF1 that exhibits hallmarks of the syndrome, including NFs and CALMs.

Methods: Using transcription activator-like effector nucleases (TALENs) and homology-dependent repair (HDR) constructs, we generated NF1+/- (NF1) minipigs
harboring a premature termination codon found recurrently in human patients. A cohort of NF1 minipigs was examined for phenotypes associated with NF1
syndrome, and cells were isolated from lesions for genetic and biochemical analysis in vitro. A separate cohort of juvenile WT and NF1 littermates was enrolled
in pharmacological studies to evaluate the potential differences in drug metabolism and target inhibition that may be present in NF1 patients. We evaluated
pharmacokinetics and pharmacodynamics of MEK inhibitors that have shown promise in patients with NF1-associated nervous system tumors.

Results: We have observed full penetrance of CALMs, a phenotype that has not been demonstrated in other animal models. A subset of NF1 minipigs also
develop NFs that closely resemble those seen in human patients. Primary Schwann cells cultured from NFs and melanocytes cultured from CALMs exhibit
spontaneous loss of heterozygosity of the remaining wild type allele, with variable corresponding levels of Ras hyperactivity. A single oral dose of Selumetinib
resulted in clinically relevant plasma concentrations in all minipigs. Moreover, pharmacodynamic analysis of MEK inhibition in peripheral blood mononuclear
cells revealed significant suppression of ERK phosphorylation, a downstream effector in the Ras/MAPK pathway. Characterization of NF1 minipig tumors and
pharmacologic analysis of NF1-targeted therapies is ongoing.

Conclusions: We have shown that NF1 minipigs develop CALMs and NFs with spontaneous LOH and can be dosed successfully with MEK inhibitors. Our results
suggest that NF1 minipigs and the cell lines generated from their tissues will be useful in answering prevailing questions in the field of NF1 and may facilitate
development of new therapies for NF1-associated nervous system tumors in humans.

Full List of Authors: Sara H. Isakson*1, Tony Rizzardi2, Alex Coutts2, Daniel Carlson2, Mark Kirstein1, James Fisher1, Jeremie Vitte3, Nancy Ratner4, Scott Fahrenkrug2, Marco
Giovannini3, Christopher Moertel1, David Largaespada1, Adrienne Watson2
1
University of Minnesota, Minneapolis, 2Recombinetics, St. Paul, 3University of California, Los Angeles, Los Angeles, 4Cincinnati Children’s Hospital, Cincinnati, United States

Disclosure of Interest: S. Isakson: None Declared, T. Rizzardi: None Declared, A. Coutts: None Declared, D. Carlson: None Declared, M. Kirstein: None Declared, J. Fisher: None
Declared, J. Vitte: None Declared, N. Ratner: None Declared, S. Fahrenkrug: None Declared, M. Giovannini: None Declared, C. Moertel: None Declared, D. Largaespada has a
conflict with: DL is the co-founder and co-owner of several biotechnology companies, specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc. (recently acquired by
Immunsoft, Inc.), and B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc., (a subsidiary of Recombinetics, Inc.) and Genentech, Inc. is funding some of his research.
This abstract describes collaborative research between Surrogen and The University of Minnesota. This work was supported by the Children's Tumor Foundation Synodos for NF1
Award, the National Institutes of Health under Award Number T32OD0100993 (SHI), the American Cancer Society Research Professor Award (DAL), and the Children's Cancer
Research Fund., A. Watson: None Declared

98 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Muscle Stem Cell Intrinsic Role of Merlin in NF2 Associated Muscle Atrophy

Marie Juliane Jung, Research Group von Maltzahn - Stem cells in Regeneration of Skeletal Muscle

Background: NF2 patients often present a slow but progressive distal muscle atrophy and paresis in later stages of the disease. In a previous study we showed
that Merlin-deficient neurons after sciatic nerve crush injury showed a proper re-innervation concomitant with no difference in muscle weight and a regular
architecture of muscle fibers. Therefore we started to investigate the role of Merlin in muscle stem cells (MSCs) after injury and during in vitro myogenesis.

Methods: Murine MSCs were isolated via FACS and microarray analysis was performed to evaluate Merlin expression during aging. Furthermore quantitative
real time-PCR of a differentiation time course of primary MSCs derived myoblasts was performed to assess Merlin expression during in vitro differentiation.
We also examined the effect of Merlin levels on differentiation and proliferation of myoblasts by using ectopic overexpression and siRNA mediated knockdown.
Additionally we performed immuno blot and Immunofluorescence analysis to assess Merlin’s effect on crucial signaling pathways involved in myogenesis. To
further investigate the functional relevance of Merlin in MSCs during regeneration of the skeletal muscle (SkM) in vivo we injured the tibialis anterior (TA) muscle
of Pax7-creER;Nf2flox/flox with cardiotoxin. The regeneration of the TA was analyzed 7 days post injury (dpi). Additionally MSC numbers were quantified using the
MSC marker Pax7.

Results: Strikingly only Merlin isoform 2 expression increased during differentiation of myoblasts. Knockdown of Merlin expression during myogenic
differentiation resulted in changed fusion index and myotube diameter. Moreover we could show that loss of Merlin results in deregulation of a major signaling
pathway during in vitro myogenesis. To further investigate the functional relevance of Merlin in MSCs in vivo investigated the regeneration of the TA muscle of
Pax7-creER;Nf2flox/flox mice after injury. Loss of Merlin expression in adult MSCs affected the regeneration and MSC numbers at 7 dpi.

Conclusions: This is the first study that could show that Merlin isoform 2 plays a crucial role in MSCs and during myogenic differentiation. Furthermore our study
suggests that Merlin is important for the regeneration of the SkM after injury and MSC function. Moreover we could already identify one crucial signaling hub
involved in MSC activation and maintenance of muscle mass, to be regulated by Merlin in MSCs. This pathway could be one potential target for the treatment of
muscle atrophy in NF2 patients.

Full List of Authors: Marie Juliane Jung*1, Dario-Lucas Helbing2, Helen Morrison2, Julia von Maltzahn1
1
Research Group von Maltzahn - Stem cells in Regeneration of Skeletal Muscle, 2Research Group Morrison - Nerve Regeneration, Leibniz Institute of Aging-Fritz Lipmann Insitute,
Jena, Germany

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 99
 Targeting the Hyaluronan-Rich Peripheral Nerve Sheath Tumor Microenvironment to Improve Drug Efficacy
and Delivery

Bryant J. Keller, University of Minnesota – Twin Cities, Minneapolis

Background: Malignant Peripheral Nerve Sheath Tumors are aggressive soft tissue sarcomas that manifest at a high rate in individuals with the genetic cancer
predisposition syndrome Neurofibromatosis Type 1. Although many therapeutic avenues have been explored, little improvement has been seen in the poor
prognosis. Surgical resection and nonspecific chemotherapeutics remain the only standard of care. Evidence is mounting that a desmoplastic reaction involving
high deposition of the glycosaminoglycan Hyaluronic Acid (HA) in the Extra-Cellular Matrix (ECM) can lead to poor drug penetrance and efficacy. Human MRI
examples display large regions of poor uptake of contrasting agent, and a human tissue microarray shows 100% positivity of HA deposition through all samples,
including plexiform neurofibromas. Breaking down this physical barrier is a promising potential avenue to improve drug penetrance, perfusion, and efficacy.

Methods: To accurately model high grade Peripheral Nerve Sheath tumors (PNST), we implemented our previously described genetically engineered mouse
(GEM)-PNST consisting of Dhh-Cre; Nf1fl/fl; PTEN fl/fl (Keng V., et al., Cancer Research. 2012). This model manifests multi-focal, 100% penetrant peripheral
nerve sheath tumors with a median lifespan of 18 days post birth. These tumors also display elevated HA levels in the ECM as well as collapsed and sparse
vasculature. This model is amenable for treatment with a human pegylated hyaluronidase (PEGPH20) from Halozyme (San Diego, CA) to deplete HA and improve
drug delivery.

Results: The GEM-PNST model, when treated with PEGPH20, displayed a dose-dependent decrease in tumor HA levels with no obvious toxicities to the animals.
Taking advantage of the natural fluorescence of the broad-spectrum chemotherapeutic doxorubicin, sections cut from animals that had received PEGPH20
treatment before doxorubicin showed a quantifiable increase in perfusion. Furthermore, CD31 staining suggests an increase in blood vessel patency post-
PEGPH20 treatment. Dual treatment of PEGPH20 with doxorubicin slightly improved longevity of these animals as compared to the monotherapy. Preliminary
results also show an exciting increase of life when PEGPH20 is combined with the targeted MEK inhibitor PD0325901.

Conclusions: PEGPH20 shows promising therapeutic benefit in targeting physical barriers in MPNSTs. Improved drug delivery and efficacy will open avenues to
further drug combinations in this currently incurable malignancy.

Full List of Authors: Bryant J. Keller*1, Adrienne Watson2, Kyle Williams1, Stephen Scully1, Rory Williams1, Leah Anderson1, Justin Knight1, Colleen Forster3, Kwangmin Choi4,
Marjorie Carlson1, Nancy Ratner4, Paolo Provenzano1, David Largaespada1
1
University of Minnesota – Twin Cities, Minneapolis, 2Recombinetics Inc., St. Paul, 3BioNet, Minneapolis, 4Cincinnati Children's Hospital, Cincinnati, United States

Disclosure of Interest: B. Keller: None Declared, A. Watson: None Declared, K. Williams: None Declared, S. Scully: None Declared, R. Williams: None Declared, L. Anderson: None
Declared, J. Knight: None Declared, C. Forster: None Declared, K. Choi: None Declared, M. Carlson: None Declared, N. Ratner: None Declared, P. Provenzano: None Declared, D.
Largaespada has a conflict with: DL is the co-founder and co-owner of several biotechnology companies, specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc.
(recently acquired by Immunsoft, Inc.), and B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc., and Genentech, Inc. is funding some of his research. The business of
all these companies is unrelated to the contents of this abstract. Other authors have no conflict of interest to disclose. Supported by W81XWH-15-1-0114 DOD Grant awarded to
D.A.L and P.P.P

100 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Synergism between Topo I and mTOR inhibitors in Malignant Peripheral Nerve Sheath Tumors Affecting NF1
Patients

Dong Hyuk Ki, Pediatric Oncology, Dana-Farber Cancer Institute/ Harvard Medical School, Boston, United States

Background: Malignant peripheral nerve sheath tumors (MPNSTs) are very aggressive and often metastatic soft tissue sarcomas, which are often found
in patients who have neurofibromatosis type 1 (NF1). Currently, surgical excision is the only curative therapy for MPNST, although many patients have
unresectable or metastatic tumors at diagnosis and the recurrence rate after surgery is high. Chemotherapy regimens are only partially effective and
associated with significant toxicity that can severely reduce the quality of life. Therefore, identification of new genetic dependencies and drugs with anti-tumor
activity will be critical to advance MPNST therapy.

Methods: To identify effective compounds, we assessed the response of primary NF1-mutant zebrafish MPNST cells that had been implanted into embryos
as a novel system to assay drug activity in vivo. Primary MPNSTs were harvested from sox10:mCherry; nf1a+/-;nf1b-/-;p53m/m zebrafish and the cells
were mechanically dispersed. Approximately 100-120 MPNST cells were implanted into the pericardial cavity of embryos at 2 days post-fertilization (dpf).
Twenty-four hours after implantation, the fluorescent tumor cross-sectional area was imaged. The embryos were arrayed in 96-well plates and were incubated
for 4 days in either vehicle or each individual drug. Quantitative assessment of the cross-sectional area of remaining fluorescent tumor cells was performed at
7 dpf and the fish were raised in the absence of drug and monitored to assess the durability of the response. The compounds showing drug response in the
embryonic implantation assay were evaluated in human MPNST cells.

Results: After testing a series of drugs, we identified inhibitors of topoisomerase I and mTOR as the most effective single agents in inducing apoptosis in
MPNST cells. Furthermore we show that the topoisomerase I inhibitor irinotecan and the mTOR kinase inhibitor AZD2014 act synergistically induce cell death
in human NF1-mutant MPNST cell lines by isobologram analysis. Mechanistic studies implicate mTOR inhibition and DNA damage induced by inhibiting
topoisomerase I in synergistically blocking phosphorylation of 4E-BP1, leading to arrest of protein synthesis and tumor cell death.

Conclusions: Co-inhibition of topoisomerase I and mTOR induces synergistic cell death in NF1-associated MPNSTs. Our study provides compelling preclinical
evidence that this drug combination is highly active against MPNST at dosages tolerated by normal tissues in vivo.

Full List of Authors: Dong Hyuk Ki*1, Felix Oppel1, Adam Durbin1, A Thomas Look1
1
Pediatric Oncology, Dana-Farber Cancer Institute/ Harvard Medical School, Boston, United States

Genome-Wide Mapping Reveals New Nuclear Functions for YAP in Schwann Cells

Joseph Kissil, Molecular Medicine, The Scripps Research Institute, Jupiter, United States

Background: The Hippo pathway regulates cell number and organ size during development through the control of the YAP transcriptional regulator. Significantly,
the pathway has emerged as a major driver of tumorigenesis in many human cancers. We have previously shown that YAP function is required for NF2-null
Schwann cell survival, proliferation, and tumor growth in vivo. Upon extracellular stimuli such as cell-cell contact, the Hippo pathway negatively regulates YAP
through cytoplasmic sequestration. At low cell density, YAP is nuclear and functions as a transcriptional activator of genes involved in cell proliferation and
survival.

Methods: Using genome-wide approaches such as ChIP-seq.

Results: We identified that in addition to its role as an activator, YAP can act as a transcriptional repressor through recruitment of EZH2, a member of the
Polycomb repressive complex (PRC2). YAP co-localizes with EZH2 on the genome to transcriptionally repress a broad network of genes mediating a host of
cellular functions, including a number of central regulators of cell cycle progression and contact inhibition of cell proliferation.

Conclusions: This work unveils a broad and underappreciated aspect of YAP function as a transcriptional repressor and suggests that Ezh2 inhibition might be
a viable therapeutic approach in the treatment of NF2.

Full List of Authors: Joseph Kissil*1, Sany Hoxha1

1
Molecular Medicine, The Scripps Research Institute, Jupiter, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 101
 Targeting a Novel RABL6A-RB1 Pathway Suppresses MPNST Pathogenesis

Jordan Kohlmeyer, Pharmacology, University of Iowa, Iowa City, United States

Background: Malignant peripheral nerve sheath tumors (MPNSTs) are deadly sarcomas that arise in ~10% of patients with neurofibromatosis type I (NF1). There
are no effective chemotherapies for MPNSTs, which are the leading cause of death in NF1 patients. In most MPNSTs, the retinoblastoma (RB1) tumor suppressor
pathway is inactivated by hyperactivation of CDK4/6 kinases, commonly through loss of cell cycle inhibitors such as p27. RABL6A is an oncogenic GTPase that
inhibits RB1 signaling, but its role in MPNST has not been studied. We hypothesized that RABL6A drives MPNST pathogenesis through inactivation of RB1.

Methods: RABL6A and p27 expression in human MPNSTs and cell lines were determined by immunohistochemistry (IHC) and western blotting. RNAi was
used to silence RABL6A in MPNST and normal human Schwann (NHSC) cells. Effects of RABL6A loss on MPNST biology and RB1 signaling were measured
by cell proliferation/survival assays, western blotting, and orthotopic xenograft tumor studies in mice. Dose response curves and in vivo drug treatments
assessed effects of RABL6A expression on MPNST responses to RB1 targeted drugs.

Results: Tissue microarray analyses showed dramatic upregulation of RABL6A coincident with p27 loss in human MPNSTs compared to patient-matched
plexiform neurofibromas. Likewise, RABL6A was upregulated in human MPNST lines compared to primary NHSCs. In vitro cell-based assays showed that
RABL6A is essential for MPNST cell proliferation and survival. Loss of RABL6A caused significant MPNST cell death and G1 phase arrest concurrent with
p27 upregulation and accumulation of active, hypo-phosphorylated RB1. Conversely, RABL6A overexpression enhanced MPNST cell proliferation and RB1
phosphorylation. We tested if MPNST viability would be reduced by drugs targeting the RB1 pathway and found that a selective inhibitor of CDK4/6 kinases,
palbociclib (PD0332991), killed MPNST cells in a RABL6A-dependent manner. Ongoing mouse studies are evaluating the in vivo efficacy of palbociclib against
MPNST tumors that express varying levels of RABL6A.

Conclusions: Our data uncovered RABL6A as a new oncogenic driver of MPNST proliferation and survival. RABL6A activity promoted p27 downregulation,
inactivation of RB1, and increased response to RB1 targeted drugs. These findings establish a critical role for RABL6A in MPNST pathogenesis and identify
RABL6A-RB1 signaling as a novel, clinically relevant target for MPNST therapy using FDA-approved CDK4/6 inhibitors.

Full List of Authors: Jordan Kohlmeyer*1, Courtney Kaemmer2, Allison Moreno Samayoa2, Chandra Maharjan1, Vickie Knepper-Adrian2, David Gordon2, Rebecca Dodd2, Benjamin
Darbro2, Munir Tanas2, Dawn Quelle2
1
Pharmacology, 2University of Iowa, Iowa City, United States

Delivering on the Vision of Bench to Bedside: A Funding Community Effort to Develop Effective Therapies for
Neurofibromatosis Type 1 Tumors

Salvatore La Rosa, Children's Tumor Foundation, New York

Background: In recent years drug discovery research in rare diseases has proved to be a valuable strategy for companies to enrich their pipeline and
diversify their portfolio of therapeutic areas. This has been sustained in large part by the opportunities created by regulators for additional revenue and market
exclusivity. However, much of the funding for rare disease research is still provided by the not-for-profit organizations, including federal agencies and private
medical foundations. The current analysis showcases the collaboration between funders of neurofibromatosis type 1 (NF1) research to develop effective
therapies for NF1-associated tumors. The authors use the example of the investigation of MEK (mitogen-activated protein kinase kinase) inhibitors for NF1-
associated tumors as a case study to highlight the unique research and funding landscape that contributed to the ultimate launch of several promising clinical
trials for various NF1 tumors leading to the results of the SPRINT study involving the use of MEK inhibitor selumetinib in inoperable plexiform neurofibromas.
By presenting an analysis of the grants from the National Institutes of Health (NIH), the Department of Defense (DoD) Congressionally Directed Medical
Research Program (CDMRP), and major philanthropic organizations that have funded NF1-MEK research over the last 10 years, the authors highlight the value
of having a collaborative funding strategy between key partner funders to create a natural flow of research that has directly contributed to the development of
effective therapies for a rare tumor syndrome.

Full List of Authors: Vidya Browder1, Annette Bakker1, Jaishri Blakeley2, Sharad Verma2, Jill Morris3, Ling Wong3, Naba Bora4, Salvatore La Rosa*1 and Children's Tumor Foundation
1
Children's Tumor Foundation, New York, 2Johns Hopkins University, Baltimore, 3National Institute of Neurological Disorders and Stroke, Bethesda, 4Congressionally Directed Medical
Research Programs, Fort Detrick, United States

102 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Habituation in Children with Neurofibromatosis Type 1 Measured by Electroencephalography

Eve Lalancette, Neurosciences of Early Development Lab, CHU Sainte-Justine Research Center; Psychology, Université de Montréal

Background: The purpose of this research is to explore the electroencephalographic (EEG) markers of habituation in children with neurofibromatosis type 1
(NF1). Studies on animal models of NF1 have identified deficits in habituation, a simple form of learning that is conserved across species and crucial for the
development of higher cognitive functions (Larkin et al., 2010; Wolman et al., 2014). However, habituation deficits in humans with NF1 have not yet been
investigated. Repetition suppression (RS), the reduction of brain activity in response to the repeated presentations of a single stimulus, is considered the
neurophysiological equivalent of habituation and can be measured using EEG. In this study, we looked at RS in the time-frequency domain through the variation
of event-related spectral power. We hypothesized that habituation deficits would be seen in the NF1 group reflected by lower RS compared to controls.

Methods: A sample of 40 participants (20 NF1 and 20 controls) between 4 and 18 years of age will be recruited. To date, EEG recordings were performed on
8 participants with NF1 and 8 controls at CHU Sainte-Justine in a soundproof room with a 128-electrodes EEG system. The task presented was composed
of thirty pseudowords repeated six times each, which allowed us to observe auditory RS in a previous study using a similar design (Knoth et al., 2018).
Timeframes of 500 ms for each presentation of the pseudowords were analyzed separately in the time-frequency domain (Morlet decomposition). Repetition
effects were observed through the variation of spectral power between the first two presentations of the pseudowords at the left temporal region of interest for
each frequency band of interest (i.e. Theta, Alpha 1 and 2).

Results: Preliminary analyses were conducted using the data of 16 participants (8 NF1, 8 controls) between 4 and 16 years of age. Mixed-design ANOVA
(groups x repetitions) analyses revealed a tendency for an interaction between repetitions and groups (F(1, 14) = 3.60, p = 0.079) in the Alpha 1 band (8-
10Hz), suggesting that while the control group shows repetition suppression between the first and second presentation of a pseudoword, the NF1 group shows
repetition enhancement.

Conclusions: The study of repetition suppression alterations in NF1 could help us understand learning disabilities in NF1. Also, EEG markers of habituation
promise to be highly relevant translational measures of basic learning mechanisms for clinical trials aimed at learning disabilities in NF1.

Full List of Authors: Eve Lalancette*1, 2, Fanny Barlaam1, Audrey-Rose Charlebois1, 2, Kristian Agbogba1, 3, Jean-Marc Lina3, 4, Inga Sophia Knoth1, Sébastien Perreault5, Sarah Lippé1, 2
1
Neurosciences of Early Development Lab, CHU Sainte-Justine Research Center, 2Psychology, Université de Montréal, 3Electrical Engineering Department, École de Technologie
Supérieure, 4Center for Advanced Research on Sleep Medicine, Hôpital du Sacré-Coeur de Montréal , 5Division of Child Neurology, CHU Sainte-Justine, Montréal, Canada

Characterization of a Novel Patient Specific Neurofibromatosis Type I (NF1) Rat Model

Laura Lambert, Genetics, UAB, Birmingham, United States

Background: Although multiple NF1 mouse models exist, there is a need for additional preclinical models to test new therapeutics. As rats are a traditional
animal model for pharmacological, toxicological, and neurological studies, using this model organism to express NF1 patient alleles will allow examination of
the efficacy and safety of pharmacological modulators as well as cognition and behavior. Therefore, we created an NF1 rat model with the patient missense
allele c.3827G>A, p.R1276Q associated in humans with spinal NF1.

Methods: Two CRISPR guides and a dually compatible repair template were designed to target exon 28 of rat Nf1. Single-cell Sprague Dawley rat
embryos were injected with ribonucleoprotein complexes of guide RNA and Cas9 protein along with a single-strand DNA repair template and transferred to
pseudopregnant recipients. Pups were genotyped by tail biopsy ten days postpartum. Animals were euthanized upon tumor mass ulceration or inhibition of
normal movement. Tissues were collected for histology and evaluated by a board certified veterinary pathologist.

Results: Three of seven pups born were positive for CRISPR activity from which independent Nf1 knock-out (P1220fs*1223) and missense knock-in
(R1276Q) colonies were established. Unmated heterozygous Nf1 R1276Q female rats do not display any spontaneous tumors at least out to 5 months of age;
however, of four females that were mated, all rapidly developed tumors within two weeks of pregnancy. Heterozygous Nf1 P1220fs*1223 females also develop
tumors post-mating but, spontaneous tumors have also been observed in an unmated female. Homozygous mutant offspring have not been detected in litters
born from incrossing of heterozygous Nf1 mutant rats, although a homozygous mutant pup was detected at embryonic day 9.5 following Cesarean section.

Conclusions: Two novel mutant Nf1 alleles, patient mutation c.3827G>A, p.R1276Q and deletion c.3661_3674del, p.P1220fs*1223, have been generated
in the rat. Initial histological evaluation indicates that the mammary tumors are consistent with cribriform mammary gland adenocarcinoma. Restriction of
tumor development to pregnancy in Nf1 R1276Q females suggests hormone induction plays a major role in tumor development. The divergence in phenotype
between patient and null alleles may be due to residual function of R1276Q missense NF1 protein. Lack of full-term homozygous mutant pups indicates that
these alleles are embryonic lethal, although statistical significance has not yet been reached.

Full List of Authors: Laura Lambert*1, Deeann Wallis1, Mick Edmonds1, Bruce Korf1, Robert Kesterson1
1
Genetics, UAB, Birmingham, United States

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 103
 Generation of Patient-Specific Neurofibromatosis Type I (NF1) Rodent Models

Laura Lambert, Genetics, University of Alabama at Birmingham, Birmingham, United States

Background: Animal models are critical for preclinical studies of therapeutics. While previously limited to null and conditional knockout alleles, advances
in genetic engineering technologies have allowed the creation of personalized animal models of NF1. Models being developed represent different types of
mutations occurring in patients including nonsense, missense, insertions, deletions, splicing, and frameshifts. These models recapitulate missense mutations
associated with familial spinal NF (p.Gly848Arg and p.Arg1276Gln), as well as the less severe genotype-phenotype correlations associated with p.delM992
and p.Arg1809Cys.

Methods: Mouse and rat Nf1 genes are targeted by pronuclear microinjection or electroporation of CRISPR/Cas9 and Cpf1 reagents with repair templates to
generate founder animals. Both traditional targeting vectors and CRISPR reagents are used to modify mouse Nf1 in embryonic stem cells. Founder animals
and chimeras are outcrossed to establish germline transmission and independent colonies. Mutant alleles are tested for function by assessing viability when
homozygous, as well as tumor formation when placed in the proper genetic context.

Results: Mouse models developed to date include c.2041 C>T (p.Arg681X), c.2542G>C (p.Gly848Arg), c.2393_2408del16 and c.2919_2920insTT.
Additional model development is currently underway in the mouse including c.5425C>T (p.Arg1809Cys), c.1466A>G (p.Tyr489Cys), c.2970-2971delAAT
(p.delM992), c.2446 C>T (p.Arg816X), and c.499_502delTGTT. Rat models created to date include the c.3827G>A (p.Arg1276Gln) mutation and a 14bp
deletion model (c.3661_3674del).

Conclusions: Patient-specific NF1 alleles have been successfully generated in rodents using ES cell and CRISPR based approaches. The c.2041 C>T; p.Arg681X
and c.2393_2408del16 mouse models are embryonic lethal when homozygous, whereas the homozygous c.2542G>C; p.Gly848Arg mice are viable with no
gross phenotype despite reduced neurofibromin levels. Creating alleles from different mutation classes will allow in vivo testing of gene-based therapeutics
such as nonsense suppression, antisense-mediated splicing modulation, gene replacement via cDNA, or nuclease-mediated gene editing approaches.

Full List of Authors: Laura Lambert*1, Ke Hu1, Deeann Wallis1, Bruce Korf1, Robert Kesterson1
1
Genetics, University of Alabama at Birmingham, Birmingham, United States

The Role of Hippo Signalling in Merlin Null Schwannomas and Meningiomas

Liyam J. Laraba, PUPSMD, Plymouth University, Plymouth, United Kingdom

Background: Loss of the NF2 gene product Merlin is the most common cause of meningiomas and schwannomas. Merlin loss leads to increased activity of
the Hippo pathway co-transcriptional activators YAP/TAZ which drive tumour phenotypes in numerous cancers. Previous studies in schwannoma have shown
that Merlin loss leads to increased degradation of the Hippo pathway kinase LATS1/2 resulting in nuclear accumulation of YAP/TAZ. This study aims to identify
novel Hippo pathway targets in Merlin negative meningiomas and schwannomas.

Methods: Primary meningioma and schwannoma cells were cultured following surgical resection and informed patient consent. Cultured cells were subjected to
drug inhibition and lentivirus mediated knockdown of YAP or TAZ, and analysed for protein expression and proliferation by Western Blot and immunocytochemistry.
An RNA sequencing analysis was conducted on the sciatic nerves of Schwann cell conditional knockout mice that were either WT, NF2-/- or NF2-/-/YAP-/-.
Differentially regulated genes were subjected to validation.

Results: Merlin negative meningiomas display a greater proportion of nuclear YAP/TAZ compared to meningiomas with functional Merlin, and are also
more proliferative following Ki-67 analysis. Both chemical inhibition (verteporfin) and lentiviral knockdown of YAP or TAZ result in decreased proliferation of
meningioma cells. Verteporfin treatment did not alter the subcellular distribution of YAP/TAZ. Lentiviral knockdown of YAP/TAZ in schwannoma resulted in
a morphological change to a bipolar Schwann cell shape and this was coupled to a decrease in levels of phospho-FAK. RNA sequencing of WT and NF2-/-
Schwann cell conditional knockout mice identified differentially regulated genes and in the case of the NF2-/-/YAP-/- genotype genes can be identified which return
to expression levels of WT cells. This analysis provides a range of putative Hippo pathway dependent genes following loss of NF2 for validation.

Conclusions: This study has shown that Merlin negative meningiomas exhibit increased YAP/TAZ activity and as a result are more proliferative. Inhibition or
knockdown of YAP/TAZ can partially ameliorate the elevated proliferation seen in these tumours. In addition, differentially regulated genes following loss of
Merlin and YAP may reveal novel mechanisms by which Merlin negative meningiomas and schwannomas proliferate. Overall, Merlin/YAP dependent genes as
well as YAP/TAZ inhibition may lead to new therapies for Merlin negative meningioma and schwannoma.

Full List of Authors: Liyam J. Laraba*1, Emanuela Ercolano1, Claire Adams1, Vasilis Lenis1, Matthias Futschik1, David Hilton1, Sara Ferluga1, CO Hanemann1, David Parkinson1
1
PUPSMD, Plymouth University, Plymouth, United Kingdom

104 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Modeling an MPNST-Like State in Immortalized Human Schwann Cells for High-Throughput Drug Screening

Alex Larsson, Masonic Cancer Center, University of Minnesota, Minneapolis, United States

Background: Neurofibromatosis Type 1 (NF1) patients are predisposed to development of many types of cancers and aggressive benign tumors. The
most highly penetrant of these are tumors of Schwann cell in origin and located in the peripheral nervous system. These include plexiform neurofibromas
and atypical neurofibromas, which are precursors of malignant peripheral nerve sheath tumors (MPNST). MPNSTs are the most deadly of the soft tissue
sarcomas, with 5-year survival rates as low as 20 percent. Loss of function mutations in Polycomb Repressive Complex 2 (PRC2) genes, such as SUZ12, are
commonly identified in NF1-associated and sporadic MPNSTs. This is highly suggestive that perturbation of epigenetic homeostasis plays a role in malignant
transformation of neurofibromas. Dysregulated chromatin remodeling caused by the loss of PRC2 in MPNSTs likely confers novel vulnerabilities, that if
identified could be exploited therapeutically.

Methods: To identify new therapeutic targets in an MPNST context, our lab created NF1 and SUZ12 deficient immortalized human Schwann cells
(iHSC1lambda) for high-throughput drug screening. Using CRISPR/Cas9, we designed guide RNAs targeting NF1 and SUZ12. We isolated clones and screened
for cells that were wildtype (WT) or had heterozygous or homozygous deletions of NF1, and/or SUZ12. The cells were subjected to an epigenetic compound
library in a drug screening experiment.

Results: Isolated clones were verified to harbor loss of function mutations in NF1 and SUZ12 as well as reduced (or loss) of NF1 or SUZ12 protein
expression via immunoblotting analysis. High-throughput drug screening revealed several compounds, some of which fell into mechanisms, that showed
selective inhibition of SUZ12-/-;NF1-/- cells compared to NF1-/- or WT cells. Two of these compounds were vorinostat and azacitidine, an HDAC and DNA
methyltransferase inhibitor respectively.

Conclusions: Testing cells of an appropriate genetic background is fundamental in assaying drug compounds that could have clinical relevance. Our cells
reflect such a need by recapitulating an MPNST genotype whereby the epigenetic machinery has been disrupted. Our initial drug discoveries of an HDAC and
DNA methyltransferase inhibitor offers confidence that MPNSTs harbor specific vulnerabilities which can be targeted. The top compounds from our screens will
be tested further in human MPNST cell lines and some may move forward to animal models with the overall goal of informing future clinical trials.

Full List of Authors: Alex Larsson*1, Samuel Finnerty1, Mark Sokolowski1, Rory Williams1, Kyle Williams1, David Largaespada1
1
Masonic Cancer Center, University of Minnesota, Minneapolis, United States

Disclosure of Interest: A. Larsson: None Declared, S. Finnerty: None Declared, M. Sokolowski: None Declared, R. Williams: None Declared, K. Williams has a conflict with: K.B.W. is
supported by Children's Tumor foundation Young Investigator Award., D. Largaespada has a conflict with: DL is the co-founder and co-owner of several biotechnology companies,
specifically NeoClone Biotechnologies, Inc., Discovery Genomics, Inc. (recently acquired by Immunsoft, Inc.), and B-MoGen Biotechnologies, Inc. He consults for Surrogen, Inc.,
and Genentech, Inc. is funding some of his research. The business of all these companies is unrelated to the contents of this abstract. Other authors have no conflict of interest to
disclose. This work was supported by the NF Research Initiative Pre-clinical MPNST Research Grant Program, The National Institutes of Health (1R01-NS086219), the American
Cancer Society Research Professor Award (to DAL), The Children's Cancer Research Fund, and The Children's Tumor Foundation Synodos for NF1 Award.

2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018 | 105
 JHU395, A Nervous Tissue Penetrant Glutamine Antagonist, Restricts Growth of Malignant Peripheral Nerve
Sheath Tumor Through Inhibition of Biosynthesis

Kathryn Lemberg, Oncology, Johns Hopkins School of Medicine, Baltimore, MD; Pediatric Oncology Branch, Center for Cancer Research, National
Cancer Institute, National Institutes of Health, Bethesda, MD; Johns Hopkins Drug Discovery

Background: Malignant peripheral nerve sheath tumor (MPNST) is a sarcoma associated with nerve tissue that occurs in ~10% of patients with
neurofibromatosis type I (NF1). When incompletely surgically resected at diagnosis the 4-year event-free survival is <30%, and improved treatments are
needed. Recent evidence suggests that perturbing glutamine utilization warrants further exploration as a potential therapeutic strategy for MPNST. Our group
described JHU395, a nervous system penetrant glutamine antagonist (GA). JHU395 delivers active GA preferentially to nervous tissue which may result in less
gastrointestinal toxicity than was observed with competitive glutamine inhibitors in past clinical trials. The primary goal of this study was to evaluate JHU395 in
preclinical models of MPNST.

Methods: We investigated glutamine antagonism on growth of MPNST cells in culture and on growth of murine flank MPNST. JHU395 was administered orally
for 14 days to mice bearing inoculated tumors derived from the NPcis (NF1+/-;p53+/-) genetically engineered model of MPNST. Tumors were measured every
other day and tumor volume was calculated as the primary endpoint. GA was detected in tumor and plasma using a previously validated bioanalytical method.
Targeted liquid chromatography-mass spectrometry (LC-MS)-based metabolomics and MS-based flux analysis with stable isotope labeled 15N2- or 13C5-
glutamine were performed on MPNST cells and murine tumors.

Results: Compared to immortalized Schwann cells, growth of MPNST was preferentially inhibited by GA (IC50=8 micromolar versus >30 micromolar).
JHU395 delivered GA to MPNST cells with >4-fold higher cell-to-plasma ratio compared to native GA and maintained ~2-fold higher tumor-to-plasma levels
in vivo. Mice treated with JHU395 had mean tumor volume >40% smaller compared to vehicle controls with limited signs of toxicity. Targeted metabolomics
of GA-treated MPNST cells demonstrated multiple effects on glutamine-dependent metabolites. In vivo flux analysis showed that JHU395-treated tumors
decreased glutamine-derived nitrogen flux into several pathways including de novo nucleotide synthesis.

Conclusions: JHU395 inhibits growth of MPNST with effects on multiple biosynthetic pathways. Nervous tissue penetrant glutamine antagonism is a feasible
and effective therapeutic approach for MPNST. Future studies will investigate how JHU395 affects driver pathways in MPNST and evaluate JHU395 in
combination with other antitumor agents.

Full List of Authors: Kathryn Lemberg*1, 2, 3, Ying Wu3, Liang Zhao1, Jesse Alt3, Alexandra Gadiano3, Rana Rais3, 4, Pavel Majer5, Jaishri Blakeley4, Barbara Slusher3, 4
1
Oncology, Johns Hopkins School of Medicine, Baltimore, MD, 2Pediatric Oncology Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health,
Bethesda, MD, 3Johns Hopkins Drug Discovery, 4Neurology, Johns Hopkins School of Medicine, Baltimore, MD, United States, 5Institute of Organic Chemistry and Biochemistry,
Czech Academy of Sciences, Prague, Czech Republic

Disclosure of Interest: K. Lemberg: None Declared, Y. Wu: None Declared, L. Zhao: None Declared, J. Alt: None Declared, A. Gadiano: None Declared, R. Rais has a conflict with:
Dracen Pharmaceuticals, P. Majer has a conflict with: Dracen Pharmaceuticals, J. Blakeley: None Declared, B. Slusher has a conflict with: Dracen Pharmaceuticals

106 | 2018 Joint Global Neurofibromatosis Conference · Paris, France · November 2-6, 2018
Schwannomatosis Schwannomas Harbor Distinct DNA Methylation Profiles

Sheila Mansouri, Princess Margaret Cancer Center, University Health Network, Toronto, Canada

Background: Schwannomas are characteristic manifestations of NF2 and schwannomatosis syndromes. However, the majority of schwannomas are solitary
and sporadic. It is unclear whether and to what extent sporadic and syndrome-associated schwannomas or their histologic subtypes represent distinct
biological groups. Clinically, although schwannomatosis schwannomas are considered benign, the majority of patients experience unmanageable pain;
however, the underlying mechanism of this pain is not well understood. There is increasing evidence for DNA methylation profiling being able to disting