Clinical Chemistry 64:3

489–491 (2018) Point/Counterpoint

COUNTERPOINT
The Potential Harms of Human Gene
Editing Using CRISPR-Cas9
Françoise Baylis*

The anticipated benefits of human gene editing with the As such, the overarching risks with human gene
use of CRISPR (clustered interspaced short palindromic editing by use of CRISPR-Cas9 are two-fold. First,
repeats)-Cas9 are both familiar and contested. First and there is the risk that certain social, economic, and
foremost is the expectation of cures for blood disor- political forces will come to bear on those deemed
ders, lung diseases, cancers, and other maladies as “unfit” in an effort to pressure them to change their
clinician-scientists master various insertion, disrup- genetics so that they might better conform to certain
tion, and deletion techniques. In addition to these external norms or expectations. Second, there is the
potential therapeutic benefits, for some, there are the risk that those who resist pressure to conform will
potential benefits of human enhancement as investiga- experience (further) oppression.
tors learn to modify specific genetic traits in an effort to The literature on the ethics of human germ line gene
improve healthy individuals. Importantly, these proxi- editing and gene editing for enhancement purposes,
mate and distant potential benefits might be obtained which details the concerns noted above, is rich and ex-
through somatic cell or germ line gene modification. pansive. And, since 2015, when a team of Chinese scien-
With germ line gene modification (which involves insert- tists led by Junjiu Huang first reported limited success in
ing, deleting, or replacing the DNA of human sperm, editing the genome of human embryos (1 ), the ethics
eggs, or embryos), there is the added potential benefit debate has spilled onto the pages of print and online
that changes made to the human genome (especially newspapers and magazines, often with headlines her-
those aimed at correcting disease-causing and sometimes alding the arrival of “designer babies.” With every new
life-limiting genetic mutations) will be inherited by fu- policy report and every new scientific development,
ture generations. This would obviate the need for repeat this debate intensifies. Such was the case recently with
somatic cell modifications from one generation to the the news report (2 ), followed a week later by the sci-
next. ence article in Nature (3 ), that Shoukhrat Mitalipov
The potential benefits of gene editing, however, are and his team had successfully created genetically mod-
neither guaranteed nor risk-free. The potential harms in- ified human embryos while minimizing the harms of
clude off-target changes (as might happen with the inac- off-target effects and mosaicism (a “success” that has
tivation of essential genes), the inappropriate activation since been contested (4 )).
of cancer-causing genes, and the rearrangement of chro- By comparison, the literature on the potential harms
mosomes. Additionally, there are the risks of on-target of somatic cell gene editing for preventive and therapeu-
changes with unintended consequences, the creation of tic purposes is meager. The underlying assumption seems
mosaics of altered and unaltered cells, and the introduc- to be that this technology will eventually be proven safe
tion of changes that generate an immune response. In and effective and that relevant research should proceed
addition to these potential medical harms, there are also apace subject to existing ethical governance frameworks
potential social harms. There is, for example, the risk that and oversight mechanisms. This perspective is evident in
the introduction and eventual wide utilization of gene several back-to-back policy documents including On
editing technology will exacerbate existing inequalities Human Gene Editing: International Summit Statement,
resulting in human rights abuses, a new wave of eugenics, published in December 2015; the Nuffield Council on
increased discrimination and increased stigmatization. Bioethics report Genome Editing: An Ethical Review, pub-
lished in September 2016; and the US National Academy
of Science and National Academy of Medicine report
Human Genome Editing: Science, Ethics and Governance,
Novel Tech Ethics, Dalhousie University, Halifax, Canada.
published in February 2017.
* Address correspondence to the author at: 1379 Seymour St., Halifax, NS, Canada B3H On Human Gene Editing, authored by the organiz-
4R2. Fax 904-494-2924; e-mail [email protected]. ing committee of the first international summit on hu-
Received August 9, 2017; accepted November 20, 2017.
Previously published online at DOI: 10.1373/clinchem.2017.278317 man gene editing (sponsored by the US National Acad-
© 2017 American Association for Clinical Chemistry emy of Science, the US National Academy of Medicine,

489
Point/Counterpoint

the UK Royal Society, and the Chinese Academy of Sci- ratio. These limitations suggest that the proposed re-
ence), states unequivocally that clinical trials involving search is premature and may prove to be unsafe or inef-
somatic cell gene editing “can be appropriately and rig- fective (14 ).
orously evaluated within existing and evolving regulatory For illustrative purposes, consider the problems with
frameworks for gene therapy, and regulators can weigh internal validity. Internal validity concerns “the ability
risks and potential benefits in approving clinical trials to make causal inferences from an experimental result”
and therapies” (5 ). The Nuffield Council Genome Edit- [(15 ), p. 119]. Following Kimmelman and Henderson,
ing report affirms that “it is unlikely that, for the most the extent to which preclinical studies have minimized
part, therapies based on genome editing will raise distinctive threats to internal validity can be effectively assessed by
issues for the handling of safety and efficacy considerations” asking the following sorts of questions:
((6 ), p. 44). Moreover, in the discussion of genome editing
(1) Was the sample size sufficient to exclude random
in the context of biosecurity and dual use, the report con-
variation as an explanation of effect sizes?
firms that “the UK research councils . . . recognise the pos-
(2) Were animals randomly assigned to treatment?
sibility for misuse of research but express confidence in ro-
(3) Were outcome assessors blinded?
bust governance procedures for the research that they
(4) Do studies explain the flow of animals from inclu-
support and the applicability of existing regulatory frame-
sion through to analysis?
works” ((6 ), p. 103). The Human Genome Editing: Science,
(5) Were appropriate controls included?
Ethics and Governance report concludes that “existing regu-
(6) Was a dose–response relationship demonstrated?
latory infrastructure and processes for reviewing and evalu-
[(13 ), p. 51]
ating somatic gene therapy to treat or prevent disease and
disability should be used to evaluate somatic gene therapy Answers to these questions with reference to the pre-
that uses genome editing” [(7 ), p. 61]. clinical data supporting the proposed first-in-the-US,
While I am hopeful that clinical research to develop first-in-human Phase 1 CRISPR-Cas9 gene-editing can-
somatic cell gene editing therapies will yield effective cer trial are disconcerting, to say the least. The answers
treatments for those with serious genetic diseases, I worry suggest that there is considerable risk of “biases or ran-
that in the haste to move somatic cell gene editing from dom errors that lead to spurious inferences” [(13 ), p. 51]:
bench to bedside, well-established research ethics stan-
(1) There was only one preclinical CRISPR-Cas9 gene
dards for trial design will be compromised. Here, I am
editing study in mice and the sample size was small—17
mindful of warnings from CRISPR experts “that rushing
animals in three different groups. Only 5 of the 17 ani-
gene editing into clinical trials so soon after its develop-
mals were assigned to the treatment group examining the
ment poses ethical issues and that hype around the tech-
effects of three CRISPR-Cas9 genome edits.
nique could damage its prospects” (8 ). For example, ge-
(2) It appears that animals were not randomly assigned
neticist John Doench at the Broad Institute in the US is
to treatment.
reported to have said: “If [CRISPR] fails the first time,
(3) It appears that outcome assessors were not blinded.
the pendulum might swing against it and a perfectly good
(4) It appears that there was one data point missing in
technology could end up on the shelf” (8 ).
the preclinical study and no robust statistical analysis was
These concerns are well placed. Consider, for
undertaken.
example, the proposed first-in-the-US, first-in-human
(5) It is unclear whether appropriate controls were included.
Phase 1 CRISPR-Cas9 gene-editing trial for cancer
(6) A T-cell infusion dose–response assessment was not
(9, 10, 11, 12 ). This trial— cosponsored by the MD
undertaken (14 ).
Anderson Cancer Center in Texas, the University of Cal-
ifornia in San Francisco, and the University of Pennsyl- In closing, while there is good reason to be enthusi-
vania, in collaboration with Parker Institute for Cancer astic about the prospect of developing safe and effective
Immunotherapy—is to involve 18 research participants— genetic treatments for persons with serious diseases, this
6 with melanoma, 6 with synovial sarcoma, and 6 with enthusiasm must be tempered so as not to visit unneces-
multiple myeloma. The plan is to remove T cells from the sary harms on those who are invited to participate in
research participants, perform three CRISPR-Cas9 ge- research aimed at developing such therapies. Moreover,
nome edits, and then return the modified T cells to the as progress is made in the realm of therapeutic somatic
cancer patients. In June 2016, the proposed research was cell gene editing, we should be mindful of the implica-
reviewed and endorsed by the NIH’s Recombinant DNA tions for germ line gene editing and gene editing for
Advisory Committee. Arguably, however, at least 2 fea- enhancement purposes. These are discrete realms of ac-
tures of trial design appear to have been compromised— tivity with different ethical challenges that need to be
the requirement for scientific validity (which includes openly discussed in international forums.
internal validity, construct validity, and external validity The 2015 statement On Human Gene Editing high-
(13 )) and the requirement for a favorable harm– benefit lights the need for an ongoing forum that is inclusive among

490 Clinical Chemistry 64:3 (2018)

nations and engages a wide range of perspectives and exper- ments: (a) significant contributions to the conception and design, acquisi-
tise so that we might work together toward a broad societal tion of data, or analysis and interpretation of data; (b) drafting or revising
the article for intellectual content; and (c) final approval of the published
consensus about the appropriateness of any clinical use of
article.
germ line gene editing. We would do well to heed this call.
Authors’ Disclosures or Potential Conflicts of Interest: No authors
declared any potential conflicts of interest.

Author Contributions: All authors confirmed they have contributed to Acknowledgment: Thanks are owed to Marcus McLeod for comments
the intellectual content of this paper and have met the following 3 require- on an earlier draft and to Tim Krahn for assistance with references.

References
1. Liang P, Xu Y, Zhang X, Ding C, Huang R, Zhang Z, et al. newsitem.aspx?RecordID=12032015a (Accessed Au- review (Accessed August 2017).
CRISPR/Cas9-mediated gene editing in human tripro- gust 2017). 11. Phase 1Trial of NY-ESO-1 redirected CRISPR Edited T
nuclear zygotes. Protein Cell 2015;6:363–72. 6. Nuffield Council on Bioethics. Genome editing: Cells (NYCE cells) engineered to express NY-ESO-1
2. Connor S. First human embryos edited in U.S. MIT Technol- a n ethical review. September 2016. http:// TCR and gene edited to eliminate endogenous TCR
ogy Review; July 26, 2017. https://www.technolo nuffieldbioethics.org/wp-content/uploads/Genome- and PD-1. Protocol # 1604 –1524, NIH OBA RAC Pub-
gyreview.com/s/608350/first-human-embryos-edited- editing-an-ethical-review.pdf (Accessed August 2017). lic Review; June 21, 2016. https://osp.od.nih.gov/
in-us/?set=608352&utm_source=MIT+Technology+ 7. National Academies of Sciences, Engineering, Medicine. wp-content/uploads/2016/08/1524_RAC_Briefing
Review&utm_campaign=0b6e65d565-The_Download& Human genome editing: science, ethics, and governance. _Slides.pdf (Accessed August 2017).
utm_medium=email&utm_term=0_997ed6f472- Washington (DC): The National Academies Press; 2017. 12. Reardon S. First CRISPR clinical trial gets green light
0b6e65d565-154818457 (Accessed August 2017). https://www.nap.edu/catalog/24623/human-genome-
from US panel. Nature News; June 22, 2016. http://
3. Ma H, Marti-Gutierrez N, Park S-W, Wu J, Lee Y, Suzuki editing-science-ethics-and-governance (Accessed August
www.nature.com/news/first-crispr-clinical-trial-gets-
K, et al. Correction of a pathogenic gene mutation in 2017).
green-light-from-us-panel-1.20137 (Accessed August
human embryos. Nature 2017;548:413–9. 8. King A. Fears that gene editing cancer trials are prema-
2017).
4. Egli D, Zuccaro M, Kosicki M, Church G, Bradley A, Jasin M. ture. Chemistry World; January 9, 2017. https://
Inter-homologue repair in fertilized human eggs? [Pre- www.chemistryworld.com/news/fears-that-gene- 13. Kimmelman J, Henderson V. Assessing risk/benefit for
print] bioRxiv August 28, 2017 as doi: https://doi.org/ editing-cancer-trials-are-premature/2500206.article trials using preclinical evidence: a proposal. J Med Eth-
10.1101/181255. https://www.biorxiv.org/content/early/ (Accessed August 2017). ics 2015;42:50 –3.
2017/08/28/181255 (Accessed November 2017). 9. Cyranoski D. CRISPR gene-editing tested in a person for 14. Baylis F, McLeod M. First-in-human phase 1 CRISPR
5. Organizing Committee for the International Summit on the first time. Nature 2016;539:479. gene editing cancer trials: are we ready? Curr Gene
Human Gene Editing. On human gene editing: interna- 10. Kaiser J. First proposed human test for CRISPR Ther 2017;17:309 –19.
tional summit statement. National Academies of Sci- passes initial safety review. Science; June 21, 2016. 15. Kimmelman J. Gene transfer and the ethics of first-in-
ences, Engineering, and Medicine; December 3, 2015. http://www.sciencemag.org/news/2016/06/first- human research: lost in translation. Cambridge: Cam-
http://www8.nationalacademies.org/onpinews/ proposed-human-test-crispr-passes-initial-safety- bridge University Press; 2009.

Clinical Chemistry 64:3 (2018) 491