Content uploaded by Françoise Baylis
Author content
All content in this area was uploaded by Françoise Baylis on Jan 07, 2018
Content may be subject to copyright.
Current Gene Therapy
ISSN: 1566-5232
eISSN: 1875-5631
Impact
Factor:
2.78
SCIENCE
BENTHAM
Françoise Baylis* and Marcus McLeod
Novel Tech Ethics, Faculty of Medicine, Dalhousie University, P.O. Box 15000, 1379 Seymour Street Halifax, NS,
Canada
A R T I C L E H I S T O R Y
Received: June 29, 2 017
Revised: October 04, 2017
Accepted: November 15, 201 7
DOI:
10.2174/1566523217666171121165935
Abstract: A prospective first-in-human Phase 1 CRISPR gene editing trial in the United States for pa-
tients with melanoma, synovial sarcoma, and multiple myeloma offers hope that gene editing tools
may usefully treat human disease. An overarching ethical challenge with first-in-human Phase 1 clini-
cal trials, however, is knowing when it is ethically acceptable to initiate such trials on the basis of
safety and efficacy data obtained from pre-clinical studies. If the pre-clinical studies that inform trial
design are themselves poorly designed – as a result of which the quality of pre-clinical evidence is de-
ficient – then the ethical requirement of scientific validity for clinical research may not be satisfied. In
turn, this could mean that the Phase 1 clinical trial will be unsafe and that trial p articipants will be ex-
posed to risk for no potential benefit. To assist sponsors, researchers, clinical investigators and re-
viewers in deciding when it is ethically acceptable to initiate first-in-human Phase 1 CRISPR gene ed-
iting clinical trials, structured processes have been developed to assess and minimize translational dis-
tance between pre-clinical and clinical research. These processes draw attention to various features of
internal validity, construct validity, and external validity. As well, the credibility of supporting evi-
dence is to be critically assessed with particular attention to optimism bias, financial conflicts of inter-
est and publication bias. We critically examine the pre-clinical evidence used to justify the first-in-
human Phase 1 CRISPR gene editing cancer trial in the United States using these tools.
We conclude that the proposed trial cannot satisfy the ethical requirement of scientific validity be-
cause the supporting pre-clinical evidence used to inform trial design is deficient.
Keywords: CRISPR, Phase 1, Cancer, Gene editing, Research ethics, Scientific validity.
1. INTRODUCTION
Select cancer patients in China and in the United States,
who are in relapse or have treatment-refractory tumors, are
now eligible to participate in first-in-human Phase 1 CRISPR
gene editing can cer trials. The trial in China (currently un-
derway) is for patients with stage IV metastatic non-small
cell lung cancer [1]. The prospective trial in the United
States is for patients with melanoma, synovial sarcoma, and
multiple myeloma [2]. For some, these are exciting times
given the potential of gene editing tools to treat human dis-
ease. Others, however, are somewhat more reserved in their
enthusiasm. Among the skeptics are those who wonder
whether CRISPR gene editing cancer trials are premature –
likely to be proven unsafe, and to have put trial participants
at risk for no potential benefit.
*Address correspondence to this author at Novel Tech Ethics, Faculty of
Medicine, Dalhousie University, P.O. Box 15000, 1379 Seymour St Halifax,
NS, Canada; Tel: 1-902-494-6458; Fax: 1-902-494-2924;
E-mail: francoise.baylis@dal.ca
In this article, we briefly describe the proposed first-in-
human Phase 1 CRISPR gene editing trial for cancer in the
United States. This trial was reviewed in June 2016 by the
Recombinant DNA Advisory Committee (RAC) at the
United States National Institutes of Health (NIH). The RAC
is an advisory committee tasked with providing the NIH Di-
rector with advice and recommendations. Any proposed gene
editing research that falls within the scope of the NIH guide-
lines is subject to RAC review. Next steps before research
participants can be enrolled include review and approval
from the United States Food and Drug Administration
(FDA), and from the University of Pennsylvania Institutional
Biosafety Committee (IBC), Conflict of Interest Standing
Committee, and Institutional Review Board (IRB).
Following this description of the proposed clinical trial,
we critically evaluate the quality of the pre-clinical evidence
presented to the RAC in support of the submission to initiate
clinical research.1 We highlight features of the pre-clinical
1 We examined the “Preclinical Data Package Supporting the Clinical Use of NY-ES0-
1-redirected TCRendo and PD1 edited T cells” (Protocol 1604-1524) submitted to the
Recombinant DNA Advisory Committee (RAC). The Preclinical Data Package contai-
1875-5631/17 $58.00+.00 © 2017 Bentham Science Publishers
Send Ord ers for Reprints to reprints@benthamscience.ae
Current Gene Therapy, 2017, 17, 309-319
309
RESEARCH ARTICLE
First-in-human Phase 1 CRISPR Gene Editing Cancer Trials:
Are We Ready?
310 Current Gene Therapy, 2017, Vol. 17, No. 4 Baylis and McLeod
evidence that, in our estimation, threaten the ethical require-
ment of scientific validity. Ezekiel Emanuel and colleagues
David Wendler and Christine Grady suggest that scientifi-
cally valid research “must have a clear scientific objective;
be designed using accepted principles, methods and reliable
practices; have sufficient power to definitively test the objec-
tive; and offer a plausible data analysis plan” [3]. In more
colloquial terms, Kirstin Borgerson writes:
“What the ethical requirement of scientific validity is
meant to get at is that the science in question is, to put it sim-
ply, good science. It follows the norms of science appropriate
in the particular discipline. It models good scientific inquiry. It
is well-designed [to advance knowledge].” [4: p.394].
We conclude that the one pre-clinical study in mice used
to justify the f irst-in-human Phase 1 CRISPR gene editing
cancer trial in the United States does not satisfy the ethical
requirement of scientific validity. Moreover, the translational
distance between the pre-clinical study and the proposed
clinical trial is unnecessarily wide – the quality of pre-
clinical evidence is seriously deficient. As a direct conse-
quences of this, the ethical requirement of scientific validity
for clinical research may not be satisfied. This conclusion
suggests that there is reason to question current enthusiasm
for proceeding with human somatic cell gene editing using
existing national regulatory frameworks and oversight
mechanisms (e.g. laws, regulations, policies, guidelines,
standards, and professional norms and practices).
2. U.S. FIRST-IN-HUMAN PHASE 1 CRISPR GENE
EDITING CANCER TRIAL
In 2016, investigators at the MD Anderson Cancer Center
in Texas, the University of California in San Francisco and
the University of Pennsylvania in collaboration with the
Parker Institute for Cancer Immunotherapy announced plans
to proceed with a CRISPR gene editing trial using autolo-
gous T cells entitled “Phase I Trial of Autologous T Cells
Engineered to Express NY-ESO-1 TCR and Gene Edited to
Eliminate Endogenous TCR and PD-1”. The usual purpose
of a Phase 1 clinical trial is to assess the safety and dosage of
a novel intervention [5, 6].2 Some Phase 1 trials, however,
also seek preliminary evidence of efficacy, [7] and this is
mostly the case with cancer trials.
The proposed Phase 1 clinical trial uses a competitive re-
population strategy that is basket designed to test several end-
points in eighteen (18) research participants with refractory tu-
mors, including melanoma (n=6), synovial sarcoma (n=6) and
multiple myeloma (n=6), for whom there are no effective thera-
pies. The primary endpoints for this first-in-human combination
of immunotherapy and gene editing are patient safety and feasi-
ned redacted material due to proprietary rights held by the investigators. We also
examined the webcast (https://osp.od.nih.gov/calendar-3/action~agenda/exact_date~
21-6-2016/cat_ids~21/), meeting agenda, briefing material and briefing slides submit-
ted to the RAC for the June 21, 2016 meeting, entitled “Phase I trial of NYESO1
redirected CRISPR Edited T cells (NYCE cells) engineered to express NYESO1 TCR
and gene edited to eliminate endogenous TCR and PD1”. The single pre-clinical
animal study using NY-ESO-1 redirected CRISPR Edited T Cells that was utilized by
the investigators to support the submission to the RAC had not been peer reviewed nor
published in a scientific journal prior to the June 21, 2016 meeting.
2 According to Lo and colleagues [6], the aims of Phase 1 clinical trials are ‘‘to assess
the safety and f easibility of the investigational intervention and to determine dosages
for subsequent clinical tr ials. Direct therapeutic benefit, although hoped for , is unlikely
in early trials, particularly if the first participants receive low doses.”
bility, as well as manufacturing feasibility. The secondary end-
points are a clinical assessment of anti-tumor responses and
survival, as well as an examination of T cell bioactivity, immu-
nogenicity, engraftment, persistence, and trafficking. The inves-
tigators acknowledge (in the consent documents) that the re-
search participants may not get any personal medical benefit
from participating in this Phase 1 clinical trial.
The basket design consists of an open label pilot study
where peripheral blood lymphocytes will be collected from
research participants. The T cells, which are a sub-type of
lymphocyte in peripheral blood involved in cell-mediated
immunity (i.e. the immune response), will then undergo
transduction using a lentiviral vector3 to express a new high
affinity T cell receptor with specificity for the NY-ESO-1
peptide (NY-ESO-1 TCR). NY-ESO-1 is a highly immuno-
genic antigen expressed on human tumors. For example,
NY-ESO-1 is expressed on melanoma, synovial and myxoid
sarcoma and advanced myeloma tumors at a rate of 28-45%,
greater than 70% and approximately 50% respectively [2]. In
addition to allowing the transduced T cells to target the NY-
ESO-1 peptide expressed on human tumors, engineering
NY-ESO-1 TCR aims to overcome an incapacity to isolate
and propagate large numbers of T cells with a defined speci-
ficity and phenotype [8].
The T cells will also be gene edited using CRISPR tech-
nology to knock-out the gene loci for the α and β chains of
the endogenous T cell receptor (TCR α and TCR β respec-
tively) as well as th e Programmed Cell Death Protein-1 (PD-
1). The rationale for gene editing TCR α and TCR β is to: (1)
reduce TCR mispairing with NY-ESO-1 TCR and the possi-
ble formation of neo-reactivity [see, e.g., 8, 9]; (2) reduce the
possibility of autoimmunity; as well as (3) promote the ex-
pression of exogenous NY-ESO-1 TCR. It is hypothesized
that gene editing PD-1 will prevent exhaustion and therefore
maintain T cell activity in the presence of the checkpoint
ligands, PD-L1 and PD-L2, including those expressed by
tumor cells or with cells within the tumor micro-environment
[see, e.g. 10, 11].4
Following transduction and gene editing, the T cells –
which are known as NY-ESO-1 Redirected CRISPR Edited
T Cells – will be expanded ex vivo. Meanwhile, the trial par-
ticipants will undergo lympho-depleting chemotherapy, and
thereafter will receive a single dose of 1x108 autologous NY-
ESO-1 Redirected CRISPR Edited T Cells/kg. The current
investigators (as well as other investigators) have used T
cells expressing NY-ESO-1 TCR in other pre-clinical studies
and in human clinical trials and have shown safety and effi-
cacy [2, 10; see, e.g., 12, 13].5 The current trial is novel,
3 A lentivirus is a virus-like particle that can be used as a vector in both clinical and
non-clinical research to deliv er recombinant DNA molecules known as transgen es into
host cells.
4 The data from this Phase I/II clinical trial indicated that although T cells remained
functional in some research participants for up to a year after infusion, there was ex-
haustion over time [11].
5 A research paper by Schietinger and Greenberg [13] is included in the RAC Public
Review in the subsection titled “Limitations of Current NY-ESO-1 Transduced T Cells
That Retain Endogenous TCR Expression”. This research paper explores the issue of
CD8 T cell dysfun ction, including T cell tolerance to self-antigens (self-tolerance), T
cell exhaustion during chronic infections, and tumor-induced T cell dysfunction.
Importantly, this paper is not a study, and does not disclose or address the issue of T
cell dysfunction in the context of NY-ESO-1 tran sduced T cells that retain endog enous
TCR expression.
First-in-Human Phase 1 C RISPR Ge ne Editing Cancer Trials Current Gene Therapy, 2017, Vol. 17, No. 4 311
however, as it includes the creation of NY-ESO-1 TCR in
conjunction with the use of CRISPR technology to triple edit
TCR α, TCR β and PD-1 genes. The investigators believe
that autologous NY-ESO-1 Redirected CRISPR Edited T
Cells will have improved antitumor activity and enhanced
persistence [e.g., 2, 13].
3. VALIDITY AND PRE-CLINICAL EVIDENCE
An overarching ethical challenge with first-in-human
Phase 1 clinical trials is knowing when it is ethically accept-
able to initiate such trials. This is not a move to be mad e on
the basis of intuition, institutional pressure, career advance-
ment, prospective financial rewards, anticipated accolades,
or international dueling. A key resource used to justify this
move is safety and efficacy data from pre-clinical studies.
In Gene Transfer and the Ethics of First-in-Human Re-
search: Lost in Translation [14], Jonathan Kimmelman de-
veloped a four-part framework to assist investigators and
reviewers in deciding when it would be ethically acceptable
to initiate first-in-human clinical trials. The framework
aimed to ensure modest translational distance (i.e. a narrow
inferential gap) between pre-clinical and clinical research.
This 2010 framework called for an assessment of the internal
and external validity of pre-clinical studies used to support
Phase 1 clinical trials, as well as an assessment of the corre-
spondence between the experimental design of the pre-
clinical studies and subsequent clinical trials. It also called
for a critical appraisal of the credibility of the supporting
evidence with particular attention to optimism bias, financial
conflicts of interest and publication bias.
More recently, in 2014, following a systematic review of
guidelines for the design and execution of in vivo animal
studies [15], the original framework has been refined in col-
laboration with Valerie Henderson. Together Henderson and
Kimmelman propose a structured process for evaluating pre-
clinical evidence in terms of potential threats to internal va-
lidity, construct validity, and external validity [16]. Below,
we evaluate the validity and the credibility of the pre-clinical
evidence presented in support of the proposed first-in-human
Phase 1 CRISPR gene editing cancer trial in the United
States.
3.1. Internal Validity
Kimmelman defines internal validity as “the ability to
make causal inferences from an experimental result” [14:
p.119]. Of concern with pre-clinical studies as supporting
evidence for clinical studies are the risks of “biases or ran-
dom errors that lead to spurious causal inferences” [16:
p.51]. Minimizing these risks requires close attention to
various elements of trial design such as sample size (a-priori
power calculations), randomized allocation of animals to
treatment, blinding of outcome assessment, dose-response
relationships and selection of appropriate controls groups.
According to Kimmelman, frequently little attention is paid
to these elements at the pre-clinical stage [e.g. 14: pp.110-
131].
Our review of the (only) one pre-clinical gene editing
study using NY-ESO-1 Redirected CRISPR Edited T Cells
in mice presented to the RAC suggests problems with these
methodological elements. For a start, the sample size is sur-
prisingly small. There were only three treatment groups for a
total of 17 animals. Moreover, there were only five animals
in the treatment group using NY-ESO-1 Redirected CRISPR
T cells. In addition, there appears to have been no random-
ized allocation, no blinding of outcome assessment, no T cell
infusion dose response assessment,6 and no robust statistical
analysis. On the assumption that such information would
have been provided to the RAC, if available, the apparent
absence of such data is concerning and arguably represents a
significant lacuna with respect to internal validity. Further-
more, although the investigators disclosed that NYESO-1
Redirected CRISPR Edited T Cells could be maintained for
at least 2 months, they failed to assign a group of control
tumor negative animals with an infusion of NYESO-1 Re-
directed CRISPR Edited T Cells to evaluate adverse effects
over a longer period of time.
3.2. Construct Validity
According to Henderson and Kimmelman, construct va-
lidity concerns the “relationship between behavioral out-
comes in animal experiments and human behaviors they are
intended to model” [15: p.2]. Construct validity also extends
to the relationsh ip betw een animals and humans, as regards
the underlying condition, treatments, causal pathways, ex-
perimental operations and clinical scenarios. At issue is
whether the animal model(s), the interventions, and outcome
assessments in pre-clinical studies are good representations
of the human condition under study [17].
Construct validity may be threatened if investigators “err
in executing experimental operations” [15: p.2] or when
“physiological derangements driving human disease are not
present in the animal model” [15: p.2]. Moreover, construct
validity is further threatened when there is no interpolation
between data from published (potentially innumerable pre-
clinical or clin ical) r esearch and first-in-human trials, or
when investigators rely on pre-clinical studies that do not
validate the hypothesis being tested [e.g. 14: pp.110-131].
Henderson and Kimmelman state that threats to construct
validity “are reduced by articulating, addressing, and con-
firming theoretical presuppositions underlying clinical gen-
eralization” [15: p.2]
3.2.1. Choice of Tumor Cell Line
A first threat to construct validity relates to the cancer in-
troduced into the mouse model. The investigators introduced
a human lung cancer cell line, A549-ESO-CBG (which was
6 On September 4, 2017, the FDA placed a clinical hold on a Phase 1 study by Cellec-
tis using gene edited allogenic CAR T cells in acute myeloid leukemia and in blastic
plasmacytoid dendritic cell neoplasm. One o f the research participants in this trial died
9 days after receiving 6.25x105 gene edited allogenic CAR T cells/kg. See, “Cellectis
Reports Clinical Hold of UCART123 Studies” http://www.cellectis.com/en/press/
cellectis-reports-clin ical-hold-of-ucart123-studies/. On November 6, 2017, the FDA
lifted the clinical hold following a number of agreed upon revisions, including a “dec-
rease of the cohort dose level to 6.25x104 UCART123 cells/kg”. See, “FDA Lifts
Clinical Hold on Cellectis Phase 1 Clinical Trials with UCART123 in AML and
BPDCN” http://www.cellectis.com/en/press/fda-lifts-clinical-hold-on-cellectis-phase-
1-clinical-trials-with-ucart123-in-aml-and-bpdcn. While CAR T cells are not the same
as CRISPR Edited T Cells, it is important to note that the (original an d revised) dosa-
ges in the Cellectis study are considerably lower than the dosage of 1x108 autologous
NY-ESO-1 Redirected CRISPR Edited T Cells/kg proposed for the first-in-human
Phase 1 CRISPR gen e editing clinical trial.
312 Current Gene Therapy, 2017, Vol. 17, No. 4 Baylis and McLeod
HLA-A2+ and NY-ESO-1+) into genetically modified NOD
Scid Gamma (NSG) mice. However, the proposed Phase 1
clinical trial does not plan to recruit lung cancer patients. The
target population includes patients with melanoma, synovial
sarcoma or multiple myeloma.
The four cancer subtypes have different molecular and
phenotypic characteristics. Arguably the investigators should
have introduced a melanoma, plasma or a sarcomatous cell
line expressing HLA-A2+ and NY-ESO-1+ into the mouse
model. Alternatively, instead of using long-established cell
lines (which have their own limitations), they could have
used biopsied cells from cancer patients with melanoma,
synovial sarcoma, or multiple myeloma. For even greater
construct validity, the investigators could have introduced
biopsied cells from prospective research participants into th e
mouse model. Cells from a biopsy would more accurately
model the biology of the original tumor [18].
To be sure, there are potential shortfalls with th e use of
patient-derived biopsied cells, including the need for immu-
nodeficient hosts, the time required to grow the tumor in the
animal model, the loss of non-transformed stromal elements,
as well as the cells not exactly resembling the human disease
due to the possible process of selection pressure that may
change the clonal composition of the engrafting tumor to
more malignant cells and clones [e.g. 18-22]. These limita-
tions, however, apply equally to established cell lines.
Patient-derived biopsied cells have been shown to be
phenotypically stable across multiple transplant generations
and typically retain the principal histological, transcriptomic,
proteomic and genomic characteristics of their donor tumor
while showing comparable treatment responses to those ob-
served clinically [e.g. 18, 22-28]. For example, histological
features including gland formation and keratin deposition
have been shown to be comparable to the original tumor,
while the gene expression profile of patient-derived cells
also clusters with the original tumor [e.g. 18, 29, 30]. These
characteristics suggest that the treatment regimen in the pre-
clinical study would be more comparable to the human clini-
cal trial if biopsied patient-derived cells were used in pre-
clinical studies instead of established cell lines [e.g. 23, 31,
32].
No robust causal inferences for the first-in-human Phase
1 CRISPR gene editing trial in patients with melanoma, sy-
novial sarcoma or multiple myeloma can be made on the
basis of one pre-clinical study in mice with a human lung
cancer cell line. The risk of mischaracterizing the anti-tumor
response of NY-ESO-1 Redirected CRISPR Edited T Cells
on a human lung cancer cell line is considerable.
3.2.2. Anatomical Location of Cancer, Co-Interventions,
Sex and Age
If investigators deliver a cancer cell line in a pre-clinical
study to an anatomical location in an animal that is different
to the tumor location in humans, then construct validity is
threatened [14: pp.110-131]. In the one pre-clinical study
presented to the RAC, NSG mice were injected with a hu-
man lung cancer cell line in the right flank. Information
about the location of tumors in research participants was not
disclosed. It is likely, however, that the right flank is a dif-
ferent anatomical location to the three sub-types of cancer in
the proposed Phase 1 clinical trial.
For example, melanoma is usually located on the skin but
has a high propensity to metastasize to the brain [33, 34].
Given that the target population includes patients with mela-
noma, the pre-clinical study should have investigated
whether the NY-ESO-1 Redirected CRISPR Edited T Cells
could migrate across the blood-brain barrier. The investiga-
tors included no statements regarding the po tential ability of
NY-ESO-1 Redirected CRISPR Edited T Cells to cross the
blood-brain barrier in the proposed Phase 1 clinical trial.
This is an important consideration in the event that the mela-
noma has brain metastasized in the research participants.
This speaks to the value of having more than one pre-clinical
experiment with cancers in anatomical locations that better
mimic the disease under study.
Also relevant is the fact that trial participants will un-
dergo chemotherapy prior to the infusion of NY-ESO-1 Re-
directed CRISPR Edited T Cells. There is no equivalent in-
tervention in the pre-clinical study [18]. While it is true that
NSG mice lack a complete functional immune system so
their inflammatory immune cell response is deficient, it is
possible that the chemotherapy (an active drug) is a con-
founding factor and thus should have been controlled for in
pre-clinical studies. As such, the NSG mice were not
matched to co-interventions. Moreover, as NSG mice lack
mature T cells, B cells, natural killer cells and are deficient
in multiple cytokine signalling pathways, the pre-clinical
study cannot assess the possibility of cytokine release syn-
drome [35]. Cytokine release syndrome is a life-threatening
complication that can cause brain edema, neurological da-
mage and death. This has been shown to occur in clinical
trials using anti-CD19 CAR-T cell therapies [36, 37].
Finally, the investigators did not design the pre-clinical
study using mice from both sexes to match the animal model
to the sex of patients in clinical setting, nor did they use dif -
ferent age groups.
No robust causal inferences for the first-in-human Phase
1 CRISPR gene editing cancer trial can be made on the basis
of one pre-clinical study in mice with a cancer in an ana-
tomical location differ ent from where the cancer is likely to
be in a human population. Additional confounding factors
include the absence of an equivalent to chemotherapy in the
pre-clinical study as well as a failure to study both sexes in a
range of ages. As such, the risk of mischaracterizing the effi-
cacy of NY-ESO-1 Redirected CRISPR Edited T Cells on a
tumor is considerable.
3.2.3. Choice of Donor T Cell Source
Another issue with constru ct validity relates to the reli-
ance on data from in vitro experiments using T cells from
two healthy donors to confirm the efficacy of lentiviral
transduction and gene editing. The problem with this strategy
is twofold, (1) an exceedingly small cohort (two cell sources)
and (2) the exclusive use of T cells from healthy donors.
Thereafter, the investigators used the same T cells in the one
pre-clinical study in mice.
The investigators recognize the limitation of using T cells
from a small cohort to determine the efficacy of lentiviral
First-in-Human Phase 1 C RISPR Ge ne Editing Cancer Trials Current Gene Therapy, 2017, Vol. 17, No. 4 313
transduction and gene editing. They have plans to expand the
in vitro analysis using cells from a greater number of healthy
donors in the future. This should have been done prior to the
RAC submission, however, as information from additional
cell sources could have usefully informed clinical trial de-
sign.
The investigators should also have tested the efficacy of
lentiviral transduction and gene editing in T cells from indi-
viduals with cancer, including patients with melanoma,
synovial sarcoma, or multiple myeloma, and thereafter used
the same T cells in pre-clinical animal studies to determine
the two study endpoints of (1) tumor size and (2) percent
survival. Information about the efficacy of lentiviral trans-
duction and gene editing in cancer patients would have been
a prudent step prior to the RAC submission. Although the
route/method of treatment delivery (i.e. infusion) in the pre-
clinical and clinical research appear to match, as do the two
study endpoints (i.e. clinical assessments of anti-tumor re-
sponses and survival), the use of cancer patient derived T
cells in the pre-clinical study in mice (and subsequently in
larger animal models that more closely mimic the human
disease, such as canines) would have improved construct
validity between the pre-clinical and clinical research.
No robust causal inferences for the first-in-human Phase
1 CRISPR gene editing cancer trial can be made on the basis
of the pre-clinical data obtained using T cells from only two
donors that were not sourced from a cancer patient popula-
tion. The risk of mischaracterizing the efficacy of T cells
from healthy donors is considerable.
3.2.4. Heterogeneous Transduced and Gene Edited Cell
Populations
The proposed Phase 1 CRISPR gene editing cancer trial
may create as many as 16 populations of autologous NY-
ESO-1 Redirected CRISPR Edited T Cells.7 This is because
of the possible permutations and combinations resulting
from the plan to first introduce a lentiviral vector to express
NY-ESO-1 TCR and second to delete the gene loci for the
TCR α, TCR β and PD-1.
At one extreme, there could be wild-type endogenous T
cells that have not been successfully modified with the lenti-
viral vector and CRISPR technology. This cell population
likely would be safe but not efficacious. At the other ex-
treme, th ere could be T cells expressing NY-ESO-1 TCR
with endogenous TCR α, TCR β and PD-1 disruptions. This
successfully modified cell population presumably would
have enhanced anti-tumor effects and be less susceptible to
exhaustion by PD-L1 and PD-L2. In between these two phe-
notypic extremes there would be various cell populations. To
date, limited available data indicate that 49% and 62% (re-
spectively) of cells from the two healthy donors expressed
NY-ESO-1 TCR after lentiviral transduction. However, a
7 There are sixteen (16) potential genotypes of autologous T cells expressing either
disrupted or intact genes for NYESO1 TCR, TCR α, TCR β, and PD1 , namely , (1)
PD1+ TCR A+B+ , (2) PD1- TCR A+B+, (3) PD 1+ TCR AB+, (4) PD1+ TCR BA+,
(5) PD1+ TCR AB-, (6) PD1- TCR AB+, (7) PD1- TCR BA+, (8) PD1- TCR AB-,
(9) PD1 + TCRA+B+, (1 0) PD1+ TCR AB+, (11) PD1+ TCR BA+, (12) PD1+ TCR
AB, (13) PD1- T CRA+B+, (14) PD1- TCR AB+, (15) PD1- TCR BA+, and (16)
PD1- TCR AB.
significant proportion (approximately 43%) of all the trans-
duced cells, were not PD-1 edited. Despite evidence of sig-
nificantly varying levels of genetic modification (transduc-
tion and gene editing), there is no plan to screen for a spe-
cific cell population prior to infusion into research partici-
pants.
In the event of any observed effect in the one pre-clinical
study in mice, no robust causal inferences for the first-in-
human Phase 1 CRISPR gene editing cancer trial could be
made about the safety and efficacy of any research interven-
tion given the potential heterogeneity of the cell populations.
3.2.5. Summary of Concerns Regarding Construct Validity
In our view, construct validity is threatened as the pre-
clinical study did not validate the hypothesis that is to be
tested in the Phase 1 clinical trial.
Firstly, the investigators should have tested the efficacy
of lentiviral transduction and gene editing in more than two
sources, as well as used T cells from cancer patients ( includ -
ing perhaps prospective research participants) with mela-
noma, synovial sarcoma, or multiple myeloma. There should
have been toxicology and efficacy studies of the various in-
vestigational cell populations to address the limitations with
respect to the potential heterogeneity of the cell populations.
In the event of any observed effect in the pre-clinical re-
search, no robust causal inferences can be made about the
safety and efficacy of any research intervention given the
potential heterogeneity of the cell populations. Ideally, the
transduced and gene edited cells should have been cell sorted
in order to obtain a homologous population of T cells ex-
pressing NY-ESO-1 TCR with endogenous TCR α, TCR β
and PD1 disruptions. Thereafter, the investigators should
have used this homologous population of T cells to deter-
mine the two endpoints of (1) tumor size and (2) percent
survival in animals with tumors formed from either mela-
noma, plasma or sarcomatous cell lines expressing HLA-
A2+ and NY-ESO-1+, or from biopsied cells from cancer
patients with melanoma, synovial sarcoma, or multiple mye-
loma. Instead, they used a human lung cancer cell line.
Moreover, they introduced the human lung cancer cell line
into the right flank of the mouse model, which is a different
anatomical location from where the cancers of interest would
be expected in the target patient populations. As well, there
is failure to take into account a range of potential confound-
ing factors including co-interventions, sex and age.
In the event of any observed effect in the pre-clinical re-
search, no robust causal inferences can be made about the
safety and efficacy of any research intervention given the
potential heterogeneity of the cell populations.
3.3. External Validity
Kimmelman describes external validity with reference to
the importance of “conducting replication studies that vary
experimental conditions” [14: p.120] in order to effectively
test whether (and if so, to what extent) “cause and effect re-
lationships hold up under varied conditions.” [17]. Among
the many possible confounders are the replication of differ-
ent models of the same disease, independent replication, rep-
lication of different species, standardization of methods, and
314 Current Gene Therapy, 2017, Vol. 17, No. 4 Baylis and McLeod
reproducibility of results [38]. More recently Henderson and
Kimmelman have suggested that external validity threats can
be tempered by ensuring that the pre-clinical research is con-
ducted in “(1) more than one model, (2) more than one labo-
ratory and (3) more than one species” [16: p.51].
Now, some commentators argue that non-human animal
studies, as currently conducted, cannot reasonably predict the
outcome of human trials [39]. This is because non-human
animals are poor models for the majority of human diseases
[see, e.g. 40, 41] due to genetic, molecular, physiological,
immunologic and cellular differences [see, e.g. 39-44], in-
cluding varied antigen distribution, processing and presenta-
tion.
Though we believe there is merit to this argument, debate
about the value and validity of pre-clinical research is be-
yond the scope of this article. Here, for the sake of argument,
we accept current research ethics norms (as enforced by re-
search oversight bodies) according to which evidence of suc-
cessful pre-clinical research in non-human animals should
precede first-in-human clinical trials.
The prospective Phase 1 CRISPR gene editing cancer
trial is supported by (only) one non-peer reviewed pre-
clinical study in genetically modified NSG mice infused with
NY-ESO-1 Redirected CRISPR Edited T Cells. The problem
with this pre-clinical study is threefold, (1) use of a single
(small) animal model, (2) in only one laboratory, and (3) in
only one species.
In our view, there should have been several pre-clinical
studies with independent replication by research teams that
are financially disinterested in the outcome [44-46]. As well,
some of these studies should have been in different (includ-
ing larger) species of non-human animals. For example, the
investigators could have validated the pre-clinical study us-
ing canines, which provide an attractive translational model
and share with humans many features, including tumor ge-
netics, molecular markers, histology, biological behavior,
tumor progression and response to conventional therapies
[47, 48]. In addition, pre-clinical studies using canines may
allow for long-term assessment of efficacy and toxicity [48].
Finally, some of these animal models should have had can-
cers more similar to those afflicting patients with melanoma,
synovial sarcoma, or multiple myeloma [44].
Importantly, the investigators stated that they had not
performed toxicology studies in the one pre-clinical study in
mice to test for safety, including clinical observation, weight,
mortality, clinical pathology, organ weight, gross pathology,
and histopathology, as well as to test for efficacy and bio-
distribution of the investigational cell product. Given that the
primary endpoints of the first-in-human Phase 1 CRISPR
gene editing cancer trial include patient safety, the investiga-
tors should have performed these tests prior to the RAC
submission in order to generate a robust assessment of
safety, rather than inform the RAC that these tests would be
performed as a later date prior to th e IND submission to the
FDA.
No robust causal inferences for the first-in-human Phase
1 CRISPR gene editing cancer trial can be made on the basis
of (only) one pre-clinical study using a single (small) animal
model, in only one laboratory, and in only one species.
4. CREDIBILITY
Leaving aside, for the moment, concerns about the ro-
bustness of the pre-clinical evidence, there is reason to ques-
tion the trustworthiness of the available information on the
basis of which “relevant experts” are expected to assess the
likely predictive value of the pre-clinical evidence mar-
shalled to justify the move to first-in-human clinical trials
[14: p.122].
According to Kimmelman, matters of potential concern
include: (1) optimism bias which can result in the skewing of
pre-clinical research findings; (2) financial conflicts of inter-
est as when investigators hold patents related to the proposed
clinical trial (th is may result in bias, information non-
disclosure, or a premature move to the clinical setting); and
(3) publication bias resulting in the non-publication of results
from pre-clinical research.
4.1. Optimism Bias
Kimmelman describes optimism bias as a conscious or
subconscious tendency on the part of investigators to present
their data in a favorable light. This bias may be unintentional
as a result of excessive enthusiasm; or intentional as a result
of decision-making about the management of data outliers
and missing data [14: pp.110-131].
The pre-clinical study involved three discrete groups of
mice injected intravenously with: (1) NY-ESO-1 transduced
T cells alone (investigators name these NY-ESO-1.TCR); (2)
NY-ESO-1 TCR and PD1/TCR α/TCR β triple knockout T
cells (investigators name these NY-ESO-1 TCR, CRISPR);
and (3) T cells alone [2]. The two study endpoints were (1)
tumor size and (2) percent survival, with no assessment of
behavioral outcome.
Examination of the graphical data presented to the RAC
reveals that approximately 33% of the animals in the NY-
ESO-1 TCR treatment group survived past 80 days. How-
ever, the corresponding graphical data on the study endpoint
tumor size at 80 days in this same NY-ESO-1 TCR treatment
group reveals no data point. It appears that the investigators
have selectively reported outcomes with respect to tumor
size in the NY-ESO-1 TCR treatment group, which is sug-
gestive of optimism bias.
4.2. Financial Conflicts of Interest
Biomedical research conducted in academic institutions
is now commonly intertwined with pharmaceutical and bio-
technology industries as part of an innovation ecosystem
[49]. In this way, academic institutions and investig ators
have embraced a new kind of entrepreneurship in which fi-
nancial conflicts of interest may arise [50]. In an effort to
manage such conflicts of interest, academic institutions and
professional organizations have developed policies govern-
ing academic-industrial collaborations. For example, the
2000 “Policy of The American Society of Gene Therapy
Financial Conflict of Interest in Clinical Research” stipulates
that,
First-in-Human Phase 1 C RISPR Ge ne Editing Cancer Trials Current Gene Therapy, 2017, Vol. 17, No. 4 315
“all investigators and team members directly responsible
for patient selection, the informed consent process and/or
clinical management in a trial must not have equity, stock
options or comparable arrangements in companies sponsor-
ing the trial” [51].
Financial conflicts of interest, however, are not limited to
conflicts resulting from relationships with research sponsors.
Technology transfer endeavors, mainly through patenting
and licensing to commercialize academic research, now play
a major role in biomedical sciences [e.g. 49, 52-56]. This
may give rise to additional conflicts of interest as per the
2013 FDA document “Guidance for Clinical Investigators,
Industry, and FDA Staff: Financial Disclosure by Clinical
Investigators” [57]. The FDA advocates extensive disclosure
by investigators about compensation received, proprietary
interests in the tested product (including a patent or licensing
agreement), equity interests in any of the research sponsors,
significant payments (including grants), and reimbursements
such as retainers for ongoing consultation or honoraria.
Dr. Carl H. June of the University of Pennsylvania is the
scientific advisor for the proposed Phase 1 CRISPR gene
editing cancer trial. Importantly, the draft consent form pre-
sented to the RAC (which includes information about con-
flicts of interest) disclosed that Dr. June “invented the tech-
nology used to expand your cells for this study and he re-
ceives significant financial benefit related to this. This tech-
nology is licensed to a biotechnology company called Life
Technologies, and has been sub-licensed to Novartis.”
During the RAC proceedings, there was confusion on the
part of RAC committee members as to whether Life Tech-
nologies and Novartis were funding this Phase 1 clinical
trial. Dr. June confirmed that they were not funding the trial
and that the investigators needed to “clean that language up.”
Dr. June explained that the Parker Institute for Cancer Im-
munotherapy was the funding sponsor. There was no unam-
biguous disclosure statement, however, addressing the rela-
tionship between Life Technologies, Novartis, and the Parker
Institute for Cancer Immunotherapy. As such, details regard-
ing potential financial conflicts of interest resulting from
relationships with research sponsors were not fully addressed
during the RAC proceedings.
As well, Dr. June did not provide information regarding
registered patents (or filed patent applications) related to
experimental agents to be utilized in the proposed Phase 1
clinical trial. Nor did he divulge to the RAC the exact na-
ture of the “significant financial benefit” he receives as a
result of his invention. An independent search reveals that
Dr. June is the inventor of a significant number of patents
registered with the United States Patent and Trademark s
Office in the field of methods and compositions for treat-
ment of cancer in humans, including the administration of
genetically modified T cells [58, 59]. Moreover, many of
these registered patents are assigned to the global
healthcare company Novartis AG.
Further, there was no disclosure to the RAC about
whether Dr. June has equity, stock options or comparable
arrangements in companies sponsoring or funding the trial,
whether he receives gifts, ongoing consultation fees or hono-
raria, or whether he will be paid with respect to research par-
ticipant recruitment. As well, there was no mention of any
mitigation str ategies that might be in place.
In an effort to address patent-related conflicts of interest,
Kimmelman has recommended use of restrictive policies
requiring: (1) disclosure of registered patents on experimen-
tal agents held by investigators to IRBs and research partici-
pants; (2) curtailed responsibilities for investigators who
hold registered patent interests in an experimental agent (up
to and including being presumptively barred from certain
activities in the clinical trial); (3) management (including
disclosure) of institutional interests in registered p atents on
experimental agents; and (4) disclosure to IRBs of all patent
filings related to the experimental agent. The stringency of
such policies are to be adjusted according to the trial and the
patent [60].
Following on these recommendations, it would have been
prudent for Dr. June to voluntarily curtail his responsibilities
as scientific advisor of this Phase 1 clinical trial. In the alter-
native, the University of Pennsylvania could have presump-
tively barred Dr. June from certain activities such as patient
interactions and selection, study design, data analysis, and
other activities. Instead, the University of Pennsylvania Con-
flict of Interest Standing Committee has yet to determine
whether Dr. June can participate in his proposed capacity, or
in the alternative, if a management plan should be issued and
agreed to by Dr. June.
In addition to individual conflicts of interest, there may
be institutional conflicts of interest. Institutional conflicts of
interest may arise when a financial interest of an employee
of an institution , or of the institution itself, could affect or
appear to affect the conduct, review, or oversight of research
in ways that are potentially harmful to the obligations, the
mission, or the values of the institution [61]. The draft con-
sent form presented to the RAC stated that the University of
Pennsylvania had a significant financial interest in the tech-
nologies being evaluated in the proposed Phase 1 clinical
trial. As such, if these technologies were proven safe and
effective, the University of Pennsylvania would financially
benefit. The RAC was not provided with detailed informa-
tion about institutional interests, financial or otherwise, in
registered patents related to the experimental agent, or any
mitigation str ategies that migh t be in place.
4.3. Publication Bias
Publications that demonstrate a strong relationship be-
tween small sample sizes and larg e treatment effect or that
withhold negative results may be a frequent occurrence in
pre-clinical efficacy studies [e.g. 14: pp.110-131; 62].
These forms of publication bias, which distort the efficacy
of pre-clinical interventions and manif est a lack of trans-
parency, invariably result in an inability to reproduce scien-
tific results. This not only complicates interpretation of the
medical literature, it also violates the ethical obligation to
only involve persons in research that contributes to knowl-
edge [62].
According to the investigators, the one pre-clinical study
in a small number of immunodeficient NSG mice (used to
justify the first-in-human gene editing cancer trial) showed a
large treatment effect and proved to be safe and efficacious.
316 Current Gene Therapy, 2017, Vol. 17, No. 4 Baylis and McLeod
This study has not yet been published in a peer reviewed
journal, however. Peer review would have allowed for an
independent scientific assessment of the relevant data. It is
widely recommended that the publication of pre-clinical
studies follow the Animal Research Reporting In Vivo Ex-
periments (ARRIVE) criteria [63: p.19].
4.4. Summary of Concerns Regarding Credibility
Optimism bias, potential financial conflicts, competing
interests between investigators, institutions and pharmaceuti-
cal companies, as well as a publication bias, raise serious
concerns regarding the credibility of the pre-clinical study
that provided an evidentiary basis for the move to the first-
in-human Phase 1 gene editing cancer trial.
5. DISCUSSION
As noted at the outset, there are those who wonder (our-
selves included) whether the proposed first-in-human Phase
1 CRISPR gene editing cancer trials are premature [64]. In
an effort to address this issu e we cr itically examined the
quality of the pre-clinical evidence used to justify the first-
in-human Phase 1 CRISPR gene ed iting cancer trial in the
United States using available evidence presented to the RAC.
Our analysis relied heavily on the four-part framework de-
veloped by Kimmelman for assessing translational distance
between pre-clin ical and clinical research and on the struc-
tured process subsequently developed by Henderson and
Kimmelman for evaluating pre-clinical evidence in terms of
potential threats to internal validity, constru ct validity, and
external validity. In our estimation the scarcity and poor
quality of pre-clinical evidence and the novelty of NY-ESO-
1 Redirected CRISPR Edited T Cells warrants further study
prior to any first-in-human trial.
Our analysis of internal validity suggests problems with
methodological elements of the pre-clinical study including
the small sample size, and the lack of a-priori power calcula-
tions, randomized allocation of animals to treatment, blind-
ing of outcome assessment, dose-response relationships, and
selection of appropriate control groups.
Construct validity is threatened as the pre-clinical study
did not validate the hypothesis to be tested in the first-in-
human Phase 1 clinical trial. The investigators should have
tested the efficacy of lentiviral transduction and gene editing
in T cells from cancer patients (including perhaps prospec-
tive research participants) with melanoma, synovial sarcoma,
or multiple myeloma and thereafter used the same T cells in
the pre-clinical animal study to determine the two study end-
points of (1) tumor size and (2) percent survival. Conse-
quently, not all of the features of the pre-clinical study were
held constant in the move from pre-clinical research to the
Phase 1 clinical trial. There is no principled justification for
not proceeding with an expanded pre-clinical analysis in-
volving more than two cell sources and using cells that were
from the target populations, namely, patients with mela-
noma, synovial sarcoma, or multiple myeloma, when testing
the efficacy of the lentiviral transduction and gene editing.
The potential heterogeneity of the cell populations further
negates the possibility of robust causal inferences about the
safety and efficacy of any research intervention.
External validity analysis indicates that limitations are
further compounded by the use of no animal models other
than mice and no independent replication. The fallacy of
drawing inferences from (only) one pre-clinical study in
mice to justify the move to a first-in-human clinical trial
cannot be over-emphasized. Addressing these limitations
would have been of potential scientific value and could use-
fully have informed clinical trial design.
Kimmelman has stated that the most effective way to
minimize translational distance (i.e. to ensure the narrowest
possible inferential gap) between pre-clinical research and
clinical research is to have all features of the relevant pre-
clinical studies held constant in the move to first-in-human
trials. Importantly, our analysis highlights significant prob-
lems with internal validity, construct validity, and external
validity as these relate to features of the one pre-clinical
study in mice. As regards the proposed first-in-human gene
editing cancer trial in the United States, in an ideal contex t,
the pre-clinical research should have proceeded in the fol-
lowing step wise fashion. There should have been research to
transduce and gene edit T cells from healthy donors, fol-
lowed by research using a similar or improved protocol to
transduce and gene edit T cells from cancer patients (ideally
patients with melanoma, synovial sarcoma, or multiple mye-
loma). Thereafter , transduced and gene edited T cells from
cancer patients should have been cell sorted to obtain a ho-
mologous population of T cells expressing NY-ESO-1 TCR
with endogenous TCR α, TCR β and PD1 disruptions. Next,
this homologous population of T cells from cancer patients
should have been introduced into a mouse model that had
either melanoma, synovial sarcoma, or multiple myeloma
like tumors in an anatomical location that mimics the human
condition in order to assess safety and efficacy. In turn, this
should have been followed by similar research in other (in-
cluding larger) animal models. As best we can tell, these
steps were not followed prior to seeking RAC approval to
initiate the first-in-human Phase 1 CRISPR gene editing can-
cer trial.
Finally, an appraisal of credibility brings into question
the trustworthiness of the publicly available information. For
example, small sample size and a missing data point in the
analysis of tumor size may indicate an optimism bias. Poten-
tial conflicts of interest cannot be overlooked given the sig-
nificant individual and institutional financial interests related
to patents, licensing arrangements with industry and mile-
stone payments. Finally, a publication bias arises due to the
small sample size and the reported large treatment effect in
the pre-clinical study, as well as the lack of peer review.
As such, all that can be concluded from the one pre-
clinical study in mice is that the investigators’ methods were
lacking in scientific rigor and did not meet the ethical re-
quirement of scientific validity.
CONCLUSION
The myriad problems of scientific validity with the pre-
clinical study presented to the RAC in support of the move to
a first-in-human Phase 1 CRISPR gene editing clinical trial
for cancer is concerning. This arguably calls into the ques-
tion the confidence expressed by international groups and
organizations in existing governance mechanisms for human
somatic cell gene editing.
First-in-Human Phase 1 C RISPR Ge ne Editing Cancer Trials Current Gene Therapy, 2017, Vol. 17, No. 4 317
For example, in December 2015, the organizing commit-
tee of the first internation al summit on human gene editing
(sponsored by the United States National Academy of Sci-
ence, the United States National Academy of Medicine, the
Royal Society and the Chinese Academy of Science) re-
leased On Human Gene Editing: International Summit
Statement at the close of the meeting [65] [F.B. was a co-
author]. This statement unequivocally endorsed the view that
regulators are capable, on the basis of past experience with
research involving gene transfer, to weigh the risks and po-
tential benefits of clinical trials involving somatic cell gene
editing:
“[b]ecause proposed clinical uses are intended to affect
only the individual who receives them, they can be appropri-
ately and rigorously evaluated within existing and evolving
regulatory frameworks for gene therapy, and regulators can
weigh risks and potential benefits in approving clinical trials
and therapies.” [Emphasis added] [65: conclusion no.2]
Similarly, fourteen months later, the February 2017 re-
port Human Genome Editing: Science, Ethics and Govern-
ance [66] authored by the United States National Academy
of Science and the United States National Academy of
Medicine assuredly concluded:
“… that clinical trials of genome editing in somatic cells
for the treatment or prevention of disease or disability
should continue, subject to the ethical norms and regulatory
frameworks that have been developed for existing somatic
gene therapy research and clinical use to treat or prevent
disease and disability.” [Emphasis added] [67: p.2]
This conclusion is formalized as recommendation 4-1:
“Existing regulatory infrastructure and processes for re-
viewing and evaluating somatic gene therapy to treat or pre-
vent disease and disability should be used to evaluate so-
matic gene therapy that uses genome editing.” [66: p.61]
Our analysis, however, suggests that confidence in existing
regulatory processes may be misplaced.
One possible explanation for the misplaced confidence in
existing regulatory processes has to do with context. Those
responsible for the interpretation and application of the regu-
latory framework in the United States, no doubt, are mindful
of the fact that Chinese researchers have a head start. Their
first-in-human gene editing trial for cancer is well underway,
the first patient having been enrolled in October of 2016. In
the Fall of 2016, Dr. June, the scientific adviser for the trial
in the United States, said explicitly “I think this is going to
trigger ‘Sputnik 2.0’, a biomedical duel on progress between
China and the United States” [64]. This prediction, not sur-
prisingly, made the headlines [68, 69]. Although Dr. June
believes that this form of competition “usually improves the
end product”, the pivotal ethical question is “at what cost?”
If the move to first-in-human clinical trials is premature, as
our analysis suggests, then trial participants will have been
put at risk for no potential benefit.
Here we side with the American geneticist and Nobel
Laureate Hermann Joseph Muller who believed that individ-
ual countries should not be encouraged to develop genetic or
biomedical sputniks. In a 1957 address at the University of
New Hampshire entitled, "Man's Responsibility for His Ge-
netic Heritage" [70], Muller hopefully asserted,
“Fortunately, men will in all probability have joined into
one world community before these techniques come into
widespread use. For if the people of one nation were to ap-
ply them intelligently and extensively even a few decades
before the rest of the world did so, they would be able soon
afterwards to rise to a so much higher level of capability as
to make them virtually invincible. The world cannot afford to
allow to individual countries their separate genetic sput-
niks!” [as reprinted in: 71: p.3]
As for the proposed first-in-human Phase 1 CRISPR gene
editing cancer trial in the United States, we side with Mil-
dred Cho – one of the RAC reviewers for this trial. Cho as-
serts in her review:
“… it seems to me that the novelty of triple editing to dis-
rupt endogenous TCR and PD-1 in combination with adop-
tive T cell transfer warrants more animal studies in order to
better anticipate the effects of PD-1 disruption and of the
bulk transfer/“pick the winner” strategy.” [72: p.12]
Subsequently, in a commentary on this proposed re-
search, Cho is reported to have said that there is only so
much investigators can learn from animal models and that at
some point “we have to take a leap of faith” [cited in: 73].
The critical issue, however, is how far should investigators
be willing to jump… The smaller the distance, the greater the
likelihood of a safe landing. This explains the ethical obliga-
tion imposed on investigators to do everything reasonably
possible to narrow the inferential gap between what is known
from pre-clinical research and what is proposed for first-in-
human clinical trials. Initiating first-in-human CRISPR gene
editing cancer trials is an important step on the path to de-
veloping safe and effective preventive and therapeutic inter-
ventions for current and future patients. It is not a step to be
taken lightly. The hoped-for knowledge is not to be obtained
at any and all costs to current patients who consent to be-
come research particip ants.
In our view, the move to first-in-human Phase 1 CRISPR
gene editing cancer trials in the United States, on the basis of
pre-clinical evidence presented to the RAC, is premature
insofar as it makes the leap of faith a leap too far. Moreover,
this leap cannot be justified by claims of urgent medical
need.
ETHICS APPROVAL AND CONSENT TO PARTICI-
PATE
Not applicable.
HUMAN AND ANIMAL RIGHTS
No Animals/Humans were used for studies that are base
of this research.
CONSENT FOR PUBLICATION
Not applicable.
CONFLICT OF INTEREST
The authors declare no conflict of interest, financial or
otherwise.
318 Current Gene Therapy, 2017, Vol. 17, No. 4 Baylis and McLeod
ACKNOWLEDGEMENTS
We thank members of the Novel Tech Ethics research
team at Dalhousie University for helpful comments during
preparation of this manuscript. As well, specific thanks are
owed to Jonathan Kimmelman for his feedback on the penul-
timate draft and to Tim Krahn for his diligent assistance with
preparation of the manuscript for publication.
REFERENCES
[1] National Institutes of Health. CPD-1 knockout engineered T cells
for metastatic non-small cell lung cancer. ClinicalTrials.gov 2016.
Available from: https://clinicaltrials.gov/ct2/show/NCT02793856?
term=crispr&rank=4 [cited: 14 June 2017].
[2] Phase I trial of NY-ESO-1 redirected CRISPR edited T cells
(NYCE CELLS) engineered to express NY-ESO-1 TCR and gene
edited to eliminate endogenous TCR and PD-1, Protocol # 1604-
1524. NIH OBA RAC Public Review 2016. Available from:
https://osp.od.nih.gov/calendar-3/action~agenda/exact_date~ 21-6-
2016/cat_ids~21/ [cited: 23 June 2017].
[3] Emanuel EJ, Wendler D, Grady C. What makes clinical research
ethical? JAMA 2000; 283(20): 2701-11.
[4] Borgerson K. Redundant, secretive, and isolated: When are clinical
trials scientifically valid? Kennedy Inst Ethics J 2014; 24(4): 394-
411.
[5] United States Food & Drug Administration. Step 3: Clinical re-
search. United States Department of Health and Human Services
2017. Available from: https://www.fda.gov/forpatients/ approv-
als/drugs/ucm405622.htm [cited: 14 June 2017].
[6] Lo B, Zettler P, Cedars MI et al. A new era in the ethics of human
embryonic stem cell research. Stem Cells 2005; 23(10): 1454-1459.
[7] United States Food & Drug Administration. Considerations for the
design of early-phase clinical trials of cellular and gene therapy
products, guidance for industry. United States Department of
Health and Human Services, Food and Drug Administration, Cen-
ter for Biologics Evaluation and Research 2015. Available from:
https://www.fda.gov/downloads/BiologicsBloodVaccines/Guidance
ComplianceRegulatoryInformation/Guidances/CellularandGene
Therapy/UCM564952.pdf [cited: 14 June 2017].
[8] Heemskerk MH. T-cell receptor gene transfer for the treatment of
leukemia and other tumors. Haematologica 2010; 95(1): 15-9.
[9] van Loenen MM, de Boer R, Amir AL et al. Mixed T cell receptor
dimers harbor potentially harmful neoreactivity. Proc Natl Acad
Sci USA 2010; 107(24): 10972-7.
[10] Rapoport AP, Stadtmauer EA, Binder-Scholl GK et al. NY-ESO-1-
specific TCR-engineered T cells mediate sustained antigen-specific
antitumor effects in myeloma. Nat Med 2015; 21(8): 914-21.
[11] Postow MA, Callahan MK, Wolchok JD. Immune checkpoint
blockade in cancer therapy. JCO 2015; 33(17): 1974-82.
[12] Robbins PF, Morgan RA, Feldman SA et al. Tumor regression in
patients with metastatic synovial cell sarcoma and melanoma using
genetically engineered lymphocytes reactive with NY-ESO-1. JCO
2011; 29(7): 917-24.
[13] Schietinger A, Greenberg PD. Tolerance and exhaustion: defining
mechanisms of T cell dysfunction. Trends Immunol 2014; 35(2):
51-60.
[14] Kimmelman J. Gene transfer and the ethics of first-in-human re-
search: lost in translation. Cambridge, Cambridge University Press,
2009.
[15] Henderson VC, Kimmelman J, Fergusson D, Grimshaw JM,
Hackam DG. Threats to validity in the design and conduct of pre-
clinical efficacy studies: A systematic review of guidelines for in
vivo animal experiments. PLOS Med 2013; 10(7): e1001489.
[16] Kimmelman J, Henderson V. Assessing risk/benefit for trials using
preclinical evidence: a proposal. J Med Ethics 2015; 42(1): 50.
[17] Kimmelman J, London AJ. Predicting harms and benefits in trans-
lational trials: ethics, evidence, and uncertainty. PLOS Med 2011;
8(3): e1001010.
[18] Aparicio S, Hidalgo M, Kung AL. Examining the utility of patient-
derived xenograft mouse models. Nat Rev Cancer 2015; 15(5):
311-6.
[19] Garrido-Laguna I, Uson M, Rajeshkumar NV et al. Tumor en-
graftment in nude mice and enrichment in stroma-related gene
pathways predict poor survival and resistance to gemcitabine in pa-
tients with pancreatic cancer. Clin Cancer Res 2011; 17(17): 5793.
[20] Eirew P, Steif A, Khattra J et al. Dynamics of genomic clones in
breast cancer patient xenografts at single-cell resolution. Nature
2015; 518(7539): 422-426.
[21] DeRose YS, Wang G, Lin YC et al. Tumor grafts derived from
women with breast cancer authentically reflect tumor pathology,
growth, metastasis and disease outcomes. Nat Med 2011; 17(11):
1514-1520.
[22] Julien S, Merino-Trigo A, Lacroix L et al. Characterization of a
large panel of patient-derived tumor xenografts representing the
clinical heterogeneity of human colorectal cancer. Clin Cancer Res
2012; 18(19): 5314.
[23] Hidalgo M, Amant F, Biankin AV et al. Patient-derived xenograft
models: an emerging platform for translational cancer research.
Cancer Discov 2014; 4(9): 998.
[24] Ding L, Ellis MJ, Li S et al. Genome remodelling in a basal-like
breast cancer metastasis and xenograft. Nature 2010; 464(7291):
999-1005.
[25] Bertotti A, Migliardi G, Galim i F et al. A molecularly annotated
platform of patient-derived xenografts ("xenopatients") identifies
HER2 as an effective therapeutic target in cetuximab-resistant colo-
rectal cancer. Cancer Discov 2011; 1(6): 508.
[26] Marangoni E, Vincent-Salomon A, Auger N et al. A new model of
patient tumor-derived breast cancer xenografts for preclinical as-
says. Clin Cancer Res 2007; 13(13): 3989.
[27] Fichtner I, Rolff J, Soong R et al. Establishment of patient-derived
non-small cell lung cancer xenografts as models for the identifica-
tion of predictive biomarkers. Clin Cancer Res 2008; 14(20): 6456.
[28] Sivanand S, Peña-Llopis S, Zhao H et al. A validated tumorgraft
model reveals activity of dovitinib against renal cell carcinoma. Sci
Transl Med 2012; 4(137): 137ra75.
[29] Zhang X, Claerhout S, Prat A et al. A renewable tissue resource of
phenotypically stable, biologically and ethnically diverse, patient-
derived human breast cancer xenograft models. Cancer Res 2013;
73(15): 4885.
[30] Reyal F, Guyader C, Decraene C et al. Molecular profiling of pa-
tient-derived breast cancer xenografts. Breast Cancer Res 2012;
14(1): R11.
[31] Hidalgo M, Bruckheimer E, Rajeshkumar NV et al. A pilot clinical
study of treatment guided by personalized tumorgrafts in patients
with advanced cancer. Mol Cancer Ther 2011; 10(8): 1311.
[32] Morelli MP, Calvo E, Ordoñez E et al. Prioritizing phase I treat-
ment options through preclinical testing on personalized tumor-
graft. J Clin Oncol 2012; 30(4): e45-e48.
[33] Fazakas C, Wilhelm I, Nagyõszi P et al. Transmigration of mela-
noma cells through the blood-brain barrier: Role of endothelial
tight junctions and melanoma-released serine proteases. PLoS One
2011; 6(6): e20758.
[34] Cohen JV, Kluger HM. Systemic immunotherapy for the treatment
of brain metastases. Front Oncol 2016; 6: 49.
[35] Sentman ML, Murad JM, Cook WJ et al. Mechanisms of acute
toxicity in NKG2D chimeric antigen receptor T cell-treated mice. J
Immunol 2016; 197(12): 4674.
[36] Maude SL, Barrett D, Teachey DT, Grupp SA. Managing cytokine
release syndrome associated with novel T cell-engaging therapies.
Cancer J 2014; 20(2): 119-122.
[37] Hu Y, Sun J, Wu Z et al. Predominant cerebral cytokine release
syndrome in CD19-directed chimeric antigen receptor-modified T
cell therapy. J Hematol Oncol 2016; 9(1): 70.
[38] Editorial. Troublesome variability in mouse studies. Nat Neurosci
2009; 12(9): 1075.
[39] Garner JP. The significance of meaning: Why do over 90% of
behavioral neuroscience results fail to translate to humans, and
what can we do to fix it? ILAR J 2014; 55(3): 438-456.
[40] Mak IW, Evaniew N, Ghert M. Lost in translation: animal models
and clinical trials in cancer treatment. Am J Transl Res 2014; 6(2):
114-118.
[41] Seok J, Warren HS, Cuenca AG et al. Genomic responses in mouse
models poorly mimic human inflammatory diseases. Proc Natl
Acad Sci USA 2013; 110(9): 3507-12.
[42] Schuh JCL. Trials, tribulations, and trends in tumor modeling in
mice. Toxicol Pathol 2004; 32(1 Suppl): 53-66.
[43] Gawrylewski A. The trou ble with anim al mode ls. Th e Sc ien tist 1 July
2007. Available from: http://www.the-scientist.com/?articles. view/
First-in-Human Phase 1 C RISPR Ge ne Editing Cancer Trials Current Gene Therapy, 2017, Vol. 17, No. 4 319
articleNo/25184/title/The-Trouble-with-Animal-Models/ [cited: 14
June 2017].
[44] Hayden EC. Misleading mouse studies waste medical resources.
Nature News 2014. Available from: http:// www.nature.com/news/
misleading-mouse-studies-waste-medical-resources-1.14938 [cited:
14 June 2017].
[45] Begley CG, Ellis LM. Drug development: Raise stand ards for
preclinical cancer research. Nature 2012; 483(7391): 531-3.
[46] Prinz F, Schlange T, Asadullah K. Believe it or not: how much can
we rely on published data on potential drug targets? Nat Rev Drug
Discov 2011; 10(9): 712.
[47] Paoloni M, Khanna C. Translation of new cancer treatments from
pet dogs to humans. Nat Rev Cancer 2008; 8(2): 147-56.
[48] Park JS, Withers SS, Modiano JF et al. Canine cancer immunother-
apy studies: linking mouse and human. J Immunother Cancer 2016;
4: 97.
[49] Angell M. Is academic medicine for sale? N Engl J Med 2000;
342(20): 1516-8.
[50] Sherkow JS. CRISPR: Pursuit of profit poisons collaboration. Na-
ture 2016; 532: 172-3.
[51] Policy of the American Society of Gene Therapy financial conflict
of interest in clinical research: adopted April 5 2000. American
Society of Gene & Cell Therapy 2000. Available from:
http://ccnmtl.columbia.edu/projects/rcr/rcr_conflicts/misc/Ref/AGS
T_CoI.pdf [cited: 14 June 2017].
[52] Johns ME, Barnes M, Florencio PS. Restoring balance to industry-
academia relationships in an era of institutional financial conflicts
of interest: promoting research while maintaining trust. JAMA
2003; 289(6): 741-746.
[53] Oliver AL, Sapir A. Shifts in th e organization and profession of
academic science: The impact of IPR and technology transfer. J
Professions Organization 2017; 4(1): 36-54.
[54] Wilan KH. Opening up the ivory tower. Cell 2007; 129(5): 847-
850.
[55] Katsnelson A. Cashing in on science. Nature 2016; 533(7601): S6-
S7.
[56] Xiangyu F, Yanping L, Guojian L, Xie J. Overview of patents on
targeted genome editing technologies and their implications for in-
novation and entrepreneurship education in universities. Hereditas
2015; 37(12): 1258-62.
[57] United States Food & Drug Administration. Guidance for clinical
investigators, industry, and FDA staff: Financial disclosure by
clinical investigators. United States Department of Health and Hu-
man Services February 2013. Available from: https://www.fda.gov/
downloads/regulatoryinformation/guidances/ucm341008.pdf [cited:
14 June 2017].
[58] Patents by inventor Carl H. June. Justia Patents 2017. Available from:
http://patents.justia.com/inventor/carl-h-june [cited: 15 June 2017].
[59] United States Patent Collection. Results of search in US Patent
Collection db for: IN/"June; Carl H" AND ABST/cell. United
States Patent and Trademarks Office 2017. Available from:
http://bit.ly/2swiUp7 [cited: 15 June 2017].
[60] Kimmelman J. Inventors as investigators: The ethics of patents in
clinical trials. Acad Med 2007; 82(1): 24-31.
[61] Conflicts of interest, responsible conduct in research. Columbia
University 2016. Availab le from: http://ccnmtl.columbia.edu/ pro-
jects/rcr/rcr_conflicts/foundation/ [cited: 15 June 2017].
[62] Scott D. Califf's big idea: Build a database for research don e before
clinical trials. StatNews.com 10 June 2016. Available from:
https://www.statnews.com/2016/06/10/califf-database-preclinical-
trials/ [cited: 15 June 2017].
[63] ISSCR. Guidelines for stem cell research and clinical translation.
International Society for Stem Cell Research 2016; 1-37. Available
from: http://www.isscr.org/docs/default-source/guide lines/isscr-
guidelines-for -stem -cell-resea rch-and-clinical-transla tion.pdf?s fvrsn=
2 [cited: 21 June 2017].
[64] Cyranoski D. CRISPR gene-editing tested in a person for the first
time. Nature 2016; 539(7630): 479.
[65] Organizing Committee for the International Summit on Human
Gene Editing. On human gene editing: International summit state-
ment. National Academies of Scie nces, Engineering, and Medi-
cine 2015. Available from: http://www8.nationalacademies.org
/onpi news/newsitem.aspx?RecordID=12032015a [cited: 15 June
2017].
[66] National Academies of Sciences • Engineering • Medicine. Human
genome editing: science, ethics, and governance. Washington, DC:
The National Academies Press 2017. Available from: https://
www.nap.edu/catalog/24623/human-genome-editing-science-ethics
-and-governance [cited: 15 June 2017].
[67] National Academy of Sciences and National Academy of Medi-
cine. Human genome editing: science, ethics, and governance – re-
port highlights. National Academy of Sciences • Engineering •
Medicine 2017; 1-4. Available from: https://www.nap.edu/
html/24623/Human-Genome-Editing-highlights.pdf [cited: 15 June
2017].
[68] Szaniszlo M. China, U.S. in 'Sputnik' race over gene editing. Bos-
ton Herald 16 November 2016. Available from: http://
www.bostonherald.com/lifestyle/health/2016/11/china_us_in_sputn
ik_race_over_gene_editing [cited: 15 June 2017].
[69] Paladhi A. Gene editing the new 'Sputnik' race as CRISPR technique
begins clinical trial. Live Mint 2016. Available from:
http://www.livemint.com/Science/lh87EZMRRInpmAOb9 FhmMM/
Gene-editing-the-new-Sputnik-race-as-CRISPR-technique-begi.html
[cited: 15 June 20 17].
[70] Muller HJ. Man's future birthright; an address at the University of
New Hampshire, 1957. Library of Congress Online Catalog.LCCN
58003608 1958. Available from: https://catalog.loc.gov/
vwebv/search?searchCode=LCCN&searchArg=58003608&search
Type=1&permalink=y [cited: 16 June 2017].
[71] Muller HJ. In: Carlson EA Ed, Man's future birthright: essays on
science and humanity by H.J. Muller. Albany, NY, State University
of New York Press, 1973; pp.3-7.
[72] Cho M. Review NIH OBA 1604-1524 : Phase I trial o f autologous T
cells engineered to express NY-ESO -1 TCR and gene edited to
eliminate endogenous TCR and PD-1. SCRIBD: RAC Review Penn
Medicine CRISPR NYESO-1 Immunetherapy study 2017; 10-13.
Available from: https://www.scribd.com/doc/316280454/RAC-
Review-Penn-Medicine-CRISPR-NYESO-1-Immunetherapystudy
[cited: 15 June 20 17].
[73] Reardon S. First CRISPR clinical trial gets green light from US
panel. Nature News 2016. Available from: http://
www.nature.com/news/first-crispr-clinical-trial-gets-green-light
from-us-panel-1.20137 [cited: 14 June 2017].
