THE GORDON WILSON LECTURE: THE ETHICS OF HUMAN GENOME EDITING

Abstract

Human genome editing has undergone major technological advances, raising the possibility of treating or preventing many illnesses. Somatic (nonheritable) genome editing, both in vitro and in vivo, is already being employed under a robust regulatory and ethical framework developed for human gene therapy. In contrast, the prospect of germline (heritable) genome editing is much more contentious, and there is currently no consensus on the proper path forward. The 2017 National Academy of Sciences (NAS) and National Academy of Medicine (NAM) report proposed a series of requirements designed to minimize ethical objections while allowing couples to accept the risks of genome editing in order to have a biologically related child without passing on a known genetic disorder. It is vital to prevent gene editing from resulting in unintended negative consequences for individuals with genetic variants. The utilization of genome editing to enhance human function is highly contentious; it may be better to focus on whether an edit creates an “unfair advantage” rather than an enhancement.

INTRODUCTION

I am honored to deliver the 2019 Gordon Wilson lecture, and I enjoyed reading about his important role in the history of this society in Dr. A. McGee Harvey's history of the society (1). I will divide my presentation into several sections that will cover (1) the technology of human genome editing, (2) potential applications, (3) governance, and (4) ethical concerns. I will conclude by discussing the language of human genome editing and the issue of “enhancement.”

THE TECHNOLOGY OF HUMAN GENOME EDITING

We now have the ability to selectively target just one out of the three billion base pairs that make up our genome (2,3). If a double-strand break is made at that point, it will likely result in gene inactivation as a result of nonhomologous end joining upsetting the sequence. If a donor DNA is supplied with homology to the region, the sequence can be altered.

A variety of different agents can accomplish the targeted double-strand break, including, for example, zinc finger nucleases, transcription activator-like effector nucleases (TALENS), and most recently the clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 system (4). It took enormous human ingenuity to design and perfect the first two technologies, whereas evolution had already solved this problem millions of years ago as bacteria developed methods to ward off the attack of phage viruses (3). The specificity of the combined tracker and guide RNAs now used in the CRISPR/Cas9 system, and the relative ease of their synthesis, has dramatically simplified genome editing (2). That said, the underlying biology and biochemistry of DNA repair after a double-strand break is complex, and improving our understanding of the molecular mechanisms may result in greater efficiency and specificity (5).

In addition to gene inactivation and modification initiated by a double-strand break in the DNA, the CRISPR/Cas9 system can now be used in other ways. For example, if Cas9's enzymatic activity is inactivated while its targeting ability is preserved, it can be used to ferry other molecules to the site. This has been exploited to perform base editing, in which an enzyme directly changes one DNA base to another (6,7). We do not yet have the ability to make all potential base edits, but a sizeable percentage of gene variants associated with disease are theoretically amenable to this approach. The expression of a gene can also be turned on or off (8). It may even be possible to use inactivated Cas9 to perform “search and replace” prime editing without double-strand breaks or donor DNA (9).

POTENTIAL APPLICATIONS

In theory, human genome editing can be used to treat or prevent a serious disease or disability, extend human life, enhance human function, or maybe even perfect human nature. In considering potential applications of genome editing, historically one of the major divisions has been between somatic (nonheritable) versus germline (heritable) gene editing. The former involves editing cells other than germ cells (eggs, sperm, and early-stage embryos), whereas the latter involves editing one or another germ cell. Somatic edits are not, therefore, passed on to subsequent generations, whereas germline edits may be passed on to subsequent generations.

Somatic genome editing can be further subdivided into ex vivo genome editing and in vivo genome editing (10). Ex vivo genome editing involves removing cells from the individual's body, making the edit, and then returning the edited cells to the individual. Bone marrow cells lend themselves to this approach because there are established methods for harvesting these cells, modifying them, and then returning them to the same, or even a different, person in a way that allows them to partially or completely repopulate the bone marrow. Thus, prime candidates for this type of human genome editing are hematologic disorders such as sickle cell anemia and thalassemia (11). Many other genetic disorders fit into this category (12). Alternatively, genome editing can be used to arm a blood cell to attack a tumor cell (13). In addition, although perhaps counterintuitive, inactivating a gene may also be valuable: one example would be inactivating a gene that reduces fetal hemoglobin after birth, resulting in higher levels of fetal hemoglobin in children and adults, which ameliorates the severity of sickle cell disease (14). Inactivating one of the co-receptors for HIV, CCR5, in CD4 positive T-cell lymphocytes may also offer some protection against contracting HIV and may ameliorate the progression of the disease in those who are already infected (15). Data from extensive studies using this strategy at the University of Pennsylvania demonstrate that the procedure is well tolerated; in some HIV-infected patients whose virus utilizes CCR5 as a co-receptor, it results in both long-term survival of the edited cells and a decrease in the viral set point during drug treatment interruption (16). Oversight of the conduct and ethics of these studies has been extensive (16-20), including review by the National Institutes of Health (NIH) and the Food and Drug Administration (FDA) at the federal level and multiple school of medicine and university committees at the local level. A recent study of an HIV positive patient who developed acute leukemia in China used genome editing of the CCR5 receptor in the donor cells used to reconstitute the patient's bone marrow after receiving marrow ablative therapy (21). While the benefits were only modest, the therapy appeared to be safe.

One of the CRISPR technology's strengths is the ability to modify multiple genes at the same time by utilizing a combination of guide RNAs directed at different sites on the genome along with the Cas9 protein. The University of Pennsylvania group has exploited this opportunity to arm a patient's T-cells to attack tumor cells (16). Specifically, they first use CRISPR to inactivate three genes in the T-cells. The first two remove the patient's own T-cell receptor, to reduce the likelihood of producing autoimmunity and to avoid competition in targeting the tumor. The third inactivates PD-1, a “checkpoint” inhibitor that can prevent a T-cell from attacking a cancer cell. Targeting the tumor is accomplished by adding a lentivirus to insert a T-cell receptor designed to target an antigen, NY-ESO-1, expressed by the tumor cells of ∼30% of multiple myeloma patients. The group recently reported that they have safely treated three patients and that after infusion the edited cells showed in vivo expansion and stable persistence (16).

In vivo genome editing, in which the editing machinery is delivered directly to the cells of interest, eliminates the need for ex vivo manipulation, but delivery to the proper cells can be challenging (10). Potentially, delivery to liver cells could treat hemophilia due to factor VIII or IX deficiency; delivery to lung cells could treat cystic fibrosis; delivery to muscle cells could treat muscular dystrophy; and delivery to brain cells could treat Huntington's disease. In the last case, the editing could be used to inactivate the disease-associated gene rather than modifying a sequence to restore gene function.

Companies are now working hard to advance the new genome editing technology. For example, the company Sangamo recently reported that it is applying ex vivo genome editing to treat sickle cell anemia and beta thalassemia and in vivo genome editing to treat two forms of mucopolysaccharide metabolism disorders, as well as hemophilia B. In addition, it is applying similar in vivo methods to modify gene regulation to treat Huntington's disease (22).

In view of the broad applicability of somatic genome editing and the robust structure of regulatory oversight that has been developed to ensure that the benefit-risk ratio is favorable (18-20,23-25), why should the performance of germline gene editing, which raises many concerns, even be considered? For many couples considering having a child, avoiding the transmission of a genetic disorder can be achieved by combining in vitro fertilization with pre-implantation genetic diagnosis (PGD) (26). This technique involves genetic analysis of a single cell at the blastomere stage or multiple trophectoderm cells at the blastocyst state, followed by embryo selection prior to uterine implantation (27). PGD has been used successfully for many couples to avoid transmitting a genetic disorder, but it requires harvesting eggs, a medical procedure with risks, and it is expensive, time consuming, and inconvenient. Moreover, more than 50% of embryo transfers do not result in a live birth. Performing germline genome editing on an egg or embryo would have these same risks and uncertainties, in addition to those related to the editing procedure itself. It theoretically would, however, increase the percentage of candidate eggs or embryos relative to PGD because eggs with the affected gene variant would be edited rather than discarded.

Even if PGD was a perfect method, it would not help a couple prevent transmission of a genetic disorder in certain situations. This occurs, for example, if both parents are homozygous for a recessive disorder or if one of the parents is homozygous for a dominant disorder. With ∼3,000 genetic disorders now being catalogued (12,28), and with individuals having these disorders increasingly interacting with each other through advocacy groups and social media, this currently small group of couples is likely to expand. Germline genome editing has the potential to help such couples have a biologically related child without transmitting the genetic disorder.

One can also contemplate the potential benefits of inactivating some genes in an attempt to reduce the risk of contracting a disease. For example, inactivating the HIV cofactor receptor CCR5 may reduce the risk of contracting HIV or lessen its severity (29); inactivating a PCSK9 gene will likely reduce one's cholesterol level and protect against developing cardiovascular disease (30); inactivating the ApoE4 gene may reduce one's risk of developing Alzheimer's disease (31); inactivating any one of three genes involved in renal sodium metabolism (SCL12A3, SLC12A1, or KCNJ1) will likely result in a reduction in blood pressure and decrease the risk of developing cardiovascular disease (32); and inactivating the G1 and G2 variants of the gene ApoL1 may decrease the risk of developing renal disease in African Americans (33).

GOVERNANCE

In the United States, the FDA has regulatory authority over human genome editing. The FDA recently reported that it received its first submissions for genome editing products in 2008 and that it was reviewing 26 INDs, 34 Pre-INDs, and 16 Pre-pre-INDs under its INTERACT initiative (34). The FDA has begun to issue guidance on how it will perform benefit-risk analysis, which will focus on the potential to correct causes of disease; the risk of unintended (“off-target”) genome modifications, which may include edits at unintended sites or deletions or insertions of DNA sequences; and the potential unknown long-term effects of both “on-target” and “off-target” genome editing (24). An additional concern when performing ex vivo genome editing is whether there will be selection for cells that have alterations in P53, a gene that is important in preventing the growth of tumor cells, because cells lacking normal P53 may tolerate the CRISPR system better than normal cells (35). There is also concern about double-strand breaks leading to large DNA deletion and complex rearrangements (36). Thus, despite dramatic advances in genome editing technology, many aspects still require further optimization, and there is the related need for precise and highly reproducible technology to identify unexpected effects.

ETHICAL CONCERNS

A number of objections to germline human genome editing have been raised (see summary in Table 1). In response to these concerns, groups from many countries have provided abundant recommendations about the proper governance and management of ethical concerns regarding germline genome editing. A recent review summarized 61 such reports from more than 50 countries (37). Of these, 54% expressly stated that germline genome editing should not be permitted under any circumstances, 11% indicated that while they opposed it now, they were open to the possibility of approving it in the future, and 5% expressed an openness to further exploration; the remaining reports did not express an explicit view. Twenty-nine countries have, in fact, signed the Oviedo Convention and enacted laws making it a criminal offense to perform germline genome editing (38).

TABLE 1.

Concerns Common to All Applications

Concerns Specific to Applications Other Than Preventing Transmission of Genetic Variants Known to Be Associated with Serious Illness or Disability

Disrespect of DNA as human heritage Challenging God's role in creation Lack of informed consent by the child and future generations affected by the editing Negative impact on individuals with disabilities related to genetic variants Perceptions of parental negligence for deciding against performing genome editing

Commodification of children Creation of social pressure to modify children to maintain a level playing field with children modified by other parents Exacerbation of social inequality based on access to technology Unknown and unpredictable risks of creating novel genome modifications Potential to create harm that will extend to multiple generations Potential for state-imposed eugenic applications Potential for criminal applications

Clearly, this is a subject on which individuals hold strong opinions and on which a societal consensus is lacking (39-50). Thus, all physicians have an obligation, in my view, to participate in public discussions and to educate their patients about this topic since much is at stake with regard to public policy and the role of science in society. At the same time, physicians must be prepared to listen to all other voices that are expressing unique perspectives. Those with disabilities due to genetic disorders deserve special attention because they are highly vulnerable to the impact of governmental policy in this area and the downstream impact of governmental policy on societal attitudes (51-54).

I had the privilege of serving on the Human Genome Editing Committee sponsored by the National Academy of Medicine and the National Academy of Sciences from 2015 to 2017 (25). The committee worked for more than a year reviewing the relevant literature and meeting with a broad range of potential stakeholders, including patient advocacy groups, clinicians, researchers, bioethicists, policy makers, experts in sampling public opinion, members of the clergy, regulatory experts, and industry representatives. The committee included members from eight different countries, representing four continents, to ensure a wide range of international perspectives. In addition, committee members had a wide range of expertise, including law, sociology, international agreements, bioethics, public communication, and patient advocacy, in addition to medicine and basic science.

The committee concluded that it is important to respect differences in national regulations governing human genome editing based on culture and religion, but there also needs to be a core set of principles that transcend culture and religion and therefore should have universal applicability. To meet this goal, the report recommended a series of foundational principles to make up that core, namely promoting well-being, transparency, due care, responsible science, respect for persons, fairness, and translational cooperation.

The committee also recommended that basic research related to human genome editing and somatic human genome editing proceed under the regulatory framework already in existence that was developed for human gene therapy since it includes strong safeguards (17-20,55).

The committee's most controversial recommendation was to not ban germline human genome editing or call for a moratorium on its use, but rather to define a series of very stringent requirements (see list in Table 2) that must be met before research could proceed. In essence, the committee advised that the red line between somatic and germline human genome editing be moved slightly, but meaningfully, into the germline space. The recommendation to limit such research to prevent a serious disease or condition in the absence of a reasonable alternative was designed to explicitly eliminate using the technology at this time to enhance human function or select for what some people might consider favorable traits. The recommendation to limit the edit to modifying a known disease-associated sequence to one known to be prevalent in the population and also known to not be associated with disease was designed to minimize the unknowable risks associated with creating new DNA sequences as well as to minimize criticism based on the notion that humans should not be “playing God” by creating new sequences. Taken as a whole, the committee's recommended restrictions will be very challenging to meet and only time will tell whether they can ever be met.

TABLE 2.

Recommended Requirements for Potential Clinical Trials of Heritable Genome Editing by NAS/NAM Report on Human Genome Editing (25)

• Absence of reasonable alternatives; • Restriction to preventing a serious disease or condition; • Restriction to editing genes that have been convincingly demonstrated to cause or to strongly predispose to that disease or condition; • Restriction to converting such genes to versions that are prevalent in the population and are known to be associated with ordinary health with little or no evidence of adverse effects; • Availability of credible preclinical and/or clinical data on risks and potential health benefits of the procedures; • Ongoing, rigorous oversight during clinical trials of the effects of the procedure on the health and safety of the research participants; • Comprehensive plans for long-term, multigenerational follow-up that still respect personal autonomy; • Maximum transparency consistent with patient privacy; • Continued reassessment of both health and societal benefits and risks, with broad ongoing participation and input by the public; and • Reliable oversight mechanisms to prevent extension to uses other than preventing a serious disease or condition.

Returning to the objections to germline human genome editing, the requirements recommended by the committee may reduce or eliminate some, but not all of them. I would highlight the serious concern about the implications of human genome editing in Table 1, whether somatic or germline, for individuals with disabilities. In my view, it is especially important that we not compromise the many gains made in the past three decades as a result of the Americans with Disabilities Act of 1990 (51-54,56,57). The committee concluded that it would be wrong to deny a couple who wanted a healthy biologically related child access to the technology that could achieve their goal as long as they understood the experimental nature of the technology and its risks, and as long as the studies were conducted under stringent regulatory and ethical oversight (58).

The committee's report received considerable public attention. While it had critics, I would characterize the overall reception as favorable, which was in some ways remarkable given the well-established historical red line between somatic and germline human genome editing. Before the report's public release, I tried to distill its essence into a simple public statement. My recommendation was, “The goal of germline editing should be healthy babies, not designer babies.” It was therefore gratifying that the Washington Post's editorial on the report emphasized this point (59).

The recent focus on germline genome editing began in 2015 with the first report on genome editing of what was an intrinsically nonviable human embryo (60). This triggered major concern among both the scientific community and the public and led to an international summit in Washington in late 2015 and the constitution of our committee and development of its charge (25,61,62). Just before the follow-up summit in Hong Kong in late 2018, information started to spread in the scientific community that Dr. Jian-kui He had successfully edited embryos of twin girls in an attempt to modify their CCR5 receptor to diminish their risk of developing HIV infection (63). He presented his research at the Hong Kong meeting to nearly universal severe international criticism (64-68) and thereafter dropped from public sight. During his presentation, he publicly thanked his university while stating that it was unaware of his studies (63). Dr. He's rationale for editing the embryos was based on the father's having HIV infection, but there are established methods for preventing HIV transmission so his medical rationale failed to meet the criterion of not having a reasonable alternative. In fact, Dr. He's studies failed to meet any of the criteria recommended by the report. Moreover, a number of studies have raised concerns about the impact of altering CCR5, including the potential for increased risk of contracting other viral infections (69) and even alterations in brain plasticity (70). Dr. He performed his studies in China, but he received his PhD from Rice University and was a postdoctoral fellow at Stanford. As a result, his failure to perform his studies to acceptable medical, regulatory, and ethical standards reflects as much on American graduate education as on Chinese scientific oversight.

In response to Dr. He's studies, there have been calls for declaring a moratorium prohibiting germline human genome editing, creating an international registry under the auspices of the World Health Organization (WHO), and forming another international committee to make recommendations on the regulation of germline human genome editing (64-67). The regulation of in vitro fertilization across countries is an indicator of the range of options to consider. The United States has minimal regulations, whereas the United Kingdom created through legislation an elaborate regulatory and reporting structure under its Human Fertilization and Embryology Authority (HFEA) that ensures fertility clinics and research centers are properly licensed and comply with the law (71). It also collects detailed information on all procedures and provides that information to the public (https://www.hfea.gov.uk). In contrast, the U.S. Centers for Disease Control and Prevention (CDC) collects and publishes data about in vitro fertilizations as required by a 1992 act, but not all clinics report their data. Only limited data outcomes are reported, and until recently there were no official guidelines limiting which procedures can be done (72,73). In 2015, an amendment to an FDA budget authorization was passed that prevents the FDA from using any of its funds to review applications for an Investigational New Drug (IND) exemption in which an embryo's DNA is intentionally altered in a way to create a heritable genetic medication (74). Thus, for all practical purposes, human germline genome editing is banned, but not outlawed, in the United States.

A group of prominent scientific and medical leaders in the field recently called for a moratorium on germline human genome editing, which was supported by Dr. Francis Collins, the director of NIH (64,65). There is, however, controversy about the benefits and drawbacks of a moratorium (66,67). It is unclear, for example, how such a moratorium would be enforced, how and by whom the moratorium would be declared over, and what criteria would be used to declare that it was no longer needed. Dr. Alta Charo, who co-chaired the 2017 NAM/NAS report, has pointed out that there are many opportunities short of a formal moratorium for exerting pressure to prevent inappropriate use of the technology (66). These include control over intellectual property, access to the materials required to perform the edits, publication guidelines, and professional society standards.

THE LANGUAGE OF HUMAN GENOME EDITING

I personally do not think we have paid sufficient attention to the language we use to describe human genome editing. Language both reflects and directs thought, and so precision is especially important when discussing complex and emotionally charged topics. It is especially important, therefore, that those who discuss human genome editing select words that accurately represent the best scientific knowledge. The proper context for considering human genome editing language requires public appreciation that every time a cell divides it is highly likely that one or more new DNA variants will result, which means that even two skin cells in the same person are probably not identical in DNA sequence (75,76). This dynamism in our genome is intrinsic to our biology. As physicians who treat patients with genetic illnesses, we have a built-in bias against DNA alterations. From this comes our framing the process by which DNA variations occur with cell division as “errors” or “infidelity,” words with highly negative connotations. When viewed from the standpoint of an evolutionary geneticist, however, we might well consider the intrinsic rate of DNA variation not as an “error” rate, but rather as an “exploration” rate, meaning a mechanism to continue trying to better adapt to an ever-changing environment. As Dr. Thomas Lewis succinctly wrote, “The capacity to blunder slightly is the real marvel of DNA. Without this special attribute, we would still be anaerobic bacteria and there would be no music” (77).

Sickle cell anemia provides a graphic example of this point since individuals who are heterozygous for hemoglobin S are protected from dying from malaria (78,79). As a result, the DNA sequence change that first produced hemoglobin S in a malaria-endemic area was not an “error,” but rather a “benefit.” More importantly, converting the hemoglobin S DNA sequence of a heterozygous person in a malaria-endemic area isn't “correcting” a mutation—it is depriving a person of a beneficial variant. I am particularly concerned about the impact that the common use of the term “correcting” has on individuals with disabilities due to genetic variants. It is hard to escape the implication that if changing a variant sequence to the reference sequence is “correcting” something, then there must be something “wrong” with that person's sequence, and by implication, the person. An alternative and important perspective is that the variant reflects the valued and important diversity of humans.

Even the word “editing” is value-laden since it also implies “correcting” or “improving.” It will be an uphill battle to change the language we use, but we have learned that it is possible to create new gender-neutral language. I think we should try to do the same for human genome editing.

HUMAN ENHANCEMENT

The 2017 NAS/NAM report did not address in detail the complex issue of the application of human genome editing to achieve enhancement of human function because it recommended limiting the procedure to treating or preventing a serious illness or disability. My personal view is that human enhancement elicits such strong emotions because it intersects with two evolutionary forces. The first is the desire to pass along one's gene, perhaps epitomized by Richard Dawkins' concept of the selfish gene (80). In this view, we all strive to enhance our ability to compete for mates, food, and material security. At the same time, there is a second force, namely, the desire to achieve social cohesion. It is through cooperative activities that we can achieve things collectively that none of us could achieve alone. This likely was one of the great advantages we had as a species of hunter-gatherers who could communicate for hunting and other activities (81,82). Social cohesion requires enforcement of social norms, and such behavior can be identified in children as young as six years old (83). Its strength can be measured by the willingness of individuals to sacrifice their own benefit to punish those who transgress the social norms of group interactions. What emerges is a highly developed sense of what constitutes an “unfair advantage.” Experiments with six-year-olds showed that they were prepared to give up some of their own Skittles candy to prevent another child from taking advantage of a third child by unfairly distributing Skittles (83).

So how do these competing forces, focused on individual and social group aspirations, respectively, play out with regard to human genome editing? Human genetic modification to treat disease and thus enhance an individual's quality of life is broadly sanctioned and applauded. Thus, both bone marrow and kidney transplantations result in patients having tissues with genetic makeups that differ from the rest of their cells, but virtually no one views that as upsetting our DNA heritage or playing God. But there is palpable disquiet about nontherapeutic genetic modifications that enhance human capabilities beyond normal function (40,48,84,85).

The world of sports provides an example of trying to balance recognition of individual achievement while still preserving group social norms. The World Anti-Doping Code's foundational principle is, “The ethical pursuit of human excellence through dedicated perfection of each athlete's natural talents,” with “fair play and honesty,” and “respect for self and other participants” as key elements (86). Thus, sports celebrate individual excellence while refining and honing the concept of “unfair advantage” (87). Those who transgress the rules can go from hero to heel overnight, as Lance Armstrong learned (88). As a result, if germline human genome editing becomes established, I think decisions about what is and what is not allowed may be better based on whether the intervention will confer an “unfair advantage” rather than trying to assess whether it meets the definition of “enhancement.”

For many years, the bioethics community has considered genome modification based on a 2 × 2 square in which “Treatment of Disease” and “Enhancement of Capabilities” occupied one side and “Somatic” and “Germline” occupied the other (89-93). The modern reality, however, is that genetic modifications to prevent disease lie somewhere between treatment and enhancement. A few hypothetical examples illustrate this point. In particular, would the following actions be characterized as “treatment,” “prevention,” or “enhancement?” (1) Gene editing to lower the cholesterol level of a patient with severe coronary disease and three prior myocardial infarctions. (2) Gene editing of a sibling of a patient with high cholesterol who also has other risk factors for coronary artery disease. (3) Gene editing of a 21-year-old healthy son of a patient to lower cholesterol. Finally, moving further afield, what would your response be to government-dictated gene editing to lower the cholesterol of all 21-year-olds for the good of the country, so as to increase work productivity later in life and reduce health care costs?

To be sure, concerns about human genome editing for enhancement are real, with some attempts to identify the genetic basis of intelligence providing theoretical targets (94,95). Thus, it is vital for an entire society to participate in establishing the social norms that will guide the use of technology (96).

I want to conclude by saying you can join the CRISPR revolution with a do-it-yourself CRISPR kit that can be purchase online for just $159 (https://www.the-odin.com/diy-crispr-kit/). In addition, Jennifer Lopez was planning a CRISPR television series several years ago, with a new episode each week in which CRISPR is used for one or more criminal or political acts (97). Stay tuned!

DISCUSSION

Merajrer, Ann Arbor: Thank you for your talk. The Genome Editing Committee had a very interesting composition, and I noticed some of the names of people involved in the cancer field and your emphasis on healthy babies. People who carry BRCA-1, BRCA-2, and any number of cancer genes produce apparently very healthy babies; super good-looking, super cute, they nurse perfectly, they grow up fine, but they carry a 20–28% chance of developing malignancies because of incomplete penetrance. All of the known cancer syndromes have incomplete penetrance, possibly with the exception of the traditional Li Fraumeni syndrome. So your emphasis on healthy babies—how does it help the community negotiate? We are already doing preimplantation selections for BRCA-1 and BRCA-2, which are pretty widespread. How has the committee thought about that?

Coller, New York City: The committee focused on genetic variants unequivocally connected to severe diseases or disabilities rather than variants that increase risk of a disease. Going forward, the question you are raising is going to be an important one. That's why I went through the gradient between treatment prevention and enhancement, and I think that those questions must be part of the dialogue.

Garethers, Ann Arbor: Barry, that was a wonderful talk and thank you for participating and giving us that report. You know we are not only made up of our own transmitted genetics but also of our environment as well … and there has been a lot of work over the last 10 years showing that not only do genetics predict what you know may happen to us in the future but it's our environment. This includes things like the microbiome, and there's even evidence about how we inherit our microbiome and it may be transmitted from generation to generation. The microbiome can affect things like our metabolism, how fat or skinny we are, and cardiovascular disease susceptibility, which may not be genetically transmitted. Was there at least discussion in the committee on things that are outside our genetics?

Coller, New York City: No. Remember that committees are held to a statement of task and don't get to play outside of it. But I actually gave, and have continued to give, talks on the limitations of genetic determinism. A lot of funny things can happen when going from a gene to a phenotype. So I'm not a genetic determinist by any means, but we do have some compelling examples where we have very powerful single gene effects.

Feldman, Philadelphia: Barry, that was a great talk. I have a comment and a question. My comment is that Philadelphia is becoming a hotbed of gene therapy on both sides of the Schuylkill River. My colleague Kamel Khalili has recently used CRISPR/Cas9 to excise the HIV virus from infected cells both in vitro and in vivo, as recently reported in Nature. My question has to do with an area that was not discussed in your report: the ethics of using an AAV vector to drive gene expression in subjects with various levels and severities of disease. Specifically, I think an important conundrum is whether to treat sicker or less sick participants in early clinical trials. This is a tricky issue because replacement gene therapy in the heart is best tested in patients with less disease because you would expect a more robust response; however, if you get no response, you have used your “gene therapy chit” because subsequent administration of an AAV may likely result in a significant loss of vector and/or endogenous cells secondary to antigen memory in the cells of the inflammatory pathway. By contrast, if you treat an individual with extreme levels of disease, you may not be able to see signals of efficacy if the tissue is substantially encompassed by disease. I'd appreciate your thoughts regarding this therapeutic conundrum.

Coller, New York City: Needless to say, behind closed doors, there was discussion of all these issues. I think we just need to do the science. The issue of autoimmunity needs to be addressed on a gene-by-gene basis. I don't think that we have the ability at this moment to predict that, but that's why we put a whole lot about preclinical testing in the report.

Williams, Boston: Thank you so much for coming. You touched on this, but I just want to explore it a little bit more. CRISPR, the whole gene editing area as you point out, is extraordinarily complicated. On the other hand, the CRISPR revolution puts in the hands of everybody who knows how to pipette the ability to do gene editing. So the question really is: What should be done with the fact that scientists all over the world have the ability to edit the genome? What's going to be our ability as scientists to prevent that?

Coller, New York City: So, I think we get into the term “medical tourism” very quickly here … and I can tell you we were profoundly disturbed by that. That's why we have the apple pie and motherhood seven principles that you can't get out of no matter what your culture or religion is. That's why there is this international committee—even though it's a National Academy of Sciences and a National Academy of Medicine-sponsored committee. If we created the perfect oversight in the United States and there was no oversight anywhere else in the world, everybody would just move somewhere else where our regulations were not followed. We talked about treaties, and we learned that treaties are dead because they take too long to negotiate and you never know whether they are going to be ratified and enforced. Once you get into these international issues, you come up against some very profound problems. It doesn't mean that you don't have soft power, and I think that's why I kept coming back to this issue of culture. Of course, we train a lot of foreign scientists who go back to their home countries. We send them back with what we think are the appropriate cultural considerations and ethical rules for their country. That's not a trivial thing to do. I don't want to despair—I think we just need to think in a very cogent way about what it is we're doing and how to go about it.

Agarwal, Birmingham: Yes, Dr. Coller, that was an outstanding presentation. Welcome everybody, to our state. My question is: Did Dr. He, whom you mentioned, have any consequences after his presentation? Were his papers still accepted by major journals, or was he chastised by his institutions for not going through the rules and regulations?

Coller, New York City: Everything I know is hearsay, but I think he's actually under the equivalent of house arrest and has not been heard from. It isn't exactly clear. Some of you have been following this … whether he did or did not have the appropriate approvals that he said he had. I think it's probably going to be difficult for us to know exactly what is going on.