Federal Law Governing Genome Editing of Human Embryos

In the United States, the response to these promises of therapeutic certainty has incrementally, but dramatically, impacted policymaking relating to germline modifications through genome editing. In 2015, participants at International Summit on Human Gene Editing issued a proposal for a moratorium on genome editing, citing this solution as the responsible course of action.⁵⁴ This subsequently shifted into discussions of how to integrate genome edited embryos into FDA's existing regulatory framework.⁵⁵ At a Center for Biologics Evaluation and Research Board Meeting in November 2016, the FDA affirmed its jurisdiction over “efficient targeted genetic modification” for both somatic and germline genome editing applications.⁵⁶ In remarks to the 2017 FDA Science Forum, Commissioner Gottlieb proclaimed “gene editing give[s] us more opportunity to envision fundamental cures for a wider range of deadly and debilitating diseases,” building upon previous Commissioner Califf's remarks that FDA will regulate these “promising technologies” in a “science-based” manner.⁵⁷ As ethicist Francois Baylis observed, the “irresponsible” became the “permissible” within a mere fourteen months.^{Reference Baylis58}

In the United States, any clinical investigational use of genome editing on human embryos falls under the purview of the FDA under the Public Health Service Act (PHSA).⁵⁹ The FDA regulates human cell and tissue products (HCT/Ps), which refers to articles “containing or consisting of human cells or tissues that are intended for implantation, transplantation, infusion, or transfer into a human recipient.”⁶⁰ Federal regulations set forth several requirements such as registering the HCT/Ps with the FDA and promulgating standards for Good Tissue Practices, including monitoring the procedures, facilities, processing equipment, and supplies and reagents used in the manufacturing process.⁶¹ Under the HCT/P system set forth in 21 CFR §1271, the FDA classifies different types of human cells, tissues, and cellular and tissue-based products into categories for regulation based on the public health risks they pose. Unlike traditional in vitro fertilization, which FDA considers “minimal manipulation” of reproductive tissue, FDA has stated that procedures such as MRT and genome editing constitute more than minimal manipulation and will be regulated as a biological product or drug.⁶² FDA has classified genome edited embryos as a biological product or drug because the procedure is intended to prevent or treat disease.⁶³ This would require an investigator to submit an investigational new drug application with corresponding “preclinical data and information on product safety, details about technique, and proposed clinical investigation protocols.”⁶⁴

In addition to federal regulations set forth by the FDA, clinical investigation using embryos is subject to federal funding restrictions as well as state laws pertaining to research on embryos, some of which apply to genome editing.⁶⁵ At the federal level, the Dickey-Wicker Amendment prohibits the use of federal funds for research in which an embryo is created or destroyed.⁶⁶ However, some state laws as well the Dickey-Wicker Amendment contain exceptions in circumstances where the research on the embryo would provide benefit to the embryo or if the investigation is defined as therapeutic research designed to lead to gestation and birth of that embryo.⁶⁷ Repeating focused rhetoric equating genome editing to curative therapy provides a compelling narrative supporting the classification of therapeutic research. These funding restrictions apply to federally funded research, but would not prohibit privately funded research.

Finally, the Consolidated Appropriations Act of 2017 currently prohibits the FDA from using federal funds to consider applications for investigational use of a drug or biological product “in research in which a human embryo is intentionally created or modified to include heritable genetic modification.”^{Reference Castro68} Importantly, these funding restrictions would not apply to certain models of fertility tourism, where a clinic in the U.S. could advertise procedures for genome editing, circumvent FDA oversight, and perform the embryo modification and implantation in another nation.

National Academies of Science, Medicine and Engineering Report: Human Genome Editing: Science, Ethics, and Governance

The policy recommendations set forth in 2017 by the National Academies of Science, Medicine and Engineering solidified genome editing as a potential and permissible curative therapy.⁶⁹ The NAS Report titled Human Genome Editing: Science, Ethics, and Governance, concluded that human genome germline editing to prevent and treat serious disease would be ethically appropriate subject to certain conditions.⁷⁰ To qualify its assertion, the NAS cautioned “more research is needed” to overcome technical hurdles before genome editing become a realistic possibility. Yet the NAS asserted that “despite significant public discomfort” there are compelling reasons to support genome editing because there is an absence of reasonable alternatives and genome editing promises to prevent diseases and conditions for which there may not be suitable alternatives.⁷¹ This position presumes the technology will be “safe and efficient,” eventually feasible, and will permit parents to have a child “free of the mutation of concern.”

Strategically, the Report's reasoning adopted the linguistic engineering employed by the Nuffield Council on Bioethics' justification in its 2016 report on human genome editing.⁷² Both the Nuffield Council and the NAS espoused the notion editing intended to “treat” an embryo for a medical condition should be viewed as “restoring function within a range of naturally occurring genomic variation.”⁷³ The NAS further rationalized there is no “naturally occurring genome,” so genome editing would merely replace a disease causing variant with a known non-pathogenic variant to “treat” the embryo.⁷⁴ The NAS qualifies “only well understood genes would be suitable candidates for germline genome editing” and it should only be used to “treat” serious disease, which some scholars assert constitutes a prudent solution to limit potential applications. However, this assertion not only reifies the false representation of genome editing as a curative therapy, but perpetuate a scientifically incorrect characterization that genome editing functions as a “search and replace” technology where one gene alone causes a disease and scientists can efficiently exchange such gene and create a healthy future child.⁷⁵

The NAS Report asserted claims to further support the proposition that genome editing constitutes an ethical, and desirable, curative therapy. First, the NAS described the risks arising from genome editing including off target effects, where the genome editing technology affects other unintended locations in the genome, and mosaicism, where the genome editing technology does not produce the desired effects in all cells, resulting in a mix of naturally occurring cells and edited cells.⁷⁶ Despite acknowledging these significant barriers, NAS proclaimed efficiency and accuracy can be high, concluding “there are sound reasons to believe off-target effects can be reduced, and PGD of the edited embryo prior to implantation will ensure genome editing performed as expected.”⁷⁷ Not only did NAS omit its reasoning for such conclusion, but this position runs contrary to current scientific findings described in detail infra. Importantly, this position relies on parsing the component functions within the genome because it presumes — without evidence — that changing one variant would not impact the rest of genome as long as the change reflects a “naturally occurring variant.”⁷⁸ Finally, the NAS asserted that genome editing may be more effective and more ethically acceptable than PGD, reasoning that genome editing would require creating less embryos than PGD. However, this conclusion first presumes the safety and efficacy of genome editing, and again, stands contrary to current scientific literature demonstrating a high ratio of embryos required to produce one full term birth.^{Reference Chen79}

Global Prohibitions and Current Laws

Until recently, there has been an unambiguous global consensus against human germline genome editing. Numerous declarations, directives, and laws promulgated by international entities and other nations explicitly prohibit human germline modification.^{Reference Ishii80} The United Nations Universal Declaration on the Human Genome and Human Rights has declared that the “human genome underlies the fundamental unity of all member of the human family…it is the heritage of humanity.”⁸¹ In Article 5, the Declaration states “research, diagnosis, or treatment affecting an individual's genome shall be undertaken only after rigorous and prior assessment of potential risks and benefits,” this intervention requires informed consent that the procedure would be guided by the individual's best interest, and if the individual does not have the capacity to consent then the intervention may only be carried out for the direct benefit or, alternatively, “pose such minimal risk and burden” to the individual that the research is “compatible with the protection of the individual's human rights.”⁸² The Council of Europe's Convention on Human Rights and Biomedicine states “an intervention seeking to modify the human genome may only be taken for preventive, diagnostic, or therapeutic purposes, and only if its aim is not to introduce any modification in the genome of any descendants.”⁸³ Aligned with this prohibition, in 2001 the European Union promulgated a directive on clinical trials that further specified, “No gene therapy trials may be carried out which result in modifications to the subject's germline genetic identity.”⁸⁴ Each statement constitutes an unmistakably clear intent to prohibit germ-line modifications though any technological iteration including genome editing.

Dozens of nations at present including Canada,⁸⁵ Germany,⁸⁶ France,^{Reference Beaumont and Tripathi87} Switzerland,⁸⁸ Sweden,⁸⁹ and Italy⁹⁰ adopted legislation prohibiting germline intervention on embryos intended for implantation.⁹¹ Laws enacted in the aforementioned nations not only prohibit germline or heritable modification, but such actions constitute criminal violation subject to fines and or imprisonment.⁹² Unequivocal prohibition and criminalization signals the egregiousness, potential for harm, and social unacceptability of such an action in these nations.⁹³

The extensive history and global reach and history undermines Caplan's arguments that prohibitions stem from “irrational fear,” and instead demonstrate prohibition affirmed longstanding precedent based on reasoned deliberation.

Shifting Global Policy and Interpretation of the Law

Despite this historical principled prohibition against germline modifications, calculated rhetoric portraying genome editing as preventive medicine to treat disease in an embryo has begun to influence global ethical discourse and policymaking in the U.S and Europe. If scientists and the media present a managed perception of the science, this has the power to dramatically shape public opinion. Recent public opinion polls appear to reflect society's increasing acceptance of germline modifications in the therapeutic context, with some studies showing more than half of the public agrees with germline genome editing to prevent serious disease.^{Reference McCaughey94} However, these findings become less convincing if the survey instrument presents the respondent with a leading presupposition that genome editing will cure or treat serious disease without appreciating the complex risks involved. This framing places a heavy burden on the respondent to reject a medical solution that would cure disease.

Managing perception of the science with a particular outcome in mind may impact public acceptance. The INSERM Ethics Committee in France, the German National Academy of Sciences, and the European Academies Scientific Advisory Council adopt the same strategic descriptions such as “correcting” a mutation, “targeted,” and “unprecedented accuracy and precision” when referring to genome editing and discuss genome editing's ability to “eventually treat or avoid monogenic disorders.”⁹⁵

By classifying genome editing as repairing and correcting an embryo rather than a historically unjustifiable risky experiment on future humans, it also shifts perceptions of how the law should apply to such actions. The European Academies adopts the pathos of genome editing as a benevolent medical cure espoused by Pinker and Savulescu, concluding that our moral obligation to “fight disease and relieve suffering” may include permitting genome editing because delaying such use would result in a profound opportunity cost.⁹⁶ It is both disingenuous and a glaring logical fallacy to equate fighting disease and relieving suffering with sanctioning highly risky experiments on human embryos as a means to create future persons.

Both INSERM and the German National Academy of Sciences acknowledge that its nation's respective federal laws would currently prohibit genome editing, but open the door to assessing the potential opportunities, benefits, and economic value of genome editing.⁹⁷ Recent publications suggest that the European Steering Committee of the Council of Europe and the European Academies Scientific Advisory Council may consider modifying current policies prohibiting germline modifications.^{Reference Chneiweiss98} In a Letter to the Editor, the European Steering Committee proclaimed no international consensus exists pertaining to germline modifications, labeled a moratorium as “not appropriate,” and instead proposed a model for risk matrices to implement “responsible use” of a “promising new technology.”⁹⁹ The Letter to the Editor also explicitly called for nations to reassess the ban against germline modifications previously set forth in the Oviedo Convention.¹⁰⁰ As ethicist Daniel Callahan has observed, “speculative harms are treated as fear mongering while speculative benefits are allowed to run amok,” which demonstrates the power of promotional claims to directly impact policy outcomes.^{Reference Callahan101} The problem, however, arises when these governance discussions omit how the current evidence fails to support speculative benefits and instead demonstrates the significant risks.

Misleading characterization of genome editing as a certain medical cure will have a dramatic impact not only on preliminary global policy discussions, but in subsequent assessment pertaining to the applicability of the current law and how the European Union and respective nations will proceed. As legal scholar Rosario Isasi observed, some proponents may interpret a nation's laws in such a manner as to circumvent or hinder the intent of the policy by asserting that genome editing corrects or restores function, so it does not “prejudice human dignity” or constitute “a eugenic practice.”¹⁰² Indeed, bioethicist Tetsuya Ishii overtly suggests germline modification prohibitions may not apply at all in some instances because such intervention “corrects” the embryo rather than impairs its functioning.¹⁰³ Strategically incorporating promotional claims as scientific fact to inform applicability of the law not only subverts the foundation of transparency required for responsible policy, but constitutes an alarming embodiment of the principle that repeating a believable fiction often enough will lead the public to trust the truth of its message, even if the outcome leads to devastating and horrific consequences.^{Reference Stafford104}

These global discussions are significant for a number of reasons. First, calls for reforming policy and law rely upon the rhetoric and promotional claims of genome editing as a curative therapy, which eclipses the current scientific evidence demonstrating signifi-cant risks. Second, the meetings occurring within the European Union dismiss historical reasons for the prohibition on germline modifications, which exist as a matter of principle that “no individual or scientist has the moral authority” to experiment with modifying the genome of future humans.^{Reference Newman105} As many in the field have suggested, is it unlikely that all nations will reach accord on the issue of permitting or prohibiting germline modifications though genome editing. Yet the existence of these discussions suggests that the campaign to relabel genome editing as a curative therapy rather than an ethically unacceptable experiment as Annas originally suggested have succeeded. As a result, the U.S.'s decision to ethically sanction future limited uses of genome editing now constitutes as a legitimate choice about which reasonable opinions may differ rather than a criminal violation.

Problems with Efficiency

Despite claims that genome engineering is highly efficient (that the genome editing mechanism induces the intended changes), these statements require further scrutiny. Preliminary models in 2015 suggested efficiency of around 25%,^{Reference Liang107} while more recent models suggest a rate of 52-72% efficiency in human embryos.^{Reference Ma108} Midic and colleagues demonstrated potential efficiency at 80-100% using rhesus monkey embryos, but carefully qualified this finding.¹⁰⁹ First, Midic and colleagues distinguished that despite such “high efficiency” in animal models, they also reported only 52% efficiency in human embryos which supports NAS's recognition that it is not possible to accurately infer the developmental events in human embryos by studying animal models.¹¹⁰ Second, in animal models using rhesus monkey zygotes, less than 10% of embryos resulted in a full term birth.¹¹¹ Midic and colleagues noted a high rate of pregnancy loss and developmental arrest.¹¹² In human embryos, Ma and colleagues demonstrated only 50% progressed to blastocysts that were only monitored over the course of three days.¹¹³

Notably, recent literature and the NAS assume the editing process does not interfere with development and that remaining embryo development equates to viability and health. In fact, some scientists have suggested the opposite: diminishing survival rate suggests that embryos that do survive germline modifications may develop subtle latent deficits.¹¹⁴

Unpredictable Effects on Development

The NAS conclusions integrated the unsupported principle that embryo survival equates to health, dismissing the highly significant impact of the process of genome editing on the subsequent developmental trajectory.¹¹⁵ NAS's characterization of genome editing as merely replacing one pathogenic variant with a healthy counterpart invokes the notion of a genomic hit-and-run. This model presumes the editing process leaves no traces of reagents used in the genome editing, and the editing does not inflict other changes. However, in 2015, the Office of Science Policy of the National Institutes of Health published a meeting report in which it described the risks associated with genome editing, including the potential for the extra chromosomal DNA to become trapped into the double strand break where the genome editing technology interacts within the gene.¹¹⁶

The assumption that the editing process does not impact the embryo's development deserves further examination.^{Reference Carroll and Charo117} Available research provides support for the opposite conclusion: that the impact on the organism as a whole is highly risky and unpredictable.^{Reference Kang118} Legal scholar George Annas asserts germline modification will never be safe and there will always be an unpredictable effect on the offspring.¹¹⁹ Evolutionary biologist Stuart Newman maintains the possibility of germline modification is “fraught with potential error” precisely because germline interventions impact every cell in the body and will affect every cell's subsequent growth and development.^{Reference Newman120} In animal models, germline modifications induced multiple devastating effects, ranging from latent health deficits, inducing tumor growth, and producing developmental abnormalities.¹²¹ Newman asserts “there is no such thing as engineering an organism” because interactions between genes are highly integrated, designed to achieve stability and balance, and manipulation of one location risks disrupting the biological equilibrium.^{Reference Newman and Newman122} Kang and colleagues explicitly warn even modifying a target to a naturally occurring allele requires scrutiny, because we are unable to predict the genome wide effect of changing a single allele and disrupting the pre-existing genome.¹²³ This may result in creating a child who may be congenitally impaired, facing biological complications and irreversible health harms.¹²⁴ This evidence directly undermines NAS's unsupported conclusions that modifying an allele constitutes the equivalent to that same naturally occurring allele.¹²⁵

Importantly, the impact of the complex and uncertain germline modifications may only become apparent during the growth and development of the future child, as the embryo grows into specific cells and tissues, during the unfolding of the nervous system, the organs, the brain, and over the course of the child's life.^{Reference Evitt126} This principle invokes two key points of importance. First, animal models that appeared to be growing and developing according to prediction suffered tumor development later in life and encountered unexpected developmental abnormalities not previously anticipated by scientists.¹²⁷ This underscores the warning that germline modifications are neither predictable nor immediately visible. Second, germline modification by its nature is irreversible, which raises the uncomfortable question of how to address potential “mistakes.” That is, what is to be done with the experiments gone wrong — future persons — who suffer health deficits arising from such germline intervention? Despite Steven Pinker's dismissive guarantee “of course we would not create embryos with the probability of being sick and deformed,” one can imagine future persons injured by these experiments would not be assuaged by these cavalier promises.¹²⁸

Mosaicism

In addition to general risks of germline modification, genome editing may result in mosaicism, where some cells contain genetic modification and other cells do not demonstrate the intended modification.¹²⁹ The NAS Report acknowledges that genome editing in a zygote or early embryo poses a “significant chance” that some of the cells in the resulting embryo would not have the desired, or any, of the intended modifications.¹³⁰ According to current research, animal models demonstrate “substantial” genetic mosaicism because the editing mechanism may occur during different stages of cell division or exert residual influence on cells after division.¹³¹ In one study using early human embryos, the rates of mosaicism in the developing cells exceeded the measured percent of efficiency to induce the desired modification.¹³²

The results pertaining to mosaicism hold several implications for the child's future health and the ability to predict the child's health outcome. Recent claims by MIT Technology Review assure the public “better techniques” can avoid mosaicism, but these promises are unsupported.^{Reference Connor133} First, the impact of mosaicism depends on the type of tissue in which the mosaic cells exist, and second, rates of mutations may vary between different tissues.¹³⁴ The earlier the mosaicism occurs in development and the more widespread the mosaicism, the more serious of an impact on the health outcomes. Mosaicism occurring in somatic cells can result in gross structural abnormalities, major Mendelian disorders, cell and tissue degeneration associated with aging, and death.^{Reference Campbell135}

Importantly, early testing for mosaicism constitutes an ineffective predictor of future health because it may not reveal the developmental trajectory and whether subsequent cell division will result in mosaicism. Currently, scientists use PGD as a means to assess the current state of the embryo, but there are no effective methods to predict the organism's growth and development without creating fully formed human beings and observing the effects of genome editing throughout several generations. Attempting to use PGD to screen the embryo is not an effective method to even assess the present rate of mosaicism because testing a single cell with PGD does not reflect the rest of the cells in the embryo, and sampling too many cells would result in structural damage to the embryo.¹³⁶ Thus, PGD may examine one cell, proclaim the absence of mosaicism, but miss all other cells that do demonstrate mosaicism and inaccurately predict the trajectory of an embryo.

Off Target Effects

Genome editing also poses challenges related to the potential for off target effects which refer to a range of unintended outcomes: where the editing nucleases make cuts to unintended locations rather than the target, modifies the target sequence but continues making modifications after the initial target, or the guide RNA may bind to nontarget sequences.¹³⁷ Insertions, deletions, or translocations can activate or deactivate crucial sequences in the genome, which could activate oncogenes related to tumor development or produce unexpected outcome such as advanced aging.¹³⁸ Even small insertions, deletions, or translocations may have substantial effects on gene expression and the resulting health of the child.¹³⁹

Identifying off target effects poses considerable challenges. Researchers create methods to identify off target effects by using a composite reference genome, which may not account for genetic variants among population groups.^{Reference Begley140} This method assesses the potential for off target effects through checking the most frequently occurring and predicted locations for off target effects.¹⁴¹ In some instances, this investigation that may be limited to only examining a few sites.¹⁴² Some insertions or deletions may be so small that they escape detection, leading researchers to report no off target effects occurred.¹⁴³ Current literature explicitly challenges presumptions of efficacy and lack of off target effects based on failure to defect them, highlighting the proposition that failing to detect an outcome does not equate to producing direct evidence that the outcome does not exist.^{Reference Callaway144}

Both scientific publications and the NAS proclaimed a lack of off target effects and promise that improved methods will reduce the potential for off target effects, yet current evidence directly undermines these proclamations.¹⁴⁵ In a 2017 Nature Methods article, Schaefer and colleagues found genome editing resulted in over 100 insertions or deletions, and over 1000 single nucleotide variants in murine models, noting these substantial off target effects were not anticipated and the off target effects would be detrimental to key cellular processes.^{Reference Schaefer146} Notably, they would not be phenotypically apparent, so preliminary appearances would likely suggest normal development.¹⁴⁷ The scientific and biotechnology community swiftly refuted the validity of Schaefer's results in letters to Nature Methods, criticizing the study methods and calling for retraction.^{Reference Spicer148} Despite the editorial decision to accede to Intellia Therapeutic's demand for retraction, Schaefer and colleagues' research should not be dismissed.^{Reference Hasson149} Subsequent research by Kosicki and colleagues similarly found both on target and significant off target damage, such as large deletions and complex rearrangements.^{Reference Kosicki150} This research demonstrates occur-rence of off target effects at locations the researchers did not know were possible, the substantial amount of off target effects, and the probable impact on the resulting health outcomes. Based on the available research in this area, NAS's statements that off target effects “appear to be low,” and that improved methods correlate to a “small potential” for off target effects are not merely unsupported, but misleading.

Risks of Assisted Reproductive Technology

In addition to facing risks pertaining to mosaicism and off target effects from genome modifications, future children also encounter background risks arising from using assisted reproductive technology (ART).^{Reference Hunter151} From a legal and ethical standpoint, policymakers and clinicians should consider not only the anticipated projected benefit and risks of genome editing, but also consider the risks of ART as inseparable and clinically meaningful because genome editing requires using in vitro fertilization. To accurately assess the risks of a proposed intervention requires including and considering all risks involved in interrelated interventions, which I have termed the Unified Composite Risk Benefit Analysis.^{Reference Drabiak152} Accordingly, conceptualizing of germline modifications requires considering risks of genome editing described supra in addition to risks arising from ART.

Current research that has controlled for maternal age and multiple births reveals that children born through ART have a 30-40% increased rate of major congenital malformations.^{Reference Lu153} ART has also been associated with increased risk of autism,¹⁵⁴ more childhood illness,¹⁵⁵ a higher occurrence of cardiovascular conditions,^{Reference Hart156} and an increased risk of cancer, specifically leukemia and lymphoma.^{Reference Amirian and Bondy157}

Researchers have hypothesized a number of reasons for such outcomes, including drugs used by the mother during ovarian stimulation;¹⁵⁸ that impaired fertility may signal existing genetic mutations;¹⁵⁹ and the impact of damage caused to the embryo arising from physical manipulation and the processes used during ART.¹⁶⁰ Current research suggests a correlation between the amount of physical manipulation to the embryo and level of damage resulting in potentially serious health deficits.¹⁶¹ Physical damage may result from temperature shifts;¹⁶² reagents used and time the embryo spends in culture;¹⁶³ and destruction to cellular architecture.¹⁶⁴ These factors could result in damage to embryo structure, aneuploidy, or disruption of chromosomal segregation and division.¹⁶⁵

NAS Report Issued Recommendations Against Evidence Documenting Risks

Current research supports the proposition that germline modification through genome editing will fail to produce the desired changes in the majority of embryos.¹⁶⁶ The NAS has not offered evidence to support the conclusion that embryos that continue to develop are indeed healthy, and instead current research demonstrates the substantial likelihood of mosaicism and off target effects. Importantly, both the use of PGD to assess the occurrence of mosaicism and current methods of predicting and locating off target effects fail to accurately evaluate the impact of genome editing, potentially resulting in the appearance of a successful intervention that missed detecting anomalies. Inability to identify mosaicism or off target effects could result in creating a future child with structural abnormalities, advanced cell and tissue degeneration associated with aging, cancer and death, in addition to the increased risk of adverse health outcomes associated with ART. The weight of the evidence unquestionably points not merely to insufficient evidence of safety and efficacy, but barriers to predicting outcomes, and the potential to impose novel risks on the resulting child.¹⁶⁷ These extensive considerations do not support the NAS's conclusion that conducting clinical trials for genome editing subject to conditions is ethically permissible.

Complexity of Genomics

In 2016, Nature published an article stating that the average person has an average of 54 mutations classified as pathogenic, but that 41 occur so frequently in the human population that they are not likely to cause severe disease.^{Reference Lek168} This finding solidified what research in the field has previously asserted: many genetic mutations have been misclassified as harmful, and not all variants are pathogenic that could cause disease.^{Reference Hayden169} Strength of evidence for supporting an association between a particular variant to its pathogenicity varies, and the existence of a variant does not foretell the variant's functional role (if any) in disease development.¹⁷⁰ For example, some variants may appear in one population group but not another, but this variation alone may merely signal differences between populations, not whether the variant confers pathogenicity.¹⁷¹ Determinations of variants as pathogenic may be based on failing to look at differences outside the reference group or sequencing errors, which results in spurious or even incorrect associations between the classification and outcome.¹⁷² As a point of reference, Goldstein and colleagues estimated about one-third of mutations currently listed as pathogenic are in fact benign.¹⁷³

Classifying variants becomes particularly murky in cases of multifactorial disease like cancer. In the case of assessing genetic variation related to the BRCA gene, variants do not constitute a binary distinction of deleterious mutation versus no mutation.^{Reference Drabiak174} Instead, some variants are associated with a high likelihood of developing cancer, some variants are associated with a slightly increased risk of developing cancer, some variants are believed to be neutral, and around forty percent constitute variants of unknown significance (VUS).^{Reference Peres175} Finally, even appropriately classified pathogenic variants may not be deleterious in the evolutionary sense, any may confer advantage to the individual based on interaction with other areas of the genome.¹⁷⁶

Each of these points highlights cracks in the very foundation of genome editing: the presumption that scientists can properly identify and classify which variants are pathogenic and cause disease. These intricacies related to identifying, classifying, and predicting the functional consequences of variants also undermine the assumption that scientists could construct a modification that accurately corresponds to a disease and do so without undermining the genome as a whole.

Treating Disease

Causal factors underlying diseases provides a prime example of how complex interaction between areas of the genome and the external environment influences health outcomes. This applies not only to multifactorial diseases such as cancer or Parkinson's disease, but also for monogenic disease such as Cystic Fibrosis and Huntington's Disease. Current research in this area directly contradicts claims that equate a variant's existence to fully penetrant disease.

Multifactorial Disease

Despite the portrayal of disease occurrence as a product of a “genetic flaw” or “faulty genes,” genetic variants appear as one of many causal factors in multifactorial diseases. Cancer, for example, represents a significant public health issue: 41% of the U.S. population will be diagnosed with cancer at some point in their lives, and 21% of the U.S. population will die from cancer.¹⁷⁷ Despite these figures, inherited mutations that contribute to the cause of cancer account for only 5-10% of all cases of cancer, with a substantial percent of remaining risk attributed to environmental and lifestyle factors.¹⁷⁸ Even genetic variants such as BRCA1 and BRCA2 (which occur in .2% of the general population) are not 100% penetrant: a person may carry a BRCA1 variant that predisposes her to risk of breast and ovarian cancer, but about never develop the disease.^{Reference Lostrumbo179}

Other diseases, such as Parkinson's Disease also defy the reductionist genetic explanation. Ninety percent of Parkinson's Disease patients do not report a family history of the disease, and the most common genetic variant associated with Parkinson's Disease only accounts for fewer than 2% of all cases.^{Reference Goldman180} This variant only correlates to a lifetime penetrance of 30-40%, which means the majority of individuals who have the variant will never develop Parkinson's Disease.¹⁸¹ This has led physician Samuel Goldman to assert although genetic variation is involved in the development of Parkinson's Disease, the vast majority of cases cannot be ascribed to genetic factors but appear to be linked with exposure to environmental factors such as pesticides, solvents, and air pollutants.¹⁸²

Instead, research points to changes to the epigenome as one causal factor in the development of disease.^{Reference Pacchierotti and Spano183} Epigenetic changes occur above the genes through methylation, altering histone proteins, or RNA interference that alters how each gene is expressed.^{Reference Rothstein184} Epigenetic changes occur at a much higher frequency than genotoxic mechanisms, and changes to the epigenome affect a number of processes relating to growth, development, and risk for future disease.¹⁸⁵ Environmental interaction with our genome occurs from numerous sources including diet, nutrition, stress, and environmental toxicants such as pesticides, heavy metals, and antibiotics.¹⁸⁶ Current research links lifestyle factors and environmental exposures with epigenetic changes and an increased risk of disease such as Alzheimer's Disease, cardiovascular disease, and autoimmune disease.^{Reference Sears and Genius187}

Assessing the causal factors involved in disease development, it becomes clear genome editing could not effectively address multifactorial diseases such as cancer or Parkinson's Disease because the majority of cases do not arise from an identifiable and predictable genetic variant. Importantly, this research directly undercuts appeals of pathos proffered by Pinker and Savulescu, who explicitly equate genome editing as a cure for multifactorial disease. Promoting genome editing would not address the heavy burden of disease arising from lifestyle and environmental influences or ameliorate environmentally mediated epigenetic damage.^{Reference Drabiak188} Recognizing and reducing these exposure levels should constitute the focus of the inquiry, along with concurrent low risk interventions such as exercise and dietary measures, which have been shown to reduce risk of disease and enhance health outcomes.^{Reference Paraskevaidi189}

Monogenic Disease

Many have appealed to using genome editing as a method to treat monogenic disease such as Huntington's Disease and Cystic Fibrosis, but presumptions that genome editing could feasibly and sufficiently address these diseases requires careful analysis.¹⁹⁰

Not all individuals with a particular disease causing mutation will develop the disease, which geneticists refer to as reduced or incomplete penetrance.^{Reference Cooper191} In the case of Huntington's Disease, even if an individual has an allele associated with monogenic disease such as greater than 36 CAG repeats, the disease may not be fully penetrant.^{Reference Nahhas192} Huntington's Disease is a “trinucleotide repeat” disorder, which is caused by an increase in the number of CAG repeats in the HD gene.^{Reference Reiner193} Repeats of 40 or larger are associated with full penetrance and disease expression, but some individuals may have intermediate alleles, defined as between 36-39 CAG repeats in the HD gene.¹⁹⁴ Estimates vary, but recent research suggests up to 3.9% of the population may carry these intermediate alleles, and a significant minority of these individuals who harbor in termediate alleles will never exhibit clinical symptoms of Huntington's Disease.¹⁹⁵ Transmission between generations can also undergo expansion or contraction between being fully penetrant and exhibiting reduced penetrance, which means a parent who has 42 CAG repeats and suffers the clinical manifestations of Huntington's Disease could produce a child with 38 repeats, who may never exhibit clinical symptoms.¹⁹⁶ These nuances suggest patterns of inheritance, disease penetrance, and even how the disease becomes expressed are affected by other factors in addition to the mutation.

Current research supports the proposition that a number of other factors influence the transmission of the mutation, its penetrance, and how the mutation is expressed. Cooper and colleagues articulate three potential explanations for variable penetrance and expressivity (the severity of the clinical symptoms): the interaction of modifier genes, epigenetic influences on the genome, and environmental factors.^{Reference Gallati197} Modifier genes refer to other locations at the genome that interact and exert influence on a particular gene, such as the HD gene.¹⁹⁸ Interaction between areas of the genome, which is called epistasis, can modify the penetrance, expressivity, age of onset, and rate of disease progression and either enhance or reduce the effect of the variant.¹⁹⁹ The interaction of genes at other loci and external factors may exert more influence than the primary mutation itself.²⁰⁰

These findings hold tremendous implications for how we perceive the cause of monogenic disease and what measures would be effective in addressing disease prevention. Literature in this area asserts “the traditional view of genetic disorders as being either monogenic or multifactorial is no longer supported and requires a change in focus toward factors that modify gene expression and disease course.”²⁰¹ Cooper and colleagues hypothesize that many mutations are conditionally pathogenic, where the mutation alone is insufficient to cause the disease but interaction between modifier genes and the environment are key components that result in pathogenicity.²⁰² Individuals may carry hundreds of potentially disadvantageous variants without adverse health effects, which further undermines the presumption that one variant alone causes the disease, and that modifying one variant constitutes a sufficient solution.²⁰³

Upon further examination, the metaphors of deleting or replacing one faulty gene to “cure” monogenic disease appear grossly incorrect, because the mutation does not function in isolation. Even if genome editing could effectively modify one gene (which is not supported by the current evidence), this research suggests one modification would not produce the desired clinical outcome of eliminating the disease. These principles also underscore warnings issued by Newman, who has repeatedly cautioned that the genome functions as a whole and modification to one location will reverberate through the entire organism likely inducing uncertain genome wide disruption. Finally, this approach wholly ignores modifiable environmental and epigenetic influences guiding penetrance and expressivity such as diet,^{Reference Gallati204} pesticide exposure,²⁰⁵ and controlling inflammation.^{Reference Olejniczak206}

Treating Infertility

Genome editing has also been presented as means to treat infertility. Some scientists assert genome editing embryos solely for research purposes will enable a greater understanding of early human development and understand causes of infertility.^{Reference Stein207} Yet investigating the medical rationale for clinical genome editing of embryos raises a similar set of findings with research demonstrating that rising rates of impaired fertility are likely due to a variety of complex environmental and lifestyle causes including aging, not inherent genetic flaws.^{Reference Colburn208} Notably, the same research suggests a connection between infertility and rising rates of other diseases such as reproductive abnormalities, endocrine disorders, and cancer.²⁰⁹

Infertility affects both females and males, impacting 12-25% of the population.²¹⁰ A portion of infertility stems from aging, and as one gynecologist observed, trying to change biology is “incredibly difficult and expensive to alter.”^{Reference Groskop211} For females, age impacts both ovulatory frequency and oocyte quality in the form of diminished ovarian function, risk of aneuploidy, genetic segregation errors, and oocyte structural defects.²¹² To place maternal age into perspective: in a thirty year old woman, 75% of embryos appear normal and do not exhibit chromosomal abnormalities, but by age forty, only 40% of a woman's embryos do not exhibit chromosomal abnormalities.²¹³ On a population level, males are also experiencing dramatic decrease in fertility measured by reduced sperm count, impaired motility, and changes in sperm morphology.²¹⁴ An ample body of research demonstrates male fertility has been steadily declining, indicating by a 1.4-3% adjusted annual decrease in sperm count.^{Reference Levine215} One study calculated sperm count has decreased an astonishing 52% between 1973 and 2011.²¹⁶ In addition to age, research suggests lifestyle choices can directly impact both female and male fertility outcomes. Factors such as smoking, alcohol use, diet, and sedentary lifestyle have been shown to negatively correlate to fertility outcomes.^{Reference Sharma217} Some promising research suggests positive effects of dietary changes and moderate exercise as an avenue to improve fertility.²¹⁸

Decreases in fertility for both sexes has been accompanied by an increase in endocrine disorders, reproductive abnormalities, and cancer.²¹⁹ In females, infertility may result, or be accompanied by conditions such as ovarian dysgenesis, premature ovarian failure, anovulation, and morphological abnormalities in the human reproductive tract.^{Reference Marques-Pinto and Carvalho220} In males, trends demonstrate decreasing sperm quality and count along with an increase in cases of low testosterone, cryptorchidism, and an increased risk of testicular cancer later in life.²²¹ Infertility does not constitute an isolated condition, but instead signals the presence of factors increasing susceptibility to systemic disease and dysfunction.^{Reference Huo222}

Population level genomics have remained relatively stable, which leads researchers to hypothesize lifestyle and environmental factors are triggering these changes, either by primary exposure or via transgenerational influence from previous generations.²²³ In addition to the impact of lifestyle factors²²⁴ such as level of activity, exercise, diet, sleep and stress, research implicates chemical exposures as a factor contributing to both the decline in fertility and reproductive disorders. A variety of environmental toxicants including pesticides, PCBs, phthalates, parabens, and Bisphenol A present in our daily environment act as endocrine disrupting chemicals (EDCs).²²⁵

In 2012, the World Health Organization and the United Nations Environment Program published a report, “State of the Science on Endocrine Disrupting Chemicals” of the impact of EDCs on human reproduction stating there are eight hundred chemicals that are known or suspected to be capable of interfering with human reproduction and called for action to address these growing health concerns.²²⁶ Exposure to EDCs can interfere with hormone synthesis, conversion, and signaling, which can impair growth throughout the life cycle and reproductive capability.²²⁷ Exposures to EDCs throughout the course of one's life have the potential to exert epigenetic changes not only to the individual's somatic cells, but also to the germ cells.²²⁸ This means EDCs are not only changing the individual's reproductive capacities, but also transmitting altered epigenetic marks to subsequent generations and potentially compromising the offspring's fertility and health outcomes as well.²²⁹

Integrated together, this research suggests that the medical rationale of using genome editing to treat infertility contains numerous flaws. Even presuming genome editing could ever be safe and effective (which is not supported by current evidence), it detracts from cases of impaired fertility that could be prevented through social policy movements that encourage reproduction during biologically viable years and lifestyle modifications that support fertility potential. Genome editing also would not address either female or male reproductive disorders such as reproductive tract abnormalities, insufficient ovarian reserves, or declining sperm count and motility. Most importantly, genome editing fails to address the fundamental causes of infertility and the larger connection to endocrine dysfunction and associated disease such as cancer. Population level increases in infertility constitutes a warning, signaling lifestyle and environmental factors are threatening both human reproduction and health outcomes, which cannot be addressed by modifying a particular gene.