August 10, 2017

The Coming Age of Genetic Modification, Part II

In a previous post, I touched on the potential social and ethical consequences that will likely emerge in the wake of Dr. Shoukhrat Mitalipov’s recent experiment in germline-edited embryos. The short version: there’s probably no stopping the genetic freight train. However, there are steps we can take to minimize the potential costs, while capitalizing on the many benefits these advancements have to offer us. In order to do that, however, we need to turn our attention away from hyperbolic rhetoric of “designer babies” and focus on the near-term practical considerations—mainly, how we will govern the research, development, and application of these procedures.

Before addressing the policy concerns, however, it’s important to understand the fundamentals of what is being discussed in this debate. In the previous blog, I noted the difference between somatic cell editing and germline editing—one of the major ethical faultlines in this issue space. In order to have a clear perspective of the future possibilities, and current limitations, of genetic modification, let’s briefly examine how CRISPR actually works in practice.

CRISPR stands for “clustered regularly interspaced short palindromic repeats”—a reference to segments of DNA that function as a defense used by bacteria to ward off foreign infections. That defense system essentially targets specific patterns of DNA in a virus, bacteria, or other threat and destroys it. This approach uses Cas9—an RNA-guided protein—to search through a cell’s genetic material until it finds a genetic sequence that matches the sequence programmed into its guide RNA. Once it finds its target, the protein splices the two strands of the DNA helix. Repair enzymes can then heal the gap in the broken DNA, or filled using new genetic information introduced into the sequence. Conceptually, we can think of CRISPR as the geneticist’s variation of a “surgical laser knife, which allows a surgeon to cut out precisely defective body parts and replace them with new or repaired ones.”

The technology is still cutting edge, and most researchers are still trying to get a handle on the technical difficulties associated with its use. Right now, we’re still in the Stone Age of genetic research. Even though we’ve made significant advancements in recent years, we’re still a long, long way from editing human IQs of our children on-demand. That technology is much further into the future and some doubt that we’ll ever be able to “program” inheritable traits into our individual genomes. In short, don’t expect any superhumanly intelligent, disease-resistant super soldiers any time soon.

The Parallels Between Artificial Intelligence and Genetic Modification

There are few technologies that inspire fantastical embellishments in popular media as much as the exaggerations surrounding genetic modification. In fact, the only technology that comes close to comparison—and indeed, actually parallels the rhetoric quite closely—is artificial intelligence (AI).   

AI is often making headlines because we’ve been culturally-conditioned to believe that every advancement—no matter how marginal—is a potential harbinger of human civilization’s impending submission to machine overlords. Of course, as I’ve noted many times before, such stories are almost always overstated. The reality is that AI is only effective in narrow contexts. By hyping its capabilities beyond its technical constraints, we misconstrue what the technology is actually capable of and run the risk of embracing knee-jerk policy responses that could significantly curtail its most beneficial uses. The same is true with genetic modification technology.

Instead of focusing on concerns (and benefits) that lie far beyond the horizon, we should be looking at the current state of this technology and its near-term effects. Only then can we craft reasonable policies that comport with the evidence. Luckily, we already have some reasonable guidance on how to do so.

Multistakeholderism and Scientific Guidance in the Age of Genomics

A recent report issued by the National Academy of Sciences (NAS)—Human Genome Editing: Science, Ethics, and Governance—recommends a set of basic principles to help guide the governance and oversight of human genome editing technologies. They include things like embracing transparency of information, promoting the well-being of those individuals involved in research and testing, and taking due care to ensure practices are consistent with responsible scientific norms. These principles, in turn, are applied to a set of policy recommendations governing laboratory research and clinical trials related to the use of genetically modified human embryos and materials.

Notably, the report makes it clear that “[e]xtensive and inclusive public participation should precede clinical trials for any extension of human genome editing beyond treatment or prevention of disease or disability. (pg. 194) (emphasis mine) The reasoning is predicated on a need for embracing what essentially amounts to a multistakeholder governance model—soliciting public comments and engagement, as well as targeting funding to promote those efforts—before taking clinical trials to the stage at which non-therapeutic uses (i.e. “enhancements”) would be assessed.

In short, the authors argue that a regulatory governance structure for conducting clinical trials of germline edited embryos should be authorized, so long as those trials are for the purposes of treating or preventing disease or disabilities.

While the report recognizes the value of “heritable genome editing … to alleviate the suffering caused by genetically inherited diseases,” it nonetheless argues that the practice “raises concerns that extend beyond consideration of individual risks and benefits.” (pg. 188) They ultimately conclude that “[m]ore research is needed before any germline intervention could meet the risk/ benefit standard for authorizing clinical trials” (pg. 189). But a coda to that conclusion offers a positive sign for the future of experimental clinical trials using this technology:

Heritable germline genome-editing trials must be approached with caution, but caution does not mean they must be prohibited. If the technical challenges are overcome and potential benefits are reasonable in light of the risks, clinical trials could be initiated, limited to only the most compelling circumstances and subject to a comprehensive oversight framework that would protect the research subjects and their descendants; and have sufficient safeguards in place to protect against inappropriate expansion to uses that are less compelling or less well understood. (pg. 189) (emphasis mine)

However, Congress has repeatedly blocked the FDA from allowing such clinical trials to even be considered for approval. In the Consolidated Appropriations Act of 2016, for example, the agency was prohibited from using funds to make any approvals for germline embryos to be used in clinical trials. The bill states that:

None of the funds made available by this Act may be used to notify a sponsor or otherwise acknowledge receipt of a submission for an exemption for investigational use of a drug or biological product under section 505(i) of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 355(i)) or section 351(a)(3) of the Public Health Service Act (42 U.S.C. 262(a)(3)) in research in which a human embryo is intentionally created or modified to include a heritable genetic modification. Any such submission shall be deemed to have not been received by the Secretary, and the exemption may not go into effect.

As the NAS report makes clear:

The current effect of this provision is to make it impossible for U.S. authorities to review proposals for clinical trials of heritable genome editing, and therefore to drive development of this technology to other jurisdictions, some regulated and others not. (pg. 191)

That needs to change if we ever hope to advance this technology beyond the laboratory. But there are a number of other policy considerations that should also be taken into account.

Considerations Beyond Genomics

By 2025, total genomic data storage will be measured in exabytes—far surpassing the total data storage of YouTube’s online video troves—and that won’t even account for the processing power and storage necessary for analyzing all that data. Each whole genome sequence requires approximately 150 gigabytes of digital storage. However, the actual storage space required can differ wildly based on the approach taken to analyze genetic sequences. A 2016 Intel study, for example, concluded that the total size of the data resulting from sequencing the human genomes of the entire human population would be about 7.3 zettabytes. Given that the genetic variances between individuals is relatively slight, it may not be necessary to map the entire genome of each individual on the planet, but merely account for the differentials. Either way, we need a concerted effort to develop the digital and computational infrastructure to handle genomic storage if we’re to make headway on benefiting from these medical advancements.

We also need to consider how privacy protections factor into the storage of an individual’s genomic sequences—that means balancing considerations over the public interest in making genomic data publicly available for research and the risks of revealing sensitive medical information. Ensuring those privacy interests are respected will be of paramount concern, and promoting policies and technologies that can safeguard our most personal medical details should be encouraged. However, we shouldn’t let privacy absolutism blind us to the many possibilities that more open and transparent access to genetic information could portend. As with any debate, we need to balance the equities and arrive at solutions that all concerned can accept as a reasonable compromise.

Finally, we need to embrace AI as an effective and necessary technological means of analyzing the vast quantities of data associated with the human genome. That means tamping down the hyperbolic rhetoric over AI and embracing its near-term potential. Onerous regulatory barriers that hinder our ability to take advantage of the best this technology has to offer will only curtail our ability to advance a better understanding of the human genome.

Unlike humans, AI can excel at drawing inferences and correlations between large numbers of disparate variables in vast data sets. This is a task that could have profound effects for our understanding of the relationships between individual genetic variations, environmental factors that contribute to genetic mutations, and physiological responses to drug treatments. By using the power of AI to determine connections between the phenomic information (that is, the total set of observable physical characteristics that define us) and genomic variations of individuals, we can improve our understanding of the human genome and hopefully draw more informed conclusions about the factors that contribute to genetic mutations and diseases.

Policy Recommendations

There’s a lot to do if we want to ensure the United States retains the leadership mantle in genomic research and development. Rather than throw up roadblocks to genetic modification, policymakers and regulators should invest in the scientific and technological infrastructure necessary to capitalize on the best these advancements can offer us. In order to prepare for this new era, policymakers and regulators should:

  1. Lift the statutory ban barring the FDA from considering clinical trials of heritable genome editing. Until the FDA has the authority to actually make a determination on the safety and efficacy of using germline-edited embryos in clinical trials, there will be an artificial barrier to medical advancements in genomics here in the United States. Congress should grant the FDA the authority to carry out its mission, not stonewall what could be the most important medical breakthroughs in human history.
  2. Embrace a multistakeholder approach to developing standards and regulations governing the application of genetic technologies for “enhancements” to the human genome. As the NAS report notes, a multistakeholder process is the most ideal means of developing future rules to govern the use of human genome editing technology for purposes “beyond [the] treatment or prevention of disease or disability.” We need new rules for this new frontier, and given the social and economic ramifications of this technology, a process that casts a wide net, engages a broad array of public and private stakeholders, and helps to promote the public trust is the only reasonable path forward.
  3. Revive the Office of Technology Assessment. From 1972-1995, Congress once had an agency—the Office of Technology Assessment (OTA)—that provided legislators with reliable updates on technological advancements. In an age of rapid, and sometimes turbulent, technological disruption, the need for reviving the OTA is increasingly apparent. A scientifically literate legislature requires sources of nonpartisan analysis to not only better inform decision-making, but promote greater trust between the scientific community, representatives in Congress, and the public.
  4. Recommit to the Precision Medicine Initiative. In 2015, President Obama announced the Precision Medicine Initiative—a plan “to further accelerate our understanding of individual variability and its effect on disease onset, progression, prevention, and treatment.” The Trump Administration should recommit to this vision in order to improve regulatory interoperability and data sharing between agencies and reassess the strict privacy requirements of the Health Information Portability and Accountability Act. Only by improving our scientific infrastructure can the United States hope to continue competing at the cutting edge of advances in genomic science.
  5. Promote and accelerate the adoption of AI. Scientific analysis is increasingly reliant on digital technology—especially AI. Ensuring minimal barriers to its deployment will be key to researchers and scientists working on developing better, safer, and more widely-available medical innovations in genetic procedures. AI can accelerate the development timeline for new genetic technologies and advance our understanding of how various environmental and genetic factors can contribute to more effective therapies for, and cures to, the diseases that plague humanity.

The government can play a positive role in speeding us towards an era of better and more individualized medical treatments. By embracing an innovation-friendly strategy on genetic modification technology, policymakers can maximize its benefits while minimizing the costs. The best place to start would be for Congress to let the FDA do its job and develop the oversight and approval mechanisms necessary for approving clinical trials for germline-edited embryonic cells.