Germline Gene Therapy: Don’t Let Good Intentions Spawn Bad Policy

The proposed moratorium on clinical applications of gene editing technology reveals ignorance about how innovation works, and callousness about human suffering.

“Human gene therapy” has been one of the most ambitious goals of biotechnology since the advent of molecular techniques for genetic modification in the 1970s. There are two distinct approaches, which present different kinds of benefits, risks, and controversies. Somatic cell human gene therapy (SHGT) alters a patient’s genes—either by the editing of existing genes or the insertion of new ones—to correct conditions inherited or acquired later in life. Somatic cells are any cells in the body except eggs or sperm, so modifications in them are not heritable—that is, passed on to offspring.

Since a four-year-old with a genetic defect called severe combined immunodeficiency, or “bubble boy disease,” was first successfully treated at the National Institutes of Health (NIH) in 1990, SHGT has achieved several other successes, including the correction of rare genetic abnormalities that cause conditions such as recurring pancreatitis and blindness from degeneration of the retina.

Up to now, gene therapy has been of a type that affects only the patient being treated but does not create a heritable change and affect future generations; that is, it does not modify sperm, eggs, or embryos in a way that would constitute “germline gene therapy” (GLGT). However, using a gene-editing system called CRISPR/Cas9, Chinese researchers reported in May 2015 an unsuccessful proof-of-principle experiment that attempted germline gene therapy on embryos that were nonviable (and going to be discarded in any case).

The Chinese experiment precipitated a firestorm in the scientific community, with some researchers and bioethicists calling for an absolute ban on attempts to treat even imminently lethal diseases with gene-editing techniques that would affect germ cells. The move toward prohibition gained ground at a conference held in Washington, DC, in December 2015 under the auspices of national academies of science and medicine of the United States, China, and the United Kingdom. The attendees called for what amounts to a moratorium on the clinical use of germline editing, concluding that it would be “irresponsible to proceed” until the risks were better understood and until there was “broad societal consensus” about such clinical research. They did not, however, recommend a prohibition on basic or preclinical research.

Those recommendations—coming mainly from people who don’t actually treat patients—were the result of the kind of group-think that dismisses conflicting minority opinions and produces poorly reasoned outcomes.

A return to the bad old days of Asilomar?

Many of those currently opposed to germline gene therapy wax nostalgic about a historic 1975 meeting of scientists, ethicists, and members of the press held in Asilomar, California, which resulted in a temporary moratorium on recombinant DNA, or gene-splicing, research; and, ultimately, in the creation of highly restrictive, unnecessary regulation. They appear to be on the verge of repeating that blunder.

The 1974 article in Science that led to the Asilomar meeting urged that “scientists throughout the world join with the members of this committee” in halting recombinant DNA experiments “until attempts have been made to evaluate the hazards and some resolution of the outstanding questions has been achieved.” And the official “Summary Statement of the Asilomar Conference on Recombinant DNA Molecules” concluded: “Even in the present, more limited conduct of research in this field, the evaluation of potential biohazards has proved to be extremely difficult,” because “[t]he new techniques, which permit combination of genetic information from very different organisms, place us in an area of biology with many unknowns.”

Stanford University biochemistry professor Paul Berg, a prime mover in the current initiative to ban germline gene therapy, was (and remains) one of the staunchest defenders of the Asilomar undertaking. In a 2008 Nature essay modestly titled, “Meetings that changed the world: Asilomar 1975: DNA modification secured,” he recalled that at the time the greatest concerns were “that introduced genes could change normally innocuous microbes into cancer-causing agents or into human pathogens, resistant to antibiotics or able to produce dangerous toxins.” But what many have forgotten is that the research community was far from a consensus on the question of whether a moratorium was necessary at the time; indeed, many in the scientific community did not regard it as a success, either scientific or intellectual.

In fact, the Asilomar cabal misunderstood and exaggerated the potential risks of recombinant DNA technology, modern biotechnology’s core technique; gave rise to a lengthy, damaging research moratorium; and induced the U.S. National Institutes of Health (NIH) to draft and promulgate overly restrictive “biosafety” guidelines.

During the Asilomar conference, Stanley Cohen, James D. Watson, and Joshua Lederberg argued publicly (and others privately) “against the forming of any official guidelines that spelled out how we should work with recombinant DNA.” In the words of science historian José Van Dijck, “In the politicized mood of the 1970s, genetics got annexed as an environmental issue; this new configuration manifested itself in changed images of genetics, genes, and geneticists,” which were no longer altogether altruistic, or even benign.

By 1978, the regulatory obstacles slowing research in many fields and labs would induce Watson to dismiss the handwringing as “senseless hysteria” and to observe that “everyone I know who works with DNA now feels the same and the mere mention of ‘DNA Guidelines’ or ‘Memorandums of Understanding’ makes our mouths froth.” (As a laboratory scientist at NIH at the time, I shared this sentiment.)

Those process-based NIH guidelines, which were and remain focused on the use of a single technique instead of on the actual risks of experiments, have plagued genetic engineering research ever since. By assuming from the beginning that recombinant DNA-modified organisms—which have come to be commonly known as “genetically modified organisms” or GMOs—were a high-risk category that needed to have sui generis regulation, the NIH guidelines created significant duplication of oversight for many already-regulated products. Worst of all, they reinforced the misconception that recombinant DNA-modified organisms are a meaningful “category.” Although NIH gradually pared back the stringency of its guidelines, stultifying process-based approaches to regulation of this noncategory have remained at other federal agencies, including the Environmental Protection Agency, the Food and Drug Administration (FDA), and the Department of Agriculture, and in many foreign countries.

Returning to germline gene therapy: It is unethical to modify normal embryos, but nobody is proposing to do that. For diseases that are genetically dominant, which means an abnormal gene from either parent causes the disease—examples of which include Huntington’s disease, familial hypercholesterolemia, polycystic kidney disease and neurofibromatosis type 1 (the last three of which are relatively common)—one could simply perform pre-implantation genetic diagnosis to identify a normal embryo (the parents’ eggs and sperm would produce both affected and unaffected embryos), and then implant it in the uterus. There is no need to manipulate normal embryos. In fact, as explained later in this article, it may not even be necessary to manipulate abnormal embryos to perform germline gene therapy, because alternative in vitro approaches are available.

According to New York Times science reporter Nicholas Wade, some participants at the December Washington conclave “noted there was no pressing medical demand now for making heritable changes to the human genome because diseases caused by a single errant gene were rare.” Well, those participants need a refresher in human genetics. An appropriate—and, indeed, compelling—application of GLGT would be to correct debilitating and ultimately lethal sickle-cell anemia, the most common inherited blood disorder in the United States, which affects more than 100,000 patients. It is marked by the presence of atypical hemoglobin molecules that distort red blood cells into a crescent, or sickle, shape. These “sickle cells” obstruct small blood vessels, causing frequent infections, pain in the limbs, and damage to various organs, including the lungs, kidneys, spleen, and brain.

In genetics terms, sickle-cell anemia is an autosomal recessive disease, which means that an affected individual has inherited a defective hemoglobin gene from both parents, so every one of his or her sets of chromosomes carries a defective gene. (That results in a single aberrant amino acid being inserted into the hemoglobin protein.) Particularly significant is that every offspring of two patients with sickle-cell disease will be afflicted with the disease. Repair of this sort of molecular lesion has been performed successfully in monkeys with new, highly precise gene-editing techniques.

News from the innovation front

However, as discussed by Matthew Porteus of Stanford and Christina Dann of Indiana University in the June 2015 issue of Molecular Therapy, several technical obstacles may preclude successful zygote injection in humans, including the fact that “only a fraction of injected zygotes give rise to viable offspring. Tens to hundreds of zygotes would need to be injected and implanted into several surrogate mothers to generate viable, genetically modified offspring.” With current technology, such an approach would be neither ethical nor feasible in humans.

Porteus and Dann also warned that the editing of genomes to correct a disease-causing mutation must not create mutations at other sites. They suggest possible alternative approaches to zygote injection that would avoid both of those pitfalls. In contrast to the zygote-injection strategy, stem cell editing that can be propagated in vitro enables characterization of the modified stem cells before use in therapy. Recent developments in animal models have shown that spermatogonial stem cells (SSCs), which ultimately give rise to haploid sperm, can be grown as clones in culture and then transplanted back into the testis to generate sperm. Thus, a potential strategy is to isolate SSCs, use genome editing to precisely correct a disease-causing mutation, perform whole-genome sequencing of clones that have undergone gene correction, and use only the clones that are free from off-target mutations. A related strategy would be to generate sperm directly in vitro from edited SSCs to be used for in vitro fertilization.

Therefore, even if the current state of technology does not permit the therapeutic correction of genetic diseases by means of editing via zygote injection, the two approaches suggested by Porteus and Dann could be attempted, even for genetically dominant diseases. Certainly, further proof-of-concept research should proceed, even if gene editing of SSCs isn’t successful immediately.

Progress is exceedingly rapid in this field. In December 2015 , scientists at the annual American Society of Hematology conference announced a stunning, apparently successful attempt to treat leukemia in an 11-month-old girl using off-the-shelf T-cells (a subset of blood lymphocytes) that had been ingeniously gene-edited. They were modified to enable them to attack the leukemic cells; to delete a gene that codes for a receptor on certain white blood cells, to prevent the cells from recognizing the recipient’s body as foreign and attacking it; and to survive the intense therapy the girl was receiving. The patient’s physicians (in London) opted to push the boundaries of clinical research to save her life.

In December 2015, an article in Science showed that an even higher degree of precision and specificity in gene editing is possible, and another article in Nature Biotechnology reported that the frequency of erroneous (off-target) cuts in DNA made by CRISPR/Cas9 can now be reduced to fewer than one per 3 trillion base pairs of DNA. (The human genome is about 3 billion base pairs in length.) In January 2016, additional refinements that reduce even further the frequency of off-target cuts were reported in Nature.

As Harvard University professor and molecular geneticist George Church wrote in a Nature commentary about gene editing in December 2015, “Many of these technologies are improving so fast it’s hard to measure.” Therefore, he said, a ban doesn’t make sense, and the prohibition of human germline editing “could put a damper on the best medical research and instead drive the practice underground to black markets and uncontrolled medical tourism.”

The constant improvements serve as a reminder that technologies are seldom successful right out of the gate; as they’re applied and refined, they improve, sometimes with astonishing rapidity. The first mobile phones and mainframe computers were large, clunky, inefficient, and temperamental. When I was a medical student during the 1970s, bone marrow transplantation was being performed in only a few institutions and as a last resort, and the success rate was abysmal. But the discovery of potent immunosuppressants and other technical advances improved the success rate markedly, and bone marrow transplants are now routine in many institutions. Some leukemias that were once a death sentence now have cure rates of around 90%. There are many similar stories in medicine, including open-heart surgery and organ transplants, which were remarkably primitive in their earliest incarnation, but which are usually uneventful now. The reality is that successful innovation is impossible without continual learning and incremental improvements by users, and in medicine, it is in clinical settings that this process must occur.

The idea that a medical technology cannot be “perfected” without innovation and learning in the clinical setting seems to have eluded Harvard stem-cell researcher George Q. Daley, who said last year about germline gene therapy, “This is an unsafe procedure and should not be practiced at this time, and perhaps never.” Never? Maybe it has been a while since Dr. Daley has seen a patient like one I remember well—a 20-year-old with sickle-cell anemia who had suffered three strokes, been crippled by repeated bone and joint infarctions, and had become a Demerol addict due to the unrelenting pain from the arthritis that resulted.

The over-regulated gene

Interventions that involve germline gene therapy should be used with great care, as is the case with all new therapeutic interventions, but we don’t need a moratorium on clinical research. (And at the very least, we must not let skepticism about potential applications that would modify humans interfere with research-based germ cell editing.) It may be ethically warranted to intervene with as yet unproven therapies in dire situations when there are no alternatives.

Ironically, much of the controversy about moratoriums may be moot because, in effect, Appendix M of the NIH Guidelines for Research Involving Recombinant or Synthetic Nucleic Acid Molecules already creates a near-absolute moratorium on germline gene therapy in the United States:

RAC [the NIH’s Recombinant DNA Advisory Committee] will not at present entertain proposals for germline alterations but will consider proposals involving somatic cell gene transfer. The purpose of somatic cell gene transfer is to treat an individual patient, for example, by inserting a properly functioning gene into the subject’s somatic cells. Germline alteration involves a specific attempt to introduce genetic changes into the germ (reproductive) cells of an individual, with the aim of changing the set of genes passed on to the individual’s offspring.

As to the scope of its applicability, “Appendix M applies to research conducted at or sponsored by an institution that receives any support for recombinant or synthetic nucleic acid molecule research from NIH,” which would appear to rule out germline gene therapy experiments by researchers at any U.S. academic institution. Nor is there likely to be much interest in germline therapeutic interventions by companies, given the uncertainty about how the FDA might regulate these technologies, concerns about opposition from the public and from the academic scientific community, and uncertain economic prospects. In effect, a moratorium is already in place.

Appendix M’s prohibition is both puzzling and disturbing. Given that the committee can reject any proposal for any reason, its unwillingness even to consider an entire category of clinical studies seems unnecessarily intransigent and arbitrary. It’s also cruel, because children will die while potentially life-saving therapies go untested.

Sound and humane public policy would see the NIH RAC repeal Appendix M and announce its intention to consider carefully crafted human germline gene therapy proposals that meet community standards for risk-benefit. Better still, NIH should get out of the business entirely, since FDA and local institutional review boards—not the RAC or NIH officials—have experience with that standard, and they will necessarily be involved whether or not NIH has a role.

Interventions that involve germline gene therapy should be used sparingly and with scrutiny, but if we don’t take the first step of clinical application, the one certainty is that we’ll never reach the goal of applying gene editing to the reduction of human suffering.

Henry I. Miller, a physician and molecular biologist, is the Robert Wesson Fellow in Scientific Philosophy and Public Policy at Stanford University’s Hoover Institution and a fellow at the Competitive Enterprise Institute. He was the founding director of the FDA’s Office of Biotechnology.

Your participation enriches the conversation

Respond to the ideas raised in this essay by writing to [email protected]. And read what others are saying in our lively Forum section.

Cite this Article

Miller, Henry. “Germline Gene Therapy: Don’t Let Good Intentions Spawn Bad Policy.” Issues in Science and Technology 32, no. 3 (Spring 2016).

Vol. XXXII, No. 3, Spring 2016