In April 2015 a team of Chinese scientists reported in a little known journal, Protein & Cell, the use of CRISPR/Cas9 to cleave and then repair the HBB gene in nonviable human embryos. The mutated form of HBB causes B-thalassemia, a potentially fatal blood disease. To say the experiment prompted controversy is an understatement. It was published on the heels of two high-profile commentaries in Nature and Science, both of which urged caution about using CRISPR/Cas9 and other technologies to edit the human germline; the Nature authors went so far as to recommend a stop of experiments precisely like the one reported in Protein & Cell. The Liang et al. paper seemed to fly in the face of the recommendations. Though the researchers used triponuclear human embryos (an essential fact missed by some of the breathless reporting in the days after), they designed the experiment as a test of a possible therapeutic strategy. If eventually proven safe, a diseased embryo would be corrected using CRISPR/Cas9 with the intent of eventually making a healthy baby. Importantly, the authors reported notable off-target effects of CRISPR-based gene editing, low efficiency of homologous recombination directed repair (HDR), mosaicism, and unwanted mutations. They concluded:
Taken together, our data underscore the need to more comprehensively understand the mechanisms of CRISPR/Cas9-mediated genome editing in human cells, and support the notion that clinical applications of the CRISPR/Cas9 system may be premature at this stage.
In the avalanche of commentary that followed was a blink-and-you-missed-it defense of the decision to publish the paper by Xiaoxue Zhang (2014), the editor of Protein & Cell. The editorial is a fascinating read, encapsulating the social and ethical tensions surrounding germline editing. Perhaps stung by criticism the experiment was unethical , Zhang says, “With extraordinary care, consideration and deliberation, Protein & Cell has decided to publish in this issue a scientific study that reports CRISPR-based gene-editing of human tripronuclear zygotes.” Zhang goes on to justify why the research was published in the first place:
The editorial decision to publish this study should not be viewed as an endorsement of this practice nor an encouragement of similar attempts, but rather the sounding of an alarm to draw immediate attention to the urgent need to rein in applications of gene-editing technologies, especially in the human germ cells or embryos.
From an ethics and policy perspective, how should we measure the Protein & Cell experiment and the decision to publish it? A framework for social responsibility such as the model proposed by Sankar and Cho in their target article might help, but would likely need to include more stakeholders than just bench scientists. Decisions to publish do not occur in isolation. Researchers, reviewers, ethics advisors, and editorial boards all have an opportunity to dialogue during the review process, so the responsibilities to publish socially controversial research are shared. In her commentary, Bovenkirk argues for a “thick” accounting of social responsibility in the sciences and this seems to make sense here. With respect to the Protein & Cellpublications, it appears that important elements of the Sankar and Cho framework are already in play. Categories of justification, reasoning to identify harms and benefits, reflexivity, and timing—specifically, designing an experiment ostensibly to uncover the risks the authors enumerate in the conclusions—are on full view in both the research paper and in the editor’s justification. Of course we must trust that other stakeholders were part of the “extraordinary care, consideration and deliberation” and that those deliberations would be fully available to outside observers.
But there is a nuance about the Protein & Cell example that is worth a closer look. In their commentary, Mertes and Penning argue that because the Liang et al research showed the dangers of any clinical use of the technology, an outcry about whether the research should have been published is misplaced. After all, both the Science and Nature essays called for more research to determine the safety and efficacy of CRISPR/Cas9 for future clinical use. Wasn’t this precisely what the experts ordered?
It would seem so. But an argument could be made—something hinted at but not directly addressed in Greeley and Charo’s lighthearted target article—that the Chinese experiment was frivolous. The results in the Protein & Cell study were anticipated in animal studies and indeed predicted by the Nature commentary. Some CRISPR/Cas9 researchers believe that the experiment provided little scientific insight. The raises the question of whether we should publish the technical feat of human germline editing or work that moves the field forward in significant ways.
So what do we do now that the Protein & Cell study is out? Follow-on experiments become increasingly derivative and are worth exponentially less to science. At what point do such experiments become data-gathering exercises rather than answers to fundamental questions about human development and disease? Given the controversial nature of human germline research, what experiments investigating disease type, embryo stage, culture conditions and other incremental advances should go forward? For example, gene-editing technologies can either knock out (cleave a gene of interest) or knock in (insert a gene of interest). The Liang and colleagues experiment was technically a knock out study. The researchers edited the HBB gene and then asked whether genetic repair pathways recruited an endogenous gene, HBD, to replace the cleaved sequence. Suppose the next CRISPR/Cas9 experiment using human embryos attempts a knock in or gene replacement strategy. Does this represent a frivolous or an important use of the technology? There are some pathways to take with respect to this question. In a recent commentary in Nature Biotechnology, we offer examples of in vitro research that is ethically and scientifically justified, such as probing questions of early human development and infertility. Such research is covered under existing oversight mechanisms and regulation.
Can scientists, reviewers, and editors—who comprise the backbone of peer-reviewed journals—achieve consensus on a framework of social responsibility and its enforcement? In this issue of AJOB, Cadigan invokes the model of corporate responsibility as a potential solution. While self-regulation has its weaknesses, there is one prominent example where journals banded together successfully over a contentious issue in policy and ethics. In 2005, The International Committee of Medical Journal Editors (ICMJE), representing fifteen leading medical journals, published a consensus statement requiring the public registration of clinical trials as a prerequisite for consideration for publication. The ICMJE editors argued that such a resource would increase public confidence in the clinical research enterprise, increase transparency, and promote collective altruism for trials volunteers. Registration in electronic trials registries skyrocketed soon after.
Now that the first human germline editing paper has appeared, we should turn to consider the deeper social justifications for research like this. Convocations of experts such as the upcoming germline editing meeting at the National Academy of Sciences will be helpful to set general normative trajectories. But the principal stakeholders—scientists, reviewers, editors, publishers, and deeply interdisciplinary advisory boards—must decide what constitutes socially responsible research as they examine and deliberate on new discoveries. If some research is deemed off limits and never to see the light of day, then there will be little incentive to undertake it in the first place.