Reboot the Debate on Genetic Engineering Arguments About Whether Process Or Product Should Be the Focus of Regulation Are Stalling Progress, Says Jennifer Kuzma

COMMENT GOVERNANCE Don’t fear the ECONOMICS China’s fraught CONSERVATION Should half EPIDEMIOLOGY Standard DIY biologists, learn relationship with Latin of Earth be set aside as strain-naming urgently from them p.167 America p.169 wilderness? p.170 needed for Zika p.173 PAULO FRIDMAN/BLOOMBERG VIA GETTY FRIDMAN/BLOOMBERG PAULO

In the United States, engineered crops now make up more than 80% of the soya bean (pictured), maize and cotton acreage. Reboot the debate on genetic engineering Arguments about whether process or product should be the focus of regulation are stalling progress, says Jennifer Kuzma.

enetic engineering (GE) has become appropriate uses of GE are becoming more com/lhyten) to investigate pest-resistant GE increasingly contentious in recent heated. In fact, in the 20 years that I have been plants and their regulation. While working years. Thousands of citizens and involved in discussions about it, oversight of on that project, and in the years since, I have Gstakeholders in the United States are cur- GE has never seemed so much like a powder found that most people in favour of product- rently striving to pass mandatory food- keg waiting to explode. based regulation believe that there is no need labelling laws, ban certain GE products and One issue that has dominated the debate is to treat GE organisms differently from con- create GE-free zones for growing food. whether the focus of regulation should be the ventionally bred ones. Moreover, these people GE is the manipulation of an organ- process by which GE organisms are made or often claim that those who think that the pro- ism’s genome through biotechnology or the GE products themselves (the living organ- cess of engineering should be the focus of reg- modern molecular techniques. It is also isms or products derived from them). ulation — and thus, who want to see most or called genetic modification, although From 1999 to 2000, I directed a US National all GE products go through regulatory review that term is understood by scientists Academy of Sciences study (see go.nature. before they enter the marketplace — are mak- to encompass older processes such as ing arguments based on values or emotions, hybridization as well. With the wealth of CRISPR EVERYWHERE rather than science, to support their views. possibilities now offered by newly devel- A Nature special issue But framing the debate around oped gene-editing tools — particularly nature.com/crispr ‘product versus process’ is neither logical CRISPR–Cas9 — debates about the safe and nor scientific. It is stalling productive

dialogue on the development of LOOSER SCRUTINY EMERGING TECHNOLOGIES appropriate oversight in the face of rapid Because of changes to genetic-engineering (GE) The product-versus-process framing has advances in GE. processes, several GE crops have entered the US reared its confusing head again in recent

marketplace without review from the US discussions. Gene editing involves changing SOURCE: USDA IN A RUT Department of Agriculture (USDA) in recent years. DNA sequences at targeted locations, usually The United States has had a system in place 12 using site-directed nucleases (proteins that for overseeing GE products since the mid- naturally cut DNA), such as CRISPR–Cas9, 1980s: the Coordinated Framework for TALENS and zinc-finger nucleases. With 10 Regulation of Biotechnology (CFRB). The these tools, genetic engineers can intro- parties involved in the development of this duce one or a few nucleotide changes to a framework — including representatives 8 gene, make insertions or deletions in a gene from the Office of Science and Technology sequence, or insert a different gene altogether, Policy (OSTP) and various federal agen- potentially from a different species. Interna- cies — determined that it is the final product 6 tional discussions have focused on which of GE that potentially poses a risk to human types of gene-editing manipulation fall under health and the environment, not the process 4 regulatory definitions of GE organisms in by which the product is made1. different oversight regimes3. Product-led regulation was seen to be a Ironically, the same GE developers who

US marketplace without USDA review without USDA US marketplace 2 science-based approach that would pre- Estimated number of plants entering once claimed that the process of GE does clude the need for new biotechnology laws. not matter for regulatory purposes are now It meant that GE organisms could be covered 0 arguing that changes to the engineering by existing laws for products intended to be 2010 2011 2012 2013 2014 2015 process justify looser regulatory scrutiny4. used as pesticides, plant pests, toxic sub- They contend that gene editing is a safer stances and so on; engineered organisms engineered product once again becomes process than first-generation GE techniques could be channelled to particular agencies the focus when the agencies actually assess owing to its precision and the smaller point — the Environmental Protection Agency the level of risk that it poses. But from a mutations often made5. (EPA), the Food and Drug Administration scientific standpoint, a product’s traits — And some US regulatory agencies are (FDA) and the US Department of Agricul- harmful or otherwise — depend in part heeding these calls. Thanks to emerging ture (USDA) — depending on what category on the process by which it is made. (This is methods of gene delivery and gene edit- they fell into. especially evident from human gene-therapy ing, genetic engineers no longer need to use So the intended use of a product has trials, where new methods for delivering DNA sequences from plant pests to intro- dictated which agency has the authority to genes have removed the need for potentially duce engineered genes into host plants. In regulate it under the CFRB. Yet, in practice, harmful viral vectors.) And in their review part because of this change to the process it is the process of GE that has been the ‘regu- procedures, the agencies recognize that the by which the organisms are being made, the latory trigger’ used to capture products for process of engineering is important. The USDA has, for about five years, decided not pre-market review. USDA, for example, requires a “detailed to regulate about 20 engineered plants (see After the CFRB was published in 1986, description of the molecular biology of the ‘Looser scrutiny’). Several have entered the each agency produced documents that system … used to produce the regulated market without going through any formal detailed the specific protocol for the GE- article”. regulatory review — either by the USDA or product categories under its purview. Thus, product and process issues are not other agencies. For example, the EPA described the steps distinct in regulation. Indeed, it does not In Europe, crop developers are anxiously that developers would need to take if they make sense scientifically to try to value one waiting for the European Commission to were marketing plants that have pesticide- approach more highly than the other. decide how changes to GE processes should like substances engineered into them, The idea that regulating products is the affect regulatory whereas the USDA laid out how developers only ‘science-based’ way has been popular “It is impossible policy. Specifically, should handle GE plants considered to be with regulators and developers beyond the to be completely the commission is ‘plant pests’. United States. For instance, plant scientist ‘science based’ expected soon to These EPA and the USDA documents Ingo Potrykus, who led the development in a regulatory deliver a verdict on specified that organisms made by recombi- of the genetically engineered vitamin-A- system.” whether the defini- nant-DNA technologies or GE (but not their enriched ‘golden rice’ variety at the Swiss tion of GE organ- conventionally bred counterparts) must go Federal Institute of Technology (ETH) in isms covers gene-edited plants in which any through regulatory review before entering Zurich, stated in 2010 that it would be a foreign DNA used in the engineering process the marketplace. “crime against humanity” not to change has been removed through selective breed- The FDA took a different approach. It from “regulating a technology on ideological ing — and which are indistinguishable from recommended through a guidance docu- terms” to “science-based regulation, guided wild plants that might have acquired the ment — not a regulation — that developers by considerations of the risks and benefits same mutation naturally (see Nature 528, of foods derived from ‘new plant varieties’ of the trait”2. 319–320; 2015). undergo a voluntary consultation process Yet many countries go further than the GE developers and some regulators have with the agency. This guidance did not United States when it comes to process- been inconsistent in their product-versus- exclude non-GE new plant varieties. In prac- based triggers for regulation, relying on process arguments for good reason. The tice, however, developers of conventionally national laws. In Brazil, a national biosafety dichotomy doesn’t work, in practice or in bred foods seem not to have undergone such law provides safety standards and oversight theory. In fact, product-based arguments consultations, whereas the FDA has been mechanisms for GE organisms; in Australia, lead to one of two conclusions: if all prod- notified of more than 100 foods derived the Gene Technology Act mandates a regula- ucts (GE or otherwise) are to be treated the from GE plants (see go.nature.com/z78s1e). tory framework for the risk assessment and same, then either all products — GE and For the EPA, the USDA and the FDA, the management of GE organisms. conventionally bred — should be regulated,

166 | NATURE | VOL 531 | 10 MARCH 2016 © 2016 Macmillan Publishers Limited. All rights reserved COMMENT or neither should be. The first option is what evidence or risk-mitigation strategies July 2015 to clarify which regulatory authority impractical and the second inadvisable given can help to address those concerns, and what is responsible for what under the CFRB7. that some products could be harmful. classes of GE products or processes should And just last month, the USDA published receive greater regulatory scrutiny. In prac- four possible scenarios for a proposed new A FRESH START tice, regulators and other stakeholders will framework for the regulation of GE crops8. It is time to reset the debate. Product-versus- need to consider a mix of product and process Within these efforts and others, stake- process arguments reflect world views about issues to capture product groups that are likely holders could do away with polarizing the desired level of regulation for GE organ- to be of greater concern. product-versus-process and science-versus- isms. These underlying viewpoints should Several models in the social-science lit- values framings, and help to establish a gov- be made explicit, and the idea that product- erature describe how such democratic delib- ernance system that is both informed by based regulation is the only science-based eration might be achieved6. And Norway’s the science and guided by the concerns and approach rejected. decision-making about GE organisms under values of citizens. ■ In reality, it is impossible to be completely its gene-technology act demonstrates how ‘science based’ in a regulatory system. Value factors outside ‘science-based’ health or envi- Jennifer Kuzma is distinguished professor judgements are embedded in all risk and ronmental harms can be incorporated into in the social sciences and co-director of the safety assessments. For example, the dose– formal regulatory processes in practice. Since Genetic Engineering and Society Center at response curve for a certain food additive 2005, regulators in Norway making decisions North Carolina State University, USA. might be known, but such data do not by about whether a GE organism will be released e-mail: [email protected] themselves tell regulators where to set an into the environment consider the results of 1. OSTP. Fed. Reg. 51, 23302 (1986). acceptable safety limit. More often, the safety reviews, and whether participants of a 2. Potrykus, I. N. Biotechnol. 27, 466–472 (2010). dose–response curve is not well established, consultation process perceive that the organ- 3. Wolt, J. D., Wang, K. & Yang, B. Plant Biotechnol. J. or known at all. This uncertainty leads to ism provides a better option than alternatives 14, 510–518 (2015). various interpretations of the data. and contributes to sustainable agricultural 4. Huang, S., Weigel, D., Beachy, R. N. & Li, J. Nature Genet. 48, 109–111 (2016). Empirical evidence matters, but human practices(see go.nature.com/5nxzcn). 5. Kokotovich, A. & Kuzma, J. Bull. Sci. Technol. Soc. interpretation brings meaning to that There is a chance to start over, in the 34, 108–120 (2014). evidence, and multiple perspectives can United States and elsewhere. In part because 6. Ramachandran, G. et al. J. Nanopart. Res. 13, 1345–1371 (2011). strengthen understanding. Thus, an over- of advances in gene editing and a greater 7. Waltz, E. Nature Biotechnol. 33, 1221–1222 sight system should focus on what concerns diversity of GE organisms being presented (2015). a diversity of stakeholders and citizens have, to regulators, the OSTP initiated a process in 8. USDA. Fed. Reg. 81, 6225–6229 (2016).

Learn from DIY biologists The citizen-science community has a responsible, proactive attitude that is well suited to gene-editing, argues Todd Kuiken.

ne of the top science stories of 2012 expertise needed to create a deadly insect the world — received CRISPR–Cas9 plasmids involved a furore about the wisdom or virus are far beyond the capabilities of in their starting kits. These kits contain more of enhancing the transmissibility the typical DIY biologist or community lab. than 1,000 standard biological parts known Oof the H5N1 avian influenza virus in fer- Moreover, pursuing such a creation would as BioBricks, the DNA-based building blocks rets. In that same year, fears mounted that go against the culture of responsibility that that participants need to engineer a biologi- do-it-yourself (DIY) biologists would cook DIY biologists have developed over the past cal system for entering into the competition. up their own versions of the virus using five years. In fact, when it comes to thinking Other components of the CRISPR–Cas9 sys- information published in the academic press. proactively about the safety issues thrown tem are also available from the iGEM registry Now, journalists and others are again up by biotechnology, the global DIY-biology (http://parts.igem.org/CRISPR). targeting the citizen-science community — a community is arguably ahead of the scien- Yet few DIY biologists seem to be using group of people with or without formal train- tific establishment. the technology. Both Tom Burkett, founder ing who pursue research either as a hobby or of the Baltimore Under Ground Science to foster societal learning and open science EASY ACCESS Space in Maryland, and Ellen Jorgensen, — amid fears about the nascent gene-editing The equipment and reagents that are needed executive director of Genspace — a commu- technology CRISPR–Cas9. In January, the to use CRISPR–Cas9 are already readily avail- nity lab in Brooklyn, New York — say that San Jose Mercury News ran an article under able to DIY biologists. Members of the teams their users are interested in CRISPR–Cas9, a pearl-clutching headline: “Bay Area biolo- that participated in the 2015 International and Genspace will be offering a workshop gist’s gene-editing kit lets do-it-yourselfers Genetically Engineered Machine (iGEM) on it in March. But none of the projects cur- play God at the kitchen table.” And although competition — including high-school stu- rently being pursued in these spaces require they are much less alarmist, scholars are dents and users of community labs around it. Users of the La Paillasse community lab in advising policymakers to consider the poten- Paris are similarly focused on projects that tial uses of gene editing “outside the tradi- CRISPR EVERYWHERE do not need CRISPR–Cas9. tional laboratory setting” (R. A. Charo & A Nature special issue The materials might be available, but H. T. Greely Am. J. Bioeth. 15, 11–17; 2015). nature.com/crispr the knowledge and understanding needed The reality is that the techniques and to make edits that have the desired effects