Recombinant DNA: a Case Study in Regulation of Scientific Research

Home , The Andromeda Strain, Viral transformation

Recombinant DNA: A Case Study in Regulation of Scientific Research

James C Chalfant* Michael E. Harmann** Alan Blakeboro***

INTRODUCTION The time when developments in science and technology were automatically welcomed as progressive and beneficial has passed. Public confidence in the scientific community has given way to skepticism, if not distrust.I This attitude is due in part to the recent development of highly complex and sophisticated technologies which, from their incep- tion, have been recognized as entailing substantial risks as well as benefits. The public now expects the government to evaluate independently the risks of new technologies before they are introduced and, moreover, demands input into the decisionmaking processes that sanction their use. Many of these new developments, however, cannot be adequately understood without a high level of scientific expertise. Indeed, these developments are so esoteric that only those working directly with the

Copyright © 1979 by ECOLOGY LAW QUARTERLY. * 1978-79 Editor-in-Chief, ECOLOGY LAW QUARTERLY. B.A. 1974, Pomona College, J.D. 1979, University of California, Berkeley. ** 1978-79 Comments Editor, ECOLOGY LAW QUARTERLY. B.A. 1976, University of California, Los Angeles, J.D. 1979, University of California, Berkeley. *** Member Third Year Class, University of California, Berkeley. B.A. 1976, Univer- sity of California, Santa Barbara. 1. The last 20 years have marked a change in the public's view of science and its products. The environmental awareness of the 1960's, sparked by Rachel Carson's The Si- lent Spring (1962), has changed the public's mood from a trusting acceptance to a presump- tive distrust of technology and science. Note, Recombinant DNA and Technology Assessment, 11 GA. L. REV. 785, 786-87 (1977) [hereinafter cited as Note, Recombinant DNA]; Rowe, Governmental Regulation of Societal Risks, 45 GEO. WASH. L. REV. 944, 945- 47 (1977); PUBLIC HEALTH SERVICE, NATIONAL INSTITUTES OF HEALTH, U.S. DEP'T OF HEALTH, EDUCATION, AND WELFARE, PUB. No. 76-1138, 1 RECOMBINANT DNA RESEARCH 180, 341 (testimony of A. Ludwig) (1976) [hereinafter cited as 1976 HEW DOCUMENTS]. In addition, the increased power of "scientific progress" planted seeds of distrust. For example, Professor Laurence Tribe views recombinant DNA research as producing unease in "what it means to be a human being" because the research is an essential step in manipulating a person's genetic identity. Scientific advance of such impact will contribute to society's fear of technology generally. See Tribe, Technolog, Assessment and the Fourth Discontinui." The Limits of Instrumental Rationality, 46 S. CAL. L. REV. 617, 649 (1973). ECOLOGY LAW QUARTERLY [Vol. 8:55 technology are capable of understanding it; even scientists in related fields are unable to comprehend all of the complexities. Since application of knowledge from more than one field of scientific inquiry is necessary to evaluate properly all of the risks, benefits, and social consequences of the new technologies, a complete analysis can be made only through the cooperative efforts of many specialists from different fields. Yet the public not only expects the government to evaluate these complex technologies, but demands a voice in the assessment. This dilemma is compounded because those who best understand the technologies, the researchers in the field, cannot be entrusted with impartial evaluation due to the possibility of bias. Recombinant DNA experimentation presents a situation in which the public, the government, and the scientific community have at- tempted to resolve this dilemma. 2 The problems raised are unique, for this is the first instance in which the recognition and control of risk necessarily involves the regulation of pure scientific research.3 Further regulation of this kind is certain to follow as other similar technologies are developed.4 Recombinant DNA research is unique because the scientists who developed and used the technique were the first to appre- ciate its consequences. They took steps to regulate themselves, despite their role bias, until the issues came to the attention of the public. Such self-regulation is a highly fortuitous occurrence; it cannot be relied upon in the future as a method of risk control. Instead, society must develop the means to recognize the potential risks of scientific development, or, barring such foresight, the capability to deal quickly and effectively with such risks as they become known. Imposition of any system of external controls on research meets with the argument that it is in the interest of society to maintain a sci- 2. For a description of recombinant DNA experimentation see text accompanying notes 37-41 infra. The term "gene-splicing" is used synonymously to recombinant DNA throughout this Article. Several books have been written about the recombinant DNA debate. Science Maga- zine's Nicolas Wade has written a straightforward history, The Ultimate Experiment (1977); Rolling Stone's science writer Michael Roger has written a more entertaining, yet scientifically lucid account, Biohazard(1977); June Goodfield's Playing God (1977) and Ted How- ard's and Jeremy Rifkin's Who Should Play God? (1977) view the controversy from a perspective that examines the general ethical questions in scientific research and experimentation. For a shorter primer on the debate that does not oversimplify the science involved, see Grobstein, The Recombinant DNA Debate, SCIENTIFIC AM., July 1977, at 22. 3. Another, earlier developed technology is nuclear power. To some extent the problems encountered in regulating nuclear power are relevant to the recombinant DNA debate and are considered in this Article. But an important difference is that nuclear power involves the application of atom splitting research; regulation does not reach directly the pure research itself. 4. For example, science is just beginning to develop the sophistication to experiment with humans and related species at a molecular level-an area fraught with potential benefits and risks. 19791 RECOMBINANT DNA entific community free from direct regulation.5 Those who make this argument see the value of an independent scientific community as of such overriding importance that research must be absolutely free from any direct restraint. Although the absence of control may result in some costs to society, the benefits of unhampered general inquiry far outweigh such costs. For example, control over scientific inquiry may discourage the "creative risk taking . . . and opportunities for seren- dipity on which original discovery depends."' 6 Also, direct regulation 7 may result in political encroachment upon legitimate research. It could provide government with a tool for exercising political influence over what should be a nonpolitical enterprise. The effect of such influence could be to focus scientific expertise on practical short-term results rather than on long-term benefits associated with the accumulation of basic scientific knowledge." A related fear is that a regulatory system

5. See generally McGill, Science and the Law, 23 CATH. LAW. 85 (1978); Delgado & Millen, God, Galileo, and Government: Toward Constitutional Protection for Scientiic In- quiry, 53 WASH. L. REV. 349 (1978); Jonas, Freedom of Scientific Inquiry and the Public Interest, HASTINGS CENTER REPORT, Aug. 1976, at 15; Lederberg, The Freedom and the Control of Science:" Notes from the Ivory Tower, 45 S. CAL. L. REV. 596 (1972). Much has been written on the legal and policy issues of the recombinant DNA research controversy. See generally Comment, Considerations in the Regulation of Biological Re- search, 126 U. PA. L. REV. 1420 (1978); Note, Recombinant DNA, supra note 1;Berger, Government Regulation of the Pursuit of Knowledge. the Recombinant DNA Controversy, 3 VT. L. REV. 83 (1978); Comment, DNA and the Congressional Prerogatives. Proposals for a Deliberate Legislative Approach to Genetic Research, 53 IND. L. J. 571 (1978); Balmer, Re- combinant DNA: Legal Responses to a New Biohazard, 7 ENVT'L L. 293 (1977); Comment, Law vs. Science: Legal Control of Genetic Research, 65 Ky. L. J.880 (1977); and Symposium, Biotechnology and the Law- Recombinant DNA and the Control of Scientfic Reseasrch, 51 S. CAL. L. REV. 969 (1978) [hereinafter cited as Recombinant DNA Research Symposium]. For an examination of international implications, see Comment, Genetic Manipulation: Research Regulation and Legal Liability Under International Law, 7 CAL. W. INT'L L. J. 263 (1977); Comment, The Potentialfor Genetic Engineering. A Proposalfor International Legal Control, 16 VA. J. INT'L L. 403 (1976); and Kamely & Curtin, InternationalActivities in Recombinant DNA Research, RECOMBINANT DNA TECH. BULL., Fall 1977, at 28. For a focus on the National Environmental Policy Act, see Chalker & Catz, Case Analysis of NEPA Implementation: NIH and DNA Recombinant Research, 1978 DUKE L. J. 57. 6. 195 SCIENCE 939 (1977) (editorial by E. Shrieur, Pres., Biosystems Ass'n) [hereinafter cited as SCIENCE EDITORIAL]. For the view that the rubric of "freedom of inquiry" should not automatically shield experimentation, see Lapp, The Non-Neutrality of Hypothe- sis Formation, in SCIENCE ETHICS AND MEDICINE 96 (H.T. Englehardt, Jr. & D. Callahan ed's 1976), reprinted in Science Policy Implications of DNA Recombinant Molecule Research: Hearings Before the Subcomm. on Science, Research, and Technology of the House Comm. on Science and Technology, 95th Cong., 1st Sess. 1022 (1977) [hereinafter cited as 1977 House Hearings]. 7. One example of this sort of influence is Senator -William Proxmire's well-known "Golden Fleece Award," given to federally funded experiments that appear to be devoid of merit. Publicizing individual experiments distorts their ultimate worth; they can only be properly evaluated in the context of a pattern of research in the field. Consequently this award may be more harmful than beneficial, especially in light of the much greater effect on reseach decisions of this type of indirect regulation than safety guidelines. See Berger, supra note 5, at 107. 8. For example, "result" emphasis could have led polio researchers of the 1940's and ECOLOGY L4W Q UARTERL Y [Vol. 8:55 could be used to repress scientific knowledge and discovery. History provides many instances of political repression of scientific inquiry. 9 Opponents of regulation fear most, however, not the possibility of to- talitarian control, but the growing public disillusionment with science. They worry that a misinformed public might demand restrictive regulations for a new scientific technology, a demand leading to unreasonable limitations on scientific freedom.' 0 Although some of the above concerns are justified, it does not follow that new technologies should be allowed to develop without regulation. Counterbalanced against the interest in maintaining an autonomous scientific community is the public interest in controlling the risks of experimentation. As scientific concern has shifted from the observation to the control of natural phenomena, risks of harm to humans and to the environment have increased in both degree and kind. II Some new technologies, as recombinant DNA research, pose potential risks to human health and to the environment. Regulation of such research, having as its principal goal the minimization of these risks, should not be seen as an attempt to directly manipulate the course of scientific inquiry for political or other reasons. The danger of regulation lies not in the good intentions of present policymakers, but in the potential tool that is provided for those with less benevolent mo- tives. It should be possible to formulate a regulatory system which minimizes the potential for abuse by limiting restrictions on scientific merit and content of experiments. Furthermore, much of the scientific research presdntly conducted in the United States is supported by public funding. Since, in addition to bearing some of the risks of research, the public has an interest in determining how its financial resources are allocated, it accordingly should have the opportunity to provide input on its desires concerning the conduct of research and the measures taken to minimize those risks.

1950's to develop a portable iron lung instead of two polio vaccines. SCIENCE EDITORIAL, supra note 6, at 939. See also Hearingson Genetic Engineering.- Examinationof the Relation- ship of a Free Society and Its Scientfic Community Before the Subcomm. on Health of the Senate Comm. on Labor and Welfare, 94th Cong., ist Sess. 26 (1975) (testimony of Dr. H. Holmon) [hereinafter cited as 1975 Genetic Hearings]. 9. For example, during the Stalin regime, a Russian geneticist, T.P. Lysenko (Director of the Institute of Genetics of the U.S.S.R. Academy of Sciences 1944-1965), asserted that characteristics were acquired through the environment, rejecting the existence of genes and chromosomes. Since this theory fit Communist Party dogma, it was adopted as fact; genetic research stagnated as a result. L. Slesin, Recombinant DNA Research: A Chronology 8 (Nov. 10, 1976) (Occasional Paper No. 2, MIT School of Arch. & Planning). 10. See, e.g., McGill, supra note 5, at 92. 11. J. GOODFIELD, PLAYING GOD: GENETIC ENGINEERING AND THE MANIPULATION OF LIFE 152 (1977) (citing Dr. Willard Gaylin, Director of the Institute of Society, Ethics, and the Life Sciences). 19791 RECOMBINANT DNA

Hence, societal values attached to environmental risk should be considered when allocating funds to pure research.1 2 Societal values concerning risk can be characterized to some extent. 13 Firstly, society is averse to acceptance of large magnitude harms, even though the probability of widespread or catastrophic damage be extremely small. ' 4 This attitude stems from an awareness of the effects of catastrophe upon thousands, perhaps millions, of people si- multaneously. 15 If a new technology would create a substantial probability of catastrophic loss, the risk averse nature of society could doom its development. Secondly, society prefers risks that it accepts voluntarily to risks that are forced upon it. 16 Hence, when new technologies yield involuntary risks, either those risks must be relatively small or the benefits of the technology great to overcome the societal preference. Thirdly, society also prefers that environmental risks from new technologies fall on different segments of society equitably. When risks and benefits are not distributed equitably, and no adjustment of the

12. Even though "recent studies... [show] that the public ranks biomedical research high on its list of priorities," the limited resources of the federal government require that allocation decisions must still be made. Steinfels, Biomedical Research and the Public: A Reportfrom the Arlie House Conference, HASTINGS CENTER REPORT, June 1976, at 21. Bi- omedical research must compete with health care delivery, financing, and preventitive services, and the entire medical sector must compete with other social programs such as housing improvement and environmental control, which have an equal, or even greater, impact upon health. Id at 22. 13. For an in-depth discussion of the following concepts, see Page, 4 Generic View of Toxic Chemicals and Similar Rirks, 7 ECOLOGY L. Q. 207 (1978). 14. Id at 240. The concept of risk aversion is based upon reluctance to take an ac- tuarily fair (1: 1) or even favorable (2:1) gamble. For example, one may be willing to wager $5 on the outcome of a sporting event, yet be unwilling to wager $500 on the same event at the same odds. Analogously, society may be willing to accept a 100% chance of 10 annual deaths for some activity, but a 1%chance of 1000 annual deaths may be intolerable. If the event causing death or injury is one that is rare, the social aversion to risk may be even greater. Common or understood risks-for example, auto accidents-are acceptable to a greater degree than unique or bizarre risks such as radiation death or mysterious plagues. While some of this greater aversion to the bizarre risks stems from the uncertain probability and magnitude of these risks, the remainder results from public fear of the strange event. Society is more willing to accept greater cost to lower the risk of such rare events than for more common occurrences. 15. Traditional loss-spreading mechanisms such as insurance cannot deal adequately with catastrophes. Id at 213. 16. Starr, Benefit-Cost Studies in SociotechnicalSystems, reprintedin PERSPECTIVES ON BENEFIT-RISK DECISION MAKING 19, 30 (1972) (report of a Colloquim Conducted by the Committee on Public Engineering Policy, National Academy of Engineering, Apr. 26-27, 1971) [hereinafter cited as COPE]. Starr examined risk in terms of death and injury, the effect of public regulation, the length of time during which the risk has been present, and whether the risk is voluntary. He concluded that "the public appears willing to accept voluntarily risks roughly 1,000 times greater than involuntary exposure risks." Id. at 41. For a critique, see H. Otway & J. Cohen, Revealed~references."Comments on the Starr Beneft-Rik Relationships. Int'l Inst. for Applied Systems Analysis Research Memorandum (Mar. 1975) (Laenburg, Austria). ECOLOGY LAW QUARTERL Y [Vol. 8:55

distribution is possible, a decision to continue development of a new 7 technology becomes much more difficult to justify.' Lastly, temporal inequity, which arises whenever the benefits of a new technology are enjoyed presently but the costs are to be borne by a future generation,' 8 is disfavored by society. A decision to develop a new technology may result in latent and irreversible environmental costs. 19 For example, while present generations would enjoy the benefits of electricity generated by nuclear power plants, future generations 20 would bear the irreversible effects of faulty nuclear waste disposal. Taking temporal inequity into consideration when deciding whether to continue or regulate new technology development leads to a more cau- 21 tious approach to development. Even if societal values are not given much weight in determining the course of scientific research, the public should at least be informed of the risks resulting from a type of research. In the medical profession, the doctrine of informed consent requires that a patient be informed of the risks of a medical procedure before he or she is asked to consent to it. Minimally, an analogous principle should apply to scientific re- 22 search. By using the development of recombinant DNA technology as a case study, this Article demonstrates the application of societal values in deciding whether to continue the development of a new technology. It then develops a structure to regulate sophisticated new technologies. In section I, a brief description of recombinant DNA technology is given, followed by a detailed discussion of the risks and benefits of conducting such research, concluding that societal interests dictate that research should continue only if safety regulation is imposed upon the experiments. Section II examines various methods by which this safety margin can be imposed, an analysis which concludes that only direct

17. One possible method of adjusting the balance would be to require the group en- joying disproportionate benefits to insure the group suffering disproportionate risks against damage. 18. Page, supra note 13, at 241. 19. Id at 213-14. 20. See generally Lucas, Nuclear Waste Management. A Challenge to Federalism, 7 ECOLOGY L. Q. 917 (1979). 21. Page, supra note 13, at 241. 22. 1975 Genetic Hearings,supra note 8, at 17-18 (statement of Dr. H. Holman); Recom- binant DNA Regulation Act, 1977 Hearing on S. 1217 Before the Subcomm. on Health and Scientific Research of the Senate Comm. on Human Resources, 95th Cong., 1st Sess. 145 (1977) (testimony of Dr. H. Holman) [hereinafter cited as 1977 Senate Hearings]; J. GOOD- FIELD, supra note 11, at 152; SCIENCE POLICY RESEARCH DIVISION, CONGRESSIONAL RE- SEARCH SERVICE FOR THE HOUSE COMM. ON SCIENCE AND TECHNOLOGY, 94th Cong., 2d Sess., REPORT ON GENETIC ENGINEERING, HUMAN GENETICS, AND CELL BIOLOGY 36 (Supp. Rep't 11 1976) [hereinafter cited as REP'T ON GENETIC ENG'R]. See Holman & Dut- ton, A Casefor Public Participationin Science Policy Formation and Practice, Recombinant DNA Research Symposium, supra note 5, at 1505. 19791 RECOMBINANT DN4 regulation by a government agency can successfully impose the requi- site safety. Finally, section III proposes a framework for government regulation of recombinant DNA research, setting forth principles applicable to regulation of other new technologies.

I RISK-BENEFIT ANALYSIS OF RECOMBINANT DNA RESEARCH A. Description of Recombinant DNA Research 1 Basic Molecular Biology All organisms may be classified as either prokaryotes or eukaryotes according to the type of cells that constitute the organism. Pro- karyotic cells are the most primitive, i e., the simpler of the two cell types, with genetic material consisting of simple molecular threads of deoxyribonucleic acid (DNA).23 Bacteria are examples of prokaryotes. Eukaryotic cells are more complex, with the genetic material consisting of DNA organized into a set of chromosomes. 24 Some unicellular organisms, such as yeast, as well as all multicellular organisms, are eukaryotes. Although there are many important structural and bio- chemical differences between the two cell types, they also have fundamental similarities. 25 The most important similarity between the two is the universality of DNA as the genetic material that contains all of the genetic information about the particular organism. Consequently, the results of studying prokaryotes, which are more easily examined because of their simpler structure, can be applied to eukaryotes. The structure of the DNA molecule is usually a double helix, which can be imagined as a pair of spiral staircases, one on top of and parallel to the other. The two strands are linked by molecular bridges, called bases, which extend from each spiral. Each base must pair with another from the other strand to form a bridge; thus, each of the helices is a pairing image of the other. 26 The sequence of the bases on a DNA 7 strand dictate the genetic characteristics of a particular type of cell, 2 and the transmission of the DNA molecule from one generation of cells 28 to the next is responsible for heredity.

23. G. STENT, MOLECULAR GENETICS 43 (1971). 24. Id 25. Id 26. Watson & Crick, A Structurefor Deoxyribonucleic Acid, 171 NATURE 737 (1953). The bases are adenine (A), guanine (G), cytosine (C), and thymine (T). (A) can pair only with (T) and (G) only with (C). Id 27. Every set of three bases on a DNA strand codes for one of 20 amino acids that are the building blocks of proteins, the essential elements of life. The mechanism of transcrip- tion and translation of this information from the code into protein will not be considered here. For a discussion of this process, see G. STEWT, supra note 23, at chs. 16, 17; J. D. WATSON, MOLECULAR BIOLOGY OF THE GENE, chs. 11-14 (2d ed. 1970). 28. G. STENT, supra note 23, at 179-84. Heredity between cells and heredity between ECOLOGY LAW QUARTERLY [Vol. 8:55

Replication of DNA generally occurs with each strand serving as a template for the formation of a new complementary strand. The two helices separate and each base on a strand attracts a complementary base manufactured by the cell. The bases are sewn together to form a chain. By this mechanism the original parent molecule generates two identical daughter molecules; half of each is the original parent and half is new strand. 29 When the cell divides into two new cells, each 30 daughter DNA carries the genetic information for each new cell.

2. Molecular Genetics and Bacteria Most of the information concerning molecular biology has come from experiments with bacteria. 3' Eukaryotic cells are too difficult to work with and too complex to yield the quality and quantity of information provided by bacteria. Furthermore, many bacteria have a life cycle of less than an hour. Billions can be grown overnight from a single parent cell with an inexpensive amount of growth medium, thus 32 making them ideal for the study of hereditary information. E. coli has become the organism most often chosen for recombinant DNA experiments. Its natural habitat is the intestine of animals, including humans, where it aids in the digestion of food.33 E. coli is valuable for the study of gene expression because its molecular biology has been studied in detail and is well understood. One third of its three 34 to four thousand genes have now been characterized. E. coli possesses the ability to transfer a gene from itself to another bacterial cell, including cells other than E. co/i. Although not unique to E co/i, the genetic material transfer mechanisms are reasonably understood only in E. co/i. 35 These mechanisms are of use to scientists parent and child organisms are both dictated by DNA molecule transmissions, albeit through different processes. Id. ch. 1. 29. Actually, replication is not quite so simple. It has been well studied only in a few of the simple prokaryotes, specifically bacteria. For example, the species of bacteria presently used in most recombinant DNA experiments, Escherichia coi (E. col), has all its DNA in the form of one circle. The circle replicates from one "Y-fork" which moves around the circle unwinding the strands for replication. An "unwinding" protein is required to help. It has been estimated that DNA replicates at a rate of 1400 bases per second, requiring E coli DNA to unwind at 84,000 rpm. G. STENT, supra note 23, at 240-41. 30. Id at 10. 31. Scientists have and will continue to rely principally on bacteria for experiments on heredity. Recombinant DNA experiments are no exception. In particular, these experiments will use E. col K12, a laboratory strain of E co/i. National Institutes of Health, U.S. Dep't of Health, Education, & Welfare, Recombinant DNA Research, Revised Guidelines, 43 Fed. Reg. 60,080, 60,115-19 (1978) [hereinafter cited as NIH Revised Guidelines]. For a further discussion of this strain, see note 83 infra. 32. G. STENT, supra note 23, at 42. 33. REP'T ON GENETIC ENG'R, supra note 22, at 9. 34. Golden, Tinkering with Life, TIME, Apr. 18, 1977, at 34. 35. For a short description of E. coi transfer mechanisms, see H. MAHLER & G. 19791 RECOMBINANT DNA performing recombinant DNA experiments because such experiments usually require the E. coli to perform at least one of these transfers. This ability to transfer genetic information is the cause of much of the risk of recombinant DNA research as well; transfer from a harmless cell to a potentially more harmful one is always a possibility.36

3. Recombinant DNA Technology Recombinant DNA experiments involve the construction of a chimera DNA molecule, so called because of the analogy to the mytholog- ical Chimera-an animal with the head of a lion, body of a goat, and tail of a serpent.37 The chimera of molecular biology is an artificially constructed DNA molecule in which two or more DNA segments- possibly from very different species-are joined to form a single independently replicating DNA molecule. Experiments with recombinant DNA molecules consist of four essential elements: (1) a method of artificially breaking and joining DNA molecules that may be derived from different sources;38 (2) a vehicle to carry the foreign piece of DNA into the bacteria so that the vehicle-DNA complex is replicated; 39 (3) a

CORDES, BIOLOGICAL CHEMISTRY 850-54 (1971); a more detailed description is contained in G. STENT, supra note 23, chs. 7, 10, 14. 36. There are several types of genetic recombination that appear naturally in E. coli. Sexual conjugation consists of transfer of some or all of the genes of one E coli cell into a second, passive cell. Other mechanisms of gene transfer in E co/i are general and restricted transduction. These occur when a virus infects the bacteria, but does not replicate itself and kill the host. Instead, the virus is incorporated into the bacteria's DNA, or genome. The virus may break away from the genome after it has been incorporated, taking with it a portion of the E. co/i genome and leaving behind some of its own genes. If the same type of virus always carries away the same genes, the process is known as restricted transduction. More rarely, in general transduction, the virus removes genes in a random pattern. See generally G. STENT, supra note 23, ch. 14. A final gene transfer process is transformation. In this case a recipient E. coli can be genetically transformed by the addition of DNA derived from some other organism, not necessarily E. coil. One or more genes from the donor DNA are incorporated into the chromosome of the recipient cell. Although this process does not occur in nature very easily because of the difficulty in incorporating DNA through the cell wall, it is a relatively simple process in an experimental laboratory. Seegenerall, id ch. 7. 37. Cohen, The Manqtulation of Genes, 233 SCIENTIFIC AM. 22, 28 (1975). 38. There are two methods available for artificially breaking and joining DNA molecules to form a chimera. If two entirely separate double-stranded DNA molecules are brought together they can be sealed by an enzyme to form a chimera. However, it is highly improbable that two ends of DNA would randomly meet in the presence of the enzyme and be sealed to one another. Cohen, supra note 37, at 26. The solution is to use DNA with an "overhand" of several base pairs, each complementary to the base pairs in the other DNA. Then if the ends happen to meet they can stick together until the enzyme can seal them. The two methods of joining DNA strands both utilize this solution. 39. The foreign DNA is attached to a vehicle or "vector" which then reproduces itself and the foreign DNA to form large colonies of the recombinant DNA-vector complexes. Viruses and plasmids, which can replicate and maintain an independent existence in the host cell, are the most likely vehicles. A good vehicle should be able to enter a cell, replicate well in the cell, and be small in size. Cohen, supra note 37, at 25. ECOLOGY LAW QUARTERL Y. [Vol. 8:55 mechanism to introduce the carrier and its passenger, the foreign piece of DNA, into a bacterial cell where the vehicle-DNA complex can be replicated; 40 and (4) a means of identifying those cells that have re- 4 ceived the passenger and vehicle DNA. ' Recombinant DNA technology is a procedure. The purpose of the procedure is to allow experimentation with novel gene combinations. The procedure can be used in many different experiments; some that could be performed without the new technique, but less easily, and some that could not be performed at all.

B. Risk-Benefit Analysis of Recombinant DNA Research In the specific case of recombinant DNA experimentation, the initial consideration should be whether gene-splicing research should be conducted at all, and, if so, what degree of control should be imposed.

40. The mechanisms for introducing a vehicle into a host cell are fairly simple. Viruses can infect cells naturally, and so need no assistance. See note 36 supra. Plasmids can be introduced if the E coi cells are treated chemically. Although only about one in a million treated cells will accept a plasmid, each milliliter of a liquid culture contains a billion bacteria, thus producing enough E. coli cells for experimental purposes. Either device enables one to complete the coupling of genomes from two different organisms. Cohen, Chang, & Hsu, Nonchromosomal Antibiotic Resistance in Bacteria.- Genetic Transformation of Esher- ichia coli by R-factor DNA, 69 PROC. NAT'L ACAD. Sci. USA 2110 (1972). Mandel & Higa, Calcium-DependentBacteriophage DNA Infection, 53 J. MOL. BIOL. 159 (1970) (letter to the editor). 41. Cohen, supra note 37, at 25. Once the vector is introduced into a bacterial cell, the chimera cells must be selected from a population of heterogenous cells. The selection process is simple because of the nature of the vectors. For example, to isolate a particular vector known as pSC 10, the bacteria are grown on tetracycline. Only those E coli with tetracycline resistance, imparted by pSCI01, can grow. To make certain the recombinant DNA as well as the vector exist, the presence of the foreign gene is tested. For example, if the recombinant gene that is added to the vector produces resistance to a second drug, say karomycin, then the screening medium will contain both drugs. Colonies that grow must contain both the vector and the gene it carries. The result is that a foreign gene has been inserted into an E. co/i cell and will be replicated every time the cell splits into two daughter cells. Cohen, Chang, Boyer, & Helling, Construction of Biologically FunctionalBacterial Plasmids in Vitro, 70 PROC. NATL. ACAD. Sci. USA 3240 (1973). An acknowledged benefit of delay was demonstrated by the thalidomide episode of the early 1960's. Birth defects resulting from pregnant women's use of the drug had not been immediately discovered. The FDA's delay in approving use of the drug enabled researchers to analyze data and to discover the harm. Merrill, Risk-Benet Decisionmakingby the Food and Drug Administration, 45 GEO. WASH. L. REV. 994, 1004. This "delay saved hundreds, thousands of children in the United States from being malformed .... " by thalidomide. 121 CONG. REC. 14,715 (1975) (E. Kennedy). However the costs of delay must be considered. Free-market economist Milton Fried- man argues that with even extremely favorable assumptions, the loss of benefits is ten to one hundred times greater than the gain from preventing thalidomide-type episodes. M. Fried- man, FrustratingDrug Advancement, NEWSWEEK. Jan. 8, 1973, at 49 (relying on the results of researcher Samuel Peltzman). S. Peltzman, REGULATIONS OF PHARMACEUTICAL INNO- VATION: THE 1962 AMENDMENTS 55 (1974). But see the FDA's defense of this critique. Wade, Drug Regulation.- FDA replies to Charges by Economists and Industry, 179 SCIENCE 775 (1973). 1979] RECOMBINANT DNA

The risks could be so minimal and the benefits so great that any control would only reduce the flow of benefits without concomitant reduction of risk-all at great expense. On the other hand, controls of some kind may be the only means of bringing the benefit-risk ratio to an acceptable level. If this is the case, then the research should continue, and only the method of imposing the controls would remain to be determined. Before choosing a procedural option, the specific benefits and risks of genetic research must be detailed and analyzed. The correct option should become apparent once this analysis has been completed.

I. Uncertainty in the Assessment of the Risks and Benfts of Recombinant DNA Research Before considering the specific risks and benefits of recombinant DNA research, a caveat is necessary. 42 Any assessment of the risks and benefits is burdened by uncertainty stemming from ignorance of the mechanism by which genes are controlled and operated, of some of the principles of host or vector escape, of the interaction of hosts or vectors with other organisms resulting in survival and transmission of dangerous genes, and of the manner by which foreign genes can harm humans or other organisms valuable to human life. Concern about these problems, as expressed in public discussions of recombinant DNA research, has been overstated, however. A major purpose of present work in this field is to resolve fundamental scientific uncertainties and contribute to the available pool of knowledge. To- ward this end, preliminary experiments have been specifically directed at areas of doubt.43 It may be possible to design other experiments to further reduce problems surrounding current research and conduct them prior to or concurrently with normal recombinant DNA research.44 The assessment of benefits and risks of any new technology may involve great uncertainty. A procedure for dealing with the uncertainty

42. This Article does not present a theoretical discussion of environmental risk-benefit analysis. For such a discussion see Symposium: Risk-Beneit Assessment in Governmental Decisionmaking,45 GEO. WASH. L. REV. 901-1150 (1977). One notable authority does not believe formal risk-benefit analysis can be applied to recombinant DNA research because of the great uncertainty in the possible outcomes. Science Policy ImplicationsofDNA Recombi- nant Molecule Research- Hearings Before the Subcomm. on Science and Technology, 95th Cong., 1st Sess. 794 (1977). Nonetheless, the alternative of intuitive decisionmaking is unacceptable, see Page, supra, note 13 at 239 n.79. Consequently, this section will present an informal risk-benefit analysis of recombinant DNA and attempt to account for the uncertainties involved. 43. See note 85 infra. 44. The National Institutes of Health is presently supporting studies assessing the risks of DNA experimentation in the general categories of (1) conversion of E. coli to an epidemic pathogen, and (2) transmission of DNA from E. coi to other organisms. NIH Revised Guidelines, supra note 31, at 60,088. ECOLOGY LAW QUARTERLY [Vol. 8:55 must be adopted for recombinant DNA and its successor problem technologies. Several procedural options exist. The optimistic approach is to allow development to continue until additional information is obtained to establish adequate proof of the risk. In many situations, society may not want to surrender significant benefits to avoid conjectural dangers. People may hope that the materialize, that they can be minimized or that, if hazards will never 45 necessary, the activity can be halted once the risks become certain. This approach ignores the societal value of risk aversion, the bias in favor of benefits when they are relatively certain and the disadvantages are speculative 4 6 the possible irreversibility of damage after it occurs, and the fact that an absence of present harm is not proof of an absence of risk. Consequently it is a poor procedure for dealing with uncertainty. A second approach is to place the "burden of proof" on the advocates of new technologies and force them to prove that development would be "safe" despite the uncertainty. In other words, supporters of a proposal have the job of reducing uncertainty. The practical effect is 47 a rebuttable presumption against development of the technology. A third procedure is to delay development. While many would consider delay to indicate a decision that the risks are unacceptable, such an approach does serve more than just a negative purpose where the effects of an affirmative policy choice may be irreversible. The purpose of delay is to allow time for additional information gathering so that uncertainty will be reduced and a more precise benefit-cost ratio calculated. In addition, delay also provides "more time for mature consideration and/or development of approaches that may minimize adverse consequences. '48 Thus, delay is useful for some environmental

45. This approach is used, many times incorrectly, in most environmental decisions. Page, supra note 13, at 230-33. 46. Green, The Risk-Beneft Calculusin Safety Determinations,43 GEo. WASH. L. REV. 791, 804 (1975) [hereinafter cited as Green, Safety Determination]; Green, The Adversary Process in Technology Assessments, in TECHNOLOGY ASSESSMENT: UNDERSTANDING THE SOCIAL CONSEQUENCES OF TECHNOLOGICAL APPLICATIONS 51 (R. Kasper ed. 1972). 47. Cf.Environmental Defense Fund v. Ruckelshaus, 439 F.2d 584, 593, 2 ERC 1114, 1119 (D.C. Cir. 1971), where the court interpreted the Federal Insecticide, Fungicide, and Rodenticide Act §§ 4(c), 6, 7 U.S.C. §§ 135b(c), 135d (1976), as requiring the manufacturer to bear the burden of proving the safety of his product when any substantial question of safety arises as to pesticide products. The public clearly wants to place the burden on the field of science that there is no hazard. 1976 HEW DOCUMENTS, supra note 1, at 241-2 (testimony of P. Hutt). See also id. at 244; 41 Fed. Reg. 27,904 (July 7, 1976). But see 1976 HEW DOCUMENTS, supra note 1, at 314 (testimony of R. Edgell), advocating placing the burden on the regulators, and after five years, removing existing regulations unless the burden of proof is met. The rationale focuses upon the harms of delaying scientific advance- for example, a delay of ten years in the introduction of polio vaccine. See also Page, supra note 13, at 234, 236-39, and note 45, infra. 48. Green, Law and Genetic Control Public Policy Questions, in Ethicaland Scieniqfc Issues Posed by Human Uses of Molecular Genetics, 265 ANNALS N.Y. ACAD. SCI. 173, 175 (1976) [hereinafter cited as Green, Law and Genetic Control]. 1979] RECOMBINANT DNA decisions. Finally, a wide safety margin can be provided in developing the technology. This approach assumes that the risks are highly uncertain while the benefits are relatively certain. It attempts to limit the development of dangerous technologies through pessimism. If after societal values have been considered the benefit-risk ratio exceeds unity with a margin of safety, the policy is acceptable. This method may result in discarding development of a beneficial technology, but that possible consequence is inherent in the pessimism of the procedure.

2 Benefits of Recombinant DN4 Research Gene-splicing experiments are powerful tools for the study of mammalian DNA and will allow significant additions to theoretical scientific knowledge. The importance of this contribution to the study of DNA structure flows from the complexity of mammalian organisms. Because mammals have genomes a thousand times larger and exceed- ingly more complex than E. coil, 49 and because it is extremely difficult to extract more than minute quantities of DNA from a mammalian cell, traditional techniques of identifying and mapping the various genes of a mammalian genome are not productive. 50 Through recombinant DNA techniques, however, bacteria can be made to replicate mammalian DNA in sufficient quantities for easy identification and 51 mapping of genes. Recombinant techniques also contribute significantly to studies of the function and expression of mammalian genes and their overall regulation. Since E. co/i eventually will be made to produce the protein product of an inserted mammalian gene, 52 recombinant techniques will lead to a better understanding of the expression of the inserted gene and its effect upon the functioning of the protein product. The regulation of genes and their expression is the most poorly understood aspect of cellular research. Some proteins exist only in a few copies in an entire cell; others exist in larger quantities but only at certain periods in the life of a cell.53 By isolating these proteins,

49. P. Berg, Mapping the Mammalian Genome 5 (unpublished paper presented at the National Academy of Sciences Forum on Recombinant DNA Research, Mar. 7-9, 1977) (copy on file at the Ecology Law Quarterly office). Only 150 out of tens of thousands of human genes have been identified, while approximately one-third of the genes in E coi are known. Id at 3. 50. Id at 4-5. 51. A gene-splicing technique for mapping the genomes of higher organisms was recently developed. A recombinant DNA molecule is radioactively labeled and bound to the region of the genome where it belongs. The position of the bound segment is indicated by radioactive emissions that blacken a photographic emulsion. Id at 9. 52. See note 59 infra. 53. G. STENT, supra note 23, at 577. ECOLOGY L4W Q UARTERL Y [Vol. 8:55

whether found in large, small, or short-lived quantities, and studying their interaction in the relatively simple environment of E coli, much 54 can be learned about gene regulation. The practical benefits that will flow from an increase in general scientific knowledge made possible by gene-splicing research are not wholly predictable. However, in adding to the understanding of gene regulation in mammals, recombinant techniques will probably contribute to the knowledge of normal cell development, differentiation, and the abnormal cell growth characteristic of cancer. 55 Regulation of gene expression is the key to each of those processes.56 This potential contribution to cancer and aging research is the most important example of the nexus between the increased knowledge recombinant techniques will bring and possible future benefits to mankind. In any event, it is not at all clear that an increase in scientific knowledge must promise predictable and specific benefits in order to be justified. The accumulation of scientific knowledge may be an end in itself with its own cultural value. History indicates that practical benefits from research are not predictable. Therefore, the best approach is to have a broad field of scholarly inquiry, creating a base of scientific acumen. The industrial and medical spin-offs from recombinant DNA research are many. Virtually any industrial process that uses prokaryotes

54. Critics of gene-splicing experiments have suggested recombinant technology is not the tool its acolytes claim it to be. The technology is considered by some to be merely a "big sewing machine"; useful, but not essential. Wade, Gene Splicing.- Congress Starts Framing Lawfor Research, 196 SCIENCE 39 (1977) [hereinafter cited as Wade, FramingLaw for Re- search]. Most commentators agree, however, that there is not a substitute for recombinant DNA technology among traditional experimental methods. Public lecture by Dr. B. Davis, Darwin, Pasteur,and the Andromeda Strain, [hereinafter cited as Davis Lecture], as printed in REP'T ON GENETIC ENG'R, supra note 22, at 251, 253 [hereinafter cited as Davis Lecture]; NATIONAL INSTITUTES OF HEALTH, U.S. DEP'T OF HEALTH, EDUCATION, AND WELFARE, PUB. No. 1489, ENVIRONMENTAL IMPACT STATEMENT ON NIH GUIDELINES FOR RESEARCH INVOLVING RECOMBINANT DNA MOLECULES (pt. 1) (1977) [hereinafter cited as I FINAL EIS]. The tedious nature of classical techniques precludes their use even for the specific methods that are available for isolating DNA sequences. Even critics of the new technology agree that recombinant techniques are less expensive and less time-consuming than classical methods. Oversight Hearing on Implementation of NIH Guidelines Governing Recombinant DNA Research:"Joint HearingBe/ore the Subcomm. on Health of the Senate Comm. on Labor andPublic Welfare andthe Subcomm. on Administrative Practiceand Procedureof the Senate Comm. on the Judiciary,94th Cong., 2d Sess. 78 (1976) [hereinafter cited as Oversight Hear- ing]. 55. D. Nathans, Benefits of Research with Recombinant DNA 6 (unpublished paper presented at the National Academy of Sciences Forum on Recombinant DNA Research, March 7-9, 1977). No scientist claims that recombinant DNA will lead to the cure of cancer, or any other disease, but recombinant DNA methodology does offer an opportunity to conduct experiments that are not practical using classic techniques. In that sense, then, it may lead to significant developments in understanding and curing diseases. 56. Davis Lecture, supra note 54, at 253. 19791 RECOMBINANT DNA

can be improved with recombinant techniques.5 7 One particular novel application involves nitrogen-fixing plants. While most plants obtain their nitrogen from chemical fertilizers, a few plants-soybeans, alfalfa, clover, and peanuts-get nitrogen from bacteria that live in association with their roots. Using recombinant techniques, researchers can tranfer genes from these bacteria to others in the soil, or even directly to the 5 8 plants, to aid the growth of all plants. The potential medical benefits are no less significant. For example, human insulin will be producible in bacterial "factories" much less expensively than was previously possible.59 It may also be possible for bacteria to produce bloodclotting factors for hemophiliacs. Currently these factors are produced in small quantities at great expense. New production techniques could alleviate the shortage and expense of these

57. The discussion here centers on the benefits of manipulating genes in prokaryotes and lower eukaryotes only. For a discussion of direct manipulation of human genes, see text accompanying note 114 infra. The usefulness of these techniques becomes apparent when one notes that manufacturing processes using bacteria are widespread in the food and beverage industries. In addition, there are novel industrial possibilities. REP'T ON GENETIC ENG'R., supra note 22, at 35. 58. R. Valentine, Environmental Impact of Nitrogen Fixation (NIF) Genes 1-2 (unpublished paper presented at the National Academy of Sciences Forum on Recombinant DNA Research, Mar. 7-9, 1977) (copy on file at the Ecology Law Quarterly office). Another possibility is insertion into bacteria of various genes which would enable the bacteria to degrade oil. REP'T ON GENETIC ENG'R, supra note 22, at 35. In the event of an oil spill, these bacteria could be used literally to eat the oil and die out when the oil food source is gone. 59. Insulin is presently supplied from the pancreases of pigs and cows. The process is expensive and unsatisfactory because of frequent allergic reactions to insulin from these sources. Additionally, there has been a global supply problem. San Francisco Chronicle, Sept. 7, 1978, at 22, col. 5. Insulin produced in bacterial factories will be inexpensive, available, and structurally identical to human insulin. Id A breakthrough toward this end was achieved in 1977 when insulin genes from a rat were placed in an E. coli cell and were replicated. Ulrich, Shine, Chirgwin, Pietet, Tischer, Rutter, & Goodman, Rat Insulin Genes: Construction of Pla-mids Containingthe Coding Sequences, 196 SCIENCE 1313 (1977). Al- though these genes were not translated by the bacteria into protein, researchers were later able to insert genes coding for human hormones and get the bacteria to translate them. Somatostatin was the first such hormone to be produced and insulin followed about six months later. San Francisco Chronicle, Sept. 7, 1978, at 22, col. 5. The genes inserted were not human genes however. They were artificial genes which coded for exactly the same insulin protein as found in human bodies. Id Nevertheless, the result is the first pragmatic benefit to actually result from recombinant DNA research. Bacteria now can be made to produce human insulin in large quantities. The most recent development has been the successful transfer of genes for animal growth hormone into E coil. This feat is a large step toward mass production of the scarce and expensive medication for treatment of dwarfism. The tasks remaining are to substitute human for animal growth hormone and to make the E. coli cell secrete the hormone it produces into surrounding growth medium, where it can be easily isolated. San Francisco Chronicle, Jan. 13, 1979, at 4, col. 5. Moreover, it is clear that the main problem for the above and other medical applications-the failure of natural human genes to function in E. co/i hosts--will not block these benefits for very long. Natu- ral human genes have functioned in eukaryote hosts. San Francisco Chronicle, Oct. 19, 1978, at 1, col. 4. The same result for bacterial hosts is almost certain to follow. ECOLOGY LAW QUARTERLY [Vol. 8:55 factors, making them widely available.60 In addition, there is the possibility that vaccines may be manufactured in bacteria more cheaply and safely than at present.6'

3. Risks of Recombinant DNA Research For harm to result from a gene-splicing experiment, a harmful organism must be produced, escape, reproduce, and survive long enough to injure another organism. 62 While the number of frightening scenarios that may be drawn is unlimited, it is very difficult to discuss any of these events in terms of probability. Many of them are so improbable that they pose no risk.63 Furthermore, probability estimates often contain large margins of error that render the estimate meaningless or even harmful; the estimates may evoke a false sense of certainty. Although all the probabilities of risk are small on an absolute scale, 64 two factors make these probabilities important. Firstly, the magnitude of such harm might be so great that the risk, measured by the product of probability and magnitude, would be unacceptable. Secondly, the cultures used in recombinant DNA experiments contain bacterial cells in concentrations as high as a billion per milliliter;65 with enough experimentation on cultures of this concentration even some of the incredibly small probabilities existing for recombinant DNA become significant. a. The possibility ofproducing a harmful organism Scientists generally agree that deliberately introducing into a recombinant DNA molecule a gene that codes for a dangerous toxin carries a relatively high risk of producing a harmful organism. 66 For example, experimenters could introduce the gene for botulinus toxin

60. Nathans, supra note 55, at 8; Davis lecture, supra note 54, at 253. 61. Nathans, supra note 55, at 9. Many.vaccines are derived from viruses grown in cell cultures, which are often contaminated by other viruses or by components of the cell culture dangerous to a patient. Id 62. The risks of recombinant DNA research are described here in finer detail than the benefits were in the last part. The reasons for this weighting are twofold. Firstly, when dealing with new and complex technologies, the largest problem is fear of the unknown. Discussing the risks in detail mitigates the irrational aspects of this fear. The societal risk aversion aspects of the fear can still be taken into account and, indeed, are considered in this section. Secondly, in writing legal articles, there is an inherent bias towards regulation of the subject. When an article reaches the conclusion that unregulated research may bejustifi- able, though not cost-effective, the risks must be definitively articulated. 63. Some bizarre risks have been mentioned by opponents of recombinant DNA research. For example, Mayor Alfred Valluci of Cambridge, Massachusetts, has expressed the fear that "[slomething could crawl out of the laboratory, such as a Frankenstein." TIME, Apr. 18, 1977, at 45. 64. Sinsheimer, Recombinant DNA--On Our Own, 26 BIOSCIENCE 599, reprinted in REP'T ON GENETIC ENG'R, supra note 22, at 249. 65. G. STENT, supra note 23, at 54. 66. 1976 HEW DOCUMENTS, supra note 1, at 180. 1979] RECOMBINANT DNA

(which causes botulism) into E co/i. If that gene is expressed in the E. co/i bacterium, the resulting organism would be dangerous. 67 Expres- sion of such a gene is not implausible, because such toxins are derived from other prokaryotes and could be translated by the machinery of the E. coli cell.6 8 Similarly, a dangerous organism may be created inadvertently by introducing into a host cell any portion of DNA derived from an organism pathogenic to humans.69 The portion of that organism making it pathogenic could be transferred to the host cell, e.g., E. co/i, producing the same result as deliberate introduction of toxin genes. Examples of such pathogens are smallpox, yellow fever, and measles viruses. The degree of risk posed by accidental introduction of pathogenic characteristics is less than in the foregoing type of research. The exact probability of danger, however, depends upon the nature of the patho- 70 gen and the technique of experimentation. Gene-splicing experiments also risk transferring resistance to drugs from a harmless to a pathogenic bacteria. For example, if a gene coding for drug resistance not normally found in E co/i is inserted by recombinant techniques, it may be transferred to a strain of bacteria harmful to humans. If no other available drug can effectively attack the pathogenic bacterial strain, the result could be a great increase in 7 incidence of the disease caused by the bacteria ' Unlike the risks discussed above, other perils associated with recombinant DNA research do not require gene expression in order to be hazardous. Fortunately, these risks are more speculative and generally far less likely to occur. For example, if a host bacterium carries a dangerous virus to a location within an organism where that virus could be pathogenic, a risk of harm is created even though the host cell does not translate the virus genome.72 It is also possible that the transfer of genes

67. Other genes known to be pathogenic to man that could produce harmful organisms include diptheria toxin and venoms from snakes. 68. Nathans, supra note 55, at 7. 69. Of course, humans are not the only organisms that could be harmed by formation of new pathogens. Animals and plants are not discussed as potential victims of the scenarios presented in this Article solely because of space limitations. 70. There is an extra element of risk in the case of accidental genetic modification not present for deliberate alteration--that the occurrence may be unanticipated and outside of procedures designed to safeguard against escape and contamination. 71. Such transfer of drug resistance between bacterial strains is known to occur naturally. Sherratt, BacterialPlasmids, 3 CELL 189, 193 (1974). When antibiotics first were used, researchers noticed an increase in resistance to drugs such as penicillin and tetracycline among strains of bacteria. See Dulaney & Laskin, The Problem of Drug-Resistant Patho- genic Bacteria, 182 ANNALS N.Y. ACAD. Sci. (1971). The probability of a resistance transfer is higher than that for most recombinant DNA hazards. Under the present legal regime, deliberate transfer of such traits is prohibited if the resistance could compromise control of any disease. NIH Revised Guidelines, supra note 31, 43 Fed. Reg. 60,080, 60,108. 72. Assume that, by the usual recombinant techniques, a virus genome is attached to a ECOLOGY LAW QUARTERLY [Vol. 8:55 into a host could change the host's behavior by interfering with its metabolic pathways. 73 Uncertainty regarding the mechanisms of interfer- ence makes a probability assessment difficult as well. Although many metabolic alterations would either kill an . co/i cell or make it less viable in any environment,74 there is a chance that the metabolic change would only alter the optimum environment of the cell without 7 affecting its viability. " The risk of producing a harmful organism depends to a large extent upon the purity of the foreign DNA inserted into a vector to form the chimera molecule. One controversial procedure that is currently used to achieve this insertion is the "shotgun" technique. An entire genome is chopped up and recombined with a vector using standard recombinant DNA technology. The result is a mixture of many different chimera molecules containing the various genes of the foreign genome. The key advantage of a "shotgun" technique is that it can be used to identify and characterize a vast number of the genes in mammalian genomes. The risk of the technique lies in the unknown chimera it will produce.76 In contrast, using "pure" genes to form a chimera has the advantage that only a single gene will be added, providing a greater measure of control over experimental results. The disadvantage of this method, however, is that it requires isolation and purification of the foreign genes, the very difficulty currently impeding progress in the identification of mammalian genomes. Moreover, the lack of control associated with "shotgunning" is not a great risk---contamination is the recombinant DNA vehicle and placed in an E co/i cell. If that E. co/i cell entered a human intestine, it could transfer the virus to cells present there. The transfer could occur either by transformation or some modified form of transduction. See note 36 supra. The virus, if harmful to these cells, could wreak havoc. A transfer between two disparate cells is unlikely because of evolutionary barriers. See text accompanying note 104 infra. The hazard of transfer is heightened, however, since the E. coi need not express the virus genome. 73. Insertion of genes into E. co/i could result in regulatory changes that enable the bacteria to inhabit a slightly different environment in humans. This new optimum environment may be higher in the intestine than before, allowing penetration of the bacteria between the cells of the small intestine and into the lymphatic system. Although E coi is harmless to humans in the lower intestine, penetration into the lymphatic system would be pathogenic. The mechanism of such a change is presently unknown, but a minor change in E coi tolerance to acidity might be sufficient. 74. Ayala, The Stability of Biological Species 6-8 (unpublished paper presented at the National Academy of Sciences Forum on Recombinant DNA Research, Mar. 7-9, 1977) (copy on file at the Ecology Law Quartery office). 75. Actually, this possibility is virtually nonexistent for the E co/i K12 strain used in recombinant DNA experiments. This strain cannot even colonize the intestine. See note 85 infra. This risk might be greater in experiments with other hosts, however. 76. Not only can any gene from the foreign genome become part of a chimera, but any gene from the various tumor viruses present in the foreign DNA cells could be ineorporated into the chimera as well. Thus, there is no way of knowing which genes will appear in the chimeras resulting from recombination. 19791 RECOMBINANT DNA

major problem. Certainly, shotgun experiments that may be contaminated by tumor viruses entail no more risk than direct handling of those viruses. 77 If the virus is harmless, then so is that shotgun experiment. Analytically, the risks of either the shotgun or the pure technique should be divided into narrower categories depending upon the source of the foreign DNA. For example, shotgun experiments using DNA from primates are likely to be more dangerous to humans than those using DNA from birds since the genetic similarity among primates makes it more likely that added genes will be expressed if they find their way into humans.78 Viral DNA contaminating the foreign DNA and pathogenic to other primates may be pathogenic to humans as well.79 Likewise, genes isolated from adult organisms, as opposed to those obtained from embryonic tissues, are likely to entail more risk 80 because the adult genes may be contaminated by pathogenic viruses. Additionally, experiments using a prokaryotic host such as E. co/i can be characterized by whether the added foreign gene is derived from an organism that normally exchanges genes with the host.8' Presumably, if such genetic exchange occurs naturally and no pathogenic organisms have resulted, artificially placing the same genes in the same host should pose little or no risk; transferring genes from organisms that do 8 not naturally exchange with the host would carry some risk. 2 The choice of the host cell used in recombinant DNA experiments affects the degree of the risk of creating a cell dangerous to humans and animals. E. co/i K 12, the bacteria scientists have been working with for decades, is the logical choice since it is the organism best understood by scientists. Because of scientists' detailed knowledge of its structure and functions, the cell is easy to manipulate for recombinant DNA experiments, the risks of its use are easier to assess, and the methods used to contain its risks are easier to develop. In fact, the E coli K12 strain can be altered by classical genetic techniques to make escape, survival, and

77. NIH Revised Guidelines, supra note 31, at 60,099. Moreover, the inability of E co/i to express foreign DNA from eukaryotes without help makes shotgunning from those higher organisms a far lower risk than originally supposed. Id at 60,087. Researchers are making a more accurate assessment of this risk by shotgunning eukaryotic genes into E. Coli and testing the virulence of these cells in mice. Id. at 60,088. 78. 1976 HEW DOCUMENTS, supra note 1, at 189. 79. Id 80. Id 81. Id 82. National Institutes of Health, U.S. Dep't of Health, Education, & Welfare, Recom- binant DNA Research, Proposed Revised Guidelines, 43 Fed. Reg. 33,042, 33,043, 33,050 (1977) [hereinafter cited as NIH Proposed Revised Guidelines]. See also Ptashne, The De- fense Doesn't Rest, THE SCIENCES, Sept./Oct. 1976, at 6, reprinted in REP'T ON GENETIC ENG'R, supra note 22, at 245-46. ECOLOGY LAW Q UARTERL Y [Vol. 8:55 reproduction of a dangerous strain more difficult.8 3 Essential to the manipulation of E. coli K12 in recombinant DNA experiments are the variety of plasmids that the bacteria carries and its ability to transfer genes to other cells.8 4 These characteristics, coupled with the location of the ecological niche of E. coi-the human intestine--create a potential risk of producing a dangerous organism. If E coli K12 were made pathogenic or if it transferred harmful genes into another bacteria or cell, the results could be rapidly manifested. How- ever, studies of E coli conducted since the recombinant DNA controversy developed indicate that E co/i K12 cannot be inadvertently converted to a pathogen by the insertion of foreign DNA; in fact, such studies have shown that scientists may not even be capable of deliberately achieving the conversion.8 5 Likewise, the transfer of DNA from E. coli K12 appears to be extremely unlikely. 6 Consequently, there is little to fear from that bacterial organism. Still, if one is to err on the side of safety, the dangers of using a microbe endemic to humans seem to call for a different host in recombinant DNA research. The problem is that there is presently no other host suitable for these experiments. 87 A replacement bacterial host for E coli must have all the advantages of E. co/i and an ecological niche potentially less dangerous. Although several replacements for E. co/i

83. The E coi K12 strain has been manipulated to obtain mutants that are dependent upon an external source for an essential building block in their growth, are sensitive to temperature, have reduced abilities to incorporate foreign DNA into their chromosomes, and have a reduced ability to exchange genetic material with other organisms. REP'T ON GENETIC ENG'R, supra note 22, at 28. These mutations "make colonization either unlikely or impossible and. . . collectively preclude survival during passage through the intestinal tract." Letter from Dr. R. Curtiss to Dr. D. Fredrickson, Director of the National Institutes of Health, at 4 (Apr. 12, 1977) (copy on file at the Ecology Law Quarterlyoffice) [hereinafter cited Curtiss Letter]. 84. It transfers genes by the mechanisms of transformation, conjugation, and transduction. See note 36 supra. 85. Letter Concerning Falmouth Workshop on Risk Assessment from S. Gorbach to NIH Director Fredrickson (July 14, 1977), [hereinafter cited as Gorbach Letter], reprintedin NATIONAL INSTITUTES OF HEALTH, U.S. DEP'T OF HEALTH, EDUCATION, AND WELFARE, PuB. No. 1490, ENVIRONMENTAL IMPACT STATEMENT ON NIH GUIDELINES FOR RE'SEARCH INVOLVING RECOMBINANT DNA MOLECULES (pt. II) app. M, at 4 (1977) [hereinafter cited as II FINAL EIS]. The results of the Falmouth Workshop have since been published. Pro- ceedings from a Workshop on Risk Assessment of Recombinant DNA Experimentation with Esherichia coli K12, 137 J. INFECTIOUS DISEASES 613 (1978). Dr. Roy Curtiss, one of the leading authorities on biological containment for recombinant DNA research, stated that there was no evidence that foreign genes could be incorporated into E Coli K12, and thus permit the bacteria to colonize the intestinal tract and generate products harmful to the mammalian host. Curtiss Letter, supra note 83, at 9-10. This opinion now seems to be the consensus of the scientific community. NIH Revised Guidelines, supra note 31, at 60,086. 86. Gorbach Letter, supra note 85, at 5. 87. National Institutes of Health, Dep't of Health, Education, & Welfare, Recombinant DNA Research, Guidelines, 41 Fed. Reg. 27,902, 27,907 (1976) [hereinafter cited as NIH Guidelines]; NIH Proposed Revised Guidelines, supra note 82, at 33,056. 19791 RECOMBINANT DN,4 have been investigated,88 none has been characterized sufficiently to supplant it. Use of a poorly characterized host simply because it lives in a safer environment would result not only in a great increase in risk, but in a decrease in expected benefits as well.89 Thus, if recombinant DNA research is to continue, those experiments requiring a bacterial host generally must involve E. coli.90 b. The possibility of the escape and survival of a dangerous organism The escape of a dangerous organism, be it an E. co/i cell or a virus, generally occurs through the air by means of an aerosol containing the microbe. 9' Aerosols are created by such laboratory procedures as pipetting, centrifuging, blending, and innoculating. Means can be instituted to reduce the possibility of escape.92 Some of these measures, such as improved microbiological technique, are relatively inexpensive; others, such as negative pressure laboratories, are quite costly. 93 Acci- dents always happen, however, and even with the best emergency precautions some microorganisms will escape. 94 The danger of organism escape is the most difficult and expensive to reduce of all the risks of recombinant DNA research. Assuming a dangerous organism is created and escapes into the air, it must still survive until it can infect another organism. Eukaryotic 95 cell lines are unable to survive in the air, but their virus vectors could. Therefore, the risk is created if either the bacterial host or a virus vector from a bacterial or eukaryotic tissue culture survives. Survival, in turn,

88. One possibility is Bacillus Subtillus (B. subtillus). The advantage of B. subtillus is that its normal environment is the soil and not any part of the human body. It has too many disadvantages to be of great use at this time, however. NIH Guidelines, supra note 87, at 27,923. See also NIH Revised Guidelines, supra note 31, at 60,085. But see NIH Proposed Revised Guidelines, supra note 82, at 33,046. 89. See NIH Guidelines, supra note 87, at 27,907. 90. Not all the hosts for recombinant DNA experiments need be bacteria. See HEW 1976 DOCUMENTS, supra note 1,at 195. However, because of their complexity, it is difficult to work with and to design experiments for eukaryotic hosts. An additional problem lies in the animal viruses used as vectors for eukaryotic hosts. The viruses could have latent effects, infecting humans years before any manifestation of infection becomes noticeable. 91. 1976 HEW DOCUMENTS, supra note 1, at 219-20. 92. Id at 220. 93. The cost of converting a normal laboratory to one of moderate physical containment, with negative air pressure to contain escaped aerosols, ranges up to $50,000. Id at 231. 94. A careful investigator could allow estimated 100 to 10,000 microorganisms to escape in one day. Curtiss, Genetic Manipulation of Miroorganisms: PotentialBenefts and Biohazards, 30 ANN. REV. MICROBIOLOGY 507, 520 (1976). Indeed, there comes a point where too many safety precautions in a laboratory are counterproductive. Too much "red tape" becomes an irritant, and safety procedures become something to avoid rather than follow. This attitude is especially prevalent when in all probability the consequences of careless techniques will not be harmful. For a poignant illustration of this attitude, see text accompanying notes 270-74 infra. 95. NIH Proposed Revised Guidelines, supra note 82, at 33,063. ECOLOGY LAW QUARTERLY [Vol. 8:55 depends upon whether the organism is from a recently isolated wild strain, from a wild strain that has grown accustomed to the "free ride" of a laboratory environment, or from a biologically modified strain that is unable to live unless supplied with certain nutrients. 96 The third has and the first the greatest chance of survival in the open air. the least 97 None can survive long, however. Of further importance is the organism's ability to survive if poured down the drain, attached to clothing or skin, or located in the mouth or gastrointestinal tract. Indeed, since most of the risks of gene-splicing 98 experiments require the organism to be ingested for harm to result, it is essential that the organism be able to survive passage through the gastrointestinal tract to the intestine.99 Controlling survival in the intestine is, therefore, a most promising means of controlling the risks of the experiments. 100 c. The possibility of the spreadof infection Even if a dangerous organism is formed, escapes from its laboratory environment, and survives in the open air long enough to affect an individual, the danger will be limited unless the infection can spread to others. Recombinant DNA experiments could yield organisms more infectious and difficult to stop than any ever known. Fortunately, the very nature of the E. coli habitat reduces greatly the likelihood of epidemic.10 1 Modem sanitation prevents the ready transmission of E. coli from person to person. Epidemics caused by pathogenic bacteria that live in the intestine generally occur only when such sanitation breaks down or when a carrier with poor hygiene handles food--even then epidemics have always been small. Furthermore, unlike organisms that are transmitted through the respiratory tract, those that travel through the digestive system, like E. coli, pose a much smaller threat. 102 Conse-

96. Davis, Natural Selection, Virulence, and Communicability 10-11 (unpublished paper presented at the National Academy of Sciences Forum on Recombinant DNA Research, March 7-9, 1977) (copy on file at the Ecology Law Quarterly office). 97. Id 98. Ingestion is required because most of the experiments use E. coli as the host, which to be pathogenic must be spread by ingestion. Id at 11. 99. Id; Curtiss Letter, supra note 83, at 5. 100. E coi K 12 is unable to survive in the human digestive tract more than a few days, unlike its wild counterpart, E. Coli, which colonizes indefinitely. E. Co/i K 12 can be made even less likely to survive by biological manipulation. It might be made more susceptible to the acidity of the stomach, through which it must pass on its journey to the intestine. It might also be altered to give it a generation time longer than that of normal E. coli so that if it does reach the intestine, it will quickly lose out in the battle for food and die as a strain. There are, however, some factors that prolong the life of E coli K12 in the intestine. See REP'T ON GENETIC ENG'R, supra note 22, at 243; Curtiss, supra note 94, at 522. 101. Davis Lecture, supra note 54, at 256. 102. Not only is the possibility of epidemic low because of modem sanitation, but also because a laboratory worker would be the first stricken. Id. at 256. In 25 years of work at 19791 RECOMBINANT DNA quently, the danger of greatest magnitude, epidemic, has only minimal 0 3 chance of occurrence. 1 d Other risks of recombinant DNA research A different kind of risk that has worried even the most staunch advocates of recombinant DNA research is that of crossing evolutionary barriers. This hazard concerns not the individual probabilities of producing a harmful organism and an epidemic, but the overall consequences of recombinant DNA research. Nature has developed barriers against the interchange of genes between species. Species are generally isolated from each other, sometimes geographically, almost always sexually.l°4 As a result, the gene pools of any one species are unique. Gene-splicing experiments will enable researchers to cross these natural barriers and mix the genes of different species. Some scientists fear that disaster could result from this constant crossing of natural boundaries, 105 as the consequences are as yet unknown. While science has often overcome unknown perils, those associated with recombinant DNA research are potentially global in dimension. In addition, when the experiments create artificial genetic pools unknown in nature, the natural environment cannot be relied upon to absorb all the possible progeny of such research without adverse effect. These two factors, extensive unknown consequences and lack of natural buffers, increase the risks associated with crossing evolutionary barriers. i06 The likelihood of adverse effects in crossing barriers between species can only be qualitatively assessed. If a foreign gene is to have evolutionary effects on a eukaryote host, it must be inserted in the germline-the sperm and eggs. Insertion in other cells may affect individu-

United States biological laboratories using the most pathogenic organisms available, there never was a known case of secondary spread from an infected worker to someone outside the laboratory. Id But see note 252, infra. 103. Additionally, an epidemic of head colds or a new influenza is far more likely to occur than mass death. Very few pathogens are capable of overcoming the amazing immu- nological system of the human body to cause death. The risks of recombinant DNA experiments, then, generally involve minor annoyance rather than extreme hazard. Only because it is more convenient to speak in terms of the worst case does this Article discuss the potential for overwhelming societal costs. 104. There are exceptions to the sexual isolation of a species. For example, consider the breeding between horses and donkeys to produce mules. Even in that case, however, the mule offspring is sterile and no further mixing of gene pools of the two species is possible. 105. Sinsheimer, On Our Own, 26 BIOSCIENCE 599, reprinted in REP'T ON GENETIC ENG'R, supra note 22, at 249. 106. One consequence of crossing the evolutionary barriers could be that the transfer of pathogenic viruses between E coi and the human intestinal villus cells would be made easier. See note 72 and accompanying text supra. By receiving eukaryotic genes, a prokaryotic cell may get the signals that would enable it to transfer a virus to the intestinal cells. But see text accompanying note 109 infra. ECOLOGY LAW QUARTERLY [Vol. 8:55 als, but not the species. It is doubtful that gene insertion into the germline could affect the evolutionary process, because only those changes in a species that promote its survival will persist in the long run.'0 7 Furthermore, there is only a narrow range of insertions that could affect evolution. Too small a change would be of essentially no evolutionary consequence while too large a change would eliminate the organism's chance of survival. This "window of change" allows only a few genes to be added, with a likely effect of slightly speeding up the process of evolution by increasing the pool of genes within a species of 08 eukaryote. 1 Insertion of DNA into prokaryotes is another matter, however. Because they are simple and lack specialized tissues, complex regulatory mechanisms, and definition between species, the evolutionary consequences of recombinant DNA experiments on prokaryotes are potentially greater than in the case of eukaryotes. The addition of a few genes to a prokaryote carries the risk of creating species of lower organisms harmful to humans in ways that are not fully predictable. Natural selection could cause the new organisms in that species to gain preeminence over unchanged members of the species. On the other hand, there is some evidence that prokaryotes are less restricted by species barriers. They are known to exchange genes not only between species, but across genus and family lines as well. Fur- thermore, some scientists believe that bacteria in the human intestine accept eukaryotic genes to a small extent. 109 If this is true, the so-called evolutionary barriers of nature are not sharply defined and are at least occasionally crossed naturally. Since such events occur in the absence of human intervention, the risk of artificially crossing evolutionary barriers among prokaryotes must be considered small; in the passage of eons of time most of the possible permutations of genes should have already occurred. "10 The final category of risk-the misuse or overuse of recombinant DNA bacteria" '-cannot be numerically assessed."12 Generally, the

107. Sometimes the imperfect competition of species allows benign changes that do not promote the survival of the species. This qualification does not affect the genetic barrier argument, however, because the vast percentage of changes will be harmful to an organism's survival. 108. F. Ayala, supra note 74, at 4. 109. Ptashne, supra note 82, at 6, reprintedin REP'T ON GENETIC ENG'R, supra note 22, at 245, 246. The cells accept the genes through the process of transformation. See note 36 supra. i 10. While many permutations may not have been harmful when they occurred in nature, the Earth has itself evolved and the environment of such changes would be different today. Thus, a change that was harmless and of no evolutionary consequence a millenia ago may be harmful today. Crossing an evolutionary barrier must be considered, then, as at least a slim risk. 111. Overuse could occur if, for example, too many nitrogen-fixing plants are formed. 1979] RECOMBINANT DNA problem is one of creation and use with unanticipated consequences. Any decision to create a new organism and release it into the natural environment requires analysis of all possible effects of the organism in its ecosphere. The premature release of organisms, then, and not the experiments themselves, causes this risk. Consequently, this risk is associated with the methods of regulating the research, but not with the decision of whether to continue experimentation. One potential misuse of recombinant DNA experimentation itself lies in the area of human genetic engineering."i 3 Human genetic engineering is replacing defective disease-causing genes in humans with normal ones." 4 Clearly there is both great potential benefit and great risk in human genetic engineering. How one defines disease and to whom the power is given to decide to replace human genes is crucial. The problems involved are obvious and eventual regulation of such experimentation is likely. The field of genetic engineering is related to recombinant DNA research only because gene-splicing experiments are a preliminary step to experimentation in human genetic engineering. While recombinant DNA experiments are probably necessary for development of the ability to replace human genes, they are by no means sufficient. The ability to practice genetic engineering lies far in the future even if recombinant DNA experiments continue. A decision to prohibit human genetic engineering need not affect recombinant DNA experiments. Nevertheless, some of those ada- mantly opposed to genetic engineering prefer to end all chance of its occurrence by banning the path of experiments that could lead to the development of bioengineering technology. Their reasoning stems from a fear that as gene-splicing experiments gain momentum and their practical benefits are realized, it will be more and more difficult to The waste nitrous oxides of these plants could contribute to destruction of the earth's ozone layer. R. Valentine, supra note 58, at 4. Misuse could occur if researchers produced an E co/i that could break down cellulose to sugars and fatty acids. If used to help humans digest cellulose, this organism might also prove harmful by creating dietary problems. REP'T ON GENETIC ENG'R, stpra note 22, at 38. 112. There is also the possibility that someone could use some pathogenic recombinant DNA as a biological weapon. The unfortunate fact remains that there are already organisms available dangerous enough to be effectively used for this purpose. Hence, no greater risk is presented than that we already face. Moreover, given that no hazardous recombinant DNA organisms have been developed, "any discussion of potential misuse by saboteurs would be entirely speculative." I FINAL EIS, supra note 54, at 119. 113. Examples of diseases caused by abnormal genes are sickle cell anemia and hemo- philia. 114. See, e.g., the report in Golden, supra note 34, at 33, that at the National Academy of Sciences Forum on recombinant DNA held March 7-9, 1977, a group opposed to human genetic engineering "took over the state and unfurled a banner reading: WE WILL CRE- ATE THE PERFECT RACE-ADOLF HITLER." ECOL OGY LAW Q UARTERL Y [Vol. 8:55

stop a movement towards eugenics.' 15 The fallacy in this reasoning is that experiments with recombinant DNA lead to many branches of experimentation, only one of which is genetic engineering. The inertia gained from recombinant DNA experiments can be diverted into other branches if the decision is made not to pursue human engineering experiments; genetic engineering may be blocked. Since there is not yet enough information on human genetic engineering to make an intelligent decision on regulation, it is foolish to ban recombinant DNA experiments simply because they may lead to genetic engineering.

4. Analysis of the Risks and Benefits of Recombinant DNA Research The previous discussion of the benefits and risks of the research demonstrates the large magnitude of both. Fortunately, while both the benefits and harms are enormous, the probabilities of their occurrence are highly dissimilar. On an absolute scale, some of the benefits are very likely to occur-e.g., advances in scientific knowledge and in the manufacture of hormones-while the risks are remote. Certainly some of the risks carry higher probabilities than others. For example, the possibility of deliberate insertion of a dangerous toxin from a prokaryote into E. coli K12 is a significant concern. Fortunately, none is likely. Morover, the harms that are largest in magnitude are the least proba- ble. Assuming a new disease were created and that it infected the researcher or laboratory worker, the chance that it would be untreatable is slim, and the chance that it would spread to others is minimal. " 6 In addition, when biological and physical containment measures are added, the chance that harm will result from any of the risks becomes increasingly small."17 The low probability of harm, both with and without containment, is a point on which both proponents and critics of the research agree." 18 Since the measure of benefits from recombinant DNA is significant and the risks are extremely improbable, despite their substantial magnitude, a balance of the two leads to the conclusion that the benefits of recombinant DNA research far outweigh the risks. This conclusion is necessarily vague because of the uncertainty involved in quantifying these benefits and risks. Of course, a naked balance of the risks and benefits is not the end

115. Davis Lecture, supra note 54, as reprintedin REP'T ON GENETIC ENG'R, supra note 22, at 256. 116. This decrease in risk raises what is known as a "zero-infinity dilemma"; a virtually zero probability that what may be a virtually infinite catastrophe will occur. Page, supra note 13, at 211. 117. Davis Lecture, supra note 54, as reprintedin REP'T ON GENETIC ENG'R, supra note 22, at 255; REP'T ON GENETIC ENO'R, supra note 22, at 241. 118. For a discussion of the role of societal values in risk-benefit analysis, see text accompanying notes 13-20 supra. 1979] RECOMBINANT DNA

of the analysis. The various societal values must be inserted into the calculus to determine whether recombinant DNA research is acceptable as a matter of social policy' 19. An informed consideration of societal values could result in rejection of an otherwise adequate policy; such a result is not illegitimate. Societal risk aversion cuts against continued recombinant DNA research. For example, society would be particularly cautious as to the risk of crossing evolutionary barriers, because of the unknown consequences. The potentially great magnitude of harm--even though the probabilities are small-is particularly important, for society tends to avoid significant hazards whenever possible.' 2 0 In addition, recombinant DNA accidents would rarely occur and might escape detection for some time as the cause of an environmental disturbance.121 Public fear of such rare and untraceable accidents could galvanize sentiment against research. This attitude may be the primary reason that there has been so much criticism of DNA experimentation. 22 Of course, the value of risk aversion does not necessarily override the conclusion of the risk-benefit analysis. Rather, it indicates that the benefit to risk ratio must be significantly greater than even to be acceptable. The societal preference for voluntary over involuntary assumption of risks must also be considered. The most significant involuntary exposure of society as a whole is the chance of epidemic. However, epidemics are a virtually impossible result of recombinant experiments using E coli K12.123 To the extent that some risk, however small, is linked to involuntary exposure, the societal preference weighs against further research. One segment of society, laboratory assistants and support staff'of the recombinant DNA researchers, may be exposed more seriously. Although recombinant DNA researchers are fully aware of the risks they assume when experimenting, assistants and staff members may not be. Notice of the dangers does not overcome the involuntary nature of

119. Page, supra note 13, at 240. 120. While one may know that one has a cold, it would not necessarily be clear that one knows the source of that cold. Pathogenic bacteria may elicit the same or similar symptoms as "natural" pathogens. 121. This concern is mitigated somewhat by the trend in risk assessment mentioned previously. See notes 85-86 supra. While one may argue that the absence of a recombinant DNA accident in the last five years is not proof that one will not happen, this fact coupled with lower risk assessments should reduce some of the public pressure directed against recombinant DNA research. NIH Revised Guidelines, supra note 3 1, at 60,087. But consider the public reaction after the Three-mile Island nuclear accident and subsequent disclosures of other nuclear accidents. 122. See text accompanying notes 101-03 supra. 123. Indeed, they may not even perceive the risks. Experiments with any microorganism tend to lull people into a false sense of security, because microorganisms are invisible and almost all of those commonly used in laboratory work are harmless. See note 273 infra. ECOLOGY LAW QUARTERLY [Vol. 8:55 their exposure. If these workers do not understand the dangers, or understand them but have no opportunity to avoid them because no other jobs are available, they cannot be said to have voluntarily assumed the risks. 124 This problem is more readily solvable by educating the personnel and mitigating their exposure than by discontinuing the research; the latter seems draconian when the former is workable. With the exception of laboratory assistants, 25 the preference for equitable sharing of benefits and risks among segments of society is generally satisfied by recombinant DNA research. The researchers and their corporations or universities receive the benefits of profits, increased prestige, and perhaps grants for future research; they also bear most of the direct physical or health risks and the costs of accidents, with the attendant compensation to injured workers and loss of prestige. The general public receives the benefit of increased theoretical knowledge and the practical benefits that result,126 while it bears only the minimal risk of an epidemic; this burden is not disproportionate to the benefits received. Society's preference for temporal equity poses no problem for recombinant DNA experiments. The research does not involve present benefits for future costs, the classic temporal inequity problem. In fact, the reverse is true. Most of the benefits of the research, especially the practical applications, will arrive in the future. At the same time, the risks of the research are borne almost wholly in the present. A dangerous recombinant organism generally would manifest harmful effects immediately after it infects a victim. 27 Recombinant DNA research creates no irreversible problems, for no laboratory-created organism could survive in any ecological niche in sufficient numbers to cause or 28 increase environmental harm. In sum, societal risk aversion and the preference for voluntary risk-taking are factors operating against continued development of recombinant DNA technology. Together, these two values lower the benefit-risk ratio in a way that can be assessed only qualitatively. Given the recent trend of discovery that the risks have been overesti-

124. Laboratory support personnel bear a larger share of the risk than the general public, yet receive only benefits as members of the general public. Again, measures should be taken to minimize this burden. 125. See text accompanying notes 49-61 supra. 126. There are possible exceptions to this generalization. For example, a cancer-induc- ing virus inserted into E co/i that infects a person may not show its effects for a period of years. 127. Davis Lecture, supra note 54, as reprintedin REP'T ON GENETIC ENG'R, supra note 22, at 258. The chance that a released organism could become permanent in nature, however small, and the possibility that it will then become difficult to eradicate, do reduce the benefit-risk ratio, but it is not the same sort of irreversibility as the escape of radioactive waste with a 250,000 year half-life. 128. See text accompanying notes 45-48 supra. 19791 RECOMBINANT DNA

mated, these values should not control. As the hazards are reevaluated downward, the level of public aversion to them should fall. The above analysis, however, involves pervasive uncertainty. One of the procedures for dealing with uncertainty must be employed in deciding whether to continue recombinant DNA research. Though the choice of which to use is value laden, reasoned use of any of the proce- 129 dures will lead to a principled result. The procedure of continuing the research until additional information proves gene-splicing to be too risky is inadequate. It fails totally to accomodate risk aversion. It also fails to recognize that an absence of present damage is not proof of an absence of environmental risk. The converse procedure, shifting the burden of proof to advocates of the research, is also inappropriate. The trend in risk assessment and the unmarred safety record of the research so indicates. Moreover, unlike many environmental problems, the cost of stopping DNA research because it cannot be proven safe may well be greater than the cost of finding the research safe when it is not. The benefits are enormously important. Losing them would impose a large cost. Since proving lack of risk is very difficult, the research would probably be halted if advo- 30 cates bore the burden of persuasion.' Critics of gene-splicing research have called for delay, the course 13 chosen by the scientists themselves when the controversy first arose. ' Such a policy is useful when a technology is generating tremendous volumes of information and is experiencing rapid advance. A short- term delay earlier in the history of recombinant DNA research yielded large increases in certainty for the field.' 32 For several reasons, however, further restraint would be wholly improper. Firstly, large amounts of information have been accumulated recently on the risks and benefits of recombinant DNA research.' 33 Intense efforts over a

129. REP'T ON GENETIC ENG'R, supra note 22, at 36. Cf.NIH Proposed Revised Guide- lines, supra note 82, at 33,044. In fact, NIH Director Fredrickson indicates that the burden should not be on the advocates but on the critics since "no evidence has come to light of a product created by these techniques that has been harmful to man or the environment. NIH PROPOSED RE- VISED GUIDELINES, supra note 82, at 33,044. Lack of evidence of harm from the research is at least support for controls that are not the most restrictive. 130. Wade, Gene Splicing.: At Grass Roots Level a HundredFlowers Bloom, 195 SCIENCE 558, 559-60 (1977) [hereinafter cited as Wade, Hundred Flowers]. 131. NIH Guidelines, supra note 87, at 27,903. 132. For a description of the "information explosion" and the efficacy of solutions to the problems it has raised, see Glick, Regulation of Molecular Genetic Research, 282 ANNALS N.Y. ACAD. SCI. 182-3 (1976). 133. Certain experiments have been devised to reduce uncertainty in specific areas. For example, some experiments using both classic and recombinant techniques have shown both that E. col K12 cannot be made pathogenic and that it is extremely unlikely that the kind of plasmids used in recombinant DNA experiments will be transferred to a wild-type strain ECOLOGY LAW QUARTERLY [Vol. 8:55 long period of time, however, probably would be required before much more information can be obtained without conducting the experiments themselves. Secondly, the simple risk assessment experiments have already been done or are presently being conducted; an increase in gen- 134 eral knowledge will be required to enhance greatly this assessment. Thirdly, available information points to the conclusion that the risks have been overestimated. Lastly, a new delay-as opposed to an outright ban-may be unenforceable. Scientists, anxious to resolve the controversy, might not abide by any further delay in reaching a final decision. The uncertainty can be accomodated, however, by providing a margin of safety in meeting the standard of acceptability. This margin would require the benefit-risk ratio to be substantially greater than 1:1 to be acceptable. The relatively certain benefits of the research cannot be ignored; nor can the data indicating lower risk than initially thought. Moreover, new information on the risks and benefits will arrive at a slower rate since the important assessment experiments have been conducted. A margin of safety in the benefit-risk ratio is the best way of dealing with these points while still accounting for the uncertainty involved. Assuming some margin of safety is required, the question is whether recombinant DNA research should continue without regulation. The question is extremely difficult to answer without invoking while in the intestine. Thus, the dangers in using E. coli have been shown to be much smaller than originally thought. II FINAL EIS, supra note 85, app. M; NIH Revised Guide- lines, supra note 31, at 60,080. The uncertainty of risks can also be reduced by experiments that: (1) indicate whether eukaryotic DNA is expressed in microorganism hosts and the factors necessary for such expression, and (2) quantitatively assess the probability of escape of the host or vector under various types of physical and biological containment. Uncertainty associated with the benefits of recombinant DNA experiments is more difficult to reduce. Only as knowledge is gleaned from recombinant DNA experiments themselves will it become clear that particular benefits will or will not result. For example, human insulin can how be produced in bacteria through the use of an insulin gene. See note 93 supra. When and if it is shown that natural eukaryotic DNA with appropriate laboratory manipulation can be expressed in E. co/i K12, then other similar benefits, such as the production of blood-clotting factors will be certain. Once basic recombinant DNA experiments yield information on function and regulation of mammalian genes, sophisticated advances in medicine will become predictable. Without information from recombinant DNA experiments themselves, however, there is no way of reducing uncertainty associated with the benefits. 134. It is true that some experimentation has already resulted in pertinent information on E. coli K12. See notes 85-86 supra. Much of this experimentation used recombinant DNA techniques. It is impossible to determine how long it would have taken to reach these results using purely classical techniques. There are some risk-reducing experiments that can only be done with recombinant techniques-at least presently. An example would be an experiment to discover whether eukaryotic DNA is ever expressed when transferred to a microorganism host and what factors are necessary for this expression. 19791 RECOMBINANT DNA

personal value judgments. The risk-benefit assessment is just too vague for a definitive answer. Fortunately, such an answer is not required. Indirectly or directly regulated recombinant DNA research will have a benefit-risk ratio easily able to meet the margin of safety requirement, and one far superior to unregulated recombinant DNA research as well. Biological and physical controls for the research imposed by indirect or direct regulation can greatly reduce the risks. Some of these controls, e.g., genetically weakened E. co/i unable to survive outside the laboratory, have already been mentioned, while others, particularly physical containment, will be discussed later.134a These controls, if implemented cor- rectly, can greatly reduce the risks without unnecessarily delaying experiments and having only a minimal effect on benefits. The result can be a near normal flow of benefits and a great decrease in risks, thereby increasing the benefit-risk ratio considerably. Certainly, regulated recombinant DNA research is acceptable, and its greater benefit- risk ratio makes it more cost-effective than unregulated recombinant DNA research, whether or not the latter's benefit-risk ratio meets the margin of safety requirement. In sum, recombinant DNA research should continue, but should also be regulated. The discussion now turns to the best way of regulating. If regulation can be accomplished without direct government control, then present policies in this regard should be abandoned. If, however, governmental intervention into scientific research is the only means of guaranteeing that the risks will be reduced effectively, then alternative structures for governmental regulation must be investigated. The Article next discusses possible indirect methods of regulation.

II INDIRECT REGULATION OF RECOMBINANT DNA RESEARCH Scientists themselves could adopt self-regulatory procedures to reduce the risks while maintaining the benefits of recombinant DNA research. Alternatively, the activities of researchers could be controlled indirectly by the patent law or tort liability systems, which provide monetary incentives to observe safety precautions. This section focuses on these indirect controls on recombinant DNA research, concluding that none of them suffice to ensure an adequate degree of risk reduction.

A4. Self-Regulation This section considers the problems with self-regulation for both

134a. See note 237 infra. ECOLOGY LAW QUARTERLY [Vol. 8:55

academic and industry-financed research. 135 The term "self-regulation" in this Article signifies a system in which a committee of scientists engaged in recombinant DNA experimentation formulates guidelines for the safe conduct of the research. Under such a regulatory scheme experimenters would be encouraged to observe those guidelines, but 136 compliance would be voluntary. 1. Self-Regulation by Academic Researchers Academic researchers may be susceptible to a self-regulatory scheme since they normally value prestige and the esteem of their col- leagues highly. If researchers were to agree collectively to a set of safety guidelines, peer pressure could help enforce compliance. 13 7 However, scientists are subject as well to the countervailing influence of professional commitment to scientific research. Scientists depend for their livelihood upon the expansion of experimentation. Gene-splicing research in particular can quickly lead to important discoveries that significantly increase the prestige of individual researchers and the scientific community in general. Individual recognition is directly related to higher salaries, tenure, and larger grants. Thus, scientists would have an inherent conflict of interest in regulating their own livelihood.' 38 This self-interest may cause researchers to take unreasonable

135. Seventy-five percent of all recombinant DNA research is conducted by private industry or by groups not funded by the federal government. 1977 Senate Hearings, supra note 22, at 55 (statement of Sen. Metzenbaum). The corporations financing recombinant DNA research include Genetech, Inc., Cetus Corporation, Upjohn Company, General Elec- tric, Eli/Lilly & Co., Miles Laboratories, Merck, Sharp & Dohme Research Laboratories, W.R. Grace & Co., and Abbot Laboratories. REP'T ON GENETIC ENG'R, supra note 22, at 51; MOTHER JONES, Feb.-Mar. 1977, at 23; Bus. WEEK, Dec. 12, 1977, at 128. 136. A completely voluntary plan is not the only possible self-regulatory model. For example, the medical and legal professions do not rely solely on voluntary compliance with collectively developed guidelines. Disciplinary boards typically have broad powers to sus- pend or expel a rule-breaker from the practice of the profession. See, e.g., CAL. Bus. & PROF. CODE § 6078 (West 1974) (attorneys). Such a plan would probably not receive substantial support from either the public or the scientific community. The regulatory systems of the legal and medical professions are frequently criticized. See generally S. AUERBACH, UNEQUAL JUSTICE (1977). 137. One proposed plan for self-regulation through peer pressure would make use of scientific journals. David Baltimore, a recombinant DNA researcher, believes that if the journal editors insisted that researchers submitting gene-splicing articles include information concerning the safety precautions taken, this would provide the "best safeguard against experiments that do not follow the guidelines because of the strong peer pressure on anyone reporting an experiment that was done without the appropriate containment." 1976 HEW DOCUMENTS, supra note 1, at 251. NIH Director Fredrickson has recommended that "all publications dealing with recombinant DNA work include a description of the physical and biological containment procedures practiced." NIH Guidelines, supra note 87, at 27,912. Even if the journals voluntarily agreed to this requirement, however, industrial researchers, who do not rely as heavily upon publication for their livelihoods, would be little affected. 138. See generally J. GOODFIELD, supra note 11, at 178-79. Individual institutions also have an interest in promoting recombinant DNA research. Universities compete with one another for grants and scientists on the basis of their facilities and the success of their re- 19791 RECOMBINANT DNA risks, since individual biases may subconsciously influence the decisionmaking process. Well-meaning scientists may underestimate risks, overestimate benefits, and tolerate hazards deemed unacceptable by so- 39 ciety. 1 Another difficulty is that the researchers themselves are burdened by whatever safety measures are adopted. Specifically, many of the recommended safety procedures for recombinant DNA research are cumbersome, making experiments more difficult, time-consuming, and expensive. 140 This factor may prevent the adoption of rigorous guidelines and discourage thorough compliance and enforcement. Furthermore, experience suggests that those most likely to play a prominent role in a self-regulation program are the experts in the regulated area of research. While an expert may have an intimate understanding of the processes and potential benefits of an experiment, that 14 1 individual may have a poor understanding of the attendant risks. For example, only an epidemologist, and not a molecular biologist, would be able to understand fully the risks of epidemic posed by a particular gene-splicing experiment. Thus, there is no guarantee that a self-regulation plan would include all necessary scientific perspectives. The only scientific self-regulation scheme implemented to date was that which preceded the adoption of the NIH controls.'4 2 It is doubtful, however, that this system could have functioned effectively over a long period of time. Adverse public reaction to recombinant DNA research has caused a backlash in the scientific community against any sort of regulatory scheme. This attitude, coupled with the increasing numbers and diverse backgrounds of researchers, may make 43 voluntary compliance with any set of regulations unlikely.1 search. Bennett & Gurin, Science that Frightens Scientists, THE ATLANTIC MONTHLY, Feb. 1977, at 58. 139. See Green, Law and Genetic Control, supra note 45, at 173. 140. For a description of existing guidelines, see notes 230-46 and accompanying text infra.- 141. This lack of understanding has proven true for recombinant DNA research. NIH Proposed Revised Guidelines, supra note 82, at 33,043. 142. These guidelines are discussed in detail in notes 225-27 and accompanying text infra. Other examples of-scientific self-regulation are limited to applied science. For example, the NIH and the National Science Foundation rely on panels of scientists to review funding proposals and disburse funds. In the medical field, the Professional Standards Review Or- ganizations review surgical procedures to ensure quality surgery and prevent unnecessary operations. 143. Animosity has increased in the scientific community towards proposals for regulation and, indeed, towards criticism of recombinant DNA research in general. Many scientists feel the publicity received by recombinant DNA research has led to an exaggeration of the risks of gene-splicing. Critics of recombinant DNA research complain of the reaction against regulation caused, in part, by this attitude. N. WADE, THE ULTIMATE EXPEIMENT 100 (1977). Wade has also claimed that untenured faculty are pressured not to speak publi- cally against the research, the implicit threat being that tenure might be withheld. Wade, ECOLOGY LAW QUARTERLY [Vol. 8:55

2. Se/f-Regulation by Private Industry Industrial research is even less susceptible to a self-regulatory scheme than academic research. The inherent biases and limitations of academic researchers are equally present in private industry. In addition, industrial researchers are less sensitive than university scientists to peer pressure and academic prestige, because the goal of industrial experimentation is the maximization of profits rather than the pursuit of scientific knowledge. Even if individual industrial researchers want to follow voluntary guidelines, the economic goals of their employers will ultimately control the course of industrial research. 144 Many corporations have announced that they voluntarily accept existing safety measures with some minor modification.145 Subsequent actions of industry, however, have cast doubt upon the degree of its commitment to these controls. 146 Although many corporations are willing to register publicly on-going research, they have refused to register future research plans for fear of disclosing trade secrets. 147 This defeats the main purpose of registration--dissemination of information to the public regarding the nature of future research. Industry has also disagreed with a ten liter limitation on products of the research, a re-

Gene-splicing. Critics of Research Get More Brickbats than Bouquets, 195 SCIENCE 466-67 (1977). Professor Ruth Hubbard of Harvard explains that "it is not an accident that most of the present critics of the gene-splicing research are tenured." Id. at 467. Critics have also come under personal attack. Id. For example, Professor James Watson has referred to col- leagues critical of the research as "shits," "kooks," and "incompetents." TIME, Apr. 18, 1977, at 32. 144. Summary and Recommendations, in PERSPECTIVES ON BENEFIT-RISK DECISION MAKING 11 (1972) (report of a colloquium conducted by the Committee on Public Engi- neering Policy, National Academy of Engineering, Apr. 26-27, 1971). Cf. Hardin, The Trag- edy of the Commons, 162 SCIENCE 1243 (1968) (arguing that appeals to conscience will not overcome rational, free-enterprise calculations of the utility of pollution). 145. For example, both the Pharmaceutical Manufacturers Association, which represents the major drug companies working with recombinant DNA, and General Electric have stated their intentions to comply voluntarily with the NIH Guidelines. Stencel, Genetic Re- search 1, EDITORIAL RESEARCH REP. 233 (1977); Letter to the authors from Manuel Aven, manager of General Electric's Physical Chemistry Laboratory in Schenectady, N.Y. (May 18, 1977)(copy on file at the Ecology Law Quarterlyoffice). One of the prominent companies specializing in gene-splicing work, Cetus Corp., stated it also will comply with the guidelines. Wade, Hundred Flowers, supra note 130, at 559. 146. Dr. John Finklea, Director of the National Institute for Occupational Safety and Health (NIOSH), stated: We have found that private industries . . . conducting or planning recombinant DNA research, have not always voluntarily and quickly dealt with other occupational safety problems. For example, one corporation [DuPont] when contacted, assured NIOSH that medical records for recombinant DNA research workers would be made available . . . [and] at the same time took legal action disputing our authority for access to employment and medical records necessary to evaluate health risks at a plant with an apparent increase in cancer among the workers. 1977 Senate Hearings,supra note 22, at 58. 147. Wade, Hundred Flowers,supra note 130, at 559. REP'T ON GENETIC ENG'R, supra note 22, at 244. 19791 RECOMBINANT DNA

striction designed to reduce the risk that such products will escape. 148 Even if industry-promulgated regulations are adequate, there is no guarantee of continued industrial compliance. There always would be an economic incentive to disregard those restrictions that limit or reduce potential profits. This economic reality suggests that safety in private recombinant DNA research can only be assured by utilizing market incentives or a more direct regulatory plan.

B. Patent Law Patent law may provide a mechanism for imposing more uniform controls on recombinant DNA research since a great deal of such experimentation is geared toward producing patentable results. 49 Pat- ents ensure research institutions a share in the monetary benefits that result from their discoveries. For example, compliance with the NIH Guidelines could be made a condition for the issuance of patents. 50

148. Id. Industry believes that the ten liter limit imposed on recombinant DNA experiments will hinder if not prevent large volume experiments necessary for determining commercial feasibility. Moreover, a number of potential industrial uses would involve quantities of recombinant DNA products above the ten liter limit. See text accompanying note 277 infra. For the text of the ten liter requirement see NIH Guidelines, supra note 87, at 27,915. 149. Patent protection has been extended to new forms of life. The U.S. Court of Cus- toms and Patent Appeals has ruled that the Upjohn Company may patent a biologically pure strain of fungus known as Stepomyces vellosus, which the company developed and now uses in the production of an antibiotic. Application of Bergy, 563 F.2d 1031 (C.C.P.A. 1977). See 19 ENVIRONMENT 23 (Dec. 1977). The court has since ruled that General Elec- tric has the right to patent a bacteria that consumes oil more efficiently than any bacteria found in nature. Application of Chakrabarty, 571 F.2d 40 (C.C.P.A. 1978). See San Fran- cisco Chronicle, Mar. 3, 1978, at 3, col. 9. The fungus was produced by conventional genetic engineering techniques, not through the use of gene-splicing technology. The decision suggests, however, that there is no basis to distinguish between conventional and recombinant techniques. Both Stanford University and the University of California have applied for patents on recombinant techniques. San Francisco Chronicle, Mar. 3, 1978, at 14, col. 5. The NIH believes patents on the products of recombinant research should be allowed, because "there are no compelling economic, social, or moral reasons to distinguish these inventions from others involving biological substances or processes that have been patented even when par- tially or wholly developed with public funds," such as vaccines for rubella and rabies. FREDERICKSON, THE PATENTING OF RECOMBINANT DNA RESEARCH INVENTIONS DEVEL- OPED UNDER DHEW SUPPORT: AN ANALYSIS BY THE DIRECTOR, NIH, 16 (Nov. 1977) [hereinafter cited as NIH PATENT ANALYSIS], reprinted in PUBLIC HEALTH SERVICE, NA- TIONAL INSTITUTES OF HEALTH, U.S. DEP'T OF HEALTH, EDUCATION, AND WELFARE, PUB. No. 78-1139, 2 RECOMBINANT DNA RESEARCH 2 (1978) [hereinafter cited as 1978 HEW DOCUMENTS]. 150. Another means that has been suggested for controlling recombinant DNA research through patent law is the total elimination of patent protection for such research. This argument has few proponents. NIH PATENT ANALYSIS, supra note 149, at 16, reprintedin 1978 HEW DOCUMENTS, supra note 149, at 2. A related proposal calls for governmental owner- ship only of the patent, with licensing to industry. 1978 HEW DOCUMENTS, supra note 149, at 74 (statement of R. Sinsheimer), 135 (statement of D. Callahan). But see 1977 House Hearings,supra note 6, at 940 (testimony of N. Latker). The purpose of either approach is to 90 ECOLOGY LAW QUARTERLY [Vol. 8:55

Such a scheme could reduce risks since researchers would have to observe safety controls in order to gain proprietary interests in the products of their research. In January 1977, the Commerce Department's Patent and Trade- mark Office proposed a plan giving patent applications involving recombinant products and techniques accelerated processing provided that the applicant attached a statement of compliance with the NIH Guidelines.' 5' However, the plan granted exemptions where it was "considered essential to avoid disclosure of proprietary information or loss of patent rights."' 52 Such exemptions would have deprived the plan of any practical effect by enabling industry to circumvent disclosure requirements. On the other hand, if researchers are required to disclose all information concerning experiments in advance of patent protection, their proprietary interests could be undermined. 153 Largely reduce the amount of recombinant DNA research by undercutting the profit motive for its continuance. Such a plan is problematic. Firstly, it assumes that the amount of recombinant DNA research should be reduced despite the strong arguments that it should continue as long as its risks are adequately controlled. Secondly, elimination of patent protection would not necessarily reduce the risks of research. Universities do not rely on profit from patents and would be largely unaffected by the ban. It has been argued that a ban on patents for DNA research may at least limit the research to publicly funded projects. See 1977 House Hearings,supra note 6, at 940-49 (testimony of N. Latker). However, private researchers may respond not by eliminating recombinant DNA research but by employing greater secrecy in their research. This effect would only further remove private research from public control. 151. 42 Fed. Reg. 2,712 (1977). 152. Criticism by Senator Dale Bumpers in 123 CONG. REC. S2,274 (daily ed. Feb. 4, 1977). See also 7 ENVT'L L. REP. 10081 (1977). 153. One solution to this dilemma would be to make such information accessible only to the Patent Office. It may be difficult for the Patent Office to keep such information confiden- tial, however, because competitors and public interest groups may have the right to such information under the Freedom of Information Act (FOIA), 5 U.S.C. § 552, which provides for public access to information in the control of government agencies. The Patent Office, however, could refuse disclosure on the grounds of exemption for the disclosure of trade secrets. 5 U.S.C. § 552(b)(4). Protection may also be provided by 18 U.S.C. § 1905 (1976), which prevents the government from releasing trade secret information. 1977 Senate Hear- ings, supra note 22, at 200 (J. Warren). But see 1977 House Hearings,supra note 6, at 1274- 93 (A. Whale, General Patent Counsel, Eli Lilly & Co., discussing FOIA disclosure problems under present and proposed recombinant DNA research regulations. He argues § 552(b)(4) will be ineffective). However, decisions by an agency not to disclose are open to challenge. 5 U.S.C. § 552(a)(4)(B). The federal agency bears the burden of proof in showing that records fall within an exception. Committee on Masonic Homes v. NLRB, 556 F.2d 214 (3d Cir. 1977); Tax Analysists Advocates v. IRS, 505 F.2d 350 (D.C. Cir. 1974); Washington Research Pro- ject, Inc., v. HEW, 504 F.2d 238 (D.C. Cir. 1974). On the other hand, the Supreme Court recently held that the FOIA exceptions do not mandate agency non-disclosure, since the concern is agency privacy, not third party privacy. Chrysler Corp. v. Schlesinger, 565 F.2d 1172 (3d Cir. 1977), aff'dsub nom. Chrysler Corp. v. Brown, 47 U.S.L.W. 4434 (1979). In addition, there is no private cause of action under the exemptions or under 18 U.S.C. § 1905 (1976). Id. A possible avenue to protection still remains, however. If disclosure of trade secrets is authorized by interpretive regulations exempt from the notice and comment requirements of 5 U.S.C. § 553 (1976), such disclosure may not be "authorized by law" as 19791 AECOMBINANT DNA4 because of widespread criticism, the Patent Office suspended the order 54 except as applied to research involving patents on safety procedures.1 Even if a conditional patent plan could avoid disclosure problems, there is no indication that it would adequately control recombinant DNA research. Since, for the most part, university research is not motivated by profit, it would be largely unaffected by patent controls. Pat- ent law would thus have to supplement current controls over 55 government sponsored research rather than replace them. Conditional patent issuance might not have the desired effect even on privately funded research. Though such research is motivated by profit, the economic gain from patents may be ancillary to more general monetary benefits from the use of recombinant techniques. Prod- ucts might be produced more quickly and inexpensively through the use of recombinant techniques-with or without patent protection. Re- searchers would substitute secrecy for patent protection, and potentially dangerous projects would become further removed from public control. Hence, a system of patent controls would probably only have an effect on the least serious risks. Another problem with controlling research through patent law is inadequate enforcement. The Patent Office does not engage in the extensive monitoring that would be required to ensure compliance with the NIH Guidelines. Enforcement of the NIH Guidelines may be limited to checking the consistency of statements of compliance. Such a system would rely too much on good faith on the part of researchers, and would be essentially a self-regulation scheme.

C Tort Law Another indirect way of controlling recombinant DNA research is required by 18 U.S.C. § 1905. While a private right of action does not exist under section 1905, review may be possible under section ten of the Administrative Procedure Act, 5 U.S.C. § 706(2)(A) (1976), if the agency's violation of section 1905 is arbitrary, capricious, an abuse of discretion or otherwise not in accordance with law. Id. See generally, K. DAVIS, I ADMINISTRATIVE LAW TREATISE § 5:8 (2d ed. 1978). The problem also could be avoided if a statute provided that information on recombinant DNA research were immune from the FOIA and other information access laws. The Pharmaceutical Manufacturers Association (PMA) suggests a statute similar to the Federal Nonnuclear Energy Research and Development Act, 42 U.S.C. §§ 5901, 5916 (1976). 1977 Senate Hearings, supra note 22, at 177 (testimony of J. Stetler). 154. 42 Fed. Reg. 13,147 (1977). The effect of the order was suspended March 9, 1977. A federal interagency committee reviewed these events on March 29, 1977, and a majority of committee members were "favorably disposed to the reinstatement of the Commerce De- partment order" because the acceleration did not change patent policies and it would moti- vate U.S. industry and foreign researchers to comply with the NIH Guidelines. NIH PATENT ANALYSIS, supra note 149, at 12. As yet the Commerce Department has not rein- stated the order. 155. Publicly funded research is controlled by making compliance with the NIH Guide- lines a condition on the receipt of federal funding. ECOLOGY LAW QUARTERLY [Vol. 8:55

by creating remedies in tort for those injured by gene-splicing research and activities. Such a plan would serve not only to compensate injured victims, but also to minimize the risks of research through the threat of financial liability. Nonetheless, tort liability, though promising in some respects, is inadequate as a regulatory tool. 156

1. Negligence In order to establish a cause of action for negligence, an injured plaintiff must show: (1) that the defendant owes a duty to the plaintiff; (2) that the defendant has breached that duty; (3) that there is an actual injury; and (4) that the defendant's breach of duty is the proximate cause of the injury. 157 The first and third requirements should pose no substantial problems for most plaintiffs in an action based upon recombinant DNA research. Presumably a duty toward the general public on the part of the researcher and research institution could be established from traditional notions of foreseeability.158 A contention that research activities breached the duty of care owed to a particular plaintiff would have to be premised upon a showing that the defendant fell below the legal standard of conduct for the profession. In this case such a standard would be that of a reasonable and prudent person with the knowledge, training, and skill normally associated with those involved in recombinant DNA research. 15 9 In the absence of any judicial or legislative guidelines, the courts would rely upon the expert testimony of scientists to establish a standard of care in each case; this reliance would inject a possibility of bias into the process and make litigation of such claims expensive.160 Even assuming that a standard of care can be established, a plain-

156. The discussion in this section will be limited to negligence and strict liability as foundations for tort remedies. A third, possibly applicable tort doctrine, is nuisance. Re- combinant DNA research might be considered a nuisance because such experimentation poses the constant threat of escape of dangerous organisms. However, it is unlikely that a court would grant relief premised on such a theory. See Note, Recombinant DNA, supra note 1, at 799-832. 157. W. PROSSER, LAW OF TORTS 143 (4th ed. 1971). 158. Whether a duty is owed to all those injured would depend on the foreseeability both of their being injured and of the type of harm suffered. Since the potential risks of recombinant DNA research are fairly well-publicized, if not well-defined, neither the type of harm that probably would occur nor its infliction on a remote plaintiff is completely unfor- seeable. This outcome is by no means certain, however. See W. PROSSER, supra note 157, at 255-56. See also Note, Recombinant DNA, supra note 1, at 815. 159. This standard of care is that generally applied to professionals. See W. PROSSER, supra note 157, at 161-62. 160. The highly technical nature of recombinant DNA research would probably require scientists for testimony relevant to a standard of care. Within the scientific community there may develop a "conspiracy of silence" with regard to lawsuits. This attitude has been encountered in medical malpractice litigation; medical doctors are extremely reluctant to tes- tify against each other. The use of scientific experts as witnesses also may result in customary practice becom- 1979] RECOMBINANT DNA

tiff may have great difficulty proving that a researcher has failed to act in accordance with that standard. Evidence of negligence nearly always would be in the possession of the defendant. The doctrine of res ipsa loquitur could prove helpful to plaintiffs by allowing a jury to draw an inference of negligence from the fact of injury without direct evidence of negligent conduct. 161 Under this doctrine, however, the plaintiff must show that the accident would not ordinarily occur in the absence of negligence. 162 This showing would be difficult since it is not clear that injuries related to recombinant DNA research can be prevented even if a legally prescribed standard of care is strictly ob- 63 served. 1 Another obstacle to recovery in tort lies in establishing a proximate relationship between the injury and defendant's conduct. Firstly, scientists might simply be unable to isolate and identify the pathogen that causes a particular harmful effect. This problem could become particularly acute if a pathogenic recombinant product establishes only a transitory ecological niche and perishes after manifesting one harmful effect, thereby erasing any trace of the cause of the disease. Sec- ondly, where a disease has several alternate causes, as with cancer, attribution of the disease to one particular source could be impossible. Thirdly, even if the cause of the disease is identified, proof that it came from a laboratory experiment rather than from natural sources could be very difficult. 64 Botulism is botulism, whether natural or created artificially from an E. coli bacterium through introduction of botu- linum toxin genes. Fourthly, in areas where several institutions are conducting similar research, significant problems could arise in determining which laboratory actually produced the pathogenic organism in

ing the standard of care regardless of its reasonableness. Id. at 164 n.60 and sources cited therein. 161. This view is the majority rule. In some states res ipsa loquitur actually shifts the burden of proof. In a few others it shifts the burden of going forward with evidence; that is, the defendant is presumed negligent until the introduction of some evidence to the contrary. Id. at 230. 162. Other requirements of res ipsa loquitur are that the accident be caused by an agency or instrumentality within the exclusive control of the defendant and that the plaintiff be free of contributory negligence. Id. at 214. The latter requirement could cause difficulty in a suit by researchers and employees who could have had a part i the negligent activities. 163. See section III infra. Res ipsa loquitur traditionally has not been applied to activities that are new and untested. For example, res ipsa loquitur was not applied to aircraft accidents until technology rendered air travel safe to the degree that accidents seldom occurred without negligence. See W. PROSSER, supra note 157, at 216. 164. If the pathogenic organism itself is identified, as opposed to the disease-causing product, and if that organism is not typically found in nature, the proof is easier. The probability of a new pathogen developing by natural evolution in an area in which recombinant DNA research is being conducted is quite low. Furthermore, statistical evidence is admissable to show causation. See, e.g., Stubbs v. City of Rochester, 226 N.Y. 516, 124 N.E. 137 (1919) (statistical evidence used to establish causation between impure water supplies and cholera infection). ECOLOGY LAW QUARTERL Y [Vol. 8:55

question. 65 Finally, it would be virtually impossible to reconstruct an accident to determine whether negligent conduct was involved in the 166 production or escape of the organism that actually caused harm.

2. Strict Liability Another tort theory that could be used in connection with litigation over recombinant DNA research is strict liability. Unlike negligence, strict liability does not depend upon a showing that defendant breached a duty to plaintiff. The mere fact that an injury was caused by defendant's actions is sufficient to impose liability. Application of strict liability would probably depend on whether recombinant DNA research can be characterized as an ultrahazardous activity.' 67 Factors which are determinative of this question include: (1) whether the activity involves a high degree of risk of some harm; (2) whether the harm is likely to be great; and (3) whether the risk of the activity cannot be eliminated by the exercise of reasonable care.' 68 The Restatement (Second)of Torts notes that an activity may be characterized as ultrahazardous "if the potential harm is sufficiently great, [even 69 though] the likelihood that it will occur is comparatively slight."' Even if recombinant DNA experiments were deemed to be an ultrahazardous activity, however, mitigating factors, such as the value of the activity to the community and the appropriateness of its location, would have to be considered; if such factors 6utweigh the threats posed 70 by the activity, strict liability could not be imposed.' Strict liability for recombinant DNA research is probably an inadequate system of control. Plaintiffs would face the considerable problem of establishing causation, which often would be a difficult task.

165. For example, in the San Francisco Bay Area, recombinant DNA research is conducted by Cetus Corporation, Genetech, Inc., the University of California, Berkeley, the University of California, San Francisco Medical Center, and Stanford University. 166. Circumstantial evidence regarding normal laboratory procedures and safety precautions is admissible as evidence of negligence, but this information would be in the control of the defendant. 167. For detailed discussion of this point, see Note, Recombinant DNA, supra note 1, at 815-19. See generally, Friedman, Health Hazards Associated with Recombinant DNA Tech- nology Should Congress Impose Liability Without Fault, Recombinant DNA Research Sym- posium, supra note 5, at 1355. 168. RESTATEMENT (SECOND) OF TORTS § 520, Comments a, b, c, h (1977). 169. Id. § 520, Comment g. 170. Id. § 520, Comments e, f, g, h. An alternative to the adoption of strict liability by judicial decision would be an imposition of strict liability by legislation. Senator Dale Bumpers introduced a bill, S. 621, 95th Cong., 1st Sess. (1977), that would have imposed strict liability on recombinant DNA research. When questioned about the possibility that strict liability might deter research, however, he stated, "I recognize the harshness of that [strict liability provision]. . . . Quite frankly, I am not hung up on it. . . [but] everybody should comply with NIH guidelines." 1977 Senate Hearings, supra note 22, at 53. The bill was not reported out of committee and Congress is not presently considering such a plan. For more information on legislative alternatives, see text accompanying notes 308-20 infra. 19791 RECOMBINA4NT DNA

Furthermore, awarding damages in the absence of fault may unnecessarily restrict scientific research.17' The primary purpose of regulation is to minimize risks without significantly reducing potential benefits. A regulatory scheme should penalize excessive risk taking, yet encourage research that maximizes benefits at a minimal risk. Holding researchers to a standard of strict liability would place a penalty on all research that causes harm, regardless of its potential risks and benefits.

3. Problems Generic to Tort Liability In addition to problems specific to negligence or strict liability, three general difficulties arise with the use of tort liability to regulate research and experimentation. Firstly, tort law does not ensure that such activities will be carried on with an appropriate balance of risks and benefits. While potential liability may discourage some high risk activities, there is no indication that it would do so uniformly or that all high risk activities would be eliminated. For example, the tort liability system would not have much impact upon high risk activities where difficult proof problems make recovery unlikely. Furthermore, some activities that society may wish to discourage would not be deterred since litigation itself would be unlikely. One of the lesser publicized risks of recombinant DNA research is non-fatal sickness--the imposition of a small amount of harm on a large number of individuals. Few people would press claims where individual recovery of damages 17 2 would be small.

171. Strict liability could halt research because it forces a researcher to internalize the costs when the benefits are mainly external. A researcher may believe (or may be told by the university) that since there is little pecuniary gain to the research, and risk of great loss, the research must be stopped even if society as a whole would prefer the research continue. Senator Bumpers has asserted that if research is curtailed, "that would not necessarily have a bad effect if a researcher had doubts about his abilities to control the products of his work." Letter from Senator Dale Bumpers to the authors, Sept. 22, 1977 (on file in Ecology Law Quarterly office). The following colloquy occured during the 1977 Senate hearings: Senator Kennedy: [Under a strict liability provision] [clan a university. . . risk conducting this work? If some accident did occur, under strict liability, the university might close down. No private university would be able to do the research under a strict liability mechanism. That may or may not be so. But that is what the universities have said. Senator Bumpers: I understand they cannot buy insurance today [for recombinant DNA research risks]. 1977 Senate Hearings, supra note 22, at 53. Id. at 190 (the insurance coverage of pharmaceutical companies would be reassessed). Similar fears about strict liability are causing the cessation of research in the vaccine development field. See Franklin & Mais, Tort Law and Mass Immunization Programs, 65 CALIF. L. REv. 754 (1977), 7 ENVT'L L. REP. (ELI) 10,081 (1977). 172. One way to deal with this problem is through class action suits. However, many difficulties, such as stringent notice requirements, make many such suits in federal court ECOLOGY LAW QUARTERLY [Vol. 8:55

A second difficulty with a tort liability model is the high transaction costs of litigation. Court costs and legal fees constitute a significant percentage of most tort damage awards. 173 In addition, difficult proof problems that would be encountered in recombinant DNA cases increase the costs of discovery and trial preparation. In particular, plaintiffs may need to spend large sums to procure expert analysis of causation and evidence regarding a standard of care. Such high transaction costs substantially reduce the usefulness of a tort liability scheme. One of the virtues of private regulation in general is that the administrative costs of a direct regulation scheme can be eliminated. In the case of gene-splicing research, however, increased transaction costs may more than counterbalance the savings. Costs of direct regulation have been estimated to be on the order of three million dollars a year.174 If victims of research mishaps were forced to resort to litigation, the net societal cost would probably be of a similar magnitude. Finally, a tort liability model would arguably place too great a burden on the scientific community by requiring it to bear a disproportionate share of the risks of scientific research. While researchers would receive some financial gain and enhanced prestige from their work, the benefits of recombinant DNA research would be enjoyed by society as a whole. 175 A tort liability model would require researchers to bear the entire cost of the risks associated with the research, as well as the resulting transaction costs. 176 Although this result might be re- unlikely. See Eisen v. Carlisle & Jaquelin, 417 U.S. 156 (1974). If the amount of damage to each individual is small, the transaction costs of notice, litigation, and dispersal of benefits may be greater than the cost of the harm even when those damages are cumulated. A related problem, inherent to all tort liability schemes, is the inability of plaintiffs to find a financially responsible defendant. Since a large percentage of research is federally funded, defendants might plead government immunity from liability since they are providing a public benefit. The Federal Tort Claims Act contains relevant exceptions to the rule of government immunity, however. See 28 U.S.C. §§ 1346, 1402, 1504, 2110, 2401, 2402, 2411, 2671-2680 (1976); W. PROSSER, supra note 157, at 972. Also, many publicly funded research institutions have applied for patents on recombinant DNA techniques. See note 149 supra. Such attempts to reap private gain from research help thwart claims of governmental immunity. 173. For example, in the Buffalo Creek dam disaster in 1972, one law firm representing only one-sixth of the survivors, "incurred almost $500,000 in expenses before the case was settled, without taking on the substantial costs of litigation or billing any attorney fees." The case was taken on a contingent-fee basis. Rabin, Dealing with Disasters: Some Thoughts on the Adequacy of the Legal System, 30 STAN. L. REV. 281, 295 (1978). 174. The estimated cost of the 1977 Kennedy bill, S. 1217, was approximately $3.8 million per year. S. REP. No. 359, 95th Cong., 1st Sess. 16 (1977). The estimated cost of the most recent House bill, H.R. 11192, is somewhat lower, averaging $2.4 million a year. H.R. REP. No. 1005, 95th Cong., 2d Sess. 33 (1978). For a discussion of these bills, see text accompanying notes 308-20 infra. 175. See notes 49-61 and accompanying text supra. The outer layer of benefits to the public is clearly larger than the inner layer of benefits for researchers. 176. It could be argued that the benefits of recombinant DNA research are like the ben- 19791 RECOMBINANT DNA garded as a simple internalization of the costs of recombinant DNA research, it actually would represent a false distribution of the burdens. Since many of the benefits resulting from gene-splicing research will be shared by all, society should bear a share of the costs. 77 Besides being inequitable, forcing scientists to bear all the costs of recombinant DNA research could also serve to adversely affect the risk/benefit calculus of private decisions to continue research. While an objective third party would conclude that the benefits of a certain experiment outweigh its risks, a researcher compelled to accept full financial responsibility for any harm resulting might decide not to proceed. As a result, an experi- 7 ment beneficial to society would not be conducted.' 8

III REGULATION THROUGH EXISTING FEDERAL LAW The right to conduct scientific research is arguably protected by the first amendment. 79 Since scientific research is a principal part of academic inquiry, regulating such research without infringing upon freedom of speech may not be possible. Although the United States Supreme Court has never decided whether scientific research is protected by the first amendment, it has made statements indicating that academic freedom is an element of protected speech.' 80 To the extent efits of manufacturing products; thus the researchers should be liable for the harm they cause. In the case of products, however, manufacturers are able to socialize the costs by passing them on to individual consumers. With recombinant DNA research there is no mechanism for passing on costs; the benefits of research are distributed generally and researchers would have to bear all losses. When and if recombinant technology reaches the point of producing consumer products, costs may be passed on more easily and it may be wise to impose liability. Query why, however, those experiments, that produce consumer products, should be singled out for liability when other experiments not expected to result in marketable products should not. 177. Otherwise, society would be obtaining valuable benefits without paying the cost. For a discussion of the benefits of recombinant DNA research, see notes 49-61 supra and accompanying text. 178. This consequence is clearly adverse to the need for regulation that minimizes risks without largely reducing benefits. 179. Primarily due to the recombinant DNA research controversy, there has been much written recently on first amendment aspects of biological research. See, e.g., Robertson, The Scientist's Right to Research.: A Constitutional Analysis, in Recombinant DNA Research Sym- posium, supra note 5, at 1203; Davidson, First Amendment Protection for Biomedical Re- search, 19 ARIZ. L. REV. 893 (1977); Delgado & Millen, supra note 5. Delgado and Millen conclude, for example, that their "review of constitutional history, first amendment theoretical rationales, and first amendment case law suggests that a persuasive case can be made for including scientific inquiry within the field of protected expression." Id. at 402-03. 180. "The essentiality of freedom in the community of American Universities is almost self evident. . . . Scholarship cannot function in an atmosphere of suspicion and distrust. Teachers and students must always remain free to inquire, to study and to evalute, to gain new maturity and understanding. Otherwise our civilization will stagnate and die." Sweezy v. New Hampshire, 354 U.S. 234, 250 (1957). See also Keyishian v. Board of Regents, 385 U.S. 589, 603 (1967). ECOLOGY LAW QUARTERL Y [Vol. 8:55

that scientific research falls under the rubric of academic freedom, the ability of government to regulate such research may be limited. Yet, although science is academic expression to the degree that it involves the free interaction and publication of ideas,' 8' actual experimentation is "action" much more than it is "speech."' 182 The dangers associated with free expression of ideas are vague and speculative com- pared to those of actual experimentation since the latter may pose a serious risk to the public health. Even assuming arguendo that scientific research is protected by the first amendment, however, regulation of its risks still is justified under certain circumstances. A number of commentators have taken the position that the regulator has the burden of establishing such a justifica- tion;183 the government should demonstrate that there are compelling reasons for regulation t84 and that the necessary controls cannot be ef- fected by means of a less onerous alternative. 85 Since there are significant health risks associated with recombinant DNA research, the government's interest could be shown to be "compelling." However, care should be taken to assure that regulation is not overly broad. Re- strictions may be constitutionally justified only to the extent necessary to regulate the health and environmental risks of scientific experimentation; such a limitation on the reach of governmental control would prevent unnecessary encroachment on the free exchange of ideas within the scientific community. 186 Regulating to minimize the health and

181. It was a limitation on this sort of speech that the Supreme Court ruled unconstitu- tionally vague in Keyishian v. Board of Regents, 385 U.S. 589, 597-604 (1967). That case involved the New York "Feinberg Law," which prevented the appointment or retention of teachers who were "subversive." The court struck down the law insofar as it limited the legitimate expression of opinion by those in the academic community. 182. 1977 House Hearing,supra note 6, at 879 (testimony of Prof. T. Emerson). See also Science Policy Implicationsof DNA Recombinant Molecule Research, Report by Subcomm. on Science, Research and Technology of House Comm. on Science and Technology, 95th Cong., 2d Sess. 58-61 (1978) [hereinafter cited as 1978 House Science Policy Rep't] (Professor Bar- ron believes the first amendment favors change, and scientific research is protected because it contributes to change; Professors Bein and Green on the other hand are less certain of the applicability of the first amendment to scientific research). See also Stellin, Freedom of In- quiry, 189 SCIENCE 953 (1975). 183. 1977 House Hearings,supra note 6, at 929 (testimony of W. Morton); 1978 House Science Policy Rep't, supra note 182, at 59-60 (testimony of Prof. Emerson). 184. For a discussion of this test and others that might be applied to the regulation of recombinant DNA research, see 1977 House Hearings, supra note 6, at 887; 1978 House Science Policy Rep't, supra note 182, at 58, 60; Robertson, A Purposive Analysis of Constitu- tionalStandards ofJudicialReview anda PracticalAssessmentof the ConstitutionalityofReg- ulatingRecombinant DNA Research, Recombinant DN4 Research Symposium, supra note 5, at 1281. 185. "Even though the government purpose be legitimate and substantial, that purpose cannot be pursued by means that broadly stifle personal liberties when the end can be more narrowly achieved." Shelton v. Tucker, 364 U.S. 479, 488 (1960). This test was used by the court in Keyishian v. Board of Regents, 385 U.S. 589 (1967). 186. "Thus the government could not prohibit, regulate or discourage in any way [sic] 1979] RECOMBINANT DNA safety risks of recombinant DNA experiments satisfies this norm, for controls would be aimed specifically at maximizing benefits as well. Although first amendment objections to the regulation of recombinant DNA research are not the only possible legal claims, 8 7 they represent the most significant barrier to regulation. The notion of "freedom of inquiry" is perhaps the most basic objection the scientific community has raised with respect to the regulation of gene-splicing research. Even if this freedom is not completely protected by the first amendment, it still represents an important policy consideration in the formulation of a direct regulation plan. Since private systems offer only limited potential for success, and since direct regulatory schemes are permissable, one possible approach is the use of existing federal laws as a regulatory structure. 88 A federal interagency task force considered this possibility and concluded that existing statutes were inadequate for such regulation. 8 9 This Section evaluates existing federal authority and reviews the findings of the interagency report. Three acts will be considered: The Toxic Substances Control Act (TSCA), the Occupational Safety and Health (OSH) Act, and the Public Health Services (PHS) Act. 190

DNA research on the ground that mankind ought not to be pursuing ideas about ways to develop new forms of life. On the other hand, experiments that presented a substantial and serious danger to the physical health and safety of the surrounding population could be subject to regulation without infringing the guarantees of the First Amendment." 1977 House Hearings, supra note 6, at 880 (testimony of Prof. T. Emerson). Other legal commentators agree that minimally sufficient regulation is justified. "To the extent that governmental action trenches on an individual scientist's right to carry out basic research and does so on grounds of content rather than concerns legitimately falling within the state's funding or police power, such action should be examined under the more stringent model of judicial review courts have developed in cases involving infringement of fundamental interests." Delgado & Millen, supra note 5, at 403; Berger, supra note 5, at 83, 107. 187. Another possible challenge to regulation would be on fifth amendment grounds. Opponents of regulation could argue that it constitutes a taking of property without compensation, because restrictions on research would both eliminate a valuable property use and constitute confiscation of a property interest in the research itself-i.e., a patentable property right. Such an attempt would probably fail since regulation is justified as an attempt to protect the public health. There would also be difficulties in showing: (1)that controlling research significantly lowers the value of a researcher's property; and (2) that a patentable procedure is necessarily a "property right" despite the granting of patents by government largesse. 188. This Article does not assume that a federal regulatory scheme is the only way to regulate recombinant DNA research. See text accompanying notes 278-95 infra. Federal laws are reviewed here because they have been most frequently mentioned as possible vehicles for control of research. 189. INTERIM REPORT ON THE FEDERAL INTERAGENCY COMMITTEE ON RECOMBINANT DNA RESEARCH: SUGGESTED ELEMENTS FOR LEGISLATION (Mar. 15, 1977) [hereinafter cited as INTERAGENCY REP'TJ, repriniedin 1977 HEW DOCUMENTS, supra note 1,at 289. 190. This list is not exhaustive. For example, the Interagency Report also considered the Hazardous Materials Transportation Act, Pub. L. No. 93-633, § 101, 88 Stat. 2156 (1974). The three laws to be considered are the most viable and far-reaching of current options in federal law, however, and thus discussion is limited to those acts. ECOLOGY L4W QUARTERLY [Vol. 8:5 5

A. Toxic Substances Control Act One possible regulatory scheme for controlling recombinant DNA research is the Toxic Substances Control Act (TSCA),19 1 administered primarily by the Environmental Protection Agency (EPA). This Act allows the Administrator of the EPA to place restrictions on the manufacture or use of chemical substances that present an "unreasonable 92 risk" of injury to health or the environment.1 Section 6 gives EPA apparent authority over recombinant DNA research. The materials used in such research, along with the immediate products, are covered in most cases by a definition of "chemical substance" in TSCA. 193 Whether the term "manufacture" covers scientific laboratory research is more ambiguous. The interagency task force concluded research was included: However, another section of the TSCA (section 5), which requires manufacturers to give EPA advance notice of plans to manufacture a new chemical substance, contains an exemption from the notice requirement for ". . .manufacturing or processing of any chemical substance. . .only in small quantities. . .solely for the purposes of scientific experimentation ... " 15 U.S.C.A. § 2604(h)(3) (West Supp. 1978). The wording of this provision would seem to indicate that scien- 194 tific experimentation constitutes manufacturing under the TSCA. EPA has the power to regulate or prohibit gene-splicing research only if it creates an "unreasonable risk of injury to the health or the environment. .. "195 Any research conducted without safety procedures probably involves an unreasonable risk.196 EPA thus has the power to impose safety requirements. The authority to regulate is useless unless the regulator is aware of the activities to be controlled. Scientific experimentation is specifically exempted from the notification requirements of § 5 of TSCA. This exemption is the "most serious deficiency in TSCA" as authority to regulate, 197 aecause researchers using small quantities of recombinant DNA are not required to notify the EPA when they intend to manufacture a new substance. 98 TSCA's regulation scheme depends upon registra-

191. Toxic Substances Control Act, 15 U.S.C.A. §§ 2601-39 (West Supp. 1978). 192. Toxic Substances Control Act § 6, id. § 2604(f)(1). 193. INTERAGENCY REP'T, supra note 189, at 32. 194. Id. at 32-33, 33 n.2. The report also concluded that the absence of an explicit exemption for scientific research in section 6 created a negative implication that recombinant DNA research was within the scope of TSCA regulation. Id. at 33. 195. 15 U.S.C.A. § 2604(f)(1) (West Supp. 1978). 196. See text accompanying notes 115-34 upra. 197. INTERAGENCY REP'T, surpa note 189, at 33. 198. See text accompanying note 194 upra. 19791 RECOMBINANT DNA tion with the EPA under the notice requirement, and without notice, recombinant DNA research cannot be regulated. Although TSCA is not capable of controlling all gene-splicing experiments, it could be used to regulate both large-scale operations using significant quantities of recombinant DNA and the manufacture of recombinant DNA research products. TSCA also gives the EPA the authority to require manufacturers to carry on exhaustive tests to determine the safety of their proposals. 99 Since the research deals presently only with small quantities, however, TSCA is of limited utility.2oo

B. OccupationalSafety and Health Act Another possible mechanism for regulating recombinant DNA research under existing federal law is the Occupational Safety and Health (OSH) Act. 201 The OSH Act could apply to the regulation of gene- splicing research since the risks of such research are related directly to the health and safety of those working in laboratories. Just as the Oc- cupational Safety and Health Administration (OSHA) now enforces safety standards in other areas, safety precautions could be applied to recombinant DNA research under OSH Act authority. Since OSHA has never totally prohibited an activity in the past, the Interagency Task Force doubts that the agency would prohibit any recombinant DNA experiments. 20 2 A related concern is "whether OSHA should give priority to the establishment [of recombinant DNA regulations] over others that have been awaiting promulgation, taking '20 3 into account the statutory test of. . . 'urgency of the need'. These comments reflect doubt as to OSHA's present administrative workload, however, rather than the OSH Act as a regulatory tool, and thus are not roadblocks to controlling recombinant DNA research through the Act. There are difficulties, however, in assigning promulgation and enforcement responsibilities to OSHA. Recombinant DNA research regulations must be revised periodically in order to place realistic controls on a constantly changing field. OSHA does not have the expertise in molecular biology to fulfill this task effectively. Furthermore, OSHA's ability to monitor facilities and enforce standards is open to question.204

199. 15 U.S.C.A. § 2603 (West Supp. 1978). 200. The Interagency Task Force Report also argues that inevitable litigation over the ambiguities discussed would delay the application of TSCA. INTERAGENCY REP'T, srupra note 189, at 33 n.3. 201. 29 U.S.C. §§ 651-678 (1976). 202. INTERAGENCY REP'T, supra note 189, at 31. 203. Id. 204. OSHA has been criticized for setting the wrong priorities, conducting an insufficient number of inspections, not identifying serious violations during inspections, acting too ECOLOGY LAW QUARTERLY [Vol. 8:55

There are also problems associated with OSHA's jurisdiction to administer controls over all institutions conducting recombinant DNA research. Firstly, the interagency report concluded that proving recombinant DNA research is a "recognized hazard" under the OSH Act may be difficult. However, the National Institutes of Health's actions in promulgating guidelines should provide proof of such "recogni- ' 20 5 tion. Secondly, OSHA does not have jurisdiction over state and local government employees; only twenty-four states have voluntarily consented to regulation by OSHA, leaving unregulated the public universities in the remaining twenty-six states. 2°6 Since most exempt institutions are sponsored by government funding, action by the NIH or another government agency would be needed to supplement OSHA 207 controls.

C. Public Health Services Act The Public Health Services (PHS) Act provides another potential mechanism for the control of recombinant DNA research. 20 8 Section 361 of the PHS Act empowers the Secretary of Health, Education and Welfare (HEW) to "make and enforce such regulations as in his judgment are necessary to prevent the introduction, transmission, or spread of communicable diseases . . . from one State or possession into any other State or possession .... ,,209 This section could apply to recombinant DNA research, since the spread of communicable disease is one of the risks of such research. 210 Although not all experiments pose this

slowly in setting standards, and being inadequately informed on the causes and safest procedures to follow. See N.Y. Times, June 6, 1977, at 28, col. I (editorial), GENERAL ACCOUNT- ING OFFICE, WORKPLACE INSPECTION PROGRAM WEAK IN DETECTING AND CORRECTING SERIOUS HAZARDS, REPORT TO THE CONGRESS, HRD-78-34 (May 19, 1978), GENERAL AC- COUNTING OFFICE, SPORADIC WORKPLACE INSPECTIONS FOR LETHAL AND OTHER SERIOUS HEALTH HAZARDS, REPORT TO CONGRESS, HRD 77-143 (Apr. 5, 1978). Secretary of Labor Ray Marshall has admitted "[t]here is much validity to these charges." N.Y. Times, Mar. 12, 1977, at 31, col 5. The slow pace of OSHA led the GAO to conclude that "the agency has had no overall impact on the health conditions that each year are estimated to kill 100,000 workers." N.Y. Times, Mar. 24, 1977, pt. IV at 16, col. 4. However, OSHA has announced a new enforcement plan which it hopes will alleviate some of these problems. N.Y. Times, Oct. 6, 1977, pt. IV at 5, col. 1. 205. INTERAGENCY REP'T, supra note 189, at 30. See also text accompanying note 190 supra. For discussion of the Guidelines, see text accompanying notes 225-77 infra. 206. INTERAGENCY REP'T, supra note 189, at 30. 207. Problems associated with regulatory plans that involve coordination of two or more government agencies are discussed in the text accompanying note 216 infra. 208. 42 U.S.C. § 264 (1976). See Note, Recombinant DNA, supra note 1, at 836-45. 209. 42 U.S.C. § 361 (1976). 210. INTERAGENCY REP'T, supra note 189, at 58. The Report admits that section 361 "could perhaps be interpreted more broadly to serve as legal support for more comrpehen- sive regulation," but that there would "have to be a reasonable basis for concluding that the products of all recombinant DNA research. . . may cause human disease," and finds that "such a conclusion would be tenuous at best, and it is unlikely that resulting requirements could be effectively imposed and enforced." Id. at 35. 1979] RECOMBINANT DNA

threat, scientists do not know which experiments do carry the risk of disease. The danger of disease in general should suffice to create section 361 jurisdiction.2" Section 361 could be the basis for a thorough regulatory scheme since it provides that "the [Secretary of HEW] may provide for such inspection. . . disinfection. . . and other measures as in his judgment may be necessary. ' 212 However, HEW has refused to regulate recombinant DNA research under section 361.213 The use of section 361 is not, however, an adequate means of controlling recombinant DNA research. The Act is "essentially aimed just at imposing labelling, packaging, and shipping requirements . . . [an] approach [that] is in line with the statutory language which emphasizes movement. ' ' 214 It is doubtful that the relevant segment of the PHS has the expertise to promulgate, monitor, and enforce regulations. 215 For comprehensive control of recombinant DNA research, a more complete scheme must be considered.

D. Combined Application of Existing Acts One solution to the problem of inadequate existing federal authority is some combination of enforcement schemes. For example, a combination of OSHA and present NIH controls might be an effective form of regulation. Another might be a combination of PHS Act safeguards for disease risks with broader environmental regulation handled by another agency. One advantage of such a scheme would be that all research could be regulated without the need for new legislation and its ensuing costs and delays. It would also utilize the combined expertise

211. HEW's Draft Environmental Impact Statement admits communicable disease could in fact result, and states further that one of the purposes of strict physical containment safeguards for experiments is to prevent epidemics. INTERAGENCY REP'T, Supra note 189, at 58. In addition, "it has already been recognized that regulations pursuant to authority under § 361 control transportation of [recombinant] DNA materials," and the research and use of such material pose a greater public risk than transport. Id. at 59. But see the Department of Transportation's definition of etiologic agents, which may not include recombinant DNA, in the context of the Hazardous Materials Transportation Act. Id. at 34. 212. 42 U.S.C. § 361 (1976). 213. HEW dismissed a Nov. 11, 1976, petition by the Environmental Defense Fund and the Natural Resources Defense Council asking for hearings to develop regulations. INTER- AGENCY REP'T, supra note 189, at 45. 214. Id. at 35. The Interagency Report does not recommend the use of § 361 to control recombinant DNA research, Id. at 8-9. 215. Although the Center for Disease Control (CDC) has neither the expertise nor the ability to monitor experimentation, and promulgate and enforce regulations that EPA and perhaps OSHA have, section 353 of the Public Health Service Act, 42 U.S.C. § 264 (1976) gives the CDC general authority to license and control operation of clinical laboratories. Former General Counsel of the Food and Drug Administration, Peter Barton Hutt, believes that § 361 of the PHS Act would provide "one of two possible means of control if the HEW wished to use it. ...-licensing the laboratories, or licensing the crucial enzymes and foreign DNA needed for the experiments. 1976 HEW DOCUMENTS, supra note 1, at 336-38. ECOLOGY LAW QUARTERLY [Vol. 8:55 of various regulatory agencies in dealing with the related but diverse problems involved in regulating recombinant DNA research. One problem of a combined approach to regulation is that it can result in different standards for different areas of research. Even if all regulatory branches began with the same set of guidelines, variations, differing interpretations and priorities could lead to inconsistent appli- 216 cation of regulatory directives. Additionally, in designing a regulatory plan involving several government agencies, overlap between their jurisdictions is a problem. Overlapping jurisdiction leads not only to an inefficient duplication of effort, but to a situation in which one or more agencies defers to the authority of another. The result is "gaps" in enforcement, where a number of agencies have jurisdiction but none properly exercises its power to ensure that all problems are addressed. 217 The best way to avoid such inefficiencies is to grant the authority to regulate recombinant DNA research to a single government agency.

E A CurrentAdministration Proposal On December 15, 1978, Secretary of HEW Joseph Califano announced that he was taking immediate steps to require that research conducted by private companies comply with the NIH Guidelines 21 8 through the authority of the Food and Drug Administration (FDA). Califano also requested that EPA review its authority to take action to control research not regulated by the FDA.219 The FDA plans to take action by promulgating regulations requiring that "any firm seeking approval of a product requiring the use of recombinant DNA methods in its development or manufacture demonstrate the firm's compliance 0 with the requirements of the NIH [Gluidelines. ' '22 Unfortunately, the Administration's proposal is defective for several reasons. The most obvious problem is that the proposed plan is a

216. Cf Doniger, FederalRegulation of Vinyl Chloride: A Short Course in the Law and Policy of Toxic Substances Control, 7 ECOLOGY L. Q. 497, 643-53 (1978) (discussing the problem of fragmentation of authority in the context of vinyl chloride regulation); Note, The Concorde Calculus, 45 GEO. WASH. L. REV. 1037, 1062-63 (discussing the conflicting cost- benefit analyses of the Department of Transportation, EPA, Federal Energy Administration, Department of State, and National Aeronautics and Space Administration on the justifiabil- ity of the Concorde Super-Sonic Transport). 217. See generaly Doniger, supra note 216, at 643-46. 218. NIH Revised Guidelines, supra note 31, at 60,080. 219. Id. 220. FDA, Notice of Intent to Propose Regulations, 43 Fed. Reg. 60,134 (1978). The authority to impose such regulations is not discussed by the FDA Commissioner but apparently stems from 21 U.S.C. §§ 348, 355, & 360(b). The NIH Guidelines would be "incorporated in good manufacturing practice regulations, should recombinant DNA techniques be proposed for the manufacture of products for commercial distribution." 43 Fed. Reg. 60,134 (1978). 19791 RECOMBINANT DNA

"mixed" regulation scheme, the sort of plan that has been criticized above. Secretary Califano suggested not only that NIH, FDA, and EPA take regulatory responsibilities, but also that OSHA have a role. 221 Gaps and overlaps in such a plan will be numerous. Nor is it clear just how extensive the FDA's regulation of pharmaceutical manufacturers would be. Private researchers would be required to demonstrate their compliance with the guidelines for marketable products, 222 but what, if any, requirements would be imposed on basic research is unknown. As Secretary Califano himself stated only three months before his request for FDA regulations, "inas- much as recombinant DNA research has not yet reached the stage where it has yielded products to be regulated by the FDA, it was agreed that FDA probably does not have the authority to impose requirements on such research at present. '223 FDA control of privately funded recombinant DNA research may, for the time being, prove to be illusory. EPA's role is also vague. Secretary Califano's statement indicates that the EPA should regulate private research "not regulated by the Food and Drug Administration. ' 224 If the FDA cannot control basic research by pharmaceutical manufacturers, apparently the EPA should. Yet, the chances for conflicting and overlapping enforcement between the agencies are great. If the EPA has no such authority, much private recombinant DNA research would remain unregulated.

IV A PROPOSED SOLUTION Direct regulation of gene-splicing by a government agency is the appropriate choice for maximizing benefits and minimizing risks. This section outlines a proposed regulatory scheme for recombinant DNA research, using concepts applicable to regulation of other new technologies. Several aspects of a regulatory scheme are explored: the actual regulations to be used, the position of state and local governments in a regulatory system, the choice of an agency to regulate, and the mechanism by which public input into the regulatory scheme can be main- tained. The authors propose that a modified form of the NIH Guidelines be used as the basis for regulation of recombinant DNA research by a federal agency. That agency should be newly created and independent of NIH.

221. When products of recombinant DNA research reach the commercial development stage "[t]he Occupational Safety and Health Administration will exercise its regulatory authority in the work place." NIH Revised Guidelines, supra note 31, at 60,125. 222. Id. at 60,134. 223. Letter from HEW Secretary Califano to Senator Edward Kennedy, Sept. 12, 1978, published in 43 Fed. Reg. 60,080, 60,105 (1978). 224. 43 Fed. Reg. 60,080 (1978). ECOLOGY LAW QUARTERLY [Vol. 8:55

A. Choice of Regulations 1. Description of the Guidelines In February 1975, a National Academy of Sciences committee sponsored a conference at Asilomar, California, to consider the risks of recombinant DNA research and to recommend means of dealing with potential hazards.225 The result was a recommendation that recombinant DNA experiments should continue provided that appropriate safety precautions are taken. The scientists felt, however, that the risks involved in recombinant DNA experimentation warranted government involvement in the decisionmaking process. Furthermore, they wanted to avoid additional regulatory responsibilities. They therefore recom- 226 mended that the Director of the National Institutes of Health (NIH) appoint a special advisory committee on the research.227 The Recom- binant DNA Advisory Committee met four times in 1975 and 1976, resulting in the release of the NIH Guidelines for Recombinant DNA Research on June 23, 1976.228 The 1976 Guidelines have been replaced recently by the NIH Re- vised Guidelines. 229 Because the Revised Guidelines retain the same format and much of the same content as the previous safety measures, the discussion below focuses on the 1976 Guidelines with notation of the significant changes. Two fundamental changes are that the Revised Guidelines have created exemptions for many experiments and low- ered containment levels, and that the standards adopted are now more flexible and subject to ongoing modification.

225. The scientists were and are concerned with the results of experiments using recombinant DNA techniques, not the recombinant DNA techniques themselves. Cohen, Recom- binant DNA.: Fact and Fiction, 195 SCIENCE 654 (1977). 226. The NIH is a subdivision of the Department of Health, Education & Welfare, which dispenses grant money for scientific experimentation. 227. During this period, only the scientific community seemed to be concerned about the problem. The controversy did not attract initially a great deal of public attention. See Mc- Caull, Research in a Box, 19 ENVIRONMENT 31 (1977). As publicity continued, however, both the public and government became more interested. This interest grew as it became clear that there was disagreement within the scientific community. See, e.g., Cavalieri, New Strains of Lfe or Death, N.Y. TIMES MAGAZINE, Aug. 22, 1976, at 8-9. 228. The NIH Guidelines were published July 7, 1976. 41 Fed. Reg. 27,911 (1976). See note 296 infra for a discussion of the criticisms of the procedure by which the Guidelines were adopted. 229. Revisions of the 1976 Guidelines were begun early in 1977 by the NIH Recombi- nant DNA Advisory Committee (RAC). The committee's proposed revisions were published in September 1977. 42 Fed. Reg. 49,596 (1977). A meeting of the Advisory Committee to the Director (DAC) was held in December 1977, where testimony was heard concerning proposed changes in the Guidelines. As a result of this meeting, and other input to the NIH, the Proposed Revised Guidelines were released for public comment on July 28, 1978 and published at 43 Fed. Reg. 33,069 (1978). After minor modifications the Revised Guidelines were released on December 27, 1978, to take effect January 2, 1979. 43 Fed. Reg. 60,080 (1978). Subsequent references herein to the "Guidelines" will refer to both the 1976 Guidelines and the 1978 Revised Guidelines unless otherwise indicated. 1979] RECOMBINANT DNA

The stated purpose of the 1976 Guidelines was "to recommend safeguards for research on recombinant DNA molecules. ' 230 Four principles, drafted with the concurrence of the general consensus at the Asilomar conference, describe the general purpose of the Guidelines: (1) certain experiments are so hazardous that they should not be performed at the present time; (2) recombinant DNA experiments that are not banned outright "can be undertaken at the present time provided that the experiment is justifiable on the basis that new knowledge or benefits to humankind will accrue that cannot be readily obtained by use of conventional methodology and that appropriate safeguards" are used; (3) as the level of risk associated with the experimentation increases, the level of containment required must also increase; and (4) the Guidelines require annual revision to keep them abreast of current knowledge. 231 Although the second principle implies both that experiments must meet a high standard of scientific merit and that the results are unobtainable from conventional alternatives, this standard has not been applied so long as appropriate safety precautions are followed. 232 Five types of experiments are expressly forbidden by the Guide- lines: (1) experiments in which any part of the DNA is derived from organisms or viruses that are themselves biohazards of any significant risk, or which use cells infected by such viruses;233 (2) experiments involving the deliberate formation of recombinant DNA products containing genes of potent toxins or venoms; (3) any experiment that involves the deliberate creation of recombinant DNA products from plant pathogens that are likely to increase virulence and host range; (4) experiments in which organisms containing recombinant DNA are released deliberately into the environment; and (5) experiments that entail the transfer of drug-resistant traits to microorganisms that are not

230. 41 Fed. Reg. 27,902, 27,911 (1976). 231. Id. at 27,911-19; 1976 HEW DOCUMENTS, supra note 1, at 176. 232. Failure to apply the standard is appropriate in the case of NIH funded research. The scientific merit of individual experiments is considered when awarding grant monies. To require that the experiment again be evaluated for scientific merit would be duplicative. Cf.Draft of French Guidelines for Recombinant DNA Experiments, June 16, 1977 (copy on file at the Ecology Law Quarterly office), which explicitly state that there should be no decision on the merits of an experiment at the level of determining safety precautions. In other words, the safety precautions required for each class of experiments were determined by examining the possible benefits and costs of regulations, as applied to that general class. It is impracticable to examine the risks and benefits of an individual experiment within that class, for reasons of uncertainty involved in further "fine-tuning," delay, and administrative expense. Thus, once the precautions are set for each class, no further examination of the benefits of an individual experiment is made. 233. DNA from smallpox or yellow fever viruses, for example, could not be used. For a listing of organisms banned from experimentation, see, e.g., Center for Disease Control (CDC) Classes 3-5, as listedin CENTER FOR DISEASE CONTROL, U.S. PUBLIC HEALTH SERV- ICE, CLASSIFICATION OF ETIOLOGIC AGENTS ON THE BASIS OF HAZARD (4th ed. 1974); 1976 HEW DOCUMENTS, supra note 1, at 179. ECOLOGY LAW QUARTERLY [Vol. 8:55 known to acquire them naturally if such a transfer might compromise 234 the medical or agricultural use of the drug to control disease agents. In addition, experiments that are otherwise permissible under the Guidelines are prohibited if they involve more than ten liters of product.235 The rationale behind this restriction is that the risks of escape of an organism grow greater as the volume of material increases. This last ban is not absolute, however; experiments involving more than ten liters of product may be allowed where an experiment produces direct 236 scientific and social benefits at a relatively low risk. For those experiments not banned by the Guidelines, proper safety measures must be observed. The degree of precaution required for a given experiment is determined by the relative risk factor assigned to it. The two basic types of safeguards upon which the Guidelines rely are physical and biological containment. Physical containment applies to laboratory techniques and involves the use of special equipment, techi- ques, and facilities. 237 Biological containment involves restrictions on the types of organisms that may be used in experimentation. 238

234. 41 Fed. Reg. 27,902, 27,914, 27,915 (1976). When a microorganism can easily acquire the drug resistance naturally, there is no purpose in banning experiments that lead to such a result. I FINAL EIS, supra note 53, at 45. 235. 41 Fed. Reg. 27,902, 27,915 (1976). 236. The Revised Guidelines also allow such that exceptions to be granted from the five categories of prohibited experiments described above. NIH Revised Guidelines, supra note 3 1, at 60,108. An appropriate containment level is assigned at the time the exception is granted. Id. 237. There are four levels of physical containment, from the most lenient, PI, to the most stringent, P4. Laboratories using P 1 precautions are commonly used for nonhazardous microbiological research. Open bench tops are used, public access is permitted, and there is no isolation from normal traffic patterns. Good laboratory techniques are required, and mouth pipetting is prohibited. NIH Revised Guidelines, supra note 31, at 60,110. P4 containment, in comparison, is similar to containment required in facilities used for research on potentially serious epidemic viruses such as Lassa Fever and encephalitis. I FINAL EIS, supra note 54, at 52. Only one P4 laboratory has been certified by the NIH, although more may open soon. NIH Memorandum, May 24, 1978 (copy on file at the Ecology Law Quar- terly office). For P4 containment, recombinant materials must be removed in sealed and disinfected containers, while wastes and, if possible, equipment must be sterilized by auto- clave or removed through a pass-through fumigation chamber. Personnel enter and exit from clothing change and shower rooms; showers are mandatory at each exit. Complete laboratory clothing, including undergarments, must be worn. All research must be conducted by means of gloves attached to airtight biological safety cabinets. The P4 facility must at least be clearly isolated within a larger building; all ducts and conduits must be sealed. Airlocks and doubledoor autoclaves are required. Negative air-pressure is man- dated, with exhaust air discharged through high efficiency air filters. NIH Revised Guide- lines, supra note 31, at 60,112-13. Between PI and P4 are two intermediate levels. For a more detailed description, see id. at 60,110-13. 238. Biological controls consist of three safety levels which correspond to the possibility of escape from and the limited life expectancies outside of the laboratory environment. The biological containment levels are designated Host-Vector I (HVI), HV2, and HV3 in increasing order of severity. NIH Revised Guidelines, supra note 31, at 60,113. The more rigorous containment levels require the use of genetically weakened bacteria that have a limited ability to survive and reproduce. E. coi K-12 is assigned the HVI classification; it 19791 RECOMBINANT DNA

A crucial step in the Guidelines is the assignment of physical and biological containment levels to experiments. Experiments are assigned levels of containment that, at least in theory, make their risks acceptable. A listing of classified experiments and their assigned containment levels is beyond the proper scope of this Article. It is possible, however, to outline here the process by which containment levels are determined and to point out changes relating to classification made in the Revised Guidelines. One assumption made in the assignment process is that experiments involving organisms that exchange DNA in nature are relatively harmless. 239 A second assumption is that, in the case of prokaryotic hosts, the risk of an experiment is directly related to the source of the foreign DNA used.240 The Guidelines also rely on the supposition that one step in the level of physical containment is equal to one step in the level of biological containment in terms of risk reduction. For example, the use of purified foreign DNA instead of shotgunning results, under the Guidelines, in a one-step reduction in either the physical or 24 biological containment levels, which are considered interchangeable. 1 The most significant differences between the 1976 Guidelines and the Revised Guidelines are in the area of classification. In response to criticisms from the scientific community and to the new lower assessment of the risks, some classes of experiments have been exempted from the Guidelines and others have been assigned lower containment levels. 242 Under the 1976 Guidelines the NIH had very limited power to change the classification of experiments. The Revised Guidelines also provide much more flexibility in the assignment of containment levels by giving the NIH much broader discretion in making permanent changes for specific experiments without the necessity of issuing

does not colonize in the normal human bowel and exhibits little if any multiplication while passing through. I FINAL EIS, supra note 54, at 53. The HV2 level requires that the bacteria be weakened genetically to ensure that survival outside the laboratory environment will not be greater than 10-8. Id. at 60,114. The HV3 level requires the same standard and, in addition, requires that the host's or vector's inability to survive be confirmed through independent means, including tests on humans or primates. id. The use of E co/i is not necessary, but the other host-vector system must pass the same requirements. For a more detailed description, see id. at 60,113-14. 239. Decision of the Director, NIH, to Release Guidelines for Research on Recombi- nant DNA Molecules, 41 Fed. Reg. 27,908 (1976). 240. Id. 241. I FINAL EIS, supra note 54, at 59. This approximate equivalence assumption is retained in the Revised Guidelines. NIH Revised Guidelines, supra note 31, at 60,119. 242. NIH Revised Guidelines, supra note 31, at 60,108-09. According to the Environ- mental Impact Statement accompanying the Proposed Revised Guidelines, "[ajl experiments exempted from the Guidelines are of minimal speculative risk and present no significant hazard to health or the environment." NIH Proposed Revised Guidelines, supra note 82, at 33,111. ECOLOGY LAW QUARTERLY [Vol. 8:55

new guidelines. 243 The responsibility for administering the Guidelines is divided between the principal researcher, the institution sponsoring the research, and the NIH. The primary responsibility for supervision of actual experiments belongs to the principal investigator. This person must assess potential biohazards, determine and implement appropriate containment, safeguards, and procedures to minimize possible accidents, and inform and train the staff. The investigator's responsibilities also include supervising the safety performance of the staff, investigat- ing and reporting to the NIH any illness resulting from recombinant DNA research or any problems pertaining to the operation of safety practices, and correcting work errors and conditions that may result in 244 the release of recombinant DNA materials. The institution must establish an institutional biohazards committee to review and certify to the NIH that the facilities, procedures, and training of personnel are appropriate, and provide a central reference file of publications for advice on containment, training, and biohazards. The committee must be composed of experts from a broad range of fields related to recombinant DNA research and must be com- petent to determine applicability of legal and professional standards of practice.245

243. The Revised Guidelines allow changes in the containment levels for specific experiments (or the assignment of levels to experiments not explicitly considered) to be approved by the Director of the NIH upon the advice of the Recombinant DNA Advisory Committee. NIH Revised Guidelines, supra note 31, at 60,127. Major revisions require the Director to provide for "an opportunity for public and Federal agency comment." Id. at 60,126. Local Institutional Biohazard Committees (IBCs) are given discretion to lower physical or biological containment requirements one level for experiments with purified DNA and characterized clones. Id. at 60,119. Prior approval of the NIH is required to lower containment levels more than one step under these circumstances. Id. 244. NIH Guidelines, rupra note 87, at 27,920. In almost all cases, NIH grants are made to institutions, placing upon the institution itself all responsibilities of the "principal investigator." Id. The institution can delegate responsibilities to individual researchers, but cannot delegate ultimate responsibility. Id.; RECOMBINANT DNA TECH. BULL., Summer 1977, at IV-4 (NIH publication). The Revised Guidelines provide for few changes in the role of the principal investigator. They do add a requirement that the investigator be "adequately trained in good microbiological techniques." NIH Revised Guidelines, supra note 31, at 60,126. 245. The roles of the institution and the institutional biosafety committee (IBCs) have been expanded under the Revised Guidelines. The institution is required to insure the training of research personnel and determine the need for medical procedures. The institutional biosafety committees have been given the discretion to approve single step reductions in containment levels for experiments with characterized clones and purified DNA. See note 243 supra. The IBCs are required to notify the NIH of such changes. NIH Revised Guide- lines, supra note 31, at 60,119. IBC members engaged in recombinant DNA research projects under review are prohibited from participating in the review of those projects and at least two members of the committee must not be affiliated with the institution. Id. at 60,124- 25. A biological safety officer must be appointed by the institution to insure that safety precautions are followed. Id. at 124. 1979] RECOMBINANT DNA

The responsibilities of the NIH are divided among the Office of the Director, the Recombinant DNA Advisory Committee, the NIH Initial Review Groups, Office of Recombinant DNA Activities, and the NIH staff. Together they revise the Guidelines, make independent evaluations of the risks involved in specific experiments, certify biological containment systems, grant exemptions, and perform on-site facility inspections.246

2. Analysis of the Guidelines 247 a. Criticism of substantivepro visions One major substantive criticism of the NIH Guidelines has been that the actual reduction of risk achieved at each containment level is uncertain. Although the stringent precautions imposed at the higher

246. Some responsibilities are shifted within the NIH in the Revised Guidelines, but the NIH's basic external responsibilities remain unchanged. 43 Fed. Reg. 33,063 (1978). 247. See note 296 infra for a discussion of procedural criticisms of the NIH hearings and decisionmaking process. The Guidelines have also been criticized as violating the procedural requirements of the National Environmental Policy Act (NEPA) § 102(2)(c), 42 U.S.C. § 4332(2)(c) (1976). While the Guidelines were released on June 25, 1976, the draft Environmental Impact Statement (EIS) required by NEPA was not released until August 26, 1976, and the final EIS was not released until October of 1977. The early release of the Guidelines was defended by arguing that they were an improvement over the status quo, since experiments were proceeding at that time. I FINAL EIS, supra note 54, at 2. Contra, II FINAL EIS, supra note 85, at K-131 (comments of Natural Resources Defense Council that until the Guidelines were published a "voluntary moratorium was in effect"). See also Note, Recombinant DNA, supra note 1, at 845-60. Some have argued that the NIH should have prepared an EIS since its grants and contracts were being used to fund recombinant DNA research, a "major Federal action" "significantly affecting the environment." See Chalker & Catz, supra note 5, at 57; II FINAL EIS, supra note 85, at K-134 (comments of Natural Resources Defense Council); id. at K-83 (comments of New York Attorney General Lefkowitz). The adequacy of the EIS has been examined by one federal district court. In May 1977, a suit was brought in the U.S. District Court for the District of Columbia to enjoin a proposed risk-assessment experiment in the Fort Detrick, Maryland P4 containment facility. Mack v. Califano, 447 F. Supp. 668 (D.D.C. 1978). After a nine-month stay to allow finalization of the EIS, the court denied the injunction. On appeal the D.C. Circuit affirmed. Mack v. Califano, No. 78-1156 (Mar. 8, 1978). See 43 Fed. Reg. 33,048 (1978). The District Court opinion cited Kleppe v. Sierra Club, 427 U.S. 390, 410 n.21 (1976), for a deferential standard of review and then held that the EIS "does represent a 'hard look' by NIH at recombinant DNA research performed in accordance with its guidelines... [which] will insure that no recombinant DNA molecules will escape from the carefully controlled laboratory to the environment." 447 F. Supp. at 670. Mack P. Calfano is not determinative of the procedural NEPA issues. First, although the substantive content of the EIS was upheld, the procedural issues were never discussed; it is unclear whether the plaintiff even alleged them. Second, since the case involved a request for a preliminary injunction, the conclusion of law applied differs from that used in cases where no preliminary injunction is requested; the petitioner "failed to demonstrate that he would be irreparably injured in the absence of the issuance of an injunction, that he is likely to prevail upon the merits, .... and that the public interest lies in granting the requested relief." 447 F. Supp. at 671. ECOLOGY LAW QUARTERLY [Vol.V 8:55

levels of containment are more effective in reducing risk than those required at the lower levels, exactly how much risk is reduced is unclear. This fact makes assignment of experiments to the appropriate containment level extremely difficult. As a result, some experiments inevitably will be assigned to levels with excessively stringent containment requirements, while others will be improperly assigned to containment levels which are insufficient for the risks involved. The only alternatives to a multi-leveled containment system, however, are either a more general risk level assignment system or a case- by-case treatment of containment problems. A system relying upon fewer classifications of broader scope would multiply the problems of uncertainty, since fewer alternative classifications would be available. The other possibility-assessing the precautionary requirements for each experiment-would result in considerable delay and bureaucratic red tape. 248 This case-by-case approach might also tend toward relaxed standards. 249 The NIH Guidelines adjust to the uncertainties in assigning containment levels by making the Guidelines stricter than are perhaps necessary so that the inherent margin of error will fall on the side of increased safety.250 Such a safety margin is required in order to 25 1 ensure minimization of the risks. Critics have attacked the physical containment provisions of the Guidelines as ineffective. Although the higher levels of containment may appear too severe, they argue that the possibility still exists that some organisms could escape the laboratory environment. A study of worker infection rates at high security disease study labs has indicated that the strictest physical containment conditions can eliminate infection,252 but opponents argue that the study is not conclusive evidence of

248. See I FINAL EIS, supra note 54, at 55. 249. A case-by-case approach in Great Britain has resulted in less stringent standards than in the United States. II FINAL EIS, supra note 85, at K-I I (letter by J. Melnick). 250. A draft of the Proposed Revised Guidelines contained the statement that "the [Rievised [Gluidelines [have] the intent of erring on the side of caution." This language was omitted in the Proposed Revised Guidelines since "scientists should not enter into an activity with the intent of erring." 43 Fed. Reg. 33,049 (1978). 251. Theoretically, this margin of safety due to uncertainty should be incorporated in the benefit-risk ratio. See text accompanying notes 115-34 supra and text accompanying notes 257-61 infra. If implemented at the experiment stage, it accounts only for uncertainty in the risks because the safety requirements for an individual experiment cannot and do not reflect expected benefitfrom that experiment. If by mistake the margin of safety for uncertainty is incorporated in both the benefit-risk ratio and the experiment stage, the control would be excessive and the benefits reduced. 252. Wedum, The Detrick Experience as a Guide to the ProbableEfficacy of P4 Microbio- logical Containment Faclities for Studies on Microbial Recombinant DNA Molecules, 1976 HEW DOCUMENTS, supra note 1, at 372, 391-92. This study by Dr. A. Wedum concluded that when the NIH P4 containment criteria were met, infection would be prevented, but that P3 containment without gastight biological safety cabinets would probably result in infection. Human failure was cited as the major reason. A broader study encompassing 3,921 laboratory-acquired disease cases concluded that 19791 RECOMBINANT DNA the adequacy of physical containment. 25 3 Nevertheless, all available studies of containment methods involved laboratories in which physical precautions were the only safeguards used. The NIH Guidelines prescribe a safeguard system which relies not only on physical safeguards, but also biological safeguards; each is able to provide contain- 254 ment of experimental dangers. Finally, the substantive provisions of the Guidelines have been criticized for improperly classifying many recombinant DNA experiments. For example, the procedure for classification assumes that "most cells with foreign DNA from higher organisms are more hazardous than those with DNA from lower organisms. '255 But since there are many different factors that determine whether a particular organism is capable of causing harm,256 this generally valid assumption may in specific cases prove wrong. On the other hand, many scientists have argued that experiments are classified more stringently than necessary. 257 Since the field of recombinant DNA research is changing rapidly and risk assessments have been modified downward,258 researchers suggest that the classifications of many experiments should be relaxed. They claim that, as a result of overly stringent classification, the benefits of recombinant DNA research have been significantly impeded because of increased with upgraded safety precautions and modem facilities, physical containment is effective. 1976 HEW DOCUMENTS, supra note 1,at 233. See also REP'T ON GENETIC ENG'R, supra note 22, at 40; Kissling, LaboratoryAcquired Infections, in BIOHAZARDS IN BIOLOGICAL RE- SEARCH 70 (A. Hellman, M. Oxman & R. Pollack eds. 1973). 253. 1976 HEW DOCUMENTS, supra note 1,at 289-90 (the Wedum study underestimated by a factor of ten); II FINAL EIS, supra note 85, at K-7 (Dr. Wedum's favorable data due to vaccination of laboratory personnel, a technique which could not be used successfully for recombinant DNA experimentation.). 254. One should also note that the Revised Guidelines have tightened physical containment provisions in some respects. See Notice of Release of Revised NIH Guidelines for Recombinant DNA Research, 43 Fed. Reg. 60,084-85 (1978). Critics have also contended that the research should be discontinued until a host safer than E col can be found. For a rebuttal to this argument, see text accompanying notes 83-90 supra. 255. II FINAL EIS, supra note 85, at K-148 (comments of Natural Resources Defense Council). Therefore, more containment is required for experiments that take foreign DNA from primates and insert them into a host than for those that take their foreign DNA from plants. The assumption is based on the probability that (1) the product of a foreign gene would more likely be harmful if the gene were similar (homologous) to human genes, and (2) the chance of incorporating pathogenic viral DNAs able to attack human tissue increases with higher phylogenetic orders of eukaryotes as the source for foreign DNA. I FINAL EIS, supra note 54, at 57-58; NIH Guidelines, supra note 87, at 27,908. See text accompanying notes 78-82 supra. 256. 41 Fed. Reg. 38,438 (1976). 257. For example, the Guidelines have been criticized for setting containment levels stricter than necessary for recombinant DNA work on viruses and plants. Introduction to the Proposed Revised Guidelines, 43 Fed. Reg. 3,043 (1978). See notes 259-60 infra. 258. For a discussion of recent studies in risk evaluation, see text accompanying notes 85-86, 133 supra. ECOLOGY LAW Q UARTERL Y [Vol. 8:55 expense and delays due to a lack of facilities. 259 Since European standards are considerably more lax than their American counterparts some American researchers have gone so far as to move their laboratories 0 abroad.26 Although the safety margin in experiment classifications is a necessary means to minimize risks,261 the margin should not be so large as to unnecessarily limit the benefits derivable from experimentation. Classifications should not be locked into outdated risk assessments; rather they should be revised periodically to reflect changes in risk assessment.

259. Dr. Michael Bishop of the University of California San Francisco Medical Center (UCSF) claims that with only one P4 lab certified currently research on viral transformation of normal animal cells into cancer cells is "virtually stymied" by the Guidelines. San Fran- cisco Chronicle/Examiner, Oct. 29, 1978, at 1, col. 3 (combined Sunday ed.). 260. The NIH Guidelines have "forced many researchers to work in Europe where guidelines are less strict." San Francisco Chronicle, Jan. 15, 1979, at 4, col. 5. Dr. William Rutter of UCSF gave his reasons for choosing to perform some experiments under less stringent conditions in France: "We had waited more than one year [to use the certified P4 facility in the U.S.]. . . . By then, the necessity of getting on with the work led us to try to seek to do them elsewhere." San Francisco Chronicle/Examiner, Oct. 29, 1978, at 1, col. I (combined Sunday ed.). Two important questions are raised by this development. The first is whether it is proper to use NIH funding for experiments outside the United States that do not comply with the NIH guidelines. Dr. Bernard Talbot, an aide to NIH Director Fredrickson has stated that "I don't think there's anything wrong" with American scientists performing experiments under less stringent standards in Europe. Id. The NIH Policies and Procedures accompanying the Proposed Revised Guidelines would allow such experimentation as long as the experiment is conducted in a country with guidelines and in compliance with those guidelines. 43 Fed. Reg. 33,093 (1978). Though research in Europe clearly does not involve as great a risk to the American public as research within the United States, it is a questiona- ble practice to fund experiments abroad that would not be funded in the United States. The problem raised by American experimentation in Europe is whether the Guidelines are in fact too stringent. Despite the continuation of research in Europe under less stringent standards than in the United States, there is no evidence of harm as a result of recombinant DNA research. 43 Fed. Reg. 33,044 (1978). Some have argued that standards adopted in Europe are "irrelevant to our public health process." Jeremy Stone, Director of Federation of American Scientists, Letter to Council Members, as printed in II FINAL EIS, supra note 85, at K-44. There is substantial commentary suggesting that the successful operation of less stringent European standards, however, is at least some evidence that the 1976 Guidelines were too stringent when adopted. See generally Zilinskas, Recombinant DNA Research and the International System, Recombinant DNA4 Research Symposium, supra note 5, at 1483; Comment, Genetic Manipulation" Research Regulation and Legal Liability Under Interna- tional Law, 7 CAL. W. INT'L L. J. 203 (1977); Comment, The PotentialforGenetic Engineer- ing." A Proposalfor International Control, 16 VA. INT'L L. J. 403 (1976); Kamely & Curtin, supra note 5, at 28; Statement on the Security of the Recombinant DNA Researches in Japan and French Guidelines on Recombinant DNA Experiments, RECOMBINANT DNA TECH. BULL., Winter 1978, at 17, 18; [West German] Guidelinesfor the Safe Handling of Recombi- nantNucleicAcid (DNA), RECOMBINANT DNA TECH. BULL., Spring 1978, at 11; and Report of the Working Party on the Practice of Genetic Manipulation (Cmnd. 6600, Aug. 1976) (Brit- ish Guidelines). 261. See text accompanying notes 115-34 supra. 1979] RECOMBINANT DNA b. Criticism of the Guidelines as incomplete The Guidelines have been criticized as incomplete since they fail to consider risks not related to health. For example, there is no discussion of the possible long-range effects of gene alteration. 262 It has also been charged that the Guidelines ignore the possible abuses of gene- splicing research. Specifically, the Guidelines ignore the ethical consequences of human genetic engineering and the potential use of recom- 263 binant DNA technology for destructive purposes by terrorists. These criticisms are unfounded. The risks of evolutionary conse- 264 quences of recombinant DNA research are indefinite and remote. The Guidelines do place greater restrictions on experiments involving organisms not known to exchange DNA in nature than on those experiments that do not; to do more is unnecessary. 265 Potential abuses of recombinant DNA research are equally unlikely. Current recombinant DNA research is not directed at human gene alteration and will not be in the near future.266 Moreover, experimental guidelines could hardly deter terrorists from using gene-splicing technology.267 Provisions for dealing with such possibilities are beyond the proper scope of the Guidelines. The Guidelines have also been criticized because they focus on the regulation of research techniques, not the products of recombinant DNA experimentation. It is possible, as these critics point out, to create gene-splicing products through conventional microbiological methods. 268 In fact, the use of one such method recently allowed construction of a new life-form. 269 That this new life-form was not a result of recombinant techniques in no way reduces the danger that it might prove to be pathogenic. The most serious risks of the experimentation are related to research techniques, however, and not to the ultimate products. The risks of "shotgun" experiments center not upon the expected result of the experiment but upon the unexpected harmful by- products that are associated with the technique. It may be that recom-

262. For an example, see II FINAL EIS, supra note 85, at K-6 (comment by R. Sin- sheimer). 263. See id. at K-100 (comment by L. Lefkowitz). 264. I FINAL EIS, supra note 54, at 102, 118; see text accompanying notes 104-11 supra. 265. "[Where the donor exchanges DNA with E. coli in nature, it is unlikely that recombinant experiments will create new genetic combinations never tested by nature." The containment requirements are greater where donors are not known to exchange DNA with E. coi in nature because "there is a greater potential for new genetic combinations to be formed and expressed." I FINAL EIS, supra note 54, at 58. 266. Id. at 119. See text accompanying note 114 supra. 267. Id. See note 113 supra. 268. But see text accompanying notes 49-50 supra. 269. Application of Chakrabarty,discussed in note 149 supra,involved General Electric's development, though not perfection, of an oil-eating bacteria using classical genetic techniques, not recombinant DNA techniques. ECOLOGY LAW Q UARTERL Y [Vol. 8:55 binant DNA products should at some time be regulated, but the present regulation of such products is unwarranted. One of the greatest problems with the NIH Guidelines is that they do not provide for adequate enforcement mechanisms. The Guidelines rely primarily on the goodwill of researchers for enforcement. 270 How- ever, this faith in researchers is apparently not justified. A report of laboratory techniques at the University of California, San Francisco indicates a cavalier attitude toward following the NIH Guidelines.271 It has been reported that half the researchers in one lab "seemed to care little. . . and among the young graduate students and post doctorates it seemed almost chic not to know the NIH rules." 272 This lack of concern by some for physical procedures is exacerbated by a change in the personnel employed in conducting DNA research. 273 Although this attitude does not necessarily indicate a defect in the substantive provisions of the Guidelines, it does indicate an enforcement problem. Assuming the Guidelines are a sufficient barrier to risk, they will not be effective if not followed. 274 Furthermore, the Guidelines do not provide realistic standards for the regulation of all areas of recombinant DNA research. As pointed out above the present NIH Guidelines cannot control research that does not receive NIH funding. Even if industry wishes to comply, further difficulties may remain because of a basic conflict between the

270. "We can't run a policing service ..... [We] try to raise the consciousness of the individuals involved and make them respect the guidelines." Wade, Recombinant DNA: NIH Rules Broken in Insulin Gene Project, 197 SCIENCE 1343 (1977) (statement of D. Mar- tin, former Chairman, UCSF Biosafety Committee). However, some fear that researchers will "fabricate" safety precautions since there is no enforcement mechanism to audit or spot-check the workers. Rumors exist that such fabrication has occurred at UCSF, but they are unproven. Id. 271. Id. at 1342. This attitude is caused in part by there having been so far no adverse consequences from improper microbiological techniques. 1976 HEW Documents, supra note 1,at 325 (testimony of D. Handler). 272. Id. However, Dr. Charles Boyer, the principal investigator at UCSF, denies these allegations and claims the NIH rules are taken seriously. 273. One researcher has noted an enormous influx into DNA research of people who lack a rigorous microbiological background, yet exhibit "an almost contemptible familiarity with E. coIl."1976 HEW DOCUMENTS, supra note 1, at 236. This change from the "elite • . .may increase the risk to a higher degree than has been estimated on the basis of research accomplished thus far." REP'T ON GENETIC ENG'R, supra note 77, at 39. The increased risk stems from lack of experience. 1976 HEW DOCUMENTS, supra note 1, at 391 (use of trained personnel decreases risk of operator exposure). Experts such as Dr. Roy Curtiss realize the result is "that inexperience and careless personnel can so easily lead to accidents." REP'T ON GENETIC ENG'R, supra note 22, at 41. 274. See REP'T ON GENETIC ENG'R, supra note 22, at 40. Examples of failure to follow safety guidelines in scientific research abound. See San Francisco Chronicle, June 21, 1978, at 2, col. 2 (communicable diseases research); Oversight Hearing, supra note 54, at 28 (re- printing a Staff Report on the Carcinogenesis Program of the National Cancer Institute, by the Senate Subcommittee on Administrative Practice and Procedure, of the Senate Judiciary Committee) (cancer research); J. GOODFIELD supra note 11, at 117 (radioisotope research). 19791 RECOMBINANT DNA objectives of industrial research and some of the substantive provisions of the NIH controls. If industry does not comply with the Guidelines, the entire purpose of the regulatory scheme is destroyed. Since 75% of recombinant DNA research is privately conducted, 275 any regulatory scheme that failed to reach privately funded research would not reach the source of most of the risks. As the amount of privately funded research increases, the Revised Guidelines must be extended to cover all recombinant DNA research. Moreover, the Guidelines' focus on controlling academic research is in some regards incompatible with commercially oriented research. For example, the Guidelines prohibit releasing recombinant DNA products into the environment. 27 6 Applied to industry without qualification, this rule could deter much private research. An unqualified application of this provision would also limit the possible benefits of DNA research to society, changing the risk-benefit balance. Eliminat- ing such benefits is not justified without a showing that a permanent prohibition on the release of gene-splitting products into the environment is the best way to minimize risks. Finally, the ten liter limit on recombinant DNA products 277 would also be difficult for private industry. While consideration must be given to the problems associated with the creation and use of large quantities of such products, many potential industrial applications of gene-splicing techniques require quantities of recombinant DNA products in excess of the current limit. If the industrial uses of recombinant DNA technology are seriously contemplated, the ten liter limit is an impractical stricture. Proponents of the Guidelines argue that both of the above limitations are presently sensible, considering the state of scientific knowledge. They argue that until the manufacture of recombinant DNA products becomes a realistic possibility, the Guidelines should not reflect production considerations, especially since the sophistication to assess standards for regulating large scale experiments and the release of recombinant DNA products into the environment does not yet exist. To relax standards too quickly, they say, would encourage industry to take unacceptable risks which would, in turn, create the possibility of environmental disaster. However, although widescale use of recombinant DNA products

275. See note 135 supra. 276. Although such experiments are banned in the 1976 Guidelines, the Revised Guide- lines allow them where expressly approved by the Director of the NIH, with the advice of the Recombinant DNA Advisory Committee and after appropriate notice and opportunity for public comment. 43 Fed. Reg. 60,080, 60,108 (1978). However, specific criteria for decisionmaking are not discussed. 277. The Revised Guidelines do provide for exceptions to this prohibition. See notes 236, 276 supra. See also NIH Revised Guidelines, supra note 31, at 60,083. ECOLOGY L4W QU4RTERLY [Vol. 8:55

for practical purposes may not be possible at the present time, the Guidelines should reflect that future contingency. Assuming the Guidelines are sufficient for the needs of academic researchers, they should reflect a more practical orientation if they are to be applied to private industry as well. Specifically, the Guidelines should set out more detailed conditions for performance of large-scale experimentation and should provide a mechanism whereby recombinant DNA products may be released into the environment as long as rigorous testing, certification, and safety standards are met. Such forward-looking regulations could meet sudden technological advancements making manufacture of products imminent. From the utilitarian point of view, as long as the Guidelines remain limited in scope and are not made mandatory upon all researchers, it is even more important to maintain realistic standards for all segments of the research. Since industrial compliance is voluntary, the Guidelines must be consistent with the basic purposes of industry. To maintain rigorous standards that are difficult for private industry to meet will only encourage industry to abandon the Guidelines alto- gether. Under the present system the formulation of regulations should balance the interest of safety with the interest of gaining at least a reasonable amount of voluntary compliance with whatever regime is adopted.

B The Role of Slate and Local Governments in Regulation Federal, state, or local government agencies could regulate recombinant DNA research by enforcing a modified version of the NIH Guidelines. This section demonstrates that neither state nor local governments should regulate to the exclusion of the federal government. Instead, the development of new technologies should be federally controlled with allowance for additional state and local regulation consistent with federal policy.

. Exclusive Local or Exclusive State Regulation a. Local regulation Dissatisfaction with the NIH Guidelines shortly after their passage led various states and localities to consider taking their own action to control recombinant DNA research. Local governments worried about the inapplicability of the Guidelines to privately funded research and the lack of enforcement provisions.278

278. Norman, City, State and Congress at Odds on DNA Research, TECH. REV., May 1977, at 6. Several local governments have enacted or considered enacting ordinances regulating recombinant DNA research. Cambridge, Massachusetts, established the Citizen's Experi- 19791 RECOMBINANT DNA

Exclusively local regulation of recombinant DNA research has some advantages, primarily an opportunity for direct public input into the regulatory scheme since local government is highly accessible to community input. This benefit could be lost in a federal regulatory scheme where public involvement is much more difficult. Notices of public federal hearings do not reach most individuals, and hearings are frequently held far away and during working hours. Another advantage of local regulation of recombinant DNA research is that the group most likely to suffer from any adverse effects would control the research. Those communities wishing to risk possible adverse effects in order to gain scientific and economic benefits for themselves would be allowed to do so. On the other hand, communities not willing to bear the risks involved would not be forced to do so by national policy. Despite these possible benefits, local regulation is beset by a number of serious problems. Although the NIH Guidelines are a starting point for a regulatory scheme, as scientific knowledge increases these standards must be revised. If local governments are the exclusive entities for administering regulations, it is difficult to see how they could also expend the resources for annual revision. In addition, this continual reevaluation would be extremely wasteful as numerous localities would be carrying on essentially the same task. At least with regard to revising the standards imposed, "pooling" of resources on the national level seems desirable. Similarly, local governments cannot af- ford to maintain the expert personnel required for an efficient monitoring and enforcement scheme. Monitoring and enforcement only seems mentation Review Board (CERB) to hold hearings and recommend action by the city council. Although none of the eight citizens were involved in recombinant DNA research and most were not involved in the biological sciences, extensive hearings on the scientific and ethical issues were held. CERB concluded that the research should continue and recorn- mended that the NIH Guidelines, with additional but minor restrictions, be applied to recombinant DNA research in Cambridge. Wade, Gene-Splicing Cambridge Citizens OK Research But Want More Safety, 195 SCIENCE 268, 269 (1977) [hereinafter cited as Wade, Cambridge Citizens]. The CERB decision demonstrated that nonexperts could participate in the decisionmaking without irrationally prohibiting all recombinant DNA research as some had feared. At the University of Michigan, two committees were formed to examine the research. One, composed of scientists and experts, assessed the objective risks and benefits. The second, composed of nonscientists, examined the ethical and moral questions. Id. at 268. This bifurcation of roles, clearly separating the quantitative assessment of the risk from the value question of "acceptability of risk" was unique, as demonstrated by the NIH's failure to do the same during its formulation of the Guidelines. Berkeley, California, although having no power to control the University of California, passed an ordinance requiring private research to conform with the NIH Guidelines. Berke- ley Gazette, Sept. 14, 1977, at 1, col. 5. For detailed discussion of local regulation in other university towns, see 1977 Senate Hearing,supra note 22, at 72; Wade, Hundred Flowers, spra note 130, at 558 (1977). ECOLOGY LAW QUARTERLY [Vol. 8:55

practical on a state or national level where the different jurisdictions can share the costs of maintaining full time experts. An attempt to cut costs by the use of volunteer committees is problematic since members of local communities are unable to devote enough time to a long term regulatory scheme. The time consumed by the controversy in committee meetings alone already has been im- mense. 279 This cost has, at times, led to an endorsement of scientific recommendations without due consideration of alternatives. 280 Even those communities that have expressed the willingness to devote the time to devise a regulatory scheme cannot be expected to continue to do so. Nonetheless, a continuing decisionmaking process is crucial to a recombinant DNA research regulatory scheme. b. State regulation For the most part, state control of recombinant DNA research has not progressed beyond the stage of preliminary consideration. The California legislature held hearings on the subject and had a bill in committee awaiting the outcome of federal legislation. New York also drafted legislation and New Jersey has considered possible state controls.28' The general feeling among state legislatures has been to "wait and see." Although a number of states have expressed an interest in regulating recombinant DNA research, they have been unwilling to act in view of possible federal legislation. This failure may be a result of hesitance to become involved in controlling research, or concern that state legislation may be preempted by later federal law. The only state that, to date, has enacted comprehensive legislation regulating recombinant DNA research is Maryland. 282 The Maryland law, which went into effect on July 1, 1977, provides for the licensing of all recombinant DNA research within the state and requires that the NIH Guidelines, as revised, be followed in such research. 283 Failure to conform with these standards can result in revocation of the license. Further legal remedies include injunctions and fines for conducting re-

279. One scientist serving on an Ann Arbor committee was forced to give up his own research for a year to participate. Wade, Cambridge Citizens, supra note 278, at 269 (refer- ring to Robert Helling, the chairman). 280. For example, it seems obvious that with only one meeting to consider the subject, Indiana University at Bloomington did little more than rubberstamp the NIH Guidelines. See Wade, Hundred flowers, supra note 130, at 559. 281. Id. at 558-59. 282. MD. ANN. CODE, art. 43, §§ 888-910 (Supp. 1978). See also 1977 Md. Laws, ch. 847, §§ 892-904, at 3307-11. 283. The law also allows the State Secretary of Health and Mental Hygiene to promulgate "such regulations as deemed necessary for implementation" of the Act. 1977 Md. Laws § 903. 19791 RECOMBINANT DNA search without a license.284 Researchers may be held liable for damage caused by their research. 285 The State Secretary of Health and Mental Hygiene and a biohazards committee appointed by the Secretary monitor the laboratories and enforce the statutes.286 The law contains a "sunset" clause; it is automatically abrogated after five years unless re- newed by the legislature and is repealed immediately in the event of 287 passage of federal legislation. The principal advantage of state regulation of recombinant DNA research is that it provides a middle ground between local and federal regulation. Promulgation of standards, monitoring, and enforcement can be accomplished more readily and efficiently than on a local level simply because the state has more resources. This ability is illustrated by the Maryland law, which is more comprehensive than any local regulation. At the same time, a state may be more sensitive than the fed- 288 eral government to local needs. However, state regulation cannot be relied upon as the sole means for regulating recombinant DNA research. Few states will act in the near future to regulate research. Even if many do, some will always remain that have no restrictions on recombinant DNA research. If the risks and benefits of recombinant DNA research were purely a local problem, this solution might be acceptable; states would take risks in exchange for the benefits of having research institutions located within their boundaries.28 9 Unfortunately this hypothetical situation is not the case. A law protecting members of a particular state from risk will not protect them from the dangers of unrestricted research conducted in neighboring states. Additionally, since the benefits of the research are in fact not localized but will accrue to society as a whole, a state could acquire the benefits and avoid most risks by banning the research within its borders, an inequitable action.

2. Combined Regulation A combination of federal, state, and local regulation would pro-

284. Fines are "not to exceed $1000 for each violation; each day such a violation continues shall constitute a separate violation." Id. § 902(c). 285. An earlier draft held researchers strictly liable for damages, but this wording was omitted from the final version. Id. § 894. 286. The committee has the power to review all proposals for research and to "conduct periodic site visits to facilities where research is being carried out to determine compliance with this act." Id. §§ 901, 902. 287. Id. § 904. 288. These state efficiencies are not necessarily available. Public input can have more effect on the state government than on the federal government, but so can lobbyists for industrial researchers. Cf. note 299 infa. 289. It is likely that research institutions would gravitate to states that did not have controls on recombinant DNA research. See note 260 supra (migration of American scientists to Europe to take advantage of less stringent research guidelines). ECOL OGY LAW Q UARTERL Y [Vol. 8:55 vide for national standards and enforcement but still allow local input into the nature of controls as applied in each jurisdiction. The most likely approach is that of defining minimum federal standards while allowing state and local governments the option of making the regulations more stringent. 290 Creating minimum federal standards would limit variations in local regulation since the federal government would regulate all research at a uniform level in the absence of local regulations. This minimum level would prevent the prospect of unrestricted research in states and localities choosing not to regulate recombinant DNA research. It would also provide a starting place for state and local regulation. One of the reasons that state and local regulation has not been onerous is that the NIH Guidelines have served as a starting point in negotiations between state and local governments and the scientific community. An outer limit should also be placed on state and local regulation of recombinant DNA research. Recombinant DNA research is a beneficial activity and a total ban is not warranted.29 1 Furthermore, to allow state and local governments the authority to ban research entirely would be to grant them too much bargaining power in negotiations with the scientific community. Local regulation plans have succeeded because of balanced adversarial processes that resulted in reasonable compromises. There is no guarantee this success will always be re- peated. Local governments might attempt to exact a price for allowing research to continue-such an action would not be equitable. The power of state and local governments to control research should be limited to reasonable additions to national standards. At least two standards for a ceiling on state and local regulation of recombinant DNA research have been proposed. One would allow state and local regulation if the standards imposed were necessary for the protection of health and environmental conditions. 292 This standard places too great a burden on state and local governments. It would be difficult to show that state and local regulation is "necessary"; federal regulation would supposedly reduce risks to acceptable levels. Regulation probably could not be justified on the basis of local conditions because, as the National Academy of Sciences has pointed out, there is no indication that the risks of recombinant DNA research are

290. Some politicians have advocated this approach. H. REP. No. 1005, 95th Cong., 2d Sess. (1978) (dissenting views of Reps. Waxman, Mikulski, Maguire, Markey, Marks & Ot- tinger). See also 1977 Senate Hearings, supra note 22, at 69-70 (Sen. Kennedy & Gov. Dukakis of Massachusetts). 291. See text accompanying notes 134-35, 289 supra. 292. HEW may approve a state or local regulation plan if "(1) the requirement is the same as, or more stringent then, a requirement under section 102(a); and (2) the requirement is necessary to protect health or the environment." H.R. 11192, 95th Cong., 2d Sess. § 106(b) (1978). 19791 RECOMBINANT DNA greater in one geographical area than in another. 293 Another proposal is to allow state and local control where relevant and material to health and the environment. 294 This standard is advan- tageous in that it permits reasonable regulation and requires states and localities to justify regulations in terms of preventing real risks. How- ever, this standard may give too much discretion to localities by letting marginal health or environmental risks justify overly stringent regulation. Both preceding standards may allow action inconsistent with federal policy and may also prevent action not conflicting with federal regulation. The optimum standard is one that allows states and localities to exert control over recombinant DNA research within limits consistent with the purposes and goals of the federal regulatory scheme. 295 In this way, the advantages of local participation and control are not lost; yet the federal scheme is protected from excessive or insufficient state regulation.

C Other Considerations I Choosing a FederalAgency to Regulate The power to promulgate standards and supervise their enforcement must be vested in a regulatory body other than the NIH. The NIH has an inherent "institutional" conflict of interest in attempting to both promote and regulate recombinant DNA research.2 96 This con-

293. S. REP. No. 359, 95th Cong., 1st Sess. 59-60 (1977). 294. S. 1217, 95th Cong., 1st Sess. §§ 1813(b)(1) (1977). 1978 HEW DOCUMENTS, supra note 1, at 713-14. 295. Cf Hines v. Davidowitz, 312 U.S. 52 (1941), striking down a state law that "stands as an obstacle to the accomplishment and execuction of the full purposes and objectives of Congress." Id. at 67. Accord, Jones v. Rath Packing Co., 430 U.S. 519, 526 (1977). This test was elaborated by Chief Justice Warren in Pennsylvania v. Nelson, 350 U.S. 497 (1956), who put forward three criteria: "I. pervasiveness of the federal regulatory scheme; 2. federal occupation of the field as necessitated by the need for national uniformity; 3. danger of conflict between state laws and the administration of the federal program." J. NOWAK, R. ROTUNDA & J. YOUNG, HANDBOOK ON CONSTITUTIONAL LAW 268 (1978). If the standard is not articulated in the federal regulatory statute, however, the present Supreme Court will not presume or infer intent. Id. at 270. For discussions of federal preemption doctrine and particular applications to environmental problems, see generally Tribe, Calfornia Declines the Nuclear Gamble.- Is Such a State Choice Preempted?, 7 ECOLOGY L. Q. 679 (1979); Lucas, supra note 20, at 917. 296. The procedure by which NIH adopted the Guidelines was heavily criticized for reasons stemming from the promotional-regulatory conffict. The committee "did not weigh the need to continue research and development-possibly an equally significant factor [as was devising methods of containment) in the risk equation." Letter from Susan Wright, science historian at the University of Michigan, 195 SCIENCE 132 (1977). According to Prof. Wright, the committee "was operating under the assumption that the research would go on, no matter what. In other words, the acceptability was prejudged." 1976 HEW DOCUMENTS, supra note 1, at 267-68. The committee never considered such alternatives as imposition of a further moratorium on the research or a complete prohibition of experiments. ECOLOGY L4W QUARTERLY [Vol. 8:55 flict is similar to that of the original Atomic Energy Commission, which had dual duties of regulation and promotion of peaceful uses of atomic energy. The result was that the AEC promoted atomic energy and ignored its regulatory functions. Congress eventually split it into the Nu- clear Regulatory Commission, which regulates the use of atomic energy, and the Energy Research and Development Agency (now en- compassed within the Department of Energy), which promotes atomic power. 297 The division of functions, leaving promotion to NIH but vesting regulation of gene-splicing research in another agency, has at- tracted support. 298 An independent regulatory body could be constructed either as an appointed commission or as an independent subdivision of a federal agency.299 The responsibility for creating, peri-

The committee's assumption that the research should continue was not surprising since "[tihere was participation of only very modest numbers of persons outside molecular biology," Oversight Hearing,supra note 54, at 82 (testimony of H. Holman); REP'T ON GENETIC ENG'R, supra note 22, at 21, 257; and "many members... [of] the whole biological community .. simply do not believe a hazard exists." 1976 HEW DOCUMENTS, supra note i, at 444 (testimony of R. Sinsheimer). Indeed, some of the committee members were actually conducting recombinant DNA experiments at the time the Guidelines were adopted. J. GOODFIELD, supra note 11, at 125-26; see also Bennet & Gurin, supra note 138, at 138. A bias also was reportedly present against critics of this view. Those who testified in favor of recombinant DNA research had their travel expenses paid by their grants or their universities while the critics had to pay their own way. Oversight Hearing,supra note 54, at 157 (testimony of S. Garb, Citizens' Comm. for the Conquest of Cancer). Thus, as stated by Prof. J. King, a prominent critic and member of Science for the People, "the function of the NIH Committee was to protect geneticists, not the public." ATLANTIC MONTHLY, Feb. 1977, at 156. The adoption process has also been criticized because of a lack of public participation in the proceedings. See Oversight Hearing, supra note 54, at 85-86. See also note 247 supra. 297. Pub. L. No. 93-428, 88 Stat. 246 (codified in scattered sections of 5 U.S.C., 42 U.S.C.). See particularly 42 U.S.C. § 5814(a) (1976). Cf Note, The Concorde Calculus, supra note 216, at 1063 (examining conflicting cost-benefit analyses biased by promotional role of institution). 298. See comments of R. Sinsheimer, cited in Scientists' Institute for Public Informa- tion, Congressional Seminar of Recombinant DNA Research, 21 (Dec. 14, 1976). See also dissenting views of Representatives Waxman, et al., H.R. REP. No. 1005, 95th Cong., 2d Sess. 39 (1978) (NIH regulation of recombinant DNA research is analogous to the former AEC's regulation of atomic energy). 299. One possibility is to create a subdivision of the EPA, which has considerable expertise in enforcement. The independence of a regulatory agency, however, does not guarantee an absence of bias. University of Chicago economist Milton Friedman believes that "[E]very regulatory commission you can name, whatever may have been its original origin, has sooner or later been taken over by the industry it is supposed to regulate and largely conducted in the interest of the industry." Friedman, Special Interests and the Law, 51 CI. BAR REC. 434, 436 (1970). See also Coase, Economists and Public Policy, REASON, Dec. 1974, at 22-23. However, the acknowledged free-market bias of these members of the Chi- cago School must be noted as well. Others view the agency's frequent sympathy toward the regulated industry as off-setting the staff's otherwise largely uncontrolled interest in the farthest reasonable reach of their regulatory power. As these defects to perfect government will ordinarily op- erate in opposmg directions, .... we have better government than if either were absent while the other remains. W. GELLHORN & C. BYsE, ADMINISTRATIVE LAW 22-28 (1974) (citing W. Gardner). See 19791 RECOMBINANT DNA

odically revising,300 and enforcing regulations based on the NIH Guidelines should be given to the new agency. The NIH would retain responsibilities for allocating funds for scientific research on a merit basis but would no longer have the responsibility for research standards. The independent regulatory body would have no power to allo- cate funds on the basis of scientific merit, but would have the power to ''veto" the actual allocation of funds to those institutions not following power would maintain "fund cut-offs" as a regu- regulations. This veto 30 latory tool without dividing regulatory powers between two agencies. ' The new agency could also enforce regulations through the licensing of private laboratories conditioned on compliance. Civil penalties and injunctions could be used as enforcement tools in the case of noncompli- ance.

2. Defining the Public's Role in the Regulatory Scheme Assuming that a federal agency should control gene-splicing research based on the NIH Guidelines, with concommitant, more stringent, but limited regulation by state and/or local governments, there remains a problem of public participation in the scheme. Regulation of large collective risks is simply not adequate without input from constit- uents. Though society's interest in the regulation of scientific research is substantial, difficulties remain in defining the proper role of the public in a regulatory scheme. Due to a lack of information and expertise, it is hard for members of the general public to efficiently and intelli- gently make decisions concerning scientific risks.3o2 Public regulation of scientific research is difficult because information required for intelligent decisionmaking is highly technical and largely inaccessible to the nonscientist. As the earlier discussion of mo- also Marcus, The DisproportionatePower of Environmentalists, 2 HARV. ENVT'L L. REV. 582

(1977). 300. NIH Director Fredrickson has stated that "the [G]uidelines must remain flexible. It is especially important that there be opportunity to change them quickly, based on new information relating to scientific evidence, potential risks, or safety aspects of the research program." 41 Fed. Reg. 27,905 (1976). 301. For a discussion of the problems of more than one agency regulating an activity, see text accompanying note 216 supra. 302. Cf the conclusion of Richard Berger, arguing that the public should make the decisions on the recombinant DNA controversy: The ability of laymen to understand and judge the merits of issues when presented with arguments from all sides is the basis of the jury system. Indeed, legislators have always made judgments upon technical issues when legislating on behalf of public safety. Berger, supra note 5, at 96-97. This theory does not contradict the point made above, for although the public can decide some issues and a general policy adequately, it cannot solve technical and scientific problems. In the same way, the legislature often gives technical problems to regulatory agencies that work within broad policy and legal parameters set by the legislature. ECOLOGY LAW QUUARTERLY [Vol. 8:55 lecular biology and risk benefit analysis indicates, an understanding of the problems of scientific research requires more than a pedestrian interest in the ultimate goal of risk minimization. Even if all the necessary information could be made available to members of the general public at less than prohibitive cost, few laymen could or would assimi- late this data sufficiently to make informed decisions. The only way for the public to receive inexpensively information on new technology development is through the popular media. But informed decisions cannot be made by reliance on information presented in this manner. In order for the media to remain "popular," they must limit and simplify the coverage given to technical problems. Since the ultimate goal of the media is profit, stories that maximize bizarre risks sell more "copy" than those that explain problems in a calm and realis- 3 tic manner.30 Another related problem is that both the media and the general public are largely dependent upon the scientific community for information regarding the risks of experimentation. Even a scientifically sophisticated decisionmaking body must depend on those actually involved in the experimentation to provide information and identify risks.3 4 An important factor in formulating a direct regulation scheme, therefore, is the enhancement of an atmosphere of of cooperation between the regulatory body and the scientific community-to encourage the risk-spotting role of the scientist. A regulatory scheme that creates an adversary relationship between regulator and scientist may only ex- acerbate a fear of repression in the scientific community, defeating the 30 5 purpose of regulation. A resolution to this conflict between the competing goals of protection of the public interest and formulation of a workable system exists in a regulatory plan that provides public input without directly involving the populace in the decisionmaking process. Such a plan would not preclude members of the general public from playing a role in regulation, but would ensure that decisionmakers represented a broad range of opinion and expertise, including that of the scientific community. 30 6 This range could be attained by requiring that the agency personnel

303. W. LOWRANCE, OF ACCEPTABLE RISK 118 (1976). See McGill, supra note 5, at 89- 92 (1978). 304. The scientists' expertise in reducing hazards is also needed. 1975 Genetic Hearing, supra note 8, at 25 (testimony of Prof. Brown). The role of risk-spotter should not be confused with the role of decisionmaker, however, which includes incorporation of societal values and burden of proof techniques. See Green, Safety Determination, supra note 43, at 792; Bazelon, Technology and the Legal Proc- ess, 62 CORNELL L. REV. 817, 826-29 (1977). 305. McGill, supra note 5, at 91. See notes 5 & 143 supra. 306. One proposed plan that provided for public participation on a limited level was S. 19791 RECOMBINANT DNA promulgating regulations be representative of nonscientific interests as well as a diversity of scientific expertise. Additionally, local "biohazard committees" and the public hearings process can be utilized for obtaining public input into revisions of the regulations.

4. The Scope of the Regulatory Scheme Finally, a federal regulatory plan should be limited in scope to recombinant DNA research. Although new technologies to develop in the future may require control to prevent environmental harm, it is not necessary at this time to build a bureaucratic machinery to deal with hypotheticals. 30 7 Furthermore, regulation of recombinant DNA research should only continue so long as the risks of such research justify the expenditure of resources. If, in the future, the risks of recombinant DNA research can be shown to be insignificant, then regulation should not continue.

V CONCLUSION Since the beginning of the recombinant DNA controversy, numer- 308 ous bills have been introduced in Congress to regulate the research. None of these bills have been passed and it is not likely that Congress will act on this issue in 1979. Two of the bills that have been proposed in past sessions of Congress, however, are representative of differing approaches to the regulation of recombinant DNA research and de- serve examination. Introduced and amenaed by Senator Edward Kennedy, S. 1217309 would have established within HEW a "National Recombinant DNA Safety Regulation Commission" comprising eleven members appointed by the President. 310 This commission would have had the authority to promulgate regulations and enforce them by licensing and inspecting

1217, 95th Cong., 1st Sess. (1977), sponsored by Senator Edward Kennedy. For a summary of the plan, see notes 309-13 and accompanying text infra. 307. The growth of bureaucratic deadwood could be avoided by using a "sunset clause" in conjunction with "zero-based budgeting"; limiting the agency's lifetime to a set number of years, unless Congress renews its authority, and reviewing the budget from the standpoint of the effectiveness of each program. See Note, Zero-Base Sunset Review, 14 HARV. J. LEGIS. 505 (1977). 308. Since February 1977, ten bills have been introduced in the House of Representa- tives. H.R. REP. No. 1005, 95th Cong., 2d Sess. 4-5 (1978). At least three have been introduced in the Senate. For reproductions of major bills introduced during 1977 see 1978 HEW DOCUMENTS, supra note 149, at 501-740. 309. S. 1217, 95th Cong., 1st Sess. (1978), 1978 HEW DOCUMENTS, supra note 149, at 671. 310. S. 1217, 95th Cong., 1st Sess. § 1801 (1978); 1978 HEW DOCUMENTS, supra note 149, at 690. ECOLOGY LAW QUARTERLY [Vol. 8:55 facilities as well as by imposing civil penalties.31' The bill would have preempted all laws of states and their political subdivisions except where such laws are more stringent than Commission standards and are "relevant and material to the health and environmental concerns" of such state or locality.312 The bill would have required licensing of institutions carrying on recombinant DNA research and would have established local "biohazard review committees" with nonscientists constituting at least one third of the members to provide public in- 3 put.31 By contrast, H.R. 11192,3 14 the most recent House proposal, would have controlled recombinant DNA research by extending the authority of the NIH to encompass privately funded research. The bill would have given to the NIH, through the Secretary of HEW, the power to impose civil penalties or seek injunctions against violators of the NIH Guidelines. 315 Although the proposal would have preempted state and local regulation of recombinant DNA research, such regulation could have been authorized by the Secretary of HEW where it was "the same as, or more stringent than," a federal requirement and if the additional requirement were "necessary to protect health or the environment . "316 The most significant difference between these two proposals is the regulatory agency. The House bill gave NIH the power to control privately funded research. This aspect of the bill has been criticized by some members of Congress for perpetuating the role of the NIH as "being both the promoter and regulator, a situation similar to that of the now defunct Atomic Energy Commission. 317 The approach of S. 1217 resolved this conflict of interest by delegating the power to control recombinant DNA research to an independent commission not under the supervision of the NIH. NIH would have continued to promote research by allocating funds, but would not have had the responsibility for regulating research. This scheme is preferable to that taken by the

311.. Any person who knowingly, willfully, or negligently -violates a provision of section 1803 or section 1808 shall be liable to the United States for a civil penalty in any amount not to exceed $10,000. . . .Each day such a violation continues .. .constitute[s] a separate violation .... S. 1217, 95th Cong., 1st Sess. § 1809(a) (1977). 1978 HEW DOCUMENTS, supra note 149, at 713-14. 312. S. 1217, 95th Cong., 1st Sess. § 1813(b)(1) (1977). 1978 HEW DOCUMENTS, supra note 149, at 713-14. 313. S. 1217, 95th Cong., 1st Sess. § 1804 (1977). 1978 HEW DOCUMENTS, supra note 149, at 700-09. 314. H.R. 11192, 95th Cong., 2d Sess. (1978). H.R. REP.No. 1005, 95th Cong., 2d Sess. (1978). 315. H.R. 11192, 95th Cong., 2d Sess. § 103 (1978). 316. Id. § 106(b). 317. H.R. REP. No. 1005, 95th Cong., 2d Sess. 35 (1978) (dissenting'views of Rep. Waxman). 1979] RECOMBINANT DN4

House bill since it would separate responsibilities for promotion and control. The House bill also did not provide adequate public input into the decisionmaking process. The bill authorized but did not require the Secretary of HEW to establish "local entities to assist in administration and enforcement" of federal regulations. 3'8 Dissenting Representatives argued that "[it does not appear that those 'local entities' will be anything more than institutional peer review committees, with corporations or educational institutions conducting [recombinant] DNA research in essence policing themselves." 3 19 In contrast, S. 1217 would have required that at least one third of the membership of local biohazard committees be nonscientists from the general public and at least- one third researchers, thus insuring a balance of perspectives on 320 the local groups. A law similar to that proposed by S. 1217 should be enacted to regulate recombinant DNA research. The controls adopted should be based, with general exceptions as stated in this Article, on the NIH Guidelines. The control of recombinant DNA research should be removed from the NIH and placed in an agency that does not promote scientific research. Controls should be extended to cover currently unregulated, privately funded research. In addition, Congress should leg- islate a ceiling for state and local regulations instituted by Congress itself. Control in this way will guarantee the benefits of the research while minimizing societal risks. Decisions regarding development of other new technologies can be approached in the same way that this Article has viewed recombinant DNA research. Legislators and administrators must assess short and long term risks and benefits, insert societal values, rationally deal with uncertainty, allow public input, and regulate only insofar as indirect regulation is inadequate. This analytic approach is essential if lawmakers are to appease both a skeptical public and an impatient scientific community.

318. H.R. 11192, 95th Cong., 2d Sess. § 102(d) (1978). 319. H.R. REP. No. 1005, 95th Cong., 2d Sess. 35 (1978) (dissenting views of Represent- atives Waxman, et al). The Revised Guidelines mandate five member Institutional Bi- osafety Committees of which at least two members must represent the interest of the surrounding committee in health and the protection of the environment. NIH Revised Guidelines, supra note 31, at 60,124-25. 320. S. 1217, 95th Cong., 1st Sess. § 1804(a)(2)(C) (1978). 1978 HEW DOCUMENTS, supra note 149, at 704-05.