TCS Daily

The Academy
Takes a Dive

By Henry I. Miller - October 22, 2002 12:00 AM

The National Academy of Sciences, a private, non-profit, self-perpetuating society of distinguished scientists and engineers, is often in the news. Consistent with its 1863 charter from Congress - which commits it "to the furtherance of science and technology and to their use for the general welfare," and to "advise the federal government on scientific and technical matters" - the Academy conducts studies on a broad range of subjects related to science and technology. Most of these are sponsored (that is to say, paid for) by the federal government.

Recent studies include topics as diverse and weighty as: "More Study is Needed Before Utah Uranium is Sealed or Moved," "No Link Found Between Hepatitis B Vaccine and Multiple Sclerosis," "U.S. Should Restore Lands Near Lakes, Rivers and Oceans," "Heavier Trucks Should Be Permitted on Interstate Highways," and "Most Ocean Pollution Not From Tanker or Pipeline Spills."

Both the public and people who are knowledgeable about the subjects studied by the Academy generally assume that these analyses are unbiased and performed by disinterested experts and that they are, therefore, the benchmark for accuracy and objectivity. Consequently, the Academy's studies are widely cited by the media and relied upon by the Congress and by Executive Branch agencies for the formulation of legislation and policy. In this way, these reports define the "popular wisdom" about many critical public policy issues and affect the allocation and expenditure of hundreds of billions of dollars in both the public and private sectors.

Plagued By Apparent Bias

Although much of the work of the Academy is highly regarded, two recent "expert" committees working under the auspices of the National Research Council, the research arm of the Academy, that assessed the scientific integrity of governmental regulation of biotechnology have been plagued by apparent bias, and their analyses have been problematical. It appears that in this critical area, the Academy offers scientific advice to government agencies on the same terms that Burger King advertises that it offers hamburgers - you pay your money, and you "get it your way." It was largely this philosophy - or at least the widespread perception of it - that led to the demise of the highly politicized Office of Technology Assessment, which was supposed to provide independent analyses of scientific and technological issues to the Congress, but too often pandered to powerful congressional patrons.

During the past two years, the Academy has conferred its imprimatur on two questionable analyses of federal biotechnology regulatory policy toward field trials and commercialization of recombinant DNA-modified plants - the more recent on regulation by the U.S. Department of Agriculture (and paid for by USDA), "Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation"; and an earlier report concerning oversight by the EPA, "Genetically Modified Pest-Protected Plants: Science and Regulation." These studies have already been cited specifically as the basis for unscientific, inconsistent and excessively burdensome new regulatory proposals from the Bush administration's Office of Science and Technology Policy ["Proposed Federal Actions to Update Field Test Requirements for Biotechnology Derived Plants and To Establish Early Food Safety Assessments for New Proteins Produced by Such Plants. Federal Register, August 2, 2002, vol. 67, pp 50577-50580.]

The USDA has the legislative authority (primarily under the Plant Pest Act) to regulate the importation and interstate movement of plants, plant products and other organisms that may introduce plant diseases or pests. For example, there has long been a permitting system for "plant pests," defined as any organism "which can directly or indirectly injure or cause disease or damage in or to any plants or parts thereof, or any processed, manufactured, or other products of plants." The USDA's Animal and Plant Health Inspection Service (APHIS) regulations incorporate an inclusive list of organisms that are or that harbor plant pests. This approach is essentially binary: A plant that an investigator might wish to introduce into the field is either on the proscribed, inclusive list of plants pests - and therefore, requires a permit - or it's exempt. This approach is risk-based, in that the organisms that are required to undergo case by case governmental review are an enhanced-risk group, compared to plants not considered to be plant pests.

For the past 15 years, however, the USDA also has maintained a parallel regime focused exclusively on transgenic plants (those that contain DNA from more than one species, and in which the DNA was transferred with molecular techniques). In order to establish this mechanism, which regulates essentially independent of risk, APHIS tortured the original concept of a plant pest as something known to be harmful, and crafted a new category - a "regulated article," defined as "any organism or any product altered or produced through recombinant DNA technology [that is, gene-splicing], which is a plant pest, or for which there is reason to believe is a plant pest." The phrase "for which there is reason to believe is a plant pest" has been broadly interpreted by APHIS to include any organism that includes any amount of DNA from a plant pest - even a snippet of DNA that is incapable of conferring the ability to cause disease or other harm.

The USDA's case by case permitting process and costly field test design and other requirements have made gene-spliced plants disproportionately expensive to develop and test. A field trial with a gene-spliced plant may be 10-20 times more expensive than the same experiment performed with a plant that has identical characteristics, but that was modified with less precise genetic techniques. This represents a huge - and unnecessary - burden for researchers who wish to develop and test gene-spliced organisms for any of a variety of purposes.

The EPA, which regulates pesticides under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), has contrived its own version of a regulated article. Under the pesticide statute, EPA has long regulated field testing - on areas greater than 10 acres - and the commercial use of pesticides. In 1994, EPA proposed its "plants as pesticides" rule, which brought under the jurisdiction of FIFRA all the substances that mediate "host plant resistance" to pests (as well as the genetic material needed to direct the synthesis of these substances) - but only if they are introduced with gene-splicing techniques.

In a final regulation published in 2001, seven years after the rule was first proposed, in place of "pesticide" EPA coined the term, "plant-incorporated protectants (PIPs)" to describe what it would regulate. These PIPs are defined in a way that places them within the FIFRA definition of a pesticide - namely, a substance intended to prevent, repel or mitigate any pest - but only if the plant was constructed by gene-splicing techniques. Plants modified with any of the less precise, less predictable techniques of "conventional breeding" are expressly exempted. Under this rule, EPA is required to conduct a case by case review of proposed field trials when an investigator intends to perform the initial field trial of a gene-spliced plant that contains a plant-incorporated protectant, and again when trials are scaled up to larger size or to additional sites (and if minor changes are made in the genetic construct). This requires the developer of a candidate for a new plant variety that could be used in commerce to obtain permission repeatedly from EPA. For gene-spliced plants EPA eliminated the 10-acre research exemption, making the likes of gene-spliced wheat, cotton and marigolds more stringently regulated than toxic chemicals similar to DDT or parathion, which enjoy the exemption! Once again, the burden on a researcher who uses gene-splicing techniques is enormous - and enormously excessive.

Ignoring a Continuum

The concept of a "regulated article" or "plant incorporated protectant" may be inventive, but it flies in the face of the disciplines of plant pathology and biology, as well conflicting with the risk-based orientation of the statutes. Moreover, USDA's and EPA's regulatory policies fail to acknowledge that genetic modification represents a continuum - from crude, imprecise traditional practices such as hybridization and intensive mutagenesis, to more precise and predictable recombinant DNA techniques. Likewise, they fail to take into consideration the extraordinary overall safety record of genetic improvement in agricultural research throughout both the pre- and post-gene-splicing eras.

Literally millions of genetically altered, but not gene-spliced, plants are field tested each year without governmental oversight or strictures: The average plant breeder of corn, soybean, wheat, or potato, for example, may put into the field 50,000 discrete, new genetic variants per year, many or all of which may be the product of "wide crosses," hybridization in which genetic material (including that from weedy or poisonous plants) has been transferred across natural breeding barriers. The safety record of the tens of thousands of field trials of gene-spliced plants that have been performed worldwide, and of the hundreds of millions acres of cultivated commercial recombinant DNA-modified crops - virtually all of which have been performed with only the plant breeding practices standard for the parental crop - has been stunning, and the results of risk-assessment experiments likewise have been reassuring.

As to the presence of recombinant microorganisms "in the environment," a study funded by the EPA twenty years ago confirmed that substantial numbers of recombinant bacteria - approximately 108 organisms/day/technician - are incidentally released from the "containment" of standard research laboratories. Thus, a vast and varied unsupervised "release experiment" involving tens of thousands of laboratories and untold millions of discrete new genotypes of recombinant microorganisms has been in progress for two decades, with no known untoward results.

During the past two decades, national and international scientific organizations - including, repeatedly, the National Academy of Sciences and the National Research Council - have addressed the question of whether there are unique risks associated with gene-spliced organisms, with congruent conclusions. A 1987 white paper from the NAS, "Introduction of Recombinant DNA-Engineered Organisms into the Environment: Key Issues," concluded that there is no evidence of the existence of unique hazards, either in the use of gene-splicing techniques or in the movement of genes between unrelated organisms. In 2000, an NRC report on the scientific basis of EPA's regulation of gene-spliced plants, "Genetically Modified Pest-Protected Plants: Science and Regulation," concurred: "the committee agrees that the properties of a genetically modified organisms should be the focus of risk assessments, not the process by which it was produced" [emphasis in original]. Perhaps the most comprehensive and unequivocal analysis was the 1989 NRC report on the risks of recombinant plants and microorganisms, "Field Testing Genetically Modified Organisms: Framework for Decisions," which concluded that "the same physical and biological laws govern the response of organisms modified by modern molecular and cellular methods and those produced by classical methods." But this analysis went even further, emphasizing that recombinant DNA techniques are more precise, circumscribed and predictable than other methods.

In other words, recombinant DNA technology is a refinement, or improvement, over older, less precise techniques and its use generates less uncertainty, which led the committee to make this strong policy recommendation, "the nature of the process [of genetic modification] is not a useful criterion for determining whether the product requires less or more oversight."

So much for the discriminatory treatment of USDA's "regulated articles" and EPA's "plant-incorporated protectants," the case by case review of which is triggered by the use of gene-splicing.

A Variety of Specious Arguments

On the basis of the exegesis above and the recognition in the 2002 NRC report on USDA that government agencies are in the "difficult position of enforcing a higher environmental standard for transgenic plants than the standards currently used to regulate the impacts of other agricultural technologies and practices," one might logically have expected an endorsement and extension of the 1987 NAS white paper and 1989 NRC report, accompanied by a recommendation to rationalize the system and to exert oversight of recombinant and conventional plants generally no differently, except for those plants with newly-introduced traits perceived to confer higher risk. (In fact, if the use of certain techniques did dictate differences in degree of oversight, arguably it is organisms crafted with the less precise and predictable traditional techniques that should be subjected to greater scrutiny than those constructed with gene-splicing techniques. This is further discussed below.) Instead, the Academy committee recommends maintaining or even increasing the stringency of the current discriminatory, process-based regulatory system that focuses on plants modified by recombinant DNA technology. It justifies this recommendation by invoking a variety of specious arguments.

First, the committee makes a critical concession:

[this] committee agrees with previous NRC committees that there are no new categories of risk associated with transgenic plants. The categories of risks from transgenic plants include those associated with the movement of the transgenes; impacts of the whole plants through escape, and through impacts on agricultural practices, non-target organism effects, and resistance evolution. For this reason, the process of producing new plant varieties should not enter into the assessment (emphasis added).

However, they then go on to endorse the view that "risks must be assessed on a case-by-case basis with consideration for the organism, trait, and environment." The two are irreconcilable: the vast majority of non-gene-spliced plant varieties - although the techniques of genetic modification are crude and the genetic changes are poorly characterized - are not subject to case-by-case governmental review at any time during the process of development.

Consider, for example, the products of induced-mutation breeding, which has been in common use since the 1950s. This technique involves exposing crop plants to ionizing radiation or toxic chemicals to induce random genetic mutations. These treatments most often kill the plants (or seeds) or cause detrimental genetic changes, but on rare occasions, the result is a desirable mutation - for example, one producing a new trait in the plant that is agronomically useful, such as altered height, more seeds, or larger fruit. In these cases, breeders lack detailed knowledge about the nature of the genetic mutation(s) that produced the useful trait, or about the large number of other mutations that inevitably have occurred in the plant. Yet the more than 2200 mutation-bred plant varieties from a range of different species that have been marketed over the last half century have been and remain subject to no formal pre-market regulation.

Likewise, the committee's rationale makes it difficult to reconcile the exemption from regulatory review of wide crosses, hybridization in which embryo rescue or similar techniques are used to transfer what plant breeders call "alien genes" from one species or genus to another to create plants that would not exist in nature.

Consider, for example, Triticum agropyrotriticum, a man-made "species" constructed by combining genes from bread wheat and a grass called quackgrass or couchgrass. Possessing all the chromosomes of wheat and one extra whole genome from the quackgrass - thereby adding tens of thousands of genes - T. agropyrotriticum was independently produced via wide crosses in the former Soviet Union, Canada, United States, France, Germany and China, where at various times it has been grown for both forage and grain. These new genetic constructions are exempt from regulation (in spite of at least the theoretical possibility that the new gene products could make them more weedy, toxic or allergenic than parental wheat varieties), although the use of gene-splicing techniques to add a single quackgrass gene to wheat would precipitate an extensive (and expensive) pre-market review (by FDA and either USDA or EPA, depending on whether the introduced gene conferred pesticidal properties). This discrepancy is irreconcilable.

As to concerns about the origin of an introduced gene, the scientific consensus holds that the risk of an introduced gene is related primarily to its function, not its origin. Moreover, the very concept of the "origin" of a gene has become murky with the accumulation and analysis of DNA sequencing data. Nearly identical DNA sequences and biochemical pathways are found across vast phylogenetic distances. This broad conservation and sharing of gene sequences in nature weakens the argument that the origin - as opposed to the function - of newly introduced genetic material poses a safety concern (assuming that the introduced material is well-characterized, in order that one knows what has been transferred).

Second, the committee claims there is greater risk from gene-splicing than other techniques because "a much broader array of phenotypic traits can potentially be incorporated into plants than was possible two decades ago." But this is a second-order kind of concern: Greater versatility is not the same as enhanced risk. The FDA emphasized this point in its 1992 policy on foods from "new plant varieties," which defined certain potentially hazardous characteristics of new foods - such as the presence of a substance new to the food supply, increased levels of an endogenous toxin, and the introduction of an allergen - that, if present, would require greater scrutiny by the agency, and which could result in additional testing and labeling, or exclusion from commerce. During the past decade, under this policy thousands of foods in U.S. supermarkets that contain byproducts of recombinant organisms have been marketed, irrespective of whether the plant arose from the application of gene-splicing or traditional genetic engineering methods. Other risk-based approaches have been described for the oversight of plants in the field.

However, in spite of the weight of scientific consensus and empirical evidence, in a particularly infelicitous circumlocution the panel concludes "that the scientific justification for regulation of transgenic plants is not dependent on historically set precedents for not regulating conventionally modified plants. While there is a need to reevaluate the potential environmental effects of conventionally improved crops, for practical reasons, the committee does not recommend immediate regulation of conventional crops."

For practical reasons, indeed! Not a single conventional crop could (or should) meet the requirements being imposed - without scientific justification - by USDA on gene-spliced plants. Conventional plant breeding - which uses far less precise and predictable methods than gene-splicing, whose mechanisms for enhancing host plant resistance are largely unknown, and which (unlike gene-splicing) has led to the occasional inadvertent introduction of undesirable traits into commercialized products - would grind to a halt if USDA's regulatory approach were applied to them.

On the basis of such scientifically unconvincing reasoning, the NAS panel recommends continued compulsory case-by-case oversight - and even more compulsive reviews - by USDA of the field trials of all gene-spliced plants.

Third, the chairman of the committee that produced the 2002 report, Fred Gould, concedes that the risk-related characteristics of a product, rather than the techniques used to make it, should determine the need for and extent of regulation, and the NRC itself has concluded that because gene-splicing is so precise and well-characterized, "[w]ith organisms modified by molecular methods, we are in a better, if not perfect, position to predict the phenotypic expression." Gould then turns these points of consensus into a false dichotomy, concluding that there are "two options: regulate all plant varieties or regulate none," and that because both are unacceptable, a tiered approach applicable only to gene-spliced plants is the logical solution. But existing regulation and models described in the scientific literature argue otherwise.

In order to circumscribe scientifically what needs to be regulated, the long-standing USDA regulations (under the Plant Pest Act) applicable to conventionally-modified plants incorporate an inclusive list of organisms that are or that harbor plant pests. This approach is essentially binary: A plant that an investigator might wish to introduce into the field is either on the proscribed, inclusive list of plants pests - and therefore, requires a permit - or it's exempt. This approach is focused on the likely risks of products, not on the process used to craft the organism; and it is risk-based, in that the organisms that are required to undergo case by case governmental review are an enhanced-risk group - organisms that can injure or damage plants - compared to plants not known to be plant pests. This two-tiered approach could easily have been expanded to encompass gene-spliced, as well as conventionally-modified plants.

Within a Scientific Framework - Always

There are other variations on existing risk-based regulatory approaches. Several years ago, the Stanford University Project on Regulation of Agricultural Introductions described a widely applicable regulatory model for the field testing of any organism, whatever the method(s) employed in its construction. Its basis is the stratification by panels of experts of various organisms into risk categories - a multi-tiered variation on the USDA's binary approach.

This model offers regulatory bodies a highly adaptable, scientific approach to the oversight of field-testing plants and other organisms. The approach is analogous to that used by the CDC and NIH for categorizing microorganisms, in order to establish laboratory safety standards for the handling of pathogens; it is applicable whether the introduced organisms are "naturally occurring," non-indigenous "exotics," or have been genetically improved by either old or new techniques. Moreover, the model is sufficiently flexible that the stringency of regulation can be varied, according to the preferences and needs of particular regulatory authorities - but, unlike the current regulatory regimes at USDA and EPA, always within a scientific framework. Under such a system, some currently unregulated introductions of traditionally-bred cultivars and exotics considered to be of moderate or greater risk would likely become subject to review, whereas many currently reviewed gene-spliced organisms would likely become exempt.

Perhaps one scientifically indefensible, internally inconsistent, analysis could be dismissed as an anomaly, but the NRC's previous report on a parallel subject - the EPA's oversight of gene-spliced plants with enhanced pest- or disease-resistance - was similarly flawed.
The committee that produced the EPA report chose to ignore crucial aspects of its charge - namely, "to examine the existing and proposed regulations to qualitatively assess their consequences for research, development, and commercialization of [gene-spliced plants modified to enhance pest-resistance]"; and to "provide recommendations to address the identified risk/benefits, and, if warranted, for the existing and proposed regulation of [gene-spliced plants modified to enhance pest-resistance]." This point is essential because most other analyses have found the EPA's existing and proposed regulation to be unscientific, illogical and damaging to agricultural research.

Both the 1987 and 1989 NAS/NRC analyses and those from other academic groups had arrived at conclusions incompatible with the EPA approach, which, as discussed above, circumscribes only gene-spliced plants for repeated, redundant case by case review of field trials, and subjects each gene product and the requisite transgenic DNA to onerous pesticide registration procedures. A large segment of the scientific community has unequivocally condemned this approach.

A 1996 report by eleven scientific societies that represent 80,000 biologists and food professionals excoriated the EPA's approach and warned of a number of negative consequences for agriculture and consumers, if the EPA's policy were to be implemented. They predicted that it would discourage the development of new pest-resistant crops, thereby prolonging the use of synthetic chemical pesticides; increase the regulatory burden for those developing pest-resistant varieties of crops; expand federal and state bureaucracies; limit the use of biotechnology for the development of pest-resistant plants to a very few developers that can bear increased regulatory costs; and handicap the United States in competition for international markets. (They were correct on all counts.)

The 11 societies' report also offered general principles and recommendations for the oversight of new plant varieties: that federal oversight should be based on scientific principles; that it "should focus on high-probability risk rather than hypothetical or unrecognizable risk"; and that "the level of risk of a plant variety to the environment or human safety is determined by the characteristics of the plant, not by the method by which a gene for pest defense is transferred."

In 1998 the Council on Agricultural Science and Technology (CAST), an international consortium of 36 scientific and professional groups, reiterated the eleven societies' criticisms, characterizing the EPA's approach as "scientifically indefensible," and observing that treating gene-spliced plants as pesticides would "undermine public confidence in the food supply."

It was extraordinary, therefore, to find in the 2000 report from the Academy that "the committee has chosen to take EPA's proposed rule and the overarching [federal governmental] coordinated framework as given." This critical decision enabled the committee - and also the later committee that addressed USDA's policy - to produce a report which accepted a policy that had been censured repeatedly; a policy that calls into question the long, distinguished history of breeding pest resistance into plants that have yielded enormous improvements in food production worldwide; a policy that if applied to other, less precise technologies would have thwarted the Green Revolution that has been, literally, life-giving to hundreds of millions of starving people in developing countries.

That the NRC's 2000 analysis of EPA's regulatory approach contains language that reflects and endorses the scientific consensus on the nature of risk - "the committee agrees that the properties of a genetically modified organisms should be the focus of risk assessments, not the process by which it was produced" [emphasis in original] - only emphasizes the logical inconsistency of choosing to ignore the flawed, central tenet of the EPA's approach to regulation; namely, that the use of gene-splicing techniques is the trigger to regulation, a tenet that violates the regulatory principle that the degree of scrutiny should be commensurate with risk.

Taking a Dive

How could the esteemed National Academy of Sciences twice have gone so far wrong in its assessment of the scientific basis for federal regulatory policy?

The game was "fixed," as in tampered with, or corrupt. In other words, the exalted Academy took a dive.

The USDA committee was stacked with members known to harbor antagonism or skepticism toward biotechnology; moreover, unlike the 1987 and 1989 NRC committees, it contained few fellows of the Academy (only two of twelve members). The EPA committee contained none save the chairman.

The members of the committee and the invited reviewers for the EPA report were selected with disregard for apparent conflicts of interest and known bias. Three members of the twelve-person committee (Stanley Abramson, Fred Betz and Morris Levin) are former EPA staff who had helped to craft and defend a variety of process-based regulatory policies at the agency, and another (Rebecca Goldburg) has produced a succession of anti-biotechnology tracts over the past decade and half. Moreover, during the formal review process, the document was reviewed by another former senior EPA official (Lynn Goldman) who had been instrumental in crafting and defending the policy in question, and by an intractable anti-biotechnology activist (Jane Rissler). Three members of the USDA committee (its chairman, Fred Gould, and David Andow and Norman Ellstrand) ) are long-time skeptics about the safety of recombinant plants and have consistently advocated process-based regulation, and another (Ignacio Chapela) is the author of a discredited article on alleged contamination of indigenous gene-spliced varieties of corn in Mexico by gene-spliced varieties. Even the staff director of the Academy's Board on Life Sciences, Frances Sharples, is a long-time anti-biotech activist.

The report on EPA oversight had the desired result: The Academy report offered sufficient cover for EPA to issue a final rule, in 2001, after seven years of delays caused by the objections and opposition of the scientific community.

Harming the Little Guy

In a similar vein, the prestige of the Academy attached to the report on USDA's regulation virtually assures the permanence of stultifying, process-based regulation at USDA that will unnecessarily inflate the costs of research and the commercialization of new plant varieties.

Because excessive regulation acts as a market-entry barrier to smaller competitors unable to bear inflated regulatory costs, the handful of large agribusiness companies currently involved in agricultural biotechnology do not object strenuously to such excessive, extensive and expensive EPA and USDA regulatory regimes. However, academic researchers, the ultimate engine for innovation, are the most severely affected victims of excessive, ill-conceived regulation. Operating on small budgets, their ability to perform field trials of gene-spliced plants and microorganisms is markedly restricted.

As part of an analysis of misconduct in biomedical research, the late DeWitt "Hans" Stetten, esteemed biologist and scientific administrator, wrote that "Science cannot tolerate the man who takes lightly his moral obligation to report strictly what is true." It appears, however, that on certain high-profile, politically-charged subjects, the National Academy of Sciences lately has chosen to exempt itself from that axiom. Not only does it tolerate such people, but it appoints them to "expert" committees, and then confers the Academy's imprimatur on their flawed work product.

The result is that the federal government - and American society at large - finds itself without an estimable, incorruptible, reliable source of advice on scientific, technological and medical issues.
If the Academy's recent analyses of the science underlying biotechnology regulation can be so palpably flawed, how can Americans and American political institutions trust the veracity and integrity of future assessments of other critical subjects ranging from air pollution and global climate change to bioterrorism and national vaccine policy? And if the Academy is discredited, where else can we turn?

Henry I. Miller is a Fellow at the Hoover Institution. He is the author of "Policy Controversy in Biotechnology: An Insider's View," and was an FDA official from 1979 to 1994.

TCS Daily Archives