The creation of a scientific report on a contentious issue presents a special difficulty. On the one hand the drafting committee must include representatives of various constituencies with opposing views. So the committee that wrote the new report included academics involved in genetics, economics, and agriculture, a representative of a public interest environmental action group, a lawyer who helps clients to obtain regulatory approvals, and a state government environmental regulator. On the other hand, there cannot be a majority and a minority report, since after all we are dealing with Objective Science, and scientists either know the truth or they don’t. NRC reports always speak with one voice. Such reports, then, can produce only a slight rocking of the extremely well gyrostabilized ship of state, no matter how high the winds and waves. Any member of the crew who mutinies is put off at the first port of call.
While usually artfully concealed, the machinery of forced consensus is apparent in the pest-protected plant report. The economist on the committee, Erik Lichtenberg, clearly felt that the sorts of regulation recommended by the report were not worthwhile and, indeed, would have costs not justified by any claimed benefits. He and his cost-benefit analysis are quarantined in an appendix and referred to only in a footnote: “This appendix was authored by an individual committee member and is not part of the committee’s consensus report. The committee as a whole may not necessarily agree with all of the contents of appendix A.” Of course, appendix A is merely economics, while the “committee as a whole” must “necessarily agree with the contents” of the rest of the report or it wouldn’t be a scientific report. In fact, the committee could have discounted the appendix on substantive grounds. Like so much of cost-benefit analysis, it fails to take account of the fact that the costs, possible ill-health, fall on different parties than the benefits, profits to corporate entities who produce the inputs into agriculture. More fundamentally, it avoids the deep problem that to provide a quantitative balancing of the books, the costs and benefits would have to be assessed in the same currency, while it has never been possible to come to a general agreement on the dollar cost of sickness and death.
There are five general issues that are in contention in the struggle over GMOs. Three of these, threats to human health, possible disruption of natural environments, and threats to agricultural production from a more rapid evolution of resistant pests, comprise the agenda of the NRC report. The other two, disruption of third- world agricultural economies and principled objections to “unnatural” interventions, are deliberately excluded. Page 2 of the report states in italics: “The study does not address philosophical and social issues surrounding the use of genetic engineering in agriculture, food labeling, or international trade in genetically modified plants.” In analyzing the risks of GMOs the committee follows a general principle established in previous Academy reports, a principle that it regards as fundamental, namely that it is the product and not the process that matters. For the NRC it is irrelevant whether a variety has been produced by conventional genetic manipulations or by transgenic transfer of DNA. What counts is whether the new property of the resultant organism is harmful to health or the environment.
The NRC authors point out, quite properly, that the conventional methods of breeding, including sexual crosses between species that do not ordinarily cross in nature, might produce varieties with some heightened toxicity to humans or other species, or with unusual invasive abilities, or with greater resistance to pests that would hasten the evolution of more effective pest species. Jane Rissler and Margaret Mellon, in their extremely informative The Ecological Risks of Engineered Crops, give many examples of new troublesome weeds that have arisen from the hybridization of crop plants with their wild relatives and several where rare wild species have been driven to extinction by hybridization with crop plants.
Indeed, the only examples we have so far of the adverse effects of agricultural varieties on any animal or plant species in nature, including on human health, have been from conventionally bred organisms or from the introduction of invasive species from distant geographical areas, or from foods like peanuts or milk to which some people are naturally allergic. So if the usual products of agricultural practice already provide numerous examples of adverse effects, why is there the massive popular and political anxiety centered on genetically engineered crops in particular? None of the authors of the reports and books seems to have noticed that if it were really only the product and not the process that matters, then nothing has changed. The NRC report itself provides a protocol for protecting consumers against new food toxins and allergens (i.e., substances causing allergies) that applies irrespective of the genetic method used in variety development and which makes use of the already existing federal apparatus for the approval of new plant varieties.
First, one asks whether a new substance is found in parts of a plant that consumers eat or with which workers come in contact. If not, the substance is “exempt from health concerns.” If it is found in such parts, then does it have chemical properties common to many allergens? If it does, then safety assessment is needed. If not, then is it similar to other substances that people eat? If not, then again we need safety assessment. The real problem revealed in the NRC report, although it did not seem to bother the panel, is that the data on which “safety assessment” is currently based are not produced by the federal agencies themselves but are provided by the very parties who are asking for approval to distribute the new variety in the first place. Moreover, no one seems to have noticed that there is, in fact, an aspect of the process of genetic engineering that does make that process unusually likely to produce unpredictable results.
All the attention has been paid to the physiological effect of the gene that has been put into the recipient, but none to the effect of where it is inserted in the recipient’s genome. Genes consist of two functionally different adjacent stretches of DNA. One, the so-called structural gene, has information on the chemical composition of the protein that the cell will manufacture when it reads the gene. The other, the so-called regulatory element, is part of a complex signaling system that concerns where and when and how much protein will be produced. When DNA is inserted into the genome of a recipient by engineering methods it may pop into the recipient’s DNA anywhere, including in the middle of some other gene’s regulatory element. The result will be a gene whose reading is no longer under normal control.
One consequence might be that the gene is never read at all, in which case it will probably be bad for the recipient and will never be part of a useful agricultural variety. But another possibility is that the cell will now produce vast amounts of a protein that ordinarily is produced in very low amount, and this high concentration could be toxic or be involved in the biochemical production of a toxin. Yet another possibility is that a toxic substance that used to be produced only in one part of a plant, not ordinarily eaten, could now be manufactured in another part. Tomatoes are delicious, but you would be ill-advised to eat the leaves and stems because they contain toxins. It is not impossible that a genetically engineered tomato might, by bad luck, start to produce these toxins in the fruit. Thus the process of genetic engineering itself has a unique ability to produce deleterious effects and, contrary to the recommendations of the NRC report, this justifies the view that all varieties produced by recombinant DNA technology need to be specially scrutinized and tested for such effects. Exactly how one would go about doing that, in view of the unknown nature of the danger, is uncertain. Even extensive testing on a variety of animals provides no guarantee of safety since there are plant substances that are toxic to some species and not to others.
As yet no one that we know of has been poisoned by a transgenic plant. There have been a couple of close calls, however. The most widely cited case is the Brazil nut protein produced by a transgenic soybean. In some subsistence agricultural communities, for example in West Africa, diets are severely deficient in an essential amino acid, methionine. Brazil nuts produce a protein that is rich in methionine and so it was thought that inserting the appropriate gene from Brazil nuts into soybeans would provide an easy fix for West African malnutrition. Unfortunately the Brazil nut protein is known to be allergenic and the transgenic soybean proved to be so as well, so the variety was never released.
Proponents of recombinant DNA technology like Alan McHughen point to this case as a proof that self-policing by a variety developer can be counted on to avoid disaster. One’s confidence in self-policing is somewhat diminished, however, by the realization that the allergenic properties of the protein were well known before the Pioneer Hi-bred seed company ever started to develop the variety in the first place. At some point they must have realized that the Food and Drug Administration would have refused approval of the variety even under our present system of regulation. How one wishes for a transcript of the discussions in the company board room.
A major part of the NRC report and the entirety of Rissler and Mellon’s book are concerned with ecological issues in the broad sense. One anxiety is that “superweeds” will be produced, dominant plants that will spread en masse either through cultivated fields or through natural habitats. Sometimes what is meant by “weeds” is unwanted species that are growing in cultivated fields. At other times these are confused with introduced invasive species like the European purple loosestrife that has taken over so many American wetlands. There are no known examples of hybrids between cultivated plants and wild relatives becoming superweeds that have destroyed natural habitats, largely because too many of the characteristics selected during domestication make cultivarsâ€”cultivated varieties of plant speciesâ€”dependent on the tender loving care given to them by farmers. Nor will the addition of a gene conferring herbicide resistance or pest resistance change that dependence. Plants growing in natural habitats are not subject to herbicides, nor are they attacked regularly by the hordes of predatory insects attracted to the concentrated free lunch offered in cultivated fields.
On the other hand, more difficult weeds of cultivated fields certainly will evolve if herbicide resistance becomes incorporated by natural crossing into species that are already weeds. The fear of superweeds is promoted by the metaphor of “escape” used to describe the passage of an engineered gene into a wild species. The image is of the mad scientist (or not-so-mad germ warfare biologist) who has created a virulent disease organism, ready at any moment to create a major epidemic unless it is rigorously confined to the laboratory or, better yet, destroyed. But transgenes are not spread like microbes, entering the body from outside. They are transmitted by reproduction of the entire genome of an organism, and if a cross occurs between an engineered plant and a wild relative, the result is an offspring that is hybrid in every respect, including all those characteristics that make cultivated varieties so ill-adapted to survival in nature, such as their demands for unnaturally high levels of nitrogen fertilizer.