Genetically Modified Plants and Human Health Risks: Can Additional Research Reduce Uncertainties and Increase Public Confidence?

Ernest Hodgson,1

Department of Environmental and Molecular Toxicology, Box 7633, North Carolina State University, Raleigh, North Carolina 27695

Received March 28, 2001; accepted June 6, 2001


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 REFERENCES
 
So long as the risks to human health from transgenic plants remain potential rather than actual, and, in any event, appear lower than those from traditional plant breeding, hazard assessment need not be extensive. However, in view of current public attitudes to transgenic plants, it is necessary that those tests that are required, be based on logic, on sound science, and in accordance with the best scientific methodology. This is particularly the case with testing for food allergenicity. Current testing is largely indirect and based on comparisons with other known food allergens. Development of direct tests that involve interaction between the actual transgenic protein in question and the immune system is essential if confidence in the regulatory system is to be restored.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 REFERENCES
 
The author acquired some of the background information for this article while serving on the National Research Council/National Academy of Science committee that produced the report entitled "Genetically Modified Pest-Protected Plants" (NRC, 2000Go). While I am truly grateful for the opportunity to have served with this distinguished committee, and for all of the interactions with them, it must be emphasized that there is no connection, formal or informal, between this article and either the committee members or its work product. Any opinions that are expressed here, whether or not they coincide with opinions expressed in the above report, represent the authors personal knowledge and beliefs.

A brief answer to the question posed in the title would be that it is antithetical to scientific philosophy to believe that further research will not reduce uncertainty but it is also true that public confidence, in the present climate, has little if anything to do with science. Before exploring these answers in more detail, certain caveats about the process must be laid out. First, while hazard assessment at its best is relatively scientific and straightforward, risk assessment is an inexact procedure involving assumptions, uncertainty factors, and default values in profusion. Thus, at its best, risk assessment provides a reasonable estimate; at its worst it scarcely rises above the level of numerology. Further, regulation of potentially toxic materials is not carried out solely on the basis of risk assessment. In addition, regulation also includes the political process that led to the relevant legislation as well as public opinion expressed by public interest groups with a variety of motivations and molded by an often less-than-expert media. However, risk is relative, and this decidedly Byzantine process may represent the best way possible for society to indicate how much risk it is willing to take at a particular point in time. The role of the toxicologist is simpler: to analyze the existing data as unambiguously as possible, to carry out the best experimental assessment of hazard, to strive for a more science-based risk assessment process, and to present the findings in a clear and unbiased manner.

Risk assessment is generally held to consist of 4 steps (Hodgson and Levi, 1997Go; NRC, 1983Go): hazard assessment, dose-response evaluation, exposure assessment, and risk characterization. It is usually conducted so as to give a quantitative assessment, it is carried out on the product rather than the process that gives rise to the product, and it is an essential preliminary to 2 further steps: risk communication and risk management.

Since improvement of any of the 4 steps should reduce uncertainty and increase the extent to which the assessment is science-based, they can be addressed individually as to current status and future needs. At the same time, the question should be asked as to whether gene products that present a threat to human health will occur more often in transgenic plants than in plants produced by traditional genetic crossing. On the one hand, traditional genetic crosses involve recombination of large parts of genomes, including many unknown genes as well as genes of importance to the plant breeder, thus giving rise to many new combinations of genes and potential gene products. Traditional plant breeding has been carried out on a trial-and-error basis for centuries, possibly millennia, and with considerable scientific rigor for a century or more. Transgenic plants, in contrast, have been produced only in the last 2 or 3 decades and usually differ by one, or at most a small number, of genes from the parent strain, permitting the argument that they are less likely to give rise to new, potentially dangerous gene products. This argument may be countered by the observation that the genes of interest in transgenic plants may be drawn from any living organism, giving rise to gene combinations capable of producing new and unpredictable products with new and unpredictable effects. It is true, however, that molecular techniques for the production of new varieties of crop plants are, in fact, only new methodologies in a very old human activity.

Despite the fact that the hazards from transgenic plants are potential rather than real, and that the hazards associated with new plant varieties have been primarily associated with traditional non-transgenic methods of plant breeding, several bills proposing to amend the Food, Drug, and Cosmetic Act have been introduced into both houses of Congress. Goldman (2000) discusses these proposed acts in detail, pointing out both legal and constitutional problems. Her conclusion is as follows: "Both the GEFSA and the GEFRKA are inconsistent with basic principles of food regulation, as well as [with] current scientific knowledge about bioengineered foods. Laws addressing the safety and labeling of bioengineered food, or the regulation of any new technology, should be based on sound science."

The plea for decisions based on sound science seems to have fallen on deaf ears in the case of Starlink corn. This corn variety was approved for use in animal, but not human food, based on the presence of Cry9C, a Bacillus thuringiensis (Bt) protein held to be a possible human allergen. This decision was based primarily on protein stability, without any direct hazard assessment, and the fact was ignored that even under worst-case scenarios, human exposure would be orders of magnitude less than that necessary to sensitize individuals and lead to allergic reactions on subsequent exposure (Anon, 2000Go).

Given the difficulties involved in policing such a restriction, it would appear inevitable that problems would arise. In the fall of 2000, evidence of the use of StarLink corn in taco shells was discovered and, in the words of Jocelyn Kaiser (Kaiser, 2000Go), "all hell broke loose." Despite the failure to find structural similarities between the Cry9 protein and known food allergens and the opinion of an EPA-appointed expert panel, which observed that the probability of harm to sensitive people through allergic reactions was low, a massive recall has been initiated, punitive firings have been carried out, and the public has been subjected to an alarming mixture of information, misinformation, and disinformation. Clearly, given this public reaction, it will no longer be possible for the EPA to regulate corn or other food products containing Cry9 proteins on the basis of sound science alone. One can only wonder what the new basis might be.

A regulation restricting the use of a variety of a food plant as ubiquitous as corn to animal feed but not human food would seem, in retrospect, to have been a disaster waiting to happen. The enormity of the disaster has recently been made clear in an excellent summary of the current situation (Thayer, 2001Go). Thayer provides an excellent summary of the nature of Starlink corn, the history of its release, and the resultant problems. The lawsuits and the litigants are discussed, as are the opinions of an EPA expert panel on the possibility of human health effects, specifically allergenicity.

Hazard Assessment
Although, in the case of synthetic organic chemicals, much, if not all, of the hazard assessment is derived from planned experiments with controlled exposures; in the case of genetically modified and other plants it is often from records of incidents.

Secondary plant chemicals (allelochemicals) may be toxic to mammals, including humans (Senti and Rizek, 1974Go) and changes in the concentrations of such compounds, whether brought about by transgenic or traditional genetic modifications, are seen as potential hazards. Although not enough cases have been described to permit generalizations to be made, new varieties developed by traditional crossing appear somewhat more likely to show human toxicity than transgenic varieties. For example, potatoes contain toxic glycoalkaloids that, in most varieties, are at relatively harmless concentrations in the tuber (Friedman and McDonald, 1977). However, the Lenape variety, a Solanum tuberosum x S. chacoense cross developed by traditional methods (Sturckow and Low, 1961Go) for pest resistance, was not released for general planting due to illness caused by ingestion of tubers with high alkaloid content (Zitnack and Johnson, 1970Go). Another potato variety (Magnum Bonum) popular in Sweden was withdrawn from the market for similar reasons (Hellenas et al., 1995Go).

It is also possible to bring enzyme and substrate together in such a way as to produce new and possibly toxic secondary plant chemicals. Again in potato, and by traditional crossing of S. brevidens and S. tuberosum, the progeny were found to contain demissine, a toxic steroidal alkaloid. Apparently, a hydrogenase found in S. brevidens that produces tomatidine from teinamine, produced demissine from solanidine, a compound found in S. tuberosum but not in S. brevidens (Laurila et al., 1996Go).

A new celery variety, developed by traditional genetic crossing and selection for resistance to Fusarium was almost ready for commercial use when it became apparent that it caused severe contact dermatitis in field workers. The cause of the dermatitis, and probably the Fusarium resistance, was the high content of linear furanocoumarins (Diawara and Trumble, 1997Go; Trumble et al., 1990Go).

Although the principal concern with transgenic plants appears to be the possibility of introducing allergenic proteins into food plants, few attempts appear to have been made to define or require rigorous testing protocols. Potential allergenicity is determined largely by homology and stability comparisons with other food allergens. The protein used in these tests is frequently the one expressed in the organism that is the source of the gene, and not the protein expressed in the host plant; this despite the fact that the protein may be modified by secondary processes (e.g., glycosylation) after expression. Thus the Bt toxins Cry1Ab and Cry3A (EPA 1995Go, 1998aGo) are held to be non-allergenic on the grounds that they are not present in high concentrations in food, are not glycosylated by the plant, and are susceptible to gastric digestion. In contrast, Cry9C is regulated as a potential food allergen because it does not degrade rapidly in gastric fluids and is heat stable (EPA, 1998bGo).

It should be recognized that the lack of direct testing and regulation by analogy is a double-edged sword. While food allergies may be avoided, it is also likely that beneficial food products will be lost. Allergenicity represents a great difficulty in hazard analysis. While ideally the tests should involve the immune system or involve an allergic endpoint, prior exposure is necessary for an allergic reaction. An SOT workshop (Kimber et al., 1999Go) clarifies some of the issues surrounding allergenicity testing. First, food allergy is relatively common and can not only have serious clinical manifestations but also may be life threatening. However, food allergens are common in many unmodified food plants so that, no matter what tests are developed and used for transgenic food plants, it will be essential to differentiate allergy resulting from the transgenic protein from that resulting from the proteins of the host plant.

The tiered tests currently used by regulatory agencies for screening for food allergens include protein homology and stability comparisons with known food allergens, and immunoassays for certain classes of antibodies (Kimber et al., 1999Go). However as stated in the NAS/NRC report (NRC, 2000Go):

However, the tests in figure 2.1* either are indirect, do not involve adverse effects, or are otherwise problematic for testing of novel proteins that have not previously been components of the food supply. Indeed, figure 2.1* starts with a decision based on whether or not the protein is derived from a source that is known to be allergenic. This decision can usually be made clearly if the source is a food plant. For transgenic proteins such as Bt endotoxins, making such a comparison would be complicated. If we conservatively choose the "yes" decision, then it would be extremely difficult to complete all of the tests listed because test materials and previously exposed human subjects are not readily available. [*Figure 2.1 not included.]

The importance of food allergy and the potential of transgenic plants to bring food allergens into the food supply should not be minimized. The expression of a brazil nut protein in soybeans resulted in a food allergen being expressed in a widely used food plant, although the variety was not commercialized (Nordlee et al., 1996Go). It is possible, from effects observed in workers using Bt sprays, that Bt endotoxins may have the potential to interact with the human immune system (Bernstein et al., 1999Go) although, even if true, the relationship to transgenic plants and food allergy cannot readily be ascertained.

It is clear that the determination of allergenicity of transgenic proteins by analogy to other food allergens is inadequate, and that tests must be developed that involve the interaction of the transgenic protein in question with the immune system. Given the extensive recent increases in our knowledge of this important system (Selgrade et al., 2001Go), the development of such tests would appear to be well within the capabilities of the scientific community.

Due to concerns over the relatedness of B. thuringiensis to B. cereus, Tayabali and Seligy (2000) tested the effect of Bt insecticidal preparations on a number of human cell types. To quote the authors, "These data, including recent epidemiological work, indicate that spore-containing Bt products have an inherent capacity to lyse human cells in free and interactive forms and may also act as immune sensitizers." Further, they say that "To critically impact at the whole body level, the exposure outcome would have to be an uncontrolled infection arising from intake of Btk/Bti spores."

It is clear that these deleterious effects cannot be related to a single protein, including the Bt proteins that are the object of gene transfer in the creation of pest-protected crop plants. These results do, however, stress the need for rigorous testing in order to allay the public alarm caused by ill-informed attempts at risk communication.

Acute, subchronic, and chronic toxicity are routinely carried out, in the case of synthetic organic chemicals, by feeding, inhalation, or dermal studies, although predominantly by the first of these. Dietary testing of transgene products, or plants expressing transgenes, presents some unique problems, since the compound to be tested will itself be a nutrient and the maximum tolerated dose (MTD) is likely to be very high. This being the case, there may be palatability problems, and appropriate controls may be impossible to devise, inasmuch as the control diet should have the same nutritional properties as the experimental diet. It has been suggested that the best alternative would be to feed the transgenic plant to forage livestock whose normal diet could include the food plant in question, using the most closely related plant variety as the control. In many cases the variety used in the creation of the transgenic plant could be used. While this is a promising approach, considerable work will be needed in order to validate domesticated animals as test organisms, taking into account differences in digestive tract structure and physiology, etc. Another advantage of the use of domesticated animals would be that an MTD need not be determined, inasmuch as the amount necessary for normal growth and development would be both obvious and a logical substitute for an MTD.

To date, no harmful effects on mammalian health have been found by feeding commercialized transgenic crops. Ewen and Pusztai (1999) claimed that changes in the rat gastrointestinal tract were caused by feeding them potatoes containing the Galanthus nivalis agglutinin. However, both the Royal Society (1999) and Kuiper et al., (1999) pointed out significant problems with the experimental design and interpretation, and it appeared clear that any differences found, even if subsequently validated, could be attributed to variations between potato lines rather than genetic modification.

Dose-Response Evaluation, Exposure Assessment, Risk Characterization, Risk Communication and Risk Management
Since adequate data are not available for appropriate dose-response and exposure assessment, it is not yet possible to provide appropriate risk characterization, as that term is understood by the scientific community. In the absence of clearly defined toxic endpoints, dose-response data cannot be obtained and the problem of obtaining exposure data is daunting. The use of food consumption databases will give unreasonably high values if, for example, corn consumption is equated with transgenic corn consumption or if all transgenes are considered equivalent. Given the virtual impossibility of developing sound risk characterization using traditional methods, it may be that new risk paradigms will need to be developed to deal with the assessment of risks to human health from transgenic food plants. Risk communication has been left largely in the hands of nonscientists although both Science and Chemical and Engineering News have performed well in the cause of bringing dispassionate reporting to this contentious issue.


    NOTES
 
1 To whom correspondence should be addressed at the Department of Toxicology, 850 Main Campus Drive, Box 7633, NCSU, Raleigh, NC 27695. Fax: (919) 513-1012. E-mail: ernest_hodgson{at}ncsu.edu. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 REFERENCES
 
Anon (2000). Aventis presents new corn evaluation to EPA. Chem. Eng. News, October 30th, 2000, 11.

Bernstein, I. L., Bernstein, J. A., Miller, M., Tierzieva, S., Bernstein, D. I., Lummus, Z. Selgrade, M. K., Doerfler, D. L., and Seligy, V. L. (1999). Immune responses in farm workers after exposure to Bacillus thuringiensis pesticides. Environ. Health Perspect. 107, 575–582.[ISI][Medline]

Diawara, M. M., and Trumble, J. T. (1997). Linear furanocoumarins. In Handbook of Plant and Fungal Toxicants (J. P. D'Mello, Ed.), pp. 175–188. CRC Press, Boca Raton, FL.

EPA (1995). Pesticide fact sheet: Plant pesticide Bacillus subsp. tenebrionis delta endotoxin and its controlling sequences in potato. U.S. Environmental Protection Agency.

EPA (1998a). Pesticide fact sheet: Bacillus thuringiensis subsp. kurstaki CryIAc delta endotoxin and the genetic material necessary for its production in corn. U.S. Environmental Protection Agency.

EPA (1998b). Pesticide fact sheet: Bacillus thuringiensis subsp tolworthi Cry9 protein and the genetic material necessary for its production in corn. U.S. Environmental Protection Agency.

Ewen, S. W. B., and Pusztai, A. (1999). Effect of diets containing genetically modified potatoes expressing Galanthus nivalis lectin on rat small intestine. Lancet 354, 1353–1354.[ISI][Medline]

Friedman, M., and McDonald, G. M. (1997). Potato glycoalkaloids: Chemistry, analysis, safety, and plant physiology. Crit. Rev. Plant Sci. 16, 55–132.[ISI]

Goldman, K. A. (2000). Bioengineered food—safety and labeling. Science 290, 457–459.[Free Full Text]

Hellenas, K. E., Branzell, C., Johnsson, H., and Slanina, P. (1995). High levels of glycoalkaloids in the established Swedish potato variety Magnum Bonum. J. Sci. Food Agric. 23, 520–523.

Hodgson, E., and Levi, P. E. (1997). A Textbook of Modern Toxicology, 2nd ed. Appleton and Lange, Stamford, CT.

Kaiser, J. (2000). Panel urges further study of biotech corn. Science 290, 1867.

Kimber, I., Kerkvliet, N. I., Taylor, S. L., Astwood, J. D., Sarlo, K., and Dearman, R. J. (1999). Toxicology of protein allergenicity: Prediction and characterization. Toxicol. Sci. 48, 157–162.[Abstract]

Kuiper, H. A., Noteborn, H. P., and Peijnenburg, A. A. M. (1999). Adequacy of methods for testing the safety of genetically modified foods. Lancet 354, 1315–1316.[ISI][Medline]

Laurila, J., Lasko, I., Valkonen, J. P. T., Hiltunen, R., and Pehu, E. (1996). Formation of parental type and novel glycoalkaloids in somatic hybrids between Solanum brevidens and S. tuberosum. Plant Sci. 118, 145–155.

NRC (1983). Risk Assessment in the Federal Goverment. National Academy Press, Washington DC.

NRC (2000). Genetically Modified Pest Protected Plants: Science and Regulation. National Academy Press, Washington, DC.

Nordlee, J. A., Taylor, S. L., Townsend, J. A., Thomas, L. A., and Bush, R. K. (1996). Identification of a brazil-nut allergen in transgenic soybeans. N. Engl. J. Med. 334, 688–692.[Abstract/Free Full Text]

Royal Society (1999). Review of data on possible toxicity of GM potatoes. Available at http://www.royalsoc.ac.uk/st_pol54.htm. Accessed on January 5, 2000.

Selgrade, M. K., Germolec, D. R., Luebke, R. W., Smialowicz, R. J., Ward, M. D. and Sailstad, D. M. (2001). Immunotoxicity. In Introduction to Biochemical Toxicology (E. Hodgson and R.C. Smart, Eds.), pp. 561–597. John Wiley and Sons, New York.

Senti, F. R., and Rizek, R. L. (1974). An overview on GRAS regulations and their effect from the point of view of nutrition. In The Effect of FDA Regulations (GRAS) on Plant Breeding and Processing, Special Publication No. 5, pp. 7–20. Crop Science Society of America, Madison, WI.

Sturckow, B., and Low, I. (1961). The effects of some Solanum glycoalkaloids on the potato beetle. Entomol. Exp. Appl. 4, 133–142.

Tayabali, A. F., and Seligy, V. L. (2000). Human cell exposure assays of Bacillis thuringiensis commercial insecticides: Production of Bacillus cereus-like cytolytic effects from outgrowth of spores. Environ. Health Perspect. 108, 919–930.

Thayer, A. (2001). StarLink corn derails Ag chain. Chem. Eng. News, January 22, 2001, 23–33.

Trumble, J. T., Dercks, W., Quiros, C. F., and Beier, R. C. (1990). Host plant resistance and linear furanocoumarin content of Apium accessions. J. Econ. Entomol. 83, 519–525.[ISI][Medline]

Zitnack, A., and Johnson, G. R. (1970). Glycoalkaloid content of B5141–6 potatoes. Am. Potato J. 47, 256–260.[ISI]