3

Crossroads of Science and Oversight

This chapter focuses on the scientific basis of the oversight of transgenic pest-protected plants. The committee recognizes that there is an urgency to solidify the regulatory framework for transgenic pest-protected plant products because of the potential diversity of novel traits that could be introduced by transgenic methods and because of the rapid rate of adoption of and public controv1ersy regarding transgenic products.

For comparison with transgenic pest-protected plants, a case study concerning conventional pest-protected plants and a discussion of scientific issues surrounding them are presented. Then case studies of transgenic pest-protected plants are discussed; these case studies provide examples of scientific review of transgenic pest-protected plants by federal agencies.

After the case studies, EPA's proposed rule for the regulation of genetically modified pest-protected (GMPP) plant gene products as plant-pesticides is then evaluated in light of the discussion and other scientific criteria. Finally, the committee suggests guiding scientific principles for oversight of transgenic pest-protected plants and sets forth specific research needs.

3.1 CASE STUDIES OF PEST-PROTECTED CROPS AND THEIR OVERSIGHT

3.1.1 Conventional Breeding for Rust Resistance in Wheat

Three main rust diseases affect common wheat (Triticum aestivum L.) and durum wheat (T. durum Desf.): stem rust (or black rust), caused by



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION 3 Crossroads of Science and Oversight This chapter focuses on the scientific basis of the oversight of transgenic pest-protected plants. The committee recognizes that there is an urgency to solidify the regulatory framework for transgenic pest-protected plant products because of the potential diversity of novel traits that could be introduced by transgenic methods and because of the rapid rate of adoption of and public controv1ersy regarding transgenic products. For comparison with transgenic pest-protected plants, a case study concerning conventional pest-protected plants and a discussion of scientific issues surrounding them are presented. Then case studies of transgenic pest-protected plants are discussed; these case studies provide examples of scientific review of transgenic pest-protected plants by federal agencies. After the case studies, EPA's proposed rule for the regulation of genetically modified pest-protected (GMPP) plant gene products as plant-pesticides is then evaluated in light of the discussion and other scientific criteria. Finally, the committee suggests guiding scientific principles for oversight of transgenic pest-protected plants and sets forth specific research needs. 3.1 CASE STUDIES OF PEST-PROTECTED CROPS AND THEIR OVERSIGHT 3.1.1 Conventional Breeding for Rust Resistance in Wheat Three main rust diseases affect common wheat (Triticum aestivum L.) and durum wheat (T. durum Desf.): stem rust (or black rust), caused by

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION Puccinia graminis Pers. f. sp. tritici Eriks and Henn.; leaf rust (or brown rust), caused by P. recondita Rob. Ex Desm. f. sp. Tritici; and stripe rust (or yellow rust), caused by P. striiformis West. All three rust diseases are fungi which are obligate parasites in nature (that is, they require a living host to survive). They all need free moisture for infection, but they have different optimal environmental conditions for disease development, so they often do not damage wheat production concurrently in the same region (Knott 1989). For example, stem rust tends to require higher temperatures than leaf rust, which requires higher temperatures than stripe rust; hence, stem rust is usually more damaging in the northern Great Plains, leaf rust in the southern Great Plains and the East, and stripe rust in the West. Of the three diseases, stem rust can cause the more devastating epidemics; for example, in 1916, a stem rust epidemic was estimated to have reduced total US wheat production by 38% (Loegering 1967). But leaf rust is more common (Schafer 1987; table 3.1). Because wheat is used primarily as a food grain, losses in total production underestimate the true economic loss when wheat is damaged so severely that it must be sold as a feed grain. Research to Reduce Losses Caused by Wheat Rusts Losses due to rusts can be reduced by cultural practices, removal of an alternative host, chemical control, and genetic protection from the pathogens (also called host-plant resistance) (Knott 1989; Schafer 1987). The goal is to break the life cycle of the pathogen. After the 1916 stem rust epidemic, a major barberry-eradication program was started in North America (Roelfs 1982). Roelfs suggested four benefits from the success of the program: disease onset was delayed by 10 days, initial inoculum was reduced, the number of pathogen races was reduced, and the pathogenic races of stem rust were stabilized. The reduction in the number of races and their stabilization were due to eliminating the sexual cycle of P. graminis. P. recondita also has alternative hosts, but none is known for P. striiformis (Schafer 1987). Chemical control of rust diseases with fungicides has been successful, but the cost of the fungicides, the economics of US wheat production, and concerns about chemicals in food grains have limited their use in the United States (Rowell 1985). Fungicides are widely used in Europe and the Pacific Northwest to control other wheat diseases. By far the most common approach to the control of rust diseases in wheat is the use of conventionally bred resistant plants because it costs less than fungicide applications and there are numerous sources of genes for pest-protection (for example, McVey 1990; Cox et al. 1994). The most important aspect of breeding for rust-protection is that not only the genet-

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION TABLE 3.1 Wheat Yield Losses Due to Stem, Leaf, and Stripe Rust in United States, 1995-1998     Yield loss, % of harvested bushels     Common Year Disease Winter Spring Durum 1995 Stem Rust 0.01 0.01 0.0   Leaf Rust 2.36 0.10 0.0   Stripe Rust 0.11 0.03 0.0 1996 Stem Rust 0.23 0.00 0.0   Leaf Rust 0.78 0.03 0.0   Stripe Rust 0.26 0.05 0.0 1997 Stem Rust 0.00 0.02 0.0   Leaf Rust 2.85 1.10 0.0   Stripe Rust 0.07 0.04 0.0 1998 Stem Rust 0.09 0.03 0.0   Leaf Rust 1.60 0.83 0.0   Stripe Rust 0.27 0.17 0.0 Average, 1995-1998 Stem Rust 0.08 0.01 0.0   Leaf Rust 1.90 0.52 0.0   Stripe Rust 0.18 0.07 0.0 Source: USDA (1999g). ics of the host but also the genetics of the pathogen must be considered. Both are subject to change—the pathogen by mutation and sexual hybridization, the host by plant breeding. Because of changes in the pathogen, protective genes in the host are overcome by new virulence genes in the pathogen. Kilpatrick (1975) used an international testing program to estimate that the average lifetime of a gene for protection from leaf, stem, or stripe rust was 5-6 years. The rapid loss of genetic pest-protection due to new virulence genes led researchers to look for new protective genes and for durable resistance (for example, Line 1995). R. Johnson (1984) has defined durable resistance as the “resistance that remains effective during prolonged and widespread use in an environment favorable to the disease.” Considering the definition, genes for durable resistance are identified only after they have been deployed in widely grown cultivars. New genes for pest-protection are constantly being searched for in wheat and its wild relatives, and plant breeders try to create new combi-

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION nations of protective genes. Increasingly, new genes are being identified and transferred from wild relatives of wheat (for example, Cox et al. 1993 and 1994; Sharma and Gill 1983). Wild relatives of cultivated plants have coevolved with the crop pathogens and so are often extremely useful sources of protective genes (Leppik 1970; Wahl et al. 1984). Although many of the wild relatives of wheat are Triticum spp., many are more distant (McIntosh et al. 1995). For example, the protective genes Lr24 and Sr24 came from tall wheat grass, Thinopyrum ponticum (Podp.) (Barkw. & Dewey); and Lr26, Sr31, and Yr9 came from rye, Secale cereale L. Little is known about the biochemistry of genetically based rust-protection, so most breeding programs use phenotypic selection (the presence or absence of the disease) and some use molecular markers to track protective genes. The main phases of any wheat-breeding program are introduction of genetic variation, inbreeding and selection of useful variants, and extensive field testing of selected variants to determine their agronomic or commercial worth (Baenziger and Peterson 1992). All the standard plant breeding methods are well documented (for example, Fehr 1987; Stoskopf et al. 1993), as are the methods specifically applied to breeding for rust-protection (Knott 1989; MacIntosh and Brown 1997). The most common breeding method for moving one or a few genes into an elite line or cultivar, especially when the genes are being transferred from a wild relative or an unadapted line, is backcrossing. It has been widely used to introduce protective genes into cultivated wheat whether those genes are derived from Triticum spp. or from more distant but sexually compatible relatives. The Agricultural Result As mentioned previously, the effect of rusts can be devastating when susceptible wheat cultivars are grown. However, estimating the value of crop resistance to rust accurately is difficult because the widespread growth of resistant cultivars affect the yield-loss estimates. A well-documented estimate (based on potential yield losses due to the disease-infecting susceptible cultivars) of the annual value of having stem rust resistance in wheat grown in western Canada was Can$217,000,000 (Green and Campbell 1979). The annual yield losses due to stem rust have averaged between 15% in Saskatchewan to 25% in Manitoba. in the United States, epidemics are localized, but the yield losses due to plant susceptibility to stem rust were as high as 56.5% in North Dakota and 51.6% in Minnesota in 1935 (Roelfs 1979). For leaf rust, the yield losses due to susceptibility were estimated at 50% in Georgia in 1972. For stripe rust, yield losses due to susceptibility were estimated at 30% in 1961 in Washington. The losses

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION should be viewed as high estimates because the weather conditions, pathogen, and host susceptibility were optimal. The average annual US losses are probably more similar to those estimated by Green and Campbell for Canada (1979). The three rusts combined reduce the annual US wheat crop by about 2% (table 3.1); most of the losses are caused by leaf rust. The low level of rust losses is attributable mainly to the use of resistant cultivars. Health and Environmental Impacts No formal assessment of the health or environmental impact of conventionally breeding wheat for resistance to rust has been undertaken by regulatory agencies, inasmuch as the products of conventional plant breeding have generally not required their oversight. Rust-resistant wheat cultivars, regardless of the source of their resistance genes, have been widely grown and consumed in food products with no history of causing health problems. Little is known about the biochemical basis or gene products for plant protection against rust, but these genes are likely to be similar to other genes in the large class of race-specific resistance genes isolated from other plants. The presence of pest-resistance genes can affect end-use quality by affecting the grain protein content. For example, leaf rust detrimentally affects leaves reducing their potential for nitrogen remobilization to the grain and reducing grain protein content (Cox et al. 1997). Lower grain protein content is generally considered a detrimental effect in hard wheats but a beneficial effect in soft wheats (Finney et al. 1987). Stem rust tends to reduce the flow of photosynthate and nitrogen to the grain; but because nitrogen is mobilized early in grain development, the overall result of stem rust is generally an increase in protein content in the grain, possibly including shriveled kernels. Environmentally, the use of rust-resistant wheat cultivars has reduced the use of fungicides, but the extent of this reduction is not well documented, because effective pest-protection has been widely deployed for many years and the economics of wheat production often preclude the widespread use of fungicides that are effective against rust. 3.1.2 Bt Crops The most widely used transgenic pest-protected plants are cultivars that express insecticidal proteins derived from the bacterium Bacillus thuringiensis (Bt). Cotton and corn are protected from some of their lepidopteran pests by Bt proteins in the Cry1A and Cry9C groups. The po-

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION tato cultivars are protected against the Colorado potato beetle by a Cry3 Bt protein. There is a tendency to consider Bt toxins as all biochemically similar, but the DNA sequence similarity among toxins can be less than 25% (Feitelson et al. 1992) and the biochemical properties of the more than 100 different Bt toxins vary widely. Potato Transgenic potato was the first Bt crop variety approved for commercial use (EPA 1995a). The target pest for transgenic Bt potatoes is the Colorado potato beetle. This pest is not a major problem in all areas of potato production, but the need for an alternative to conventional insecticides for controlling it by conventional farming techniques was apparent in the years before approval because the beetle had become resistant to all available classes of conventional insecticides. Just as Bt potatoes reached the market, a novel insecticide, imidicloprid, also reached the market. The new insecticide was so effective on a number of potato pests that it competed effectively with Bt potatoes that controlled only the beetle pest. In 1998, Bt potatoes in the United States were planted in 50,000 acres, which is 3.5% of the total US potato acreage (Idaho Statesman 1998). Strains of Colorado potato beetle resistant to imidicloprid are already evolving in a number of locations (for example, Suffolk County, NY), so Bt potatoes may soon be planted on a much larger scale. However, low rates of adoption of Bt potato may ultimately be due to the need for potato growers to use chemicals to control insect pests other than the Colorado potato beetle. In those cases, protection from the Colorado potato beetle may not offset the cost of the chemicals and the transgenic seed (Gianessi and Carpenter 1999). Some of the Cry3 protein produced in Bt potatoes is coded from the full length bacterial gene. However, a significant fraction of the Cry3 toxin produced in Bt potato is a smaller, truncated form of molecule (Perlak et al. 1993). EPA documentation indicates that the potential for the smaller Cry3 molecule to induce a food allergy is similar to that of the larger Cry3 molecule. EPA (1995c) indicates that “despite decades of widespread use of Bacillus thuringiensis as a pesticide there have been no confirmed reports of immediate or delayed allergic reactions from exposure”. Bt toxin's history of use (microbial Bt sprays have been registered since 1961) and rapid digestion in simulated gastric fluids (in less than 30 seconds; EPA 1995c) are considered evidence of safety by the EPA. In addition, in acute toxicity studies, no adverse effects were exhibited (Lavrik et al. 1995) (see section 3.1.3). However, it must be recognized that the microbial Bt toxins that have been widely used for decades to

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION control lepidopteran pests have less than 60% molecular similarity to the toxin which is active against Colorado potato beetle (Feitelson et al. 1992). This beetle-active toxin has only been used in spray form since the mid-1980s on limited acreage of potatoes. Also, the sprayed beetle toxin is not applied to the tuber; whereas, in the transgenic potato varieties, it is present in the tuber (Rogan et al. 1993). As indicated in section 2.6, field studies that compared the biodiversity of insects in fields with transgenic pest-protected potatoes and in fields with nontransgenic potatoes treated with synthetic insecticides found higher densities of above ground beneficial arthropods in the transgenic Bt fields. USDA referred to the findings in its positive response to Monsanto 's request for nonregulated status of transgenic Bt potato (USDA 1995a). EPA's pesticide fact sheet for Bt potatoes (EPA 1995a) did not refer to those field data but concluded that no negative ecological effects of Bt potatoes were expected on the basis of a series of laboratory tests conducted by Monsanto (for example, Sims 1993; Keck and Sims 1993). Details of methods and results of the laboratory tests were voluntarily provided to the committee by Monsanto. Examination of this information indicated that most of the procedures and conclusions were valid. However, in some cases, the approach to testing seemed inefficient. For example, tests for nontarget effects on honeybee larvae used a bioassay in which 5 uL of a Bt-toxin solution was pipetted into the bottom of larval cells and observations for potential mortality were made (Maggi 1993b); this approach would be better for a contact toxin than for toxins such as Bt toxin, which must be ingested. In the tests for adult honeybees (Maggi 1993a), the amount ingested was estimated by weighing the solution before and after presentation to honeybees and controlling for evaporative loss; however in the larval study (Maggi 1993b) the amount ingested was not estimated and it is unclear how much of the solution was consumed by the larvae. A positive control (that is, a group of larvae presented with a solution that will definitely kill them) was not mentioned in the study provided to the committee. For some other tests, the absence of information made interpretation of results less clear. For example, tests with ladybird beetles used adults and provided the Bt toxin in a honey solution (Hoxter and Smith 1993). Consumption was measured by comparing the weight of the test diet before and after presentation to the beetles. However, the measurement of consumption was not useful, because there was no control for evaporative loss. Without knowledge of the amount consumed, it would be better to gather data on egg production which is more sensitive to stress. Tests of larval ladybird beetles would also offer a more sensitive toxicity test for Bt. The ladybird beetle test and other tests for EPA used Bt toxin produced by bacteria instead of plants. In some of the nontarget testing, that

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION seems to have been done to increase the toxin concentration to more than 100 times the concentration in pollen1 or nectar. That is justified, but it would also be ecologically relevant to determine effects of the actual pollen and nectar produced by the plants under field conditions. In soil-degradation studies in which the bacterially produced toxin is used at concentrations that could be obtained from the plant itself, there seems to be less justification for not using the plant itself. It is surprising that in the Bt soil-degradation studies, either bacterially produced toxin or freezedried and highly pulverized Bt-potato plant material is used (Keck and Sims 1993). An ecologically more realistic approach was used in peer-reviewed studies by Donegan et al. (1995) and by Palm et al. (1996): where Bt and non-Bt plant material was placed in the field and monitored for decomposition, microbial diversity, and Bt-toxin titer. Donegan et al. (1996) also used an ecologically realistic system to test for differences in rhizosphere and leaf-dwelling microorganisms associated with field-grown transgenic Bt and non-Bt potatoes. No biologically significant differences were found. Similar tests would be valuable in regulatory assessments. Overall, the data presented to EPA by Monsanto demonstrate that the transgenic Bt potatoes are likely to be environmentally much less disruptive than current chemical control practices against the Colorado potato beetle. The concentration of toxin produced in the foliage (19.1 µg/g ) of Bt potatoes (EPA 1995a) far exceeds the concentration needed to kill young Colorado potato beetles (Perlak et al. 1990 and 1993). This level of toxin is expected to fit the EPA Scientific Advisory Panel's (SAP 1998) definition of a “high dose” that can be useful in delaying evolution of resistant pest strains (see section 2.9). The concentration of toxin in the tubers themselves is low (1.01 µg/gm), but potato beetles do not typically feed on the tubers. Corn Unlike the Colorado potato beetle, which can devastate potato production in some areas when not controlled with insecticides, the European corn borer, which is the major target of transgenic Bt field corn, has not commonly been controlled with insecticides. A survey of the literature (Gianessi and Carpenter 1999) indicates that across the corn belt only 1   Some varieties of potatoes (for example, Russet Burbank) produce very little pollen, so the discussion of pollen refers to other potato cultivars that have been commericialized (for example, Superior).

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION 5.2% of the acreage is sprayed annually for corn borers and in Iowa only 2.6%. Some of the reasons for the lack of chemical control are that the perceived yield loss has always been considered small (estimated at about 4%), the cost of pesticides is high relative to the crop 's value, and typical insecticides have not been very efficient at killing the pest after it bores into the plant. In addition to the European corn borer, other insect pests are targeted by the Bt corn cultivars. For example, the southwestern corn borer, a major corn pest in some areas of the Midwest, can be controlled by Bt corn; and the corn earworm, a minor pest of field corn but a major pest of sweet corn, is also a potential target. A recent USDA study (1999d) indicates that in two of five regions mean yields were significantly higher with Bt corn than with non-Bt corn. How the increase in yield will affect farmers' profits is not evident, given increased seed cost and the increased potential for higher national production of corn which could lead to a decrease in prices. At least three Bt toxins are produced by commercial transgenic pest-protected corn cultivars. The most common, Cry1Ab, is produced either as a full-length protoxin, as produced in B. thuringiensis (Monsanto variety), or as a truncated, preactivated toxin (Novartis variety). In nature, the 130 to 140 kilodalton protoxin is converted to a 60-65 kilodalton protein when the target insect ingests Bt (Federici 1998). Cry1Ac toxin is produced by the Dekalb cultivars; and a biochemically distinct Bt toxin, Cry9C, is produced by corn developed by AgrEvo (EPA 1998c). All the corn cultivars that produce Cry1A toxins have been approved for human consumption, but currently Cry9C corn has been approved only for cattle feed (EPA 1998c). That restriction by EPA has been established because the Cry9C toxin, unlike the Cry1A toxins, does not degrade rapidly in gastric fluids and is relatively more heat-stable; these characteristics of Cry9C raise concerns including those of allergenicity. Novartis-produced Bt corn does not produce detectable levels of Bt toxin in the silks or corn kernels, so it does not effect the corn earworm, which feeds mostly on these two plant parts. The Cry9C toxin in AgrEvo corn is produced in the silk and corn ear, but it is not toxic to the corn earworm. The biological complexity of current Bt corn products is much greater than that of Bt potato and cotton; for example, only one company has commercialized Bt potato, and only one toxin type and seasonal tissue distribution are exhibited for each crop species. Environmental impacts of Bt field corn must be judged against the typical corn system in which no insecticides are applied to control lepidopterans. Peer-reviewed studies have demonstrated adverse effects of Cry1Ab on predaceous lacewings (Hilbeck et al. 1998a, b), however, none of the studies conducted for EPA by the registrant has found adverse

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION effects on lacewings or other aboveground beneficial insects (EPA 1997a and 1998a). It is difficult to reconcile the different findings of the studies conducted for EPA by Monsanto (for Cry1Ac and Cry1Ab toxins) and the studies by Hillbeck and colleagues (Cry1Ab). In the first study by Hilbeck et al. (1998a), lacewings were fed on small larvae of Bt-sensitive and Bt-nonsensitive herbivores that had eaten vegetative-stage Bt or non-Bt corn. The concentration of toxin to which the lacewings were exposed could have been above the 50 parts per million (50 ppm) expected in an ecologically realistic system. A total of 200 lacewings were used per treatment. The second Hillbeck et al. study (1998b) fed larvae purified bacterially-produced Bt at a concentration of 100 ppm in an artificial diet. In the Monsanto studies, the concentration of toxin was 20 ppm and involved coating lepidopteran eggs with bacterially produced toxin (Hoxter and Lynn 1992). In each Monsanto study, 30 lacewings were used per treatment. The Hilbeck et al. (1998a) and Monsanto studies followed larvae to pupation. The Hilbeck studies found more than a 50% increase in mortality; the Monsanto studies found no difference in mortality or lower mortality associated with Bt treatment. Because lacewings typically feed only on the internal content of the eggs, they may not have ingested much of the toxin which was deposited on the shells of the eggs in the Monsanto study. Given that Bt corn is already planted over millions of acres in the United States, it seems appropriate for EPA, USDA, or registrants to sponsor careful field tests to determine whether lacewings or other natural enemies of crop pests are adversely affected by Bt corn. One preliminary study of this type found no differences between Bt and non-Bt corn in effects on any natural enemies of crop pests (Pilcher et al. 1997), but more detailed studies would be useful. Likewise, the committee recommends that EPA should provide guidelines for determining the most ecologically relevant test organisms and test procedures for assessing nontarget effects in specific cropping systems. Peer-reviewed studies (for example, MacIntosh et al. 1990) demonstrated that the Bt toxin in corn could affect many lepidopteran species. A laboratory study showed that pollen from some Bt corn cultivars can kill and slow growth of monarch caterpillar larvae if enough pollen is placed on the milkweed leaves fed to the caterpillars (Losey et al. 1999)(see section 2.6.2). If monarchs are indeed being killed in nature by this pollen, the non-Bt corn planted as a refuge for susceptible pest insects could be planted around the edges of corn fields so that adjacent milkweed would be dusted only with pollen from non-Bt corn. It might also be possible to

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION shift to using Bt corn which does not produce biologically significant amounts of Bt in its pollen (EPA 1998b; Andow and Hutchinson 1998). The potential for resistance to Bt toxin in a number of the target pests of Bt corn has been of concern to EPA, environmentalists, and university researchers (Ostlie et al. 1997; Andow and Hutchison 1998; Matten 1998; SAP 1998). In particular, the corn earworm may be especially vulnerable to evolving Bt resistance. Corn earworm is substantially less sensitive to Bt toxins than the primary target pest of Bt corn, the European corn borer. Bt corn varieties that express the toxin in the silks or corn kernels where corn earworm feed do not produce a high enough dose for corn earworm mortality. Corn earworm is also subject to selection pressure from Bt toxins in Bt cotton, since this pest feeds on a number of crops, including cotton, where it is known as the cotton bollworm (EPA and USDA 1999). All the commercial cultivars provide substantial protection against the European corn borer (Ostlie et al. 1997). However, the Novartisproduced cultivars, which use green tissue and pollen-specific promoters to drive gene expression, have lower efficacy later in the season (Ostlie et al. 1997; Andow and Hutchison 1998). The lower late-season efficacy is also seen in the Dekalb-produced corn (Andow and Hutchison 1998). Lack of a high dose in these two types of Bt corn could undermine the high-dose refuge approach endorsed by EPA (Matten 1998; and section 2.9) and achievable with other Bt corn cultivars. Cotton Like potatoes, conventionally-grown cotton has been heavily treated with insecticides to control lepidopteran pests. Therefore, the introduction of Bt cotton can produce considerable environmental benefits. A 1998 survey indicated a general decrease in insecticide useage on Bt cotton (Mullins and Mills 1999). For example, in 66 comparisons in the Mississippi, Louisiana, and Arkansas region, the average number of insecticide sprays per field was 10.1 for non-Bt cotton and 7.9 for Bt cotton. Many of these insecticide treatments were made to control the boll weevil which is not affected by Bt. In 20 comparisons in the North Carolina, South Carolina, and Virginia region (where the boll weevil is not a pest), the average number of insecticide sprays was 3.7 for non-Bt cotton and 1.2 for Bt cotton. USDA's Economic Research Service found less clear patterns in changes in insecticide used on Bt cotton (USDA 1999d). Comparison of mean pesticide acre-treatments for 1997 showed that in only two of three regions surveyed did the adoption of Bt cotton reduce insecticide treatments normally used to control pests targeted by Bt. In one of three regions, total insecticide treatments for all other pests was higher for Bt adopters than for nonadopters (USDA 1999d). The results should be

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION of a toxicant) and given the public controversy regarding transgenic products, EPA should reconsider its categorical exemption of transgenic pest-protectants derived from sexually compatible plants. 3.2.2 Exemption of Viral Coat Proteins In addition to exempting plant-pesticides derived from sexually compatible plants, the 1994 and 1997 EPA documents propose a number of more specific exemptions. EPA generally provides more reasonable scientific justification for these exemptions. One specific class of plant products that was proposed for categorical exemption was viral coat proteins (VCPs). VCPs are already present in foods because of natural virus infections of crops and have not caused obvious medical problems, so health concerns are considered minimal. The EPA exemption of VCPs is also based on considerations that “include the low potential for adverse effects to nontarget organisms and the potential benefits (environmental and economic) of utilizing VCP (virus coat protein) mediated resistance.” The committee, in general, agrees with this assessment of the minimal health and nontarget effects posed by VCP expression in crop plants (see also section 3.1.4) and concludes that Viral coat proteins in transgenic pest-protected plants are not expected to jeopardize human health because consumers already ingest these compounds in nontransgenic food. However, the committee questions the categorical exemption of all viral coat proteins under FIFRA due to concerns about outcrossing with weedy relatives. Although ecological concerns are discussed and a more restrictive exemption that considers outcrossing is presented, the proposed rule favors complete exemption of VCPs. EPA should not categorically exempt viral coat proteins from regulation under FIFRA. Rather, EPA should adopt an approach, such as the agency 's alternative proposal (as stated below in Option 2), that allows the agency to consider the gene transfer risks associated with the introduction of viral coat proteins to plants. Option 2: Exemption of coat proteins form plant viruses produced in plant with low potential for outcrossing to wild relatives. Under this exemption the Agency would limit its exemption of VCP-mediated resistance coat proteins to those viral coat protein/plant combinations that would have the least potential to confer selective advantage on free-living wild relatives.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION In section 2.7 and section 3.1.4, the committee explains why the more restrictive exemption should be considered. 3.2.3 Exemption for Nontoxic Modes of Action The 1994 EPA document requested comments on a proposal to exempt plant-pesticides that acted primarily by affecting plants and “that act through nontoxic modes of action.” The types of substances that clearly are in this category are structural barriers such as plant hairs; substances that inactivate or resist toxins that are produced by pests; and substances that decrease chemical components needed for pest growth. As discussed in chapter 2 (section 2.4 and section 2.5), these exemptions are unlikely to result in any new human exposure to harmful substances. However, within the same category the 1994 EPA document also discusses exempting plant hormones. Plant hormones often cause multiple changes in plants, including changes in secondary metabolites that might be toxic, so the scientific basis of such an exemption is questionable. As with the exemption of VCPs, the categorical exemption of substances that act through nontoxic modes of action mostly considers human health effects. As outlined in previous sections of our report ( section 2.6 and section 2.7) there is a need to consider separately the impact of such substances on nontarget species and the potential for the genes that code for these substances to move to feral populations or weedy relatives of the crop, where they could increase recipient plants' fitness. Categorical exemption under FIFRA might not be scientifically justifiable. 3.2.4 Oversight for Pleiotropic Effects The 1994 EPA document states that any food safety questions beyond those associated with the plant-pesticide, such as those involving changes to food quality or raised by unexpected or unintended compositional changes, are under FDA's jurisdiction. Similarly, food safety issues associated with alterations in levels of a substance with pesticidal properties, or the appearance of a substance with pesticidal properties, that occur as an unintended consequence of modifications to a non-pesticidal trait would also fall under FDA's authority. That is an important statement and shifts an important component of pest-protected plant assessment to FDA. As discussed previously in this report (section 2.4.1 and section 2.5.2), genetic changes that result in production of a specific plant protectant can result in production of biologically active compounds other than the in-

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION tended plant protectants. Such pleiotropic effects are sometimes difficult to predict. Furthermore, as outlined in previous (section 2.4.2 and section 2.5), many approaches to producing plant protectants through the use of plants that are sexually compatible with the crop plant can result in crops that produce new compounds owing to linkage between the genes for the plant protectants and genes for the other compounds. FDA needs to address these “unintended compositional changes” carefully during their consultation process with the plant producers. USDA and EPA should also be aware of those unintended changes in evaluating the potential agricultural and ecological effects of pest-protected plants. The committee recommends that EPA, FDA, and USDA collaborate on the establishment of a database on natural plant compounds of potential dietary or other toxicologic concern. The database would be publicly available and updated regularly. The following guidelines should be considered: initial emphasis should be on obtaining baseline profiles for food plants that are known to have toxic constituents and on the commonest varieties; differences among varieties, developmental stages, tissues and environmental conditions are important and should be analyzed after initial average baselines have been established; only information based on state-of-the-art chemistry and analytic methods should be incorporated; and potential information should be peer-reviewed by a committee of experts before it is added to the database (see also section 3.4.1). 3.3 SUGGESTED QUESTIONS FOR OVERSIGHT Given the above concerns with the scientific basis of proposed oversight, the committee proposes that federal agencies use the following questions as a guide in developing their review process. These decision keys leave sufficient room for agency judgment case by case. For the most part, the agencies are following a similar logic in their decision-making, but there are some points where current decision-making does not agree with the following questions; these discrepancies are pointed out in the text. Because the Coordinated Framework for the Regulation of Biotechnology was designed for transgenic products (see chapter 1) and the agencies do not actively assess conventional pest-protected plant products, the following questions focus on transgenic pest-protected plant products. However, the questions could be adapted and applied to nonregulatory safety assessments of conventional pest-protected plants, as the underly-

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION ing concerns are not dependent on the method used to produce the plant (section 2.2.1). 3.3.1 Health Concerns: Guiding Principles The principles in the following questions could be used to determine when a detailed analysis of health risks is warranted for transgenic pest-protected plants. Is the substance found in plant parts that consumers3 eat or workers come into contact with? Yes or Unknown—go to 2. No—exempt from health concerns. Is the substance known to have general chemical and physical properties common to many allergens? Note: Criteria outlined in figure 2-1 could offer components for this type of evaluation. Yes or Unknown—subject to safety assessment. No—go to 3. Is the substance similar to substances that people now eat or come into contact with, and can confident predictions of safety based on the similarities be made? Yes—go to 4. No or Unknown—subject to safety assessment. Is the expected exposure to the substance substantially greater than current exposures? Yes or Unknown—subject to safety assessment. No—go to 5. Is there a reasonable chance, based on known properties of the substances, that its production will lead to harmful concentrations of toxicants or allergens that consumers eat or workers come into contact with? 4 Yes or Unknown—subject to safety assessment. No—exempt from health concerns. 3   Including human and non-human consumers, such as food animals or pets. 4   Pleiotropic effects.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION EPA exempts from FFDCA and FIFRA pesticidal substances in transgenic pest-protected plants that are derived from transgenes from sexually compatible species. The committee's questions are not in accordance with that categorical exemption. Given that transfer and manipulation of genes between sexually compatible plants could potentially result in adverse effects (for example, modulation of a pathway increases the concentration of a toxicant), the categorical exemption of pest protectants solely on the basis of derivation from sexually compatible plants could be scientifically unsound in some cases. FDA's policy for foods derived from new plant varieties is designed to address questions 1 through 5 with respect to dietary exposure to substances that are not regulated by EPA as pesticides. For pesticidal substances, EPA may consult with FDA on allergenicity issues (see chapter 4). 3.3.2 Ecological Concerns: Guiding Principles Nontarget effects and hybridization with weedy relatives are subjects of concern for transgenic pest-protected plants. The committee suggests that a particular pest-protected plant needs to be exempt from both of these ecological concerns in order to avoid safety assessments. Nontarget Effects: Guiding Principle Nontarget effects are often unknown or difficult to predict. Along with standard screens for toxicity to nontarget species, comparison with agricultural practices that would occur if the transgenic pest-protected plant were not used could be made. For example, nontarget effects of transgenic Bt cotton could be compared with nontarget effects from nontransgenic cotton and the accompanying pesticide use needed to compensate for the lack of the transgenic trait. Broader environmental consequences such as changes in soil quality, wildlife habitat, or the use of fertilizers or water could be used to determine the contribution of the new variety to the sustainability of the agricultural system in which it is grown (Cook 1999). Such general environmental considerations could have effects on nontarget organisms. However, it is important to point out that there is disagreement among scientists, including within the committee, as to whether comparison to currently used pest control practices should be the determining factor for allowing commercialization of a transgenic pest-protected plant. Most agree that it is one of many important factors. Therefore, both toxicity testing and field tests comparing agricultural methods are suggested. The committee recognizes that the question below leaves much room for agency judgment.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION Is it reasonable to expect that commercialization of plants with the transgenic resistance trait will have more substantial adverse effects on nontarget organisms than current pest control5 has on these organisms? Yes or More data needed to make a determination—subject to nontarget considerations. No—exempt from nontarget considerations. Hybridization with Wild or Weedy Relatives: Guiding Principles The following guiding principles regarding hybridization with wild or weedy relatives are suggested for reviewing transgenic pest-protected plants. These guidelines are designed for annual crop plants and may require modification in order to address perennials. EPA's categorical exemptions of transgenic plants that have sexually compatible, nontoxic, and viral coat proteins are not in agreement with these principles in some cases. USDA analyzes these concerns according to risks posed to agriculture, so weedy relatives with agricultural effects are of concern; its methods are similar to the following questions, although original data are not always used. FDA does not provide oversight for ecological concerns. Does the cultivated plant occur in feral populations or hybridize with related species in the United States?6 Yes or More data needed—go to 2. No—exempt from weedy-relative considerations. Have feral populations or wild relatives been reported as weedy or invasive in the United States or have a reasonable potential to become weedy?7 Yes or More data needed—go to 3. No—exempt from weedy-relative considerations. 5   Current pest control methods could include both the use of chemical insecticides or other non-chemically based methods. 6   Hybridization refers to any naturally occurring gene flow that results in permanent introgression of genes from cultivated plants into noncultivated populations. Annual crops that persist for 1 or 2 years as volunteers are not considered to be feral populations. 7   Applies to plants in both managed and unmanaged habitats. A species does not have to be included on the Federal Noxious Weed List to qualify as weedy or invasive, but it should be mentioned in peer-reviewed journals or other professional publications.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION Does the gene for resistance confer a specific type of resistance or a greatly enhanced degree of resistance that is not found in feral populations or sexually compatible wild relatives in the United States? 8 Yes or More data needed—go to 4. No—exempt from weedy-relative considerations. Is it reasonable to expect that this trait could have a substantial impact on the population dynamics of feral plants or wild relatives and will lead to increased abundance?9 Yes or More data needed—subject to weedy-relative considerations. No—Exempt from weedy-relative considerations. In addition to the recommendations in section 3.1.4, the committee recommends that USDA should research, publicize, and periodically revise lists of plant species with feral populations or wild relatives in the United States in order to evaluate the impacts of outcrossing. 3.4 RESEARCH NEEDS The committee realizes that there remain some uncertainties regarding the use of pest-protected plants, including transgenic pest-protected plants. These uncertainties can lead to ambiguities in regulation and often force agencies to base their decisions on minimal data sets. Additional research should continue to refine and improve the risk assessment methods and procedures and continue to develop additional data on both conventional and transgenic pest-protected plant products. Research along the following lines should be given priority to aid in decisionmaking. These categories have been chosen on the basis of the discussions in chapter 2 and this chapter. Many of these research needs are also highlighted in the executive summary (section ES.5). 8   The frequency of the resistance trait might vary among populations. If the resistance trait is regarded as rare, go to 4. Also, go to 4 if the resistance trait is found only in geographically isolated populations. 9   This will require agency judgment.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION 3.4.1 Health Effects Research Methods for more efficiently and accurately identifying potential food allergens in transgenic pest-protected plants should be developed. Criteria of digestibility and overall homology with known allergens can be good indicators of allergenicity (Metcalfe 1996a), but the identification of specific protein sequences (or epitopes) involved in allergic responses, the further development of tests with human immune-system endpoints, and the development of more-reliable animal models should be pursued (section 2.5.1). The committee suggests the establishment of a database on natural plant-defensive compounds of potential dietary or other toxicologic concern. Information needed for this database includes a clear list of what plants are used, phenotypic variation in the substances in different parts of plants, and genetic variations in different varieties. Research is needed to determine the baseline concentrations of secondary compounds in plant species of potential dietary or other toxicological concern and to determine how these compounds may vary depending on the genetic background and environmental conditions (see section 2.5.2 and recommendations in section 3.2.4). For longterm toxicity testing, research should be conducted to examine whether longterm feeding of transgenic pest-protected plants to animals whose natural diets consist of large quantities and the type of plant material being tested (for example, grain or forage crops fed to livestock) could be a useful method for assessing potential human health impacts (see section 2.5.1). 3.4.2 Plant Breeding and Molecular Biology Research Research on the mechanisms of pest-protection in both conventional and transgenic pest-protected plants should be encouraged so that we can produce crops that are only minimally affected by diseases and pests, deploy pest-protection strategies that have only minimal impact on the environment, and produce crops that can be consumed or used safely by humans and animals. A major goal of current and future development of conventional and transgenic pest-protected plants should be to decrease the potential for ecological and health problems associated with some types of pest-protected plants (section 2.2.1). That includes developing breeding approaches and assays for avoiding the development of varieties with unintended high concentrations of potential toxins or decreased concentrations of essential nutrients, controlling expression of transgenes that have potential adverse nontarget effects to only nonedible plant tissues, and

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION eliminating expression of transgenes that encode resistance factors in pollen. In addition, development of strategies that enhance the effective life span, or durability, of transgenic pest-protection mechanisms is vital. Research to develop better promoters that restrict expression of transgenes to non-edible plant tissues could lead to decreased potential for food safety problems with some pest-protected plants. Research could also lead to the more efficient use of non-constitutive promoters that result in more durable pest-protection or environmental safety. Transgenic or other techniques to decrease the potential for the spread of transgenes into wild populations should be explored. For conventional pest-protected plants and for transgenes moved by breeding to new cultivars, the linkage of pest-protection traits to other traits carried inadvertently by the breeding process should be investigated for commercial cultivars, and more research should be conducted on potential health and ecological impacts of such linkage (section 2.4.2). Recent advances in plant genomics should help to identify the biochemical and physiological function of linked genes. Similarly, research is needed to better understand potential pleiotropic effects of pest-protection genes. 3.4.3 Ecological Research Research to increase our understanding of the population biology, genetics, and community ecology of the target pests should be conducted, so that more ecologically and evolutionarily sustainable approaches to pest management with pest-protected plants can be developed ( section 2.6). Knowledge of pests' roles in the larger biological community (for example, their role as food sources for nontarget organisms or their roles as predators of other agriculturally relevant pests) will allow us to anticipate better the indirect effects of declines in the pests due to both conventional and transgenic pest-protected plants. Knowledge of the pest population biology will enable prediction of the types of pest-protection mechanisms that would most efficiently reduce a target organism' s pest status (Kennedy et al. 1987) and would help us to design more accurate resistance management plans (Gould 1998). Research to assess gene flow and its potential consequences should be conducted (section 2.7). A list of plants with wild or weedy relatives in the United States should be established in an accessible public database (see section 3.3). This database should include the geographic locations of these relatives and could be used to determine which crop-weed complexes should be regulated. For weed species of concern (plants that might hybridize with transgenic pest-protected plants), more ecological and agricultural research is needed on the following: weed distribution

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION and abundance (past and present), key factors that regulate weed population dynamics in managed and unmanaged areas, the likely impact of specific, novel resistance traits on weed abundance in managed and unmanaged areas, and rates at which resistance genes from the crop would be likely to spread among weed populations. Because it is sometimes difficult to predict ecosystem level effects from small scale laboratory and field tests, longterm monitoring of pest-protected crops should be conducted after commercialization of these crops. EPA and USDA's Agriculture Research Service and Animal Health Plant and Inspection Service should encourage long-term monitoring for ecological impacts. Also, more rigorous field comparisons should be conducted to determine the relative impacts of conventional and transgenic pest-protected crops compared to impacts of standard and alternative agricultural practices on nontarget organisms. Further studies are needed to determine the distances and densities of biologically active Bt corn pollen in the vicinity of a crop. More information is needed about the timing of pollen release, the types of insect species that would be harmed by ingesting pollen at observed concentrations, and the magnitude of mortality due to pollen versus other factors that limit nontarget populations. 3.5 RECOMMENDATIONS EPA should provide guidelines for determining the most ecologically relevant test organisms and test procedures for assessing nontarget effects in specific cropping systems. The USDA should require original data to support agency decision-making concerning transgenic crops when published data are insufficient. In cases when crucial scientific data are lacking about the potential impacts of gene flow on wild or weedy relatives (for example, squash case study), the committee recommends delaying approval of deregulation pending sufficient data (for example, surveys from several years in several regions), establishing a scientifically rigorous monitoring program in key areas to check for undesirable effects of resistance transgenes after the transgenic pest-protected plant is commercialized, or restricting the initial areas where the plants can be grown. USDA should research, publicize, and periodically revise lists of plant species with feral populations or wild relatives in the United States in order to evaluate the impacts of outcrossing.

OCR for page 104
GENETICALLY MODIFIED PEST-PROTECTED PLANTS: SCIENCE AND REGULATION The EPA, FDA, and USDA should collaborate on the establishment of a database for natural plant defensive compounds of potential dietary or other toxicological concern. Given that transfer and manipulation of genes between sexually compatible plants could potentially result in adverse effects in some cases (for example, modulation of a pathway that increases the concentration of a toxicant), and given public controversy regarding transgenic products, EPA should reconsider its categorical exemption of transgenic pest-protectants derived from sexually compatible plants. EPA should not categorically exempt viral coat proteins from regulation under FIFRA. Rather, EPA should adopt an approach, such as the agency 's alternative proposal, that allows the agency to consider the gene transfer risks associated with the introduction of viral coat proteins to plants. EPA should review exemptions of transgenic pest-protected plant products to ensure that they are consistent with the scientific principles elucidated in this report.