E-mail this article to
yourself or a friend.
Enter address:


Some see problems with renewals of Bt crop registrations

(Oct. 1, 2001 – CropChoice opinion) – The following are excerpts from the report, "Final Comments for Submission to the Environmental Protection Agency Docket No. OOP-00678B: Concerning the Revised Risks and Benefits Sections for Bacillus thuringiensis Plant-Pesticides," by Bill Freese of the environmental group Friends of the Earth. The comments discuss various problems with Bt crops and why the Environmental Protection Agency should not re-register five varieties of these transgenic crops, which produce insecticidal proteins from the soil bacterium Bacillus thuringiensis.

The EPA had approved transgenic StarLink corn, which is a Bt variety, only for livestock feed and industrial purposes, because of concerns that it was allergenic to humans. Despite that, StarLink found its way into the human food supply, forcing costly recalls, financial nightmares for farmers and grain handlers, and a steep cut in corn purchases from major export customers such as Japan.

To see the complete comments that Friends of the Earth submitted to the EPA, visit www.foe.org/safefood/btreg.html.

Key Findings

Health Assessment of Bt Crops

* The EPA has failed to establish and/or apply standards for the health assessment of the genetically engineered Bt crops being considered for registration: three varieties of Bt corn, one each of cotton and potatoes (Sections 3 & 12)

* While the EPA is supposed to evaluate three parameters of allergenicity – digestive stability, resistance to heat and structural similarity to known allergens or toxins – the Agency has either failed to demand that companies submit corresponding studies, despite the lapse of 5 years since the original registrations, or the data submitted are of poor quality (Section 3, Table 2)

+ Amino acid sequence homology: The EPA has collected virtually no data on potential similarities between the structure of Bt proteins and the structures of known allergens and toxins, even though all experts agree that this simple test should always be carried out on all genetically engineered foods, and even the EPA considers it to be "required" data (Section 4)

+ Digestive stability: Two studies submitted to the EPA demonstrate that the Bt corn protein, Cry1Ab, exhibits digestive stability similar to that of StarLink corn’s Cry9C protein. This was one of the key findings that led the EPA to reject StarLink for human consumption (Section 6)

+ Heat stability: Cry1Ab also exhibits stability to heat comparable to that of StarLink Cry9C. The EPA has accepted substandard tests for other Bt crops that fail to assess the parameter of interest – the degree to which the Bt protein is degraded by heat (Section 5)

Deficient Product Characterization

* Many of the studies submitted by companies were conducted on Bt proteins from different crop lines, representing distinct transformation events, than the one registered, leading to the dubious practice of "data bridging." (Section 9)

* Much of the safety testing was performed on truncated versions of bacterial surrogate proteins rather than the full-length, plant-produced Bt proteins people are actually exposed to. Studies supposedly demonstrating the equivalence of bacterial surrogate proteins to their plant-produced counterparts for the purpose of safety testing do not meet the standards set by EPA’s scientific advisors, and so should not be accepted (Section 10)

* The EPA relies unduly on data generated with Bt microbial pesticides, which differ in significant ways from the Bt crop-produced proteins (Sections 8 & 11)

* Monsanto’s Yieldgard corn is the result of a transformation gone awry. While Monsanto intended to insert the full-length cry1Ab gene, apparently only a fragment was incorporated, for reasons unknown. Monsanto has had difficulty even detecting the 92 kD protein supposedly produced by this gene fragment. Yieldgard also lacks the foreign genes for the herbicide-resistance trait that were supposed to be inserted and were "apparently lost during the development of the MON 810 line." (Sections 1 and 8.2)

* The Agency has no data confirming the removal of the ampicillin resistance gene from the DNA used to transform Syngenta’s Bt 11 corn (Section 1)

Procedural Violations

* The EPA has largely ignored the recommendations of its scientific advisors (SAP Bt Plant-Pesticides, SAP Mammalian Toxicity), contrary to the Agency’s claim that it has considered them (Sections 1 & 10)

* The EPA’s Biopesticides Registration Action Document shows no evidence that the EPA has considered "the most current health and ecological data" or "all available scientific information on Bt products," the supposed purpose of this re-assessment

* "Required" data that should have been collected and carefully evaluated before these crops were originally registered are still missing/deficient despite the lapse of five years, and the EPA appears poised to compound this error by re-registering the crops, perhaps without time limit, in the absence of these "required" data (Section 3)

Data Apparently Not Considered

* A 1999 study by Bernstein et al partly sponsored by the EPA suggesting that the Bt proteins Cry1Ab and Cry1Ac can elicit antibody (IgE) responses consistent with allergic reactions in farm-workers has apparently not been considered, despite the explicit recommendation by SAP Bt Plant-Pesticides that the Agency make use of serological agents developed in the course of this study to conduct "clinical assessment of exposed individuals" (Section 11)

* A series of 4 studies on mice dated 1999 and 2000 by Vazquez et al showing that Cry1Ac is a potent immunogen, increases antibody responses to other antigens, and binds to surface proteins in the mouse small intestine have also apparently been ignored. One of these studies was recommended by SAP Bt Plant-Pesticides for consideration by the EPA (Section 11)

* The Agency has had evidence since the 1994 that Bt protoxins differ immunologically from the truncated proteins used for testing purposes, yet apparently has not followed up on this important finding (Section 8)

* Evidence that the toxic portion of Cry1A proteins can have a different 3-D conformation depending on whether it is part of the protoxin or in its free state, and that DNA structurally associated with the protoxin is released during the proteolysis process that generates the toxic fragment from the protoxin has also gone unconsidered (Section 8).

Executive Summary

Over the past decade, the scientific and medical communities have become increasingly concerned about the potential of genetically engineered (GE) foods to cause allergies. GE foods produce "novel" proteins that are often new to the human diet and are most often derived from bacteria. Allergies are triggered by aberrant immune system responses, which often occur when a susceptible person is exposed to a new food (or food protein). It is thought that food allergies afflict 2-2.5% of adults and 6-8% of children (SAP StarLink II, p. 11; Sampson 1999), or about 8 million Americans children (based on US Census data). Food allergies are becoming more common, for reasons still unknown. Because GE foods introduce novel proteins, and the process of acquiring allergies is still poorly understood, a growing number of experts recommend labeling of genetically engineered foods and monitoring for potential allergic reactions after market introduction.

The Environmental Protection Agency is presently considering whether or not to re-register five genetically engineered Bt crops – three varieties of Bt corn and one each of cotton and potatoes – which produce insecticidal proteins derived from a soil bacterium known as Bacillus thuringiensis. Concern about the potential allergenicity of one variety of Bt corn, StarLink, led the EPA to deny approval of this crop for human food use. After the food supply nevertheless became contaminated with StarLink, a panel of experts that included some of the nation’s leading allergists advised the EPA that it was impossible to establish a threshold below which the StarLink protein, Cry9C, could be considered safe for human consumption. It was for this reason that the EPA refused to approve a petition by StarLink’s developer, Aventis CropScience, to allow even an extremely low level of Cry9C residues in the food supply.

In examining studies submitted to the EPA and recommendations by the Agency’s scientific advisors, Friends of the Earth has found evidence that the bacterial-derived insecticidal proteins produced by other Bt crop varieties either possess properties similar to those of Cry9C, or have not been assessed for the same. These allergenic properties include digestive stability, resistance to heat, and structural similarity to known allergens (or toxins).

In particular, studies conducted by Aventis CropScience and Dr. Hubert Noteborn demonstrate that the Cry1Ab toxin produced by the Bt corn varieties registered to Monsanto and Syngenta exhibit digestive stability that is similar to StarLink’s Cry9C toxin. Noteborn’s study also shows that Cry1Ab and Cry9C have comparable resistance to breakdown by heat. The Agency has apparently not collected any data on potential similarities between the Bt proteins and known allergens or toxins, aside perhaps from limited data on Bt potatoes (see Table 2).

Additional concerns are raised by the poor quality of the studies conducted by the registrant companies. In many cases, these studies were conducted on Bt proteins from different crop lines than the one that was registered. In addition, the bulk of safety testing was performed on truncated versions of bacterial surrogate proteins rather than the full-length plant-produced Bt proteins that people are actually exposed to. This use of surrogates for testing purposes is a controversial practice, and is supposed to be permitted only if the bacterial and plant-produced versions are shown to be equivalent. The so-called "test substance equivalence" studies submitted by registrants do not meet the standards recommended by EPA’s expert advisors. Failure to establish such equivalence in the case of the StarLink protein, Cry9C, was a major reason the EPA refused to permit even low levels of this protein in the food supply.

In addition, several of the Bt crops did not turn out as planned by the registrants, casting doubt on their ability to control the crude "gene gun" transformation (i.e. engineering) process. Monsanto’s Yieldgard corn, for example, apparently contains only a fragment of the gene that was supposed to be inserted, due to an unexplained failure in the transformation process. The company has had difficulty even detecting the protein actually produced by this gene fragment, and has conducted its testing instead on the portion of this protein that is resistant to trypsin. Yieldgard corn also lacks the foreign gene for the herbicide-resistance trait that was supposed to be inserted, again for reasons unknown.

Both the registrants and the EPA make invalid comparisons between Bt microbial pesticides (i.e. Bt sprays) and the plant-produced Bt proteins to support their contention that the latter are safe, despite the fact that there are a number of important differences between the native Bt proteins and the modified versions produced in crops.

Finally, the EPA has apparently failed to consider important evidence demonstrating that Cry1Ab elicits antibody (IgE) responses consistent with an allergic response in farm-workers, as well as other studies demonstrating that the closely related Cry1Ac induces a potent antibody response in mice, increases the antibody response to other antigens, and binds to surface proteins on the mouse small intestine.

The seriously deficient data submitted by registrants raise questions about the potential health impacts of their Bt crops. The EPA’s flaccid acceptance of poorly conducted corporate studies, and its apparent unwillingness to demand "required" data, casts doubt on the entire regulatory process, which appears to be unduly biased towards approval of Bt crops in the absence of critical data.

The unanswered questions with respect to the potential health impacts of Bt crops take on added significance in the context of an increasing incidence of food allergy, whose causes remain unknown. The EPA is urged to conduct a thorough and rigorous re-assessment of the Bt crops being considered for re-registration by demanding that missing data be supplied, requiring that deficient studies be repeated under proper conditions, and giving full consideration to the recommendations of its expert scientific advisors. Re-registration should not even be considered until complete and accurate data have been collected and carefully evaluated.

1. Introduction

The EPA’s stated objective in its reassessment of Bt crops is "to assure that the decisions on the renewal of these registrations are based on the most current health and ecological data." The reassessment is supposed to take account of "all available scientific information on Bt products," giving particular consideration to recommendations by the EPA’s Scientific Advisory Panels (e.g. SAP Bt Plant-Pesticides 2000, SAP Mammalian Toxicity 2000) as well as the National Academy of Sciences’ in-depth review of Bt crops (NAS 2000) (EPA BRAD 2001, p. I1, my emphasis).

A thorough-going reassessment is badly needed for several reasons. First, genetically engineered Bt crops are still a relative newcomer on the agricultural scene, having been approved for commercial planting barely more than half a decade ago. Since that time, there has been a great deal of research on these crops which should be considered before a decision is made on whether to allow their continued cultivation. Second, the science of genetic engineering is still primitive. The process of transgene insertion is crude and haphazard. As a result, even plant developers often lack complete data on their transgenic products. Pleiotropic (i.e. unexpected) effects are more common than once imagined, and often remain undiscovered for years, by which time the crop may have come to be cultivated on tens of millions of acres.

For example, since their introduction in 1994, evidence has accumulated that Roundup Ready (RR) soybeans, by far the most widely planted genetically engineered crop, have lower phytoestrogen levels (Lappe et al 1999); depressed root development, nodulation and nitrogen fixation; lower levels of aromatic amino acids; and lower average yields than their conventional counterparts (Benbrook 2001). In addition, the original promise of reduced herbicide use has proven false; in fact, cultivation of Roundup Ready soybeans is associated with application of greater amounts of herbicide, on average, than their conventional counterparts (Ibid). These hard facts – lower yields and increased herbicide use – contrast sharply with vague biotech industry promises that future GE crops will somehow "feed the world" and reduce agricultural chemical use. Finally, fragments of DNA undetected at the time of original registration were recently discovered in Monsanto’s product, five years after the crop was first commercialized. Two of these fragments, 72 and 250 base pairs in length, appear to be partial copies of the CP4 EPSPS gene which lends RR soybeans their herbicide resistant trait (Monsanto RR Soybeans 2000). More recently, a Belgian team reported that a 534 base pair segment of DNA flanking the previously undetected 250 bp fragment does not match soybean genomic DNA, as expected, and suggested that it could be scrambled plant DNA or DNA from an unknown source (Windels et al 2001).

Scientists are only now beginning to unravel the molecular basis of the pleiotropic effects mentioned above. In order to do so, thorough and accurate product characterization data are essential. For this reason, one could argue that no genetically engineered crop should be in the fields or on the market until it has been completely and accurately characterized at the molecular level. At the very least, it seems hard to argue with the notion that regulatory authorities should demand the best information possible with the most modern techniques that are available. Some of the improved techniques employed by Monsanto and the Belgian team in obtaining more accurate characterization of RR soybeans (e.g. genome walking, cosmid library construction and northern blot analysis) (Monsanto RR Soybeans 2000, p. 4) have apparently not been applied to the Bt products being considered for reregistration.

For instance, the Agency is set to re-register Syngenta’s Bt 11 corn despite lack of data confirming removal of the ampicillin resistance gene from the transforming DNA (EPA BRAD 2001, p. IIA5). In the case of Monsanto’s Bt corn, the EPA seems satisfied with the conclusion that: "These genes which confer glyphosate tolerance were apparently lost during the development of the MON 810 line…" (Ibid, p. IIA7). In addition, the Agency appears to be untroubled by the failed transformation event by which only a portion of the intended full-length cry1Ab gene sequence was actually incorporated into the genome of MON 810 corn, despite Monsanto’s apparent difficulty in detecting the protein encoded by this unintended fragment (Ibid, p. IIA7; see Section 9.2).

The very prevalence of genetically engineered Bt crops is still another reason for a strict and thorough reassessment. Even if impacts on health and the environment prove to be minor, it must not be forgotten that we are dealing with crops planted on tens of millions of acres, to which large numbers of people are exposed. To take a concrete example: If StarLink’s Cry9C protein is shown to be even a rare allergen, affecting just a tiny percentage of the population, the likelihood that tens of millions of people have consumed tainted corn products (Friends of the Earth 2001, pp. 14-15) suggests that a large number of people may have been affected. And if StarLink, which was never planted on more than 0.4% of total U.S. corn acreage, can spread so widely in the food system, it is reasonable to assume that the other Bt corn varieties, which constitute roughly 20% of U.S. corn production, have been consumed by nearly every American for several years now.