In this chapter, the committee examines the evidence that substantiates or negates specific hypotheses and claims about the health risks and benefits associated with foods derived from genetically engineered (GE) crops. There are many reviews and official statements about the safety of foods from GE crops (for example, see Box 5-1), but to conduct a fresh examination of the evidence, the committee read through a large number of articles with original data so that the rigor of the evidence could be assessed.
Some of the evidence available to the committee came from documents that were part of the U.S. regulatory process for GE crops conducted by the U.S. Environmental Protection Agency (EPA), the U.S. Department of Agriculture (USDA), and the U.S. Food and Drug Administration (FDA). Other evidence came from studies published by regulatory agencies in other countries or by companies, nongovernmental organizations (NGOs), and academic institutions. The committee also sought evidence from the public and from the speakers at its public meetings and webinars.1
The committee thinks that it is important to make clear that there are limits to what can be known about the health effects of any food, whether non-GE or GE. If the question asked is “Is it likely that eating this food today will make me sick tomorrow?” researchers have methods of getting quantitative answers. However, if the question is “Is it likely that eating
1 The committee has compiled publicly available information on funding sources and first-author affiliation for the references cited in this chapter; the information is available at http://nas-sites.org/ge-crops/.
this food for many years will make me live one or a few years less than if I never eat it?” the answer will be much less definitive. Researchers can provide probabilistic predictions that are based on the available information about the chemical composition of the food, epidemiological data, genetic variability across populations, and studies conducted with animals, but absolute answers are rarely available. Furthermore, most current toxicity studies are based on testing individual chemicals rather than chemical mixtures or whole foods because testing of the diverse mixtures of chemicals experienced by humans is so challenging (Feron and Groten, 2002; NRC, 2007; Boobis et al., 2008; Hernández et al., 2013).
With regard to the issue of uncertainty, it is useful to note that many of the favorable institutional statements about safety of foods from GE crops in Box 5-1 contain caveats, for example: “no overt consequences,” “no effects on human health have been shown,” “are not per se more risky,” and “are not likely to present risks for human health.” Scientific research can answer many questions, but absolute safety of eating specific foods and the safety of other human activities is uncertain.
The review in this chapter begins with an examination of what is known about the safety of foods from non-GE plants and how they are used as counterparts to those from GE crops in food-safety testing. U.S. food-safety regulatory testing for GE products and GE food-safety studies conducted outside the agency structure are then assessed. A variety of hypothesized health risks posed by and benefits of GE crops are examined, and the chapter concludes with a short discussion of the challenges that society will face in assessing the safety of GE foods that are likely to be developed with emerging genetic-engineering technologies.
An oft-cited risk of GE crops is that the genetic-engineering process could cause “unnatural” changes in a plant’s own naturally occurring proteins or metabolic pathways and result in the unexpected production of toxins or allergens in food (Fagan et al., 2014). Because analysis of risks of the product of the introduced transgene itself is required during risk assessment, the argument for unpredicted toxic chemicals in GE foods is based on the assumption that a plant’s endogenous metabolism is more likely to be disrupted through introduction of new genetic elements via genetic engineering than via conventional breeding or normal environmental stresses on the plant. The review below begins by discussing natural chemical constituents of plants in the context of food safety to provide a background on what the natural plant toxins are and how they vary in non-GE plants. The review then goes on to explain the premise used by regulatory agencies to compare GE crops with their non-GE counterparts.
Endogenous Toxins in Plants
Most chemicals of primary metabolism (for example, those involved in the formation of carbohydrates, proteins, fats, and nucleic acids) are shared between animals and plants and are therefore unlikely to be toxic. Perceived risks associated with alterations of plant compounds arise mainly from alterations of plant-specific molecules, popularly known as plant natural products and technically named secondary metabolites. Collec-
tively, there are more than 200,000 secondary metabolites in the plant kingdom (Springob and Kutchan, 2009). Crop species vary in the number of secondary metabolites that they produce. For example, potato (Solanum tuberosum) is known for its high diversity of secondary metabolites and can have more than 20 sesquiterpenes (a single group of related compounds), some of which are thought to confer resistance to diseases (Kuc, 1982). The concentrations of these secondary metabolites within some tissues in a particular plant species may vary from high—for example, chlorogenic acids alone make up about 12 percent of the dry matter of green coffee beans (Ferruzzi, 2010)—to trace amounts (many minor saponins in legumes) and may be associated with particular stages of plant development (some found only in seeds) or may increase in response to external stimuli, such as pathogen or herbivore attack, drought, or altered mineral nutrition (Small, 1996; Pecetti et al., 2006; Nakabayashi et al., 2014). Many secondary metabolites function as protective agents, for example, by absorbing damaging ultraviolet radiation (Treutter, 2006), acting as antinutrients (Small, 1996), or killing or halting insects and pathogens that damage crops (Dixon, 2001). Plant secondary metabolites that protect against pathogen attack have been classified as either phytoanticipins (if they exist in a preformed state in a plant before exposure to a pathogen) or phytoalexins (if their synthesis and accumulation are triggered by pathogen attack) (VanEtten et al., 1994; Ahuja et al., 2012). The toxic properties of some plant compounds are understood, but most of these compounds have not been studied. Some secondary metabolites and other products (such as proteins and peptides) in commonly consumed plant materials can be toxic to humans when consumed in large amounts, and examples are listed below:
- Steroidal glycoalkaloids in green potato skin, which can cause gastrointestinal discomfort or, more severely, vomiting and diarrhea.
- Oxalic acid in rhubarb, which can cause symptoms ranging from breathing difficulty to coma.
- Gossypol in cottonseed oil and cake, which can cause respiratory distress, anorexia, impairment of reproductive systems, and interference with immune function in monogastric animals.
- Nonprotein amino acid canavanine in alfalfa sprouts, which can be neurotoxic.
- Hemolytic triterpene saponins in many legume species, which can increase the permeability of red blood cell membranes.
- Cyanogenic glycosides in almonds and cassava, which can cause cyanide poisoning.
- Phototoxic psoralens in celery, which are activated by ultraviolet sunlight and can cause dermatitis and sunburn and increase the risk of skin cancer.
Friedman (2006) provided information that demonstrated that some glycoalkaloids in potato can have both harmful and beneficial effects. The Food and Agriculture Organization has recognized that foods often contain naturally occurring food toxins or antinutrients but that at naturally occurring concentrations in common diets they can be safely consumed by humans (Novak and Haslberger, 2000; OECD, 2000). The health risks associated with some secondary metabolites in common foodstuffs are generally well understood, and the plants are either harvested at times when the concentrations of the compounds are low, the tissues with the highest concentrations of toxins are discarded, or, as in the case of cassava (Manihot esculenta), the food is prepared with special methods to remove the toxic compounds. In other cases, food preparation may be the cause of the presence of a toxic compound (for example, the formation of the probable carcinogen acrylamide when potatoes are fried at high temperatures or when bread is toasted). Plant breeders have generally screened for toxins that are typical of the plant group from which a crop was domesticated and have excluded plants that have high concentrations of the compounds.
Unintended changes in the concentrations of secondary metabolites can result from conventional breeding (Sinden and Webb, 1972; Hellenas et al., 1995). In some cases, conventionally bred varieties have been taken off the market because of unusually high concentrations of a toxic compound, as in the case of a Swedish potato variety that was banned from sale in the 1980s because of high concentrations of glycoalkaloids (Hellenas et al., 1995).
Rather than being a cause of worry, many secondary metabolites are perceived as having potential health benefits for humans and are consumed in increasingly large quantities (Murthy et al., 2015). Examples include the isoflavone phytoestrogens found in a number of leguminous plants, such as soybean (Glycine max) and clover (Trifolium spp.), which have been ascribed beneficial activities, including chemoprevention of breast and prostate cancers, cardiovascular disease, and post-menopausal ailments (Dixon, 2004; Patisaul and Jefferson, 2010). Also, various perceived antioxidants, such as anthocyanins (Martin et al., 2013), and some saponins may have anticancer activity (Joshi et al., 2002). There is, however, disagreement as to whether many of the compounds are beneficial or toxic at the concentrations consumed in herbal medicines or dietary supplements (see, for example, Patisaul and Jefferson, 2010).
FINDING: Crop plants naturally produce an array of chemicals that protect against herbivores and pathogens. Some of these chemicals can be toxic to humans when consumed in large amounts.
Substantial Equivalence of Genetically Engineered and Non–Genetically Engineered Crops
A major question addressed in the regulation of GE crops is whether the concentrations of the toxic secondary metabolites are affected by genetic engineering. In addition to the plant toxins, nutrients, introduced genes, and proteins and their metabolic products in specific GE crops are assessed with a comparative approach that is generally encompassed by the concept of substantial equivalence.
The concept of substantial equivalence has a long history in safety testing of GE foods. The term and concept were “borrowed from the [U.S. FDA’s] definition of a class of new medical devices that do not differ materially from their predecessors, and thus, do not raise new regulatory concerns” (Miller, 1999:1042). No simple definition of substantial equivalence is found in the regulatory literature on GE foods. In 1993, the Organisation for Economic Co-operation and Development (OECD) explained that the “concept of substantial equivalence embodies the idea that existing organisms used as food, or as a source of food, can be used as the basis for comparison when assessing the safety of human consumption of a food or food component that has been modified or is new” (OECD, 1993:14).
The Codex Alimentarius Commission’s Guideline for the Conduct of Food Safety Assessment of Foods Derived from Recombinant-DNA Plants is careful to state that “the concept of substantial equivalence is a key step in the safety assessment process. However, it is not a safety assessment in itself; rather it represents the starting point which is used to structure the safety assessment of a new food relative to its conventional counterpart” (CAC, 2003:2). The Codex guideline also makes clear that a safety assessment of a new food based on the concept of substantial equivalence “does not imply absolute safety of the new product; rather, it focuses on assessing the safety of any identified differences so that the safety of the new product can be considered relative to its conventional counterpart” (CAC, 2003:2). The OECD (2006) came to a similar conclusion. Conflict among stakeholders often comes into play during the determination of what constitutes evidence of differences from substantial equivalence sufficient to justify a detailed food-safety assessment.
The Codex Alimentarius Commission concluded that the concept of substantial equivalence “aids in the identification of potential safety and nutritional issues and is considered the most appropriate strategy to date for safety assessment of foods derived from recombinant-DNA plants” (CAC, 2003:2). Despite some criticism of the substantial-equivalence concept itself (for example, Millstone et al., 1999) and operational problems (for example, Novak and Haslberger, 2000), it remains the cornerstone for
GE food-safety assessment by regulatory agencies. The present committee examined its use in practice and its empirical limitations.
The precautionary principle, which is described in more detail in Chapter 9 (see Box 9-2) is a deliberative principle related to the regulation of health, safety, and the environment and typically involves taking measures to avoid uncertain risks. The precautionary principle has been interpreted in a number of ways, but it is not necessarily incompatible with use of the concept of substantial equivalence. In the case of foods, including GE foods, it can be reasonably argued that even a small adverse chronic effect should be guarded against, given that billions of people could be consuming the foods. However, the degree of precaution taken in the face of uncertainty is a policy decision that varies among countries and according to the specific uncertainty being considered. For example, many European countries and the European Union (EU) as a whole generally take a more precautionary approach with GE foods and climate change whereas the United States has historically taken a more precautionary approach with tobacco products and ozone depletion (Wiener et al., 2011). The reader is directed to Chapter 9 for further discussion of how different regulatory frameworks address uncertainty in the safety of GE foods.
Some differences between a GE food and its non-GE counterpart are intentional and identifiable (for example, the presence of a Bt toxin in maize kernels) or are due to practices directly associated with the use of the GE crops (for example, increased use of glyphosate). Some of the risks posed by the intended changes can be anticipated on the basis of the physiological and biochemical characteristics of the engineered change. There are often established protocols for assessing such risks, especially when a change involves exposure to a known toxin. However, other risks have been hypothesized for GE crops because previous uses of a trait (for example, Bt as an insecticidal spray) did not have consumption of the GE plant products as the route of exposure. New routes of exposure could result in unanticipated effects.
In contrast with such intended differences, some potential differences between GE crops and their non-GE counterparts are unintentional and can be difficult to anticipate and discern (NRC, 2004). Two general sources of unintended differences could affect food safety:
- Unintended effects of the targeted genetic changes on other characteristics of the food (for example, the intended presence of or increase in one compound in plant cells could result in changes in plant metabolism that affect the abundance of other compounds).
- Unintended effects associated with the genetic-engineering process (for example, DNA changes resulting from plant tissue culture).
Much of the concern voiced by some citizens and scientists about the safety of GE foods is focused on potential risks posed by unintended differences. Some of the biochemical and animal testing done by or for government agencies is aimed at assessing the toxicity of such unintended differences, but what is adequate and appropriate testing for assessing specific toxicities is often difficult to determine. In some cases, the unintended effects are somewhat predictable or can be determined; in such cases, tests can be designed. In other cases, the change or risk could be something that has not even been considered, so the only effective testing is of the whole food itself. As discussed in Chapter 6, there is a tradeoff between costs of such testing and societal benefits of reduction in risks.
The approach of comparing new varieties to existing varieties is just as applicable to crops developed by conventional plant breeding as it is to GE crops (see Chapter 9). The discussion above on endogenous toxins (see section “Endogenous Toxins in Plants”) shows that such crops pose some risks. The 2000 National Research Council report Genetically Modified Pest-Protected Plants found that “there appears to be no strict dichotomy between the risks to health and the environment that might be posed by conventional and transgenic pest-protected plants” (NRC, 2000:4). Similarly, the 2004 National Research Council report Safety of Genetically Engineered Foods found that all forms of conventional breeding and genetic engineering may have unintended effects and that the probability of unintended effects of genetic engineering falls within the range of unintended effects of diverse conventional-breeding methods. The 2002 National Research Council report Environmental Effects of Transgenic Plants found that “the transgenic process presents no new categories of risk compared to conventional methods of crop improvement but that specific traits introduced by both approaches can pose unique risks” (NRC, 2002:5). That finding remains valid with respect to food safety and supports the conclusion that novel varieties derived from conventional-breeding methods could be assessed with the substantial-equivalence concept.
FINDING: The concept of substantial equivalence can aid in the identification of potential safety and nutritional issues related to intended and unintended changes in GE crops and conventionally bred crops.
FINDING: Conventional breeding and genetic engineering can cause unintended changes in the presence and concentrations of secondary metabolites.
Although the committee agrees that crops developed through conventional breeding could result in food-safety risks, its statement of task focuses on GE crops. Furthermore, there have been claims and counterclaims about the relative safety of GE crops and their associated technologies compared with conventionally bred crops and their associated technologies. Therefore, the remainder of this chapter examines possible risks and benefits associated with GE crops and assesses the methods used to test them in and beyond government regulatory systems.
Whether testing is done for regulatory purposes or beyond the regulatory realm, it typically involves three categories of testing: acute or chronic animal toxicity tests, chemical compositional analysis, and allergenicity testing or prediction. Although the precision, transparency, specific procedures, and interpretation of results vary among countries, criticisms about the adequacy of testing are not so much country-specific as they are method- and category-specific. For example, there may be arguments about whether a 90-day whole-food animal test is more appropriate than a 28-day test, but the bigger issue is about whether whole-food testing is appropriate. The committee uses a description of the U.S. testing methods as an example, but it mostly examines the criticism of food-safety testing more broadly.
The structure of the U.S. regulatory process for GE crops based on the Coordinated Framework for the Regulation of Biotechnology is briefly reviewed in Chapter 3 and is examined in more detail in Chapter 9. The focus in this chapter is on the testing itself. The present section provides insight into U.S. procedures by describing the risk-testing methods used for two examples of traits in commercialized GE crops: Bt toxins and crop resistance to the herbicides glyphosate and 2,4-D.
Regulatory Testing of Crops Containing Bt Toxins
EPA considers plant-produced Bt toxins to be “plant-incorporated protectants,” a class of products generally defined as “a pesticidal substance that is intended to be produced and used in a living plant, or in the produce thereof, and the genetic material necessary for the production of that pesticidal substance” (40 CFR §174.3). EPA specifically exempts plant-incorporated protectants whose genetic material codes for a pesticidal substance that is derived from plants that are sexually compatible. Bt toxin genes are not exempted because they come from bacteria (see Chapter 9 for regulatory details).
For Bt toxins produced by GE crops, EPA took into consideration that there was already toxicity testing of Bt toxins in microbial pesticides and
that the toxins were proteins that, if toxic, typically show almost immediate toxicity at low doses (EPA, 2001a; also see Box 5-2). The pesticidal safety tests mostly involved acute toxicity testing in mice and digestibility studies in simulated gastric fluids because one characteristic of food allergens is that they are not rapidly digested by such fluids.
Box 5-2 provides a verbatim example of the procedures used for testing as reported in EPA fact sheets for the Cry1F Bt toxin so that readers can see what is involved in the testing. The actual research is not typically done by EPA itself. The registrant is usually responsible for testing. Results of the tests of Cry1F show no clinical signs of any toxicity even when Cry1F protein was fed at 576 mg/kg body weight, which would be the equivalent of about ¼ cup of pure Cry1F for a 90.7-kilogram (200-pound) person. Another part of the testing described in Box 5-2 is allergenicity testing. Concerns about the EPA testing methods are discussed in sections below on each category of testing.
Regulatory Testing of Crops Resistant to Glyphosate and 2,4-D and of the New Uses of the Herbicides Themselves
The regulatory actions taken for herbicide-resistant (HR) crops are different from regulatory actions taken to assess Bt crops. With Bt crops, regulatory actions are related to the crop itself. With HR crops, there are regulatory processes for the plant itself and separate regulatory processes for the new kind of exposure that can accompany spraying of a herbicide on a crop or on a growth stage of a crop that has never been sprayed prior to availability of the GE variety.
EPA governs the registration of herbicides such as glyphosate and 2,4-D. Both glyphosate and 2,4-D were registered well before the commercialization of GE crops. However, EPA has authority to re-examine herbicides if their uses or exposure characteristics change.
A good example of such re-examination was the 2014 EPA registration of the Dow AgroSciences Enlist Duo® herbicide, which contains both glyphosate and 2,4-D for use on GE maize (Zea mays) and soybean. Because the glyphosate component of Enlist Duo had already been in use on GE maize and soybean, EPA did not conduct further testing of glyphosate alone. However, 2,4-D was registered previously only for applications to maize up to 20 centimeters tall and for preplant applications to soybean. The proposed use of 2,4-D on GE crops was expected to change use patterns and exposure and thereby triggered a safety assessment of the new use 2,4-D. Additionally, EPA compared the toxicity of the formulation that contained both herbicides to the toxicity of the individual herbicides and concluded the formulation did not show greater toxicity or risk compared to either herbicide alone.
In the human health risk assessment portion of the EPA Enlist Duo registration document, the following tests and results with 2,4-D were considered (EPA, 2014a):
- An acute dietary test in rats that found a lowest observed-adverse-effect level (LOAEL) of 225 mg/kg (about 1 ounce per 200-pound person).
- A chronic-dietary-endpoint, extended one-generation reproduction toxicity study in rats that found a LOAEL of 46.7 mg/kg-day in females and higher in males.
- Inhalation tests involving data from a 28-day inhalation toxicity study in rats that found a LOAEL of 0.05 mg/L-day.
- Dermal tests that showed no dermal or systemic toxicity after repeated applications to rabbits at the limit dose of 1000 mg/kg-day.
- Reviews of epidemiological and animal studies, which did not support a linkage between human cancer and 2,4-D exposure.
Analysis of the results of those tests and agronomic and environmental assessments resulted in the product’s registration.
EPA received over 400,000 comments in response to the initial proposal to register the new use of 2,4-D. Some of the concerns submitted to EPA were similar to ones some members of the public expressed in public comments to the committee, including questions about whether EPA had considered toxicity of only the active ingredient or of the formulated herbicide and whether it had tested for synergistic effects of 2,4-D and glyphosate. EPA (2014b:7) responded that
acute oral, dermal, and inhalation data, skin and eye irritation data, and skin sensitization data are available for the 2,4-D choline salt and glyphosate formulation for comparison with the 2,4-D parent compound and glyphosate parent compound data, and these test results show similar profiles. The mixture does not show a greater toxicity compared to either parent compound alone. Although no longer duration toxicity studies are available, toxic effects would not be expected as the maximum allowed 2,4-D exposure is at least 100-fold below levels where toxicity to individual chemicals might occur, and exposure to people is far below even that level.
The committee did not have access to the actual data from the registrant.2
EPA does not regulate the commercialization of the GE herbicide-resistant crops themselves. That is the role of USDA’s Animal and Plant Health Inspection Service (APHIS) under the Plant Protection Act. Under its
2 In November 2015, EPA took steps to withdraw the product’s registration in light of new information that indicated there could be synergistic effects of the two herbicides, which could possibly result in greater toxicity to nontarget plants (Taylor, 2015). A court ruling in January 2016 allowed the herbicide to remain on the market while EPA considered other administrative actions (Callahan, 2016).
statutory authority, APHIS controls and prevents the spread of plant pests (see Box 3-5). On the basis of a plant-pest risk assessment (USDA–APHIS, 2014a), APHIS concluded that Enlist™ GE herbicide-resistant maize and soybean engineered to be treated with the Enlist Duo herbicide (containing glyphosate and 2,4-D) were unlikely to become plant pests and deregulated them on September 18, 2014 (USDA–APHIS, 2014b). In its document on the decision to deregulate Enlist GE herbicide-resistant maize and soybean (USDA–APHIS, 2014a:ii), APHIS states a general policy that “if APHIS concludes that the GE organism is unlikely to pose a plant pest risk, APHIS must then issue a regulatory determination of nonregulated status, since the agency does not have regulatory authority to regulate organisms that are not plant pests. When a determination of nonregulated status has been issued, the GE organism may be introduced into the environment without APHIS’ regulatory oversight.”
FDA did not identify any safety or regulatory issues in its consultation with Dow AgroSciences on the Enlist maize and soybean varieties (FDA, 2013). FDA also explained the basis of Dow’s conclusion that Enlist soybean is not “materially different in composition” from other soybean varieties (FDA, 2013):
Dow reports the results of compositional analysis for 62 components in soybean grain, including crude protein, crude fat, ash, moisture, carbohydrates, [acid detergent fiber] ADF, [neutral detergent fiber] NDF, total dietary fiber (TDF), lectin, phytic acid, raffinose, stachyose, trypsin inhibitor, soy isoflavones (i.e., total daidzein, total genistein, total glycitein), minerals, amino acids, fatty acids, and vitamins. No statistically significant differences in the overall treatment effect and the paired contrasts between each of the DAS-44406-6 soybean treatment groups and the control were observed for 29 of the components. A statistically significant difference in the overall treatment effect was observed for 16 components (crude protein, carbohydrates (by difference), NDF, calcium, potassium, cystine, palmitic acid, oleic acid, linoleic acid, linolenic acid, behenic acid, folic acid, γ-tocopherol, total tocopherol, lectin, and trypsin inhibitor). However, differences between the control and the DAS-44406-6 treatment groups were small in magnitude. Differences between DAS-44406-6 soybean and the control were considered not biologically relevant because the mean values were either within the ranges generated using the reference lines, consistent with the ranges of values in the published literature, or both.
FINDING: U.S. regulatory assessment of GE herbicide-resistant crops is conducted by USDA, and by FDA when the crop can be consumed, while the herbicides are assessed by EPA when there are new potential exposures.
FINDING: When mixtures of herbicides are used on a new GE crop, EPA assesses the interaction of the mixture as compared to the individual herbicidal compounds.
Technical Assessment of Human Health Risks Posed by Genetically Engineered Crops
As explained in Chapter 2, the development and use of GE crops is governed by more than national and regional regulatory standards. In the cases of the GE crops commercially available in the United States and some other countries in 2015, inputs from many public and private institutions regarding their specific concerns have influenced the type and extent of GE crop food-safety tests conducted by companies, agencies, and other researchers. Many stakeholders have criticized the testing used by U.S. and other national regulatory agencies for lacking rigor (for example, Hilbeck et al., 2015). Researchers in companies, NGOs, and universities have sometimes conducted more extensive safety tests than are required by national agencies or have reanalyzed existing data, as described below. All testing as of 2015 fell into three categories: animal testing, compositional analysis, and allergenicity testing and prediction.
Short-Term and Long-Term Rodent Testing with Compounds and Whole Foods. One common criticism of the animal testing conducted by or for regulatory agencies in the United States and elsewhere is related to its short duration (for example, Séralini et al., 2014; Smith, 2014). Indeed, there is a range in the duration and doses within the test protocols used by regulatory agencies that depends in part on the product. Doses for subchronic and chronic toxicity studies are such that the lowest dose (exposure level), which is many times higher than expected for human exposure, is set to ensure that it does not elicit acute adverse effects that would interfere with examining the potential chronic-effect endpoints. As can be seen in the discussion above, EPA conducted an extended one-generation reproduction toxicity study in male and female rats in its assessment of 2,4-D, and it relied on previous long-term studies for the assessment of cancer risk associated with it. For assessment of the Bt toxin Cry1F and for the bacterially derived proteins in 2,4-D-resistant maize and soybean, company testing submitted to EPA, FDA, and USDA relied on acute toxicity testing. In all the cases above, the experiments were conducted by adding large amounts of a single test chemical to an animal’s diet. Tests with high concentrations of a chemical are typical of EPA testing protocols for pesticides.
What is different between GE crop evaluation and that of general agri-
cultural chemicals is the use of “whole food” tests. These tests are aimed at assessing potential hazards due to the combined intentional and unintentional changes that might have been caused by the genetic engineering of the crop. In such tests, it is not possible to use concentrations higher than what is in the crop itself because potential unintended effects are not typically known. Thus, it is impossible for a researcher to know what compounds should be increased in concentration in a fabricated diet, and the only way to assess such unintended effects is to feed the actual GE crop to test animals. For testing GE maize, soybean, and rice (Oryza sativa),3 flour from kernels or seed is added to an animal’s diet and constitutes between about 10–60 percent of the diet. The high percentages can be used because the crop products are nutritious for the animal. In the case of whole foods that are not typically part of a rodent’s diet, whether GE or non-GE, it is impossible to achieve very high concentrations of the test food because it would cause nutritional imbalance. The whole-food tests done for regulatory agencies are generally conducted for 28 or 90 days with rats, but some researchers have run tests for multiple generations.
The utility of the whole-food tests has been questioned by a number of government agencies and by industry and academic researchers (for example, Ricroch et al., 2014), and they are not an automatic part of the regulatory requirements of most countries that have specific GE food-testing requirements (CAC, 2008; Bartholomaeus et al., 2013). However, in its 2010 report A Decade of EU-Funded GMO Research (2001–2010), the European Directorate-General for Research and Innovation concluded that “the data from a well-designed 90-day rodent feeding study, together with data covering the gene insert, the compositional analysis, and the toxicity of the novel gene product, form the optimal basis for a comparative assessment of the safety of [genetically engineered] food and its conventional counterpart in the pre-market situation” (EC, 2010a:157). The European Food Safety Authority (EFSA) developed principles and guidance for establishing protocols for 90-day whole-food studies in rodents at the European Commission’s request (EFSA, 2011b), and 90-day, whole-food studies were made mandatory by the European Commission (EC, 2013). Most studies reported in the peer-reviewed literature have concluded that there was a lack of adverse effects of biological or toxicological significance (see, for example, Knudsen and Poulsen, 2007; MacKenzie et al., 2007; He et al., 2008, 2009; Onose et al., 2008; Liu et al., 2012), even though some of the studies found statistically significant differences between the GE and non-GE comparator in toxicity.
The criticisms of whole-food tests come from two perspectives. One perspective is that whole-food studies cannot provide useful tests of food
3 GE rice was not commercialized in 2015, but GE varieties in development have been tested.
safety because they are not sensitive enough to detect differences (see, for example, Bartholomaeus et al., 2013; Kuiper et al., 2013; Ricroch et al., 2013a, 2014) and that animal testing is not needed because other types of required testing ensure safety (Bartholomaeus et al., 2013; Ricroch et al., 2014). Ricroch et al. (2014) pointed to the costs of the 90-day tests, which they reported as being €250,000 (in 2013 money). The second perspective is that whole-food tests could be useful, but there is concern about their design and conduct or about the parties who conduct them (the companies commercializing the GE crops). That perspective is evident in Séralini et al. (2007), Domingo and Bordonaba (2011), Hilbeck et al. (2015), and Krimsky (2015). Boxes 5-3 and 5-4 describe some of the specific procedures and practices involved in doing these tests.
The committee heard from invited speakers (Entine, 2014; Jaffe, 2014) and members of the public who provided comments at meetings and it received a number of written public comments highlighting the work of one research group (Séralini et al., 2012, 2014) that has conducted a number of whole-food studies of GE herbicide-resistant and insect-resistant crops and
of direct consumption of glyphosate. Some comments made to the committee pointed to the publications of that research group as evidence that GE crops and foods derived from GE crops were deleterious to human health; other comments questioned the robustness and accuracy of the research. The committee also heard from the lead researcher himself at one of its meetings (Séralini, 2014). Because of the attention garnered by this specific research group, the committee examined the primary research paper from the group and many articles related to it (Box 5-5).
A general question that remains for all whole-food studies using animals is, How many animals, tested for how long, are needed to assess food safety when a whole food is tested? That question is related to the question of how large an effect the tested food would have to have on the animal for it to be detected with the experiment. The statistical procedure called power analysis can answer the first question, but the committee did not find such analyses in articles related to GE crop whole-food studies. The EFSA scientific committee (EFSA, 2011b) provided general guidance on power analysis. Figure 5-2, from the EFSA report, shows the relationship between the number of experimental units (cages with two animals) per treatment group and the power of an experiment in standard-deviation units. Standard deviations quantify how much the measurement of a trait or effect varies among animals that have been given the same diet. The report concluded that, if researchers follow OECD Test No. 408 of 10 males and 10 females per treatment (OECD, 1998a), a test should be able to detect a difference equal to about 1 standard deviation (with 90-percent confidence) unless the food has a different effect on males and females, in which case, the smallest difference that could be detected would be about 1.5 standard deviations from the experimental mean.
Because the relationship is quite abstract for the nonstatistician, the committee examined the size of the standard deviations in a number of whole-food safety articles. It found that the sizes of the standard deviations compared with the mean value of a measured trait depended heavily on the trait being measured and on the specific research article. For example, in the Hammond et al. (2004) study, the average white blood cell count for the four treatments, each with 9 or 10 female Sprague-Dawley rats, is 6.84 103/µl, and the average standard deviation is 1.89 103/µl. On the basis of rough calculations, this test would have the power to discern statistically whether the GE food caused an increase in white blood cell count of about 35 percent with about 90-percent confidence. If the male white blood cell count effects and standard deviations were similar to those in females, the test could have found about a 25-percent increase.
OECD (1998a) made general recommendations, such as those used in
Hammond et al. (2004), for the number of units (cages with two animals) per treatment. Following these guidelines leads to the assumption that less than a 25-percent change in the white blood cell count was not biologically relevant. The EU Standing Committee on the Food Chain and Animal Health adopted the mandatory use of 90-day whole-food testing of GE crops, and its protocols generally follow OECD guidelines for the testing of chemicals (EC, 2013).
EFSA also published a document (EFSA, 2011c) that focused specifically on the questions, What is statistical significance? and What is biological relevance? The accessibly written document makes clear that the two are very different and that it is important to decide how large a difference is biologically relevant before designing an experiment to test a null hypothesis of no difference. The problem in most whole-food animal studies is in determining how large a biological difference is relevant. Most of the statistically significant differences observed in the literature on the animal-testing data were around a 10- to 30-percent change, but the authors do not give detailed explanations of why they conclude that a statistically significant difference is not biologically relevant. A general statement is sometimes made that the difference is within the range for the species, but because the range of values for the species typically come from multiple laboratories, such a statement is not useful unless the laboratories, instrumentation, and health of the animals were known to be comparable.
Clearly, the European Commission relied on both expert judgment and citizen concerns in making its assessment of biological relevance of the effects of GE foods in requiring 90-day testing. It is reasonable to ask what balance of the two is the basis for this judgment. As pointed out by the 2002 National Research Council report, “risk analysis of transgenic plants must continue to fulfill two distinct roles: (1) technical support for regulatory decision making and (2) establishment and maintenance of regulatory legitimacy” (NRC, 2002:6). Fulfilling the two roles can lead to different country-specific and region-specific decisions. This issue is discussed further in Chapter 9.
One specific criticism of the 90-day whole-food studies revolves around an EU-funded project conducted by Poulsen et al. (2007) in which rice was genetically engineered to produce the kidney bean lectin, agglutinin E-form, which is known to have toxic properties. In a 90-day test, rats were fed diets of 60-percent rice with the lectin gene or 60-percent rice without the lectin gene. The researchers concluded that they did not find any meaningful differences between the two treatments. However, in a treatment in which the diets were spiked with 0.1-percent recombinant lectin (a high dose), biological effects including significant differences in weight of small intestines, stomach, and pancreas and in plasma biochemistry were found. Poulsen et al. included results from a preceding 28-day feeding study and
compositional analyses of the rice diets. The criticism involves the question, If a whole-food study with a known toxin does not demonstrate effects, how can the test be considered useful? (Bartholomaeus et al., 2013). If a whole-food study with an animal finds statistically significant effects, there is obviously a need for further safety testing, but when there is a negative result, there is uncertainty as to whether there is an adverse effect on health. In the specific case of lectin gene in rice, one could argue that the statistical power of the whole-food test was insufficient or that, when the toxin is in the structure of the food, it is no longer toxic so the food is safe.
Other Long-Term Studies with Rodents. In addition to the work of Séralini et al. (2012, 2014), there have been other long-term rodent studies, some of which included multiple generations. Magana-Gomez and de la Barca (2009), Domingo and Bordonaba (2011), Snell et al. (2012), and Ricroch et al. (2013b) reviewed the studies. Some found no statistically significant differences, but quite a few found statistically significant differences that the authors generally did not consider biologically relevant, typically without providing data on what was the normal range. In the multigeneration studies, the sire and dam are dosed via the diet before conception, and the parent generation and pups are dosed via the diet throughout the duration of the study to determine multiple generational outcomes, including growth, behavior, and phenotypic characteristics. Some studies have looked at three or four generations. For example, Kiliç and Akay (2008) conducted a three-generation rat study in which 20 percent of the diet was Bt maize or a non-Bt maize that otherwise was genetically similar. All generations of female and male rats were fed the assigned diets, and the third-generation offspring that were fed the diets were sacrificed after 3.5 months for analysis. The authors found statistical differences in kidney and liver weights and long kidney glomerular diameter between the GE and non-GE treatments but considered them not biologically relevant. Similarly, statistically significant differences were observed in amounts of globulin and total protein between the two groups. There was no presentation of standards used for judging what would be a biologically relevant difference or for what the normal range was in the measurements.
The standard deviations in measurements of the traits (that is, effects) of individual animals in a treatment in the long-term studies were similar to those of studies of shorter duration. Therefore, the power of the tests to detect statistically significant differences was in the range of 10–30 percent. The committee could not find justification for considering this statistical power sufficient. It can be argued that the number of replicates (number of units of two animals per treatment) in the studies should be substantially increased, but one argument against an increase in numbers is related to the ethics of subjecting more animals to testing (EC, 2010b). One could
also argue that it is unethical to conduct an underpowered study. However, most if not all of the rodent studies are based on widely accepted safety evaluation protocols with fixed numbers of animals per treatment. Cultural values regarding precaution for human safety and those regarding the number of animals subjected to testing are in conflict in this case. As pointed out by Snell et al. (2012), a close examination of the long-term and multigenerational studies reveals that some have problems with experimental design, the most common being that the GE and non-GE sources were not isogenic and were grown in different locations (or unknown locations). Those problems in design make it difficult to determine whether differences are due to the genetic-engineering process or GE trait or to other sources of variation in the nutritional quality of the crops.
In cases in which testing produces equivocal results or tests are found to lack rigor, follow-up experimentation with trusted research protocols, personnel, and publication outlets is needed to decrease uncertainty and increase the legitimacy of regulatory decisions. There is a precedent of such follow-up studies in the literature on GE crop environmental effects that could serve as a general model for follow-up food-safety testing (see Chapter 4 section “Genetically Engineered Crops, Milkweed, and Monarch Butterflies”). The USDA Biotechnology Risk Assessment Research Grants Program has enabled this approach in a few cases.
Beyond Rodent Studies. Mice and rats are typically used in toxicity studies because of their general physiological similarities to humans and their small size, but some farm animals are considered to be better models of human physiology than rodents. The best example is the pig, which is considered to be better than rodents as a model, especially with respect to nutritional evaluations (Miller and Ullrey, 1987; Patterson et al., 2008; Litten-Brown et al., 2010). Porcine insulin has been used for decades to control blood sugar in patients who have childhood-onset diabetes mellitus (type I diabetes). Pig heart valves are used for human mitral valve replacement, and pig skin has been investigated as a possible donor tissue. The pig is monogastric as is the human, and its gastrointestinal tract absorbs and metabolizes nutrients (lipids and micronutrients) in the same manner as in humans.
Reviews of studies with animals fed GE foods have included studies using both rodents and farm animals (Bartholomaeus et al., 2013; DeFrancesco, 2013; Ricroch et al., 2013a,b, 2014; Swiatkiewicz et al., 2014; Van Eenennaam and Young, 2014). Those animal studies have taken advantage of the fact that maize and soybean are major components of the diets of many farm animals. Some of the reported studies that used farm animals have designs similar to those of rodent studies and have variation in duration and replicates similar to that of the rodent experiments. Some of the tests were run for 28 days (for example, Brouk et al.,
The experiments with pigs are especially relevant. Most of them were conducted in one prolific laboratory (Walsh et al., 2011, 2012a,b, 2013; Buzoianu et al., 2012a,b,c,d, 2013a,b). The studies range from examination of short-term growth of piglets to multigenerational studies of sows and piglets, with mixed designs having either generation or both exposed to Bt maize and non-Bt maize. Characteristics measured included food consumption and growth, assessment of organ size and health, immunological markers, and microbial communities. The authors of the studies generally concluded that Bt maize does not affect health of the pigs, but they reported a number of statistically significant differences between Bt maize treatment and control maize treatment. In one experiment (Walsh et al., 2012a), the weaned piglets that were fed Bt maize had lower feed-conversion efficiency during days 14–30 (P > 0.007) but no significant effect over the full span of the experiment. In another experiment (Buzoianu et al., 2013b), there was lower efficiency in the Bt treatment during days 71–100 (P > 0.01) but again no effect over the full span of the experiment.
In those experiments with pigs and experiments with other farm animals and rodents, there was apparently one source of the GE food and one source of the non-GE food per study, and it is generally not clear that the food sources were isogenic or grown in the same location. That makes it difficult to determine whether any statistical differences found were due to the engineered trait or to the batches of food used, which in at least some experiments varied in nutrient content and may have differed in bioactive compounds (produced in response to plant stressors), which may have a profound effect on outcomes of nutritional studies. Another issue is that many statistical tests were performed in most studies. That could result in accumulation of false-positive results (Panchin and Tuzhikov, 2016). Although this is not a situation in which a stringent correction for doing multiple tests is called for (Dunn, 1961), there is reason to be cautious in interpretation of statistical significance of individual results because multiple tests can lead to artifactual positive results. The issue of multiple test results is common in many fields, and one approach used in genetics is to use the initial tests for hypothesis generation with follow-up experiments that test an a priori hypothesis (for example, Belknap et al., 1996). If a straightforward application of Bonferonni correction is used, each animal study that measures multiple outcomes, whether for GE crops or any other potential toxicant, could require over 1,000 animals to obtain reasonable statistical power (Dunn, 1961).
In addition to the literature on controlled experiments with livestock, Van Eenennaam and Young (2014) reviewed the history of livestock health
and feed-conversion ratios as the U.S. livestock industry shifted from non-GE to GE feed. Producers of cattle, milk cows, pigs, chickens, and other livestock are concerned about the efficiency of conversion of animal feed into animal biomass because it affects profit margins. The data examined start as early as 1983 and run through 2011. Therefore, livestock diets shifted from all non-GE feed to mostly GE feed within the duration of the study. Van Eenennaam and Young found that, if anything, the health and feed-conversion efficiencies of livestock had increased since the introduction of GE crops but that the increase was a steady rise, most likely because of more efficient practices not associated with use of GE feed. In the studies that they reviewed, the number of animals examined was large (thousands). Of course, most livestock are slaughtered at a young age, so that data cannot address the issue of longevity directly. However, given the general relationship between general health and longevity, the data are useful.
FINDING: The current animal-testing protocols based on OECD guidelines for the testing of chemicals use small samples and have limited statistical power; therefore, they may not detect existing differences between GE and non-GE crops or may produce statistically significant results that are not biologically meaningful.
FINDING: In addition to experimental data, long-term data on the health and feed-conversion efficiency of livestock that span a period before and after introduction of GE crops show no adverse effects on these measures associated with introduction of GE feed. Such data test for correlations that are relevant to assessment of human health effects, but they do not examine cause and effect.
RECOMMENDATION: Before an animal test is conducted, it is important to justify the size of a difference between treatments in each measurement that will be considered biologically relevant.
RECOMMENDATION: A power analysis for each characteristic based on standard deviations in treatments in previous tests with the animal species should be done whenever possible to increase the probability of detecting differences that would be considered biologically relevant.
RECOMMENDATION: In cases in which early published studies produced equivocal results regarding health effects of a GE crop, followup experimentation using trusted research protocols, personnel, and publication outlets should be used to decrease uncertainty and increase the legitimacy of regulatory decisions.
RECOMMENDATION: Public funding in the United States should be provided for independent follow-up studies when equivocal results are found in reasonably designed initial or preliminary experimental tests.
Compositional Analysis of Genetically Engineered Crops. As part of the regulatory process of establishing substantial equivalence, GE crop developers submit data comparing the nutrient and chemical composition of their GE plant with a similar (isoline) variety of the crop. In the United States, submitting such data to FDA is voluntary, although as of 2015 this seems to always be done by developers. Developers and regulators compare key components of the GE variety with published reference guides that list the concentrations and variabilities of nutrients, antinutrients, and toxicants that occur in crops already in the food supply.4 The section “Regulatory Testing of Crops with Resistance to Glyphosate and 2,4-D and the New Uses of the Herbicides Themselves” earlier in this chapter gives an example of the types of nutrients and chemicals that are generally measured. In the specific case of the soybean resistant to 2,4-D and glyphosate, measurements of 62 components in the soybean were submitted by Dow AgroSciences. There were statistically significant differences between the GE and comparison varieties in 16 of the 62. The differences were considered to be small and within the range of published values for other soybean varieties. They were therefore “considered not biologically relevant.” In compositional analysis, as in some of the whole-food animal testing, it is difficult to know how much of the variance and range in values for the components is due to the crop variety, the growing conditions, and the specific laboratory experimental equipment. In the United States, regulatory agencies require that the comparison be between the GE crop and its isogenic conventionally bred counterpart grown in side-by-side plots. In those cases, it is hard to attribute differences to anything but the genetic-engineering process.
FINDING: Statistically significant differences in nutrient and chemical composition have been found between GE and non-GE plants by using traditional methods of compositional analysis, but the differences have been considered to fall within the range of naturally occurring variation found in currently available non-GE crops.
4 OECD develops consensus documents that provide reference values for existing food crops (OECD, 2015). These are publicly available online at http://www.oecd.org/science/biotrack/consensusdocumentsfortheworkonthesafetyofnovelfoodsandfeedsplants.htm (accessed May 9, 2016). The International Life Science Institute (ILSI) also maintains a crop composition database at www.cropcomposition.org (accessed May 9, 2016). ILSI reports that in 2013 the database contained more than 843,000 data points representing 3,150 compositional components.
Composition of Processed Genetically Engineered Foods. General compositional analysis and the specific content of the introduced proteins are typically conducted on raw products, such as maize kernels or soybean seed. However, much of the human consumption of these products occurs after substantial exposure to heat or other processing. If in processing of foods the amounts of GE proteins substantially increase, consumers are potentially exposed to a risk that is different from that anticipated from testing the raw material. In the production of oil, for example, the goal is to separate the oil from other compounds in the raw crop, such as proteins and carbohydrates. Crude oils can contain plant proteins (Martín-Hernández et al., 2008), but in highly purified oils even sophisticated approaches have failed to find any nondegraded proteins (Hidalgo and Zamora, 2006; Martín-Hernández et al., 2008). Those results are reflected in the fact that people who are allergic to soybean are not affected by purified oils (Bush et al., 1985; Verhoeckx et al., 2015).
A few studies have searched for a means of finding DNA in plant-derived oils to identify the origin of the oil as GE or non-GE for labeling purposes (Costa et al., 2010a,b) or to identify the origin of olive oil (Muzzalupo et al., 2015). It is possible to detect DNA, but the amounts are typically diminished in purified oils to 1 percent or less of the original content. Similarly, Oguchi et al. (2009) were not able to find any DNA in purified beet sugar. Some countries exempt products from labeling if GE protein or DNA is not detectable. For example, in Japan, where foods with GE ingredients typically require labeling, oil, soy sauce, and beet sugar are excluded because of degradation of GE proteins and DNA (Oguchi et al., 2009). Australia and New Zealand have similar exemptions from labeling for such highly refined foods as sugars and oils (FSANZ, 2013).
The detection of GE protein and DNA in other processed foods depends on the type of processing. For example, the amount of the Bt protein Cry1Ab detected by immunoassay in tortillas depends on cooking time (de Luis et al., 2009). The detected amount of Cry9C protein remaining in samples of corn bread, muffins, and polenta was about 13, 5, and 3 percent of the amount in the whole-grain maize (Diaz et al., 2002). For Cry1Ab in rice, Wang et al. (2015) found that baking was more effective in lowering the detection using polyclonal antibodies of the Cry1Ab protein than microwaving, but 20 minutes of baking at 180ºC left almost 40 percent of the protein intact. Heat denaturation of proteins can lower antibody binding to epitopes and cause lower detection of GE proteins.
FINDING: The amount of GE protein and DNA in food ingredients can depend on the specific type of processing; some foods contain no detectable protein and little DNA. In a few countries that have manda-
tory labeling of GE foods, that is taken into account, and food without detectable GE DNA or GE protein is not labeled.
Newer Methods for Assessing Substantial Equivalence. As explained in Chapter 2, governance of GE crops includes regulatory governance. Although not required to by governing bodies, companies and academic researchers have moved beyond the typical measurements of food composition to newer technologies that involve transcriptomics, proteomics, and metabolomics. The new methods provide a broad, nontargeted assessment of thousands of plant characteristics, including the concentrations of most of the messenger RNAs, proteins, and small molecules in a plant or food. These methods are more likely to detect changes in a GE crop than the current regulatory approaches. If a GE crop has been changed only as intended, any changes observed in these -omics measurements theoretically should be predictable in a given environment. The science behind the methods, including the current limitations of their interpretation, is discussed in Chapter 7. The discussion here focuses on how the methods have already been applied in the assessment of risk of health effects of currently commercialized GE crops.
Ricroch et al. (2011) reviewed -omics data from 44 studies of crops and detailed studies of the model plant Arabidopsis thaliana. Of those studies, 17 used transcriptomics, 12 used proteomics, and 26 used metabolomic methods. Ricroch (2013) updated the number of studies to 60. The committee found that many more studies had been done since those reviews were published, and many of them have used multiple -omics approaches. The sophistication of the studies has increased (Ibáñez et al., 2015) and is likely to increase further. As recommended in Chapter 7, there is a need to develop further and share databases that contain detailed -omics data (Fukushima et al., 2014; Simó et al., 2014).
In some studies of GE plants in which simple marker genes were added, there were almost no changes in the transcriptome (El Ouakfaoui and Miki, 2005), but use of other -omics methods has revealed changes (Ren et al., 2009). For example, in a comparison of glyphosate-resistant soybean and non-GE soybean, García-Villalba et al. (2008) found that three free amino acids, an amino acid precursor, and flavonoid-derived secondary metabolites (liquiritigenin, naringenin, and taxifolin) had greater amounts in the GE soybean and 4-hydroxy-l-threonine was present in the non-GE soybean, but not in the GE variety. They hypothesized that the change in the flavonoids may have been because the modified EPSPS enzyme (a key enzyme of the shikimate pathway leading to aromatic amino acids) introduced to achieve glyphosate resistance could have different enzymatic properties that influenced the amounts of aromatic amino acids. The committee was not aware of such a hypothesis before this metabolomic study. (A concern was expressed in a comment submitted to the committee that
the EPSPS transgene would cause endocrine disruption. The committee found no evidence to suggest that the changes found by García-Villalba et al. would have such an effect.)
On the basis of previous experimentation, it is predicted that, when a gene for a nonenzymatic protein (such as a Bt toxin gene) is added to a plant, there will be very few changes in the plant’s metabolism (Herman and Price, 2013). However, when a gene has been added specifically to alter one metabolic pathway of a plant, a number of predicted and unpredicted changes have been found. For example, Shepherd et al. (2015) found that, when they downregulated enzymes (that is, decreased expression or activity) involved in the production of either of two toxic glycoalkaloids (alpha-chaconine and alpha-solanine) in a GE potato with RNA-interfering transgenes that regulated synthesis of one toxic glycoalkaloid, the other compound usually increased. When they downregulated production of both compounds, beta-sitosterol and fucosterol increased. Neither of these compounds has the degree of toxicity associated with alpha-chaconine and alpha-solanine. Other compounds also differed from controls in concentration, but some of the changes may have been due to products generated during the tissue-culture process used in these experiments and not to the transgenes.
Many of the studies have found differences between the GE plants and the isogenic conventionally bred counterparts, but for many components there is more variation among the diverse conventionally bred varieties than between the GE and non-GE lines (Ricroch et al., 2011, Ricroch, 2013). Furthermore, the environmental conditions and the stage of the fruit or seed affect the finding. Chapter 7 addresses the future utility of the -omics approaches in assessing the biological effects of genetic engineering.
FINDING: In most cases examined, the differences found in comparisons of transcriptomes, proteomes, and metabolomes in GE and non-GE plants have been small relative to the naturally occurring variation found in conventionally bred crop varieties due to genetics and environment.
FINDING: If an unexpected change in composition beyond the natural range of variation in conventionally bred crop varieties were present in a GE crop, -omics approaches would be more likely to find the difference than current methods.
FINDING: Differences in composition found by using -omics methods do not, on their own, indicate a safety problem.
Food Allergenicity Testing and Prediction
Allergenicity is a widespread adverse effect of foods, several plants, tree and grass pollens, industrial chemicals, cosmetics, and drugs. Self-reporting of lifetime allergic responses to each of the most common food allergens (milk, egg, wheat, soy, peanut, tree nuts, fish, and shellfish) ranges from 1 to 6 percent of the population (Nwaru et al., 2014). Allergies are induced in a two-step process: sensitization from an initial exposure to a foreign protein or peptide followed by elicitation of the allergic response on a second exposure to the same or similar agent. Sensitization and elicitation are generally mediated by immunoglobulins, primarily IgE, and the responses may range from minor palatal or skin itching and rhinitis to severe bronchial spasms and wheezing, anaphylaxis, and death. In addition to IgE responses to food allergens, IgA has been identified as an inducible immune mediator primarily in the gastrointestinal mucosa in response to foods, foreign proteins, pathogenic microorganisms, and toxins. The role of IgA in classical allergy has been investigated (Macpherson et al., 2008).
Assessment of the potential allergenicity of a food or food product from a GE crop is a special case of food-toxicity testing and is based on two scenarios: transfer of any protein from a plant known to have food-allergy properties and transfer of a protein that could be a de novo allergen. Predictive animal testing for allergens in foods (GE and non-GE) is not sufficient for allergy assessment (Wal, 2015). Research efforts are ongoing to discover or develop an animal model that predicts sensitization to allergy (Ladics and Selgrade, 2009), but so far none has proved predictive (Goodman, 2015). Therefore, researchers have relied on multiple indirect methods for predicting whether an allergic response could be caused by a protein that is either added to a food by genetic engineering or appears in the food as an unintended effect of genetic engineering. Endogenous protein concentrations with known allergic properties also have to be monitored because it is possible that their concentration could increase due to genetic engineering.
A flow diagram of the interactive approach to allergen testing recommended by the Codex Alimentarius Commission (CAC, 2009) and EFSA (2010, 2011a) is presented in Figure 5-3 (Wal, 2015); Box 5-2 describes the EPA testing of the Bt toxin Cry1F that generally follows this approach. The logic behind the approach starts with the fact that any gene for a protein that comes from a plant that is known to cause food allergies has a higher likelihood of causing allergenicity than any gene from a plant that does not cause an allergic response. If the introduced protein is similar to a protein already known to be an allergen, it becomes suspect and should be tested in people who have an allergy to the related protein. Finally, if a protein fits none of the above characteristics but is not digested by simulated gastric fluid, it could be a novel food allergen. The latter factor comes from
research demonstrating that some, but not all, proteins already known to be food allergens are resistant to digestion by gut fluid.
There is one case in which that approach was used and a GE crop with allergenicity issues was detected early and prevented from being commercialized, and a second case in which a GE crop was withdrawn from the market based on the possibly that it included a food allergen. In the first case, research was conducted on a soybean line genetically engineered to produce a Brazil nut (Bertholletia excelsa) protein, which was a known allergen. Sera from patients allergic to Brazil nut protein were available and tested positive for activity against the GE soybean protein. Because the segregation from the human food supply of GE soybean with that protein could not be guaranteed, the project was halted (Nordlee et al., 1996). The soybean variety was never commercialized.
In the second case, EPA allowed a Bt maize variety developed by Aventis CropScience with a potential for allergenicity (due to decreased digestion of the protein Cry9c in simulated gastric fluid) to be sold as cattle feed under the name StarLink™; because of the potential for allergenicity, the variety was not approved for direct human consumption. However, the Bt protein was found in human food, so the maize variety was removed from all markets. After that incident, EPA no longer distinguished between Bt proteins in human food versus in animal feed (EPA, 2001b). Bt crop varieties are approved in the United States for all markets or none.
The interactive approach for testing should work for GE crops when the testing is for a transgene that is expressed by the plant as a protein that does not affect its metabolism (for example, Bt toxins). The testing does not cover endogenous allergens whose concentrations have been increased by unintended effects of genetic engineering. In 2013, the European Commission set a requirement for assessing endogenous allergens in GE crops (EC, 2013). A number of articles since then have supported the approach (Fernandez et al., 2013) or have found it unnecessary and impractical (Goodman et al., 2013; Graf et al., 2014). Soybean is an example of a crop that has endogenous allergens. A paper on endogenous soybean allergens concluded that there is enough knowledge of only some soybean allergens for proper testing (Ladics et al., 2014). As emphasized by Wal (2015), there is considerable variation among conventionally bred varieties in the concentrations of endogenous allergens, especially when they are grown under different conditions. Therefore, the existing variation must be taken into consideration in assessing a GE variety. Of course, the issue is not only the magnitude of variation but the potential change in the overall exposure of the global human population to the allergen.
One example of an existing potential allergen of concern is gamma-zein, one of the storage proteins produced in the maize kernel that is a comparably hard-to-digest protein (Lee and Hamaker, 2006). Concern was expressed to the committee that GE maize may have higher amounts of gamma-zein, which could be allergenic (Smith, 2014). Krishnan et al. (2010) found that young pigs consuming maize generate antibodies against gamma-zein. That observation and the fact that the protein withstands pepsin digestion suggest that gamma-zein could be an allergen. In a comparison of the Bt maize line MON810 with non-Bt maize, known maize allergens, including the 27-kDa and 50-kDa gamma-zein proteins, were not found to be in significantly different amounts (Fonseca et al., 2012). On the other hand, conventionally bred Quality Protein Maize is reported to have a 2 to 3 fold higher
There can be a connection between immune response and allergenicity. One well-cited study brought up in the public comment period was that by Finamore et al. (2008), who assessed the effect of Bt maize ingestion on the mouse gut and peripheral immune system. They found that Bt maize produced small but statistically significant changes in percentage of T and B cells and of CD4+, CD8+, γδT, and αβT subpopulations at gut and peripheral sites and alterations of serum cytokines in weanlings fed for 30 days and in aged mice. However, there was no significant response in weaning mice that were fed for 90 days, which they related to further maturation of the immune system. They concluded that there was no evidence that the Bt toxin in maize caused substantial immune dysfunction. Similarly, Walsh et al. (2012a) did not find immune function changes in a long-term pig feeding study (80 or 110 days) on Bt MON810 maize compared with non-GE maize. Overall, no changes of concern regarding Bt maize feeding and altered immune response have been found.
At a public meeting that the committee held on health effects of GE foods, a question was raised about whether current testing for allergenicity is insufficient because some people do not have acidic conditions in their stomachs. Regarding that issue, digestibility of the proteins is assessed with simulated gastric fluid (0.32 percent pepsin, pH 1.2, 37ºC), under the premise that an undigested protein may lead to the absorption of a novel allergenic fragment (Astwood et al., 1996; Herman et al., 2006). Stomach fluid is typically acidic, with a pH of 1.5–3.5, which is the range at which pepsin (the digestive enzyme of the stomach) is active, and the volume of stomach fluid is 20–200 mL (about 1–3 ounces). Simulated gastric fluid was developed to represent human gastric conditions in the stomach and is used in bioavailability studies of drugs and foods (U.S. Pharmacopeia, 2000).
In general, if the pH of the stomach is greater than 5, pepsin will not be active, and less breakdown of large proteins will take place. Hence, the usefulness of simulated gastric fluid in the case of a less acidic (higher pH) stomach is questionable, whether used for non-GE foods or GE foods. Untersmayr and Jensen-Jarolim (2008:1301) concluded that “alterations in the gastric milieu are frequently experienced during a lifetime either physiologically in the very young and the elderly or as a result of gastrointestinal pathologies. Additionally, acid-suppression medications are frequently used for treatment of dyspeptic disorders.” Trikha et al. (2013) used a group of 4,724 children (under 18 years old) who had received a
5 Jung, R., W.-N. Hu, R.B. Meeley, V.J.H. Sewalt, and R. Nair. Grain quality through altered expression of seed proteins. U.S. Patent 8,546,646, filed September 14, 2012, and issued October 1, 2013.
diagnosis of gastroesophageal reflux disease (GERD) and who were treated with gastric acid-suppressive medication and matched with 4,724 children who had GERD but were not so treated. Those treated with acid-reducing medicine were more than 1.5 times as likely to have a diagnosis of food allergy as those who were not so treated. The difference between the two GERD groups was statistically significant (hazard ratio, 1.68; 95-percent confidence interval, 1.15–2.46).
The National Research Council report Safety of Genetically Engineered Foods pointed out that there were important limitations in allergenicity predictions that could be done before commercialization (NRC, 2004). Since that report was published, there have been improvements in the allergen database, and research has been funded to improve precommercialization prediction. However, as the committee heard from an invited speaker, “no new methods have been demonstrated to predict sensitization and allergy in the absence of proven exposure” (Goodman, 2015). Before commercialization, the general population will probably not have been exposed to an allergen similar enough to an allergen in a GE plant to cause cross-reactivity, so it would be useful to use the precommercialization tests only as a rough predictor. To ensure that allergens did not remain in the food system, the Safety of Genetically Engineered Foods report called for a two-step process of precommercialization testing and post-commercialization testing. Even though progress has been made on allergenicity prediction since that report was published in 2004, the committee found that post-commercialization testing would be useful in ensuring that no new allergens are introduced. There have been no steps toward post-commercialization testing since 2004. The committee recognized that such testing would be logistically challenging, as described in a scientific report to EFSA (ADAS, 2015). Post-commercialization surveillance of such specific agents as drugs and medical devices is difficult, but there is generally a well-defined endpoint to look for in patients. In the case of food, the detection of an allergic response to a particular protein would be confounded by multiple exposures in the diet. However, several region-wide human populations have been exposed to GE foods for many years whereas others have not; this could enable an a priori hypothesis to be tested that populations that have been exposed to foods from specific GE crops will not show a higher rate of allergic response to such foods.
FINDING: For crops with endogenous allergens, knowing the range of allergen concentrations in a broad set of crop varieties grown in a variety of environments is helpful, but it is most important to know whether adding a GE crop to the food supply will change the general exposure of humans to the allergens.
FINDING: Because testing for allergenicity before commercialization could miss allergens to which the population had not previously been exposed, post-commercialization allergen testing would be useful in ensuring that consumers are not exposed to allergens, but such testing would be difficult to conduct.
FINDING: There is a substantial population of persons who have higher than usual stomach pH, so tests of digestibility of proteins in simulated acidic gastric fluid may not be relevant to this population.
The overall results of short-term and long-term animal studies with rodents and other animals and other data on GE-food nutrient and secondary compound composition convinces many (for example, Bartholomaeus et al., 2013; Ricroch et al., 2013a,b; Van Eenennaam and Young, 2014) but not all involved researchers (for example, Dona and Arvanitoyannis, 2009; Domingo and Bordonaba, 2011; Hilbeck et al., 2015; also see DeFrancesco, 2013) that currently marketed GE foods are as safe as foods from conventionally bred crops. The committee received comments from an invited speaker (Smith, 2014) and from the public regarding the possible relationship between increases in the incidence of specific chronic diseases and the introduction of GE foods into human diets. Appendix F includes a representative list of the comments about GE food safety that were sent to the committee through the study’s website. The comments mentioned concerns about such chronic diseases as cancers, diabetes, and Parkinson’s; possible organ-specific injuries (liver and kidney toxicity); and such disorders as autism and allergies. Smith (2003:39) made the claim that “diabetes rose by 33 percent from 1990 to 1998, lymphatic cancers are up, and many other illnesses are on the rise. Is there a connection to [genetically modified] foods? We have no way of knowing because no one has looked for one.”
As part of the committee’s effort to respond to its task to “assess the evidence for purported negative effects of GE crops and their accompanying technologies,” it used available peer-reviewed data and government reports to assess whether any health problems may have increased in frequency in association with commercialization of GE crops or were expected to do so on the basis of the results of toxicity studies. The committee presents additional biochemical data from animal experiments but relies mostly on epidemiological studies that used time-series data. The epidemiological data for some specific health problems are generally robust over time (for example, cancers) but are less reliable for others. The committee presents the available data knowing that they include a number of sources of bias,
including changes over time in survey methods and in the tools for detection of specific chronic diseases. As imperfect as the data may be, they are in some cases the only information available beyond animal experiments for formulating or testing hypotheses about possible connections between a GE food and a specific disease. The committee points out that the lack of rigorous data on incidence of disease is not only a problem for assessing effects of GE foods on health. More rigorous data on time, location, and sociocultural trends in disease would enable better assessment of potential health problems caused by environmental factors and other products from new technologies.
A review of the American Cancer Society’s database indicates that mortality from cancers in the United States and Canada has continued to decrease or stabilized in all categories except cancers of the lung and bronchus attributable to smoking. The decreases in mortality are due in part to early detection and improved treatment, so mortality data can mask the rate at which cancers occur. For that reason, the committee sought data on cancer incidence rather than cancer mortality. Figures 5-4 and 5-5 show
changes in cancer incidence in U.S. women and men, respectively, from 1975 to 2011 (NCI, 2014). If GE foods were causing a substantial number of specific cancers, the incidence of those cancers would be expected to show a change in slope in the time series after 1996, when GE traits were first available in commercial varieties of soybean and maize. The figures show that some cancers have increased and others decreased, but there is no obvious change in the patterns since GE crops were introduced into the U.S. food system. Figures 5-6 and 5-7 show cancer incidence in women and men in the United Kingdom, where GE foods are not generally being consumed. For the specific types of cancers that are reported in both the United States and the United Kingdom, there is no obvious difference in the patterns that could be attributed to the increase in consumption of GE foods in the United States. (The absolute numbers cannot be compared because of differences in methodology.)
Forouzanfar et al. (2011) published data on breast and cervical cancer incidence worldwide from 1980 to 2010. As can be seen in Figure 5-8, the global incidence of those two cancers has increased. An examination of the plots for North America (high income) (Canada and the United States), where GE foods are eaten, compared with the plots for western
income] and western Europe are different from those in the studies above on the incidence of cancer in the United States and the United Kingdom.)
Taken together, Figure 5 through Figure 8 do not support the hypothesis that GE foods have resulted in a substantial increase in the incidence of cancer. However, they do not establish that there is no relationship between cancer and GE foods because there can be a delay in the onset of cancer that would obscure a trend, and one could hypothesize that something else has occurred with GE foods in the United States that has lowered cancer incidence and thus obscured a relationship. The committee had limited evidence on which to make its judgments, but the evidence does not support claims that the incidence of cancers has increased because of consumption of GE foods.
There is ongoing debate about potential carcinogenicity of glyphosate in humans. Assessment of glyphosate is relevant to the committee’s report because it is the principal herbicide used on HR crops (Livingston, et al. 2015), and it has been shown that there are higher residues of glyphosate in HR soybean treated with glyphosate than in non-GE soybean (Duke et al., 2003; Bøhn et al., 2014). Box 5-5 provides details about a study by Séralini et al. (2012, 2014) that concluded that glyphosate causes tumors in rats. The committee found that this study was not conclusive and used incorrect statistical analysis. The most detailed epidemiological study that tested for a relationship between cancer and glyphosate as well as other agricultural chemicals found “no consistent pattern of positive associations indicating a causal relationship between total cancer (in adults or children) or any site-specific cancer and exposure to glyphosate” (Mink et al., 2012:440; also see section below “Health Effects of Farmer Exposure to Insecticides and Herbicides”).
In 1985, EPA classified glyphosate as Group C (possibly carcinogenic to humans) on the basis of tumor formation in mice. However, in 1991, after reassessment of the mouse data, EPA changed the classification to Group E (evidence of noncarcinogenicity in humans) and in 2013 reaffirmed that “based on the lack of evidence of carcinogenicity in two adequate rodent carcinogenicity studies, glyphosate is not expected to pose a cancer risk to humans” (EPA, 2013:25399).
In 2015, the International Agency for Research on Cancer (IARC) of the World Health Organization (WHO) issued a monograph on glyphosate as part of its volume on some organophosphate insecticides and herbicides (IARC, 2015). In the monograph, IARC classified glyphosate in Group 2A (probably carcinogenic to humans). A summary and reasons for the classification were published in Lancet Oncology (Guyton et al., 2015).
The 2015 IARC Working Group found that, although there is “limited evidence in humans for the carcinogenicity of glyphosate,” there is “sufficient evidence in experimental animals for the carcinogenicity of glyphosate”
(IARC, 2015:78). Furthermore, IARC noted that there is mechanistic support in that glyphosate induces oxidative stress, which could cause DNA damage, and some epidemiological data that support the classification.
EFSA (2015) evaluated glyphosate after the IARC report was released and concluded that glyphosate is unlikely to pose a carcinogenic risk to humans. Canada’s health agency concluded that “the level of human exposure, which determines the actual risk, was not taken into account by WHO (IARC)” (Health Canada, 2015). The Canadian agency found that current food and dermal exposure to glyphosate even by those who work directly with glyphosate is not a health concern as long as it is used as directed on product labels (Health Canada, 2015). EPA (2015) found that glyphosate does not interact with estrogen, androgen, or thyroid systems.
A comment to the committee expressed concern that glyphosate breaks down to formaldehyde, which was classified as a known human carcinogen by IARC (2006). However, this hypothesis was not supported; Franz et al. (1997) used radiolabeled glyphosate and failed to show formation of formaldehyde in the normal environmental degradation of glyphosate.
FINDING: The incidence of a variety of cancer types in the United States has changed over time, but the changes do not appear to be associated with the switch to consumption of GE foods. Furthermore, patterns of change in cancer incidence in the United States are generally similar to those in the United Kingdom and Europe, where diets contain much lower amounts of food derived from GE crops. The data do not support the assertion that cancer rates have increased because of consumption of products of GE crops.
FINDING: There is significant disagreement among expert committees on the potential harm that could be caused by the use of glyphosate on GE crops and in other applications. In determining the risk from glyphosate and formulations that include glyphosate, analyses must take into account both marginal exposure and potential harm.
It has been hypothesized that kidney disease may have increased because GE proteins reached the kidney. The committee examined epidemiological data to determine whether there was a correlation between the consumption of GE foods and the prevalence of chronic kidney disease (CKD).
The total prevalence of all stages of CKD in the United States increased 2 percent from about 12 percent in 1988–1994 to 14 percent in 1999–2004, but the total prevalence has not increased significantly since then.
Figure 5-9 presents prevalence data on the five progressively more serious, recognized stages of CKD (USRDS, 2014). The greatest percent increase is seen in Stage 3, and based on the study (USRDS, 2014), a large amount of the increase occurred in people with comorbidity of cardiovascular disease. Prevalence of CKD increases substantially with age (Coresh et al., 2003), so the aging of the U.S. population may contribute to the overall increase (U.S. Census Bureau, 2014), as does the increase in diabetes and hypertension (Coresh et al., 2007).
FINDING: The available data on prevalence of chronic kidney disease in the United States show a 2 percent increase from 1988 to 2004, but the increase does not appear to be attributable to consumption of GE foods.
Obesity in humans is a complex condition associated with several genetic and environmental factors—including geography, ethnicity, socioeconomic status, lack of exercise, availability of fresh fruits and vegetables, and less nutritional meals (Thayer et al., 2012)—and an altered functioning microbiome (Turnbaugh et al., 2009).
Studies of various species examined body-weight gain when animals were fed a GE crop, a non-GE isogenic comparator, or a non-GE, nonisogenic control. The authors concluded that there were no biologically relevant differences in body-weight gain regardless of the length of the studies (Rhee et al. 2005; Hammond et al., 2006; Arjó et al., 2012; Buzoianu et al., 2012b; Ricroch et al., 2013a,b; Halle and Flachowsky, 2014; Nicolia et al. 2014).
Human population studies have shown that obesity has become more prevalent in the United States (for example, Fryar et al., 2014). An (2015) provided a graphic of the change in U.S. adults (sorted by education level) from 1984 to 2013 (Figure 5-10). As can be seen in the figure, the percentage of obese U.S. adults increased until about 2009, at which time it appears to level off. Because there is no increase in the slope after commercialization of GE crops, these data do not support the hypothesis that GE crops have increased obesity. These time-series data do not prove that there is no association, but if one is present, it is not strong.
Those statistics on obesity coincide with those on the incidence of type II diabetes in the United States (Abraham et al., 2015) and therefore do not support a relationship between GE crops and type II diabetes.
FINDING: The committee found no published evidence to support the hypothesis that the consumption of GE foods has caused higher U.S. rates of obesity or type II diabetes.
Gastrointestinal Tract Diseases
Although the gastrointestinal tract has evolved to digest dietary proteins in the stomach and small intestine effectively for absorption and use of amino acids, it is normal for some full proteins or their fragments to cross the gut barrier through a paracellular route (between cells) or damaged mucosa and for the immune system, which has a high presence at the interface of the gut wall and the internal circulation, to respond accordingly. It is also not unusual, given the high sensitivity of today’s analytical equipment, for proteins or fragments to be detected in minute amounts in different body fluids. Detection methods are not specific to transgene-produced proteins but can find any dietary protein or fragment that is able to pass from the gastrointestinal tract into the bloodstream and tissues. The presence
of a dietary protein or its fragment in the bloodstream or in tissues is not unusual or a cause for health concerns.
About 60–70 percent of the body’s immune system is in the gastrointestinal tract’s gut-associated lymphoid tissue, which has an interface with the gut luminal contents, including toxins, allergens, and the associated microbiota. For GE crops, a public concern has been that the immune system is compromised through ingested transgenic proteins. That possibility has been investigated in animal studies that examined immune system bio-markers and epithelial cell integrity (see section “Beyond Rodent Studies” above and Walsh et al., 2011).
It was suggested to the committee in presentations and public comments that fragments of transgenes may have some special properties that would result in human diseases if they were absorbed into the body through
the digestive tract. The mechanism by which such genes or proteins would affect the body is not clear, although Smith (2013) hypothesized that consuming GE foods increased gut permeability.
FINDING: The committee could find no published evidence supporting the hypothesis that GE foods generate unique gene or protein fragments that would affect the body.
Celiac disease is an autoimmune disorder that affects about 1 percent of the population of western countries. It is triggered in susceptible people by consumption of gluten-containing cereal grains (Fasano et al., 2003; Catassi et al., 2010). Symptoms of celiac disease are the result of an immune reaction that causes marked gastrointestinal inflammation in persons susceptible to gliadin, a component of gluten protein found in wheat, rye (Secale cereale), and barley (Hordeum vulgare) (Green and Cellier, 2007). In addition to exposure to gluten, the etiology of celiac disease is multifactorial and includes genetic predisposition, microbial infection of the gastrointestinal tract, antibiotic exposure, and gastrointestinal erosion (Riddle et al., 2012). Diagnosis is based on detection of serum concentrations (serotypes) of IgA tissue transglutaminase and endomysial antibody IgA, the relief of symptoms upon gluten avoidance, and tissue biopsy. The genetic changes related to the serotyped IgAs are found in about 30 percent of the Caucasian population, but susceptibility to celiac disease is found in only 1 percent of this population (Riddle et al., 2012).
The committee was able to find data on the incidence of celiac disease in the United Kingdom (West et al., 2014; Figure 5-11) and a detailed study conducted by the Mayo Clinic in one county in Minnesota (Murray et al., 2003; Ludvigsson et al., 2013). In the Minnesota and UK studies, there is a clear pattern of increase in celiac-disease incidence (or at least its detection or the extent of self-reports) that started before 1996 (Catassi et al., 2010), when U.S. citizens began to consume more GE foods and the use of glyphosate increased in the United States but not in the United Kingdom. The increases are similar in magnitude to that found in U.S. military personnel, in whom prevalence increased from 1.3 per 100,000 in 1999 to 6.5 per 100,000 in 2008 (Riddle et al., 2012). The authors cautioned that most cases of celiac disease are undiagnosed. Some of the observed increase may be related to improvements in diagnostic criteria, greater awareness of the disease in physicians and patients, better blood tests, and increases in the number of biopsies. However, recent observations point to an increase in incidence beyond those factors (J. A. Murray, Mayo Clinic, personal communication, February 1, 2016).
On the basis of data collected in the 2009–2010 National Health and Nutrition Examination Survey, Rubio-Tapia et al. (2012) reported a prevalence of celiac disease of 0.71 percent with 1.01 percent in non-Hispanic whites in a sample of 7,798 subjects. It should be noted that there has not been any commercial production of GE wheat, rye, or barley in the world. The committee found no evidence that the introduction of GE foods affected the incidence or prevalence of celiac disease worldwide.
FINDING: Celiac-disease detection began increasing in the United States before the introduction of GE crops and the increased use of glyphosate. It appears to have increased similarly in the United Kingdom, where GE foods are not typically consumed and glyphosate use did not increase. The data are not robust, but they do not show a major difference in the rate of increase in incidence of celiac disease between the two countries.
Speakers and some members of the public suggested that the prevalence of food allergies has increased because of GE crops. The committee examined records on the prevalence of food allergies in the United States over time. As is clear from Figure 5-12 and Jackson et al. (2013), the prevalence of food allergies in the United States is rising. For a rough comparator, the committee examined data on hospital admissions for food allergies in the United Kingdom over time (Figure 5-13). UK citizens eat far less food derived from GE crops. The data (Gupta et al., 2007) suggest that food allergies are increasing in the United Kingdom at about the same rate as in the United States (but the types of measurement are different).
FINDING: The committee did not find a relationship between consumption of GE foods and the increase in prevalence of food allergies.
Autism Spectrum Disorder
Autism is often described by such symptoms as difficulty in communicating, forming personal relationships, and using language and abstract concepts. According to the American Psychiatric Association (2013), autism spectrum disorder (ASD) encompasses the previous diagnoses of autism, Asperger syndrome, pervasive developmental disorder not otherwise specified, and childhood disintegrative disorder. Accurate diagnosis of ASD can be difficult, but efforts to identify children with ASD have increased in the United States over the last three decades (CDC, 2014).
In the 2010 Centers for Disease Control and Prevention (CDC) survey of ASD in 11 regions of the United States (CDC, 2014), the overall prevalence in children 8 years old was about 1 in 68 (1.47 percent), but there was wide variation among regions and sociocultural groupings of children. The CDC report stated that “the extent to which this variation might be
attributable to diagnostic practices, under-recognition of ASD symptoms in some racial/ethnic groups, socioeconomic disparities in access to services, and regional differences in clinical or school-based practices that might influence the findings in this report is unclear” (CDC, 2014:1). The degree to which the increase in ASD prevalence since 1990 is due to improved diagnosis is also unclear.
Before 1990, few children in the United States or the United Kingdom had diagnoses of ASD (Taylor et al., 2013), but the prevalence has increased dramatically in both countries. Researchers in the United States and United Kingdom wrote a report that examined prevalence of ASD in the United Kingdom over time and compared it with that in the United States (Taylor et al., 2013). They concluded that “a continuous simultaneous extraordinary rise in the number of children diagnosed as autistic began in both countries in the early 1990s and lasted for a decade. The distribution of first time diagnosis according to age and gender was the same. These similarities between countries as well as within different locations in each country point to a common etiology for this extraordinary medical case” (Taylor et al., 2013:5). There is a higher prevalence in the United States, but it is difficult to evaluate whether it is because of differences in efforts in and approaches to diagnosis and in sociocultural factors that seem to influence prevalence. The overall similarities in prevalence of ASD in the United Kingdom, where GE foods are rarely eaten, and in the United States, where GE foods are commonly eaten, suggest that the major rise in ASD is not associated with consumption of GE foods.
FINDING: The similarity in patterns of increase in autism spectrum disorder in children in the United States, where GE foods are commonly eaten, and the United Kingdom, where GE foods are rarely eaten, does not support the hypothesis of a link between eating GE foods and prevalence of autism spectrum disorder.
The committee heard from some members of the public and some invited speakers that ailments of gastrointestinal origin could be caused by GE crops or their associated technologies or by foods derived from GE crops. The committee investigated the evidence available for that hypothesis.
Gastrointestinal Tract Microbiota
The committee received comments from the public that foods derived from GE crops could change the gut microbiota in an adverse way. Three
scenarios can be considered as related to the potential effects of GE crops on the gut microbiota: the effect of the transgene product (for example, Bt toxin), unintended alteration of profiles of GE plant secondary metabolites, and herbicide (and adjuvant) residue (for example, glyphosate) and its metabolites in HR crops.
Research on the human gut microbiota (the community of microorganisms that live in the digestive tract) is rapidly evolving with recent reports (Dethlefsen and Relman, 2011; David et al., 2014) that suggest that microbiota perturbations occur fairly quickly owing to dietary components or antibiotic treatment. Microbiota composition and state are now well recognized to be linked to noncommunicable chronic diseases and other health problems, so factors that cause either beneficial or adverse changes in the microbiota are of interest to researchers and clinicians. However, the science has not reached the point of understanding how specific changes in microbiota composition affect health and what represents a “healthy” microbiota. The effect of different dietary patterns (for example, high-fat versus high-carbohydrate diets) on the gut microbiota has been linked to metabolic syndrome (Ley, 2010; Zhang et al., 2015).
As discussed above, most proteins, including those in GE and conventionally bred crops, are at least partially digested in the stomach by the action of pepsin that is maintained by the acidic pH of the stomach in most people. Further digestion and absorption are a function of the small intestine, where amino acids and dipeptides and tripeptides are absorbed. Therefore, an effect of a dietary protein on the microbiota, whether from GE or non-GE foods, is unlikely. However, there is some evidence that Bt proteins can be toxic to microorganisms (Yudina et al., 2007), and some nondegraded Bt protein is found within the lumen of the gut but not in the general circulation of pigs (Walsh et al., 2011). Buzoianu et al. (2012c, 2013a) studied the effect of Bt maize feeding on microbiota composition in pigs. In their 2012 study, 110-day feeding of Bt maize (variety MON810) and of isogenic non-GE maize diets led to no differences in cultured Enterobacteriaceae, Lactobacillus, and total anaerobes from the gut; 16S rRNA sequencing showed no differences in bacterial taxa, except the genus Holdemania with which no health effects are associated (Buzoianu et al., 2012c). In the follow-up study in which intestinal content of sows and their offspring were examined with 16S rRNA gene sequencing, the only observed difference for major bacterial phyla was that Proteobacteria were less abundant in sows fed Bt maize before farrowing and in offspring at weaning compared with the controls (Buzoainu et al., 2013a). Fecal Firmicutes were more abundant in offspring fed GE maize. There were other inconsistent differences in mostly low-abundance microorganisms. On the basis of the overall results from their studies, the authors concluded that none of the changes seen in the animals was expected to have biologically relevant health effects on the animals.
Relatively few studies have examined the influence of plant secondary metabolites from any crop on the gut microbiota. The review of Valdés et al. (2015) highlighted investigations on polyphenol-rich foods—such as red wine, tea, cocoa, and blueberries—on the microbiota. Effects were considered minor. As discussed above (see the section “Endogenous Toxins in Plants”), current commercialized GE crops do not have distinctly different secondary metabolite profiles that would lead one to think that they would affect the gut microbiota.
No studies have shown that there are perturbations of the gut microbiota of animals fed foods derived from GE crops that are of concern. However, the committee concluded that this topic has not been adequately explored. It will be important to conduct research that leads to an understanding of whether GE foods or GE foods coupled with other chemicals have biologically relevant effects on the gut microbiota.
FINDING: On the basis of available evidence, the committee determined that the small perturbations found in the gut microbiota of animals fed foods derived from GE crops are not expected to cause health problems. A better understanding of this subject is likely as the methods for identifying and quantifying gut microorganisms mature.
Horizontal Gene Transfer to Gut Microorganisms or Animal Somatic Cells
Horizontal (or lateral) gene transfer is “the stable transfer of genetic material from one organism to another without reproduction or human intervention” (Keese, 2008:123). Since GE crops were commercialized, concern has been voiced by some scientists and some members of the public that foreign DNA introduced into plants through genetic-engineering technologies might, after ingestion, be transferred to the human gut microbiota and directly or indirectly (that is, from bacteria) into human somatic cells. Although most of the concern regarding horizontal gene transfer has been focused on antibiotic-resistance genes used as markers of the transgenic event, other transgenes, such as those with Bt toxins, have also been of concern.
A prerequisite for horizontal gene transfer is that the recombinant DNA must survive the adverse conditions of both food processing and passage through the gastrointestinal tract. Netherwood et al. (2004) showed in patients with a surgically implanted exiting tube placed at the end of the small intestine (an ileostomy) that a small amount of the GE soybean transgene EPSPS passed through the upper gastrointestinal tract to the point of the distal ileum; in subjects without an ileostomy, no transgene was recovered from their feces. In their review on stability and degradation of
DNA from foods in the gastrointestinal tract, Rizzi et al. (2012) noted that recombinant plant DNA fragments were detected in the gastrointestinal tracts of nonruminant animals but not detected in blood or other tissues, although some nonrecombinant plant DNA could be found. The authors concluded that some natural plant DNA fragments persist in the lumen of the gastrointestinal tract and in the bloodstream of animals and humans.
For an event to be considered horizontal gene transfer, DNA must be in the form of a functional (rather than fragmented) gene, enter into bacterial or somatic cells, and be incorporated into the genome with an appropriate promoter, and it must not adversely affect the competitiveness of the cells; otherwise, the effect would be short-lived.
Plant DNA has not been demonstrated to be incorporated into animal cells; however, it has been shown to be transferred in prokaryotes (bacteria). Indeed, molecular geneticists had to find genetic-engineering approaches for getting DNA to be taken into eukaryote cells and incorporated into a genome. The report A Decade of EU-Funded GMO Research (2001–2010) (EC, 2010a) described a study that shows that rumen ciliates (a type of microorganism) exposed to Bt176 maize for 2 or 3 years did not incorporate the Bt176 transgene. There are no reproducible examples of horizontal gene transfer of recombinant plant DNA into the human gastrointestinal microbiota or into human somatic cells. Three independent reviews of the literature on the topic (van den Eede et al., 2004; Keese, 2008; Brigulla and Wackernagel, 2010) concluded that new gene acquisition by the gut bacteria through horizontal gene transfer would be rare and does not pose a health risk.
FINDING: On the basis of its understanding of the process required for horizontal gene transfer from plants to animals and data on GE organisms, the committee concludes that horizontal gene transfer from GE crops or conventionally bred crops to humans does not pose a substantial health risk.
Transfer of Transgenic Material Across the Gut Barrier into Animal Organs
Conflicting reports exist regarding the question of intact transgenes and transgenic proteins from foods crossing the gut barrier. Spisák et al. (2013) published results that indicate that complete genes in foods can pass into human blood. That is plausible, but Lusk (2014) examined the approach used by Spisák et al. and found it more likely that the findings were due to contaminants. Lusk emphasized the need for negative controls in such studies. Placental transfer of foreign DNA into mice was found by Schubbert et al. (1998) by detection in the mouse fetus, but a later report
from the same laboratory (Hohlweg and Doerfler, 2001) did not find the transfer in an eight-generation study.
Studies with dairy cows and goats did not find transgenes or GE proteins in milk, although chloroplast DNA fragments were detected in milk (Phipps et al., 2003; Nemeth et al., 2004; Calsamiglia, et al., 2007; Rizzi et al., 2008; Guertler et al., 2009, Einspanier, 2013; Furgał-Dierżuk et al., 2015). That makes it clear that there is no apparent potential for trangenes or transgenic proteins to be present in dairy products. However, these animals are ruminants, and their digestive systems are different from that of humans.
Walsh et al. (2012a) studied the fate of a Bt gene and protein in pigs that have digestive systems that are more similar to that of humans. They found no evidence of the gene or protein in any organs or blood after 110 days of feeding on Bt maize, but they did find them in the digestive contents of the stomach, cecum, and colon. Fragments of Cry1Ab transgene (as well as other common maize gene fragments) but not the intact Bt gene were found in blood, liver, spleen, and kidney of pigs raised on Bt maize (Mazza et al., 2005).
FINDING: Experiments have found that Cry1Ab fragments but not intact Bt genes can pass into organs and that these fragments present concerns no different than other genes that are in commonly consumed non-GE foods and that pass into organs as fragments.
FINDING: There is no evidence that Bt transgenes or proteins have been found in the milk of ruminants. Therefore, the committee finds that there should be no exposure to Bt transgenes or proteins from consuming dairy products.
OVERALL FINDING ON PURPORTED ADVERSE EFFECTS ON HUMAN HEALTH OF FOODS DERIVED FROM GE CROPS: On the basis of detailed examination of comparisons of currently commercialized GE and non-GE foods in compositional analysis, acute and chronic animal-toxicity tests, long-term data on health of livestock fed GE foods, and human epidemiological data, the committee found no differences that implicate a higher risk to human health from GE foods than from their non-GE counterparts.
There are now a number of examples of crops, either commercialized or in the pipeline toward commercialization, that have GE traits that could improve human health. Improvement of human health can be the sole moti-
vation for development of a specific crop trait, or it can be the secondary effect of a crop trait that is developed primarily for another reason. For example, the genetic engineering of rice to have higher beta-carotene has the specific goal of reducing vitamin A deficiency. GE maize that produces Bt toxins is engineered to decrease insect-pest damage, but a secondary effect could be a decrease in contamination of maize kernels by fungi that produce mycotoxins, such as fumonisins, that at high concentrations could impair human health. Beyond the direct effects of the crops on improvement of human health, there is also a potential indirect benefit associated with a decline in the exposure of insecticide applicators and their families to some insecticides because some GE plants decrease the need for insecticidal control.
Foods with Additional Nutrients or Other Healthful Qualities
Improved Micronutrient Content
According to WHO, some 250 million preschool children are vitamin A–deficient. Each year, 250,000–500,000 vitamin A–deficient children become blind, and half of them die within 12 months of losing their sight.6 Unlike children in wealthier societies, those children have diets that are restricted mostly to poor sources of nutrients, such as rice (Hefferon, 2015). Overall improvement of the diets of the children and their parents is a goal that has not been reached; measures that improve the nutritional quality of their food sources are desirable although not optimal, as a diverse, healthy diet would be.
Crop breeders have used conventional breeding to improve the concentrations of beta-carotene in maize (Gannon et al., 2014; Lividini and Fiedler, 2015), cassava, banana and plantain (Musa spp.) (Saltzman et al., 2013), and sweet potato (Ipomoea batatas) (Hotz et al., 2012a,b). There is some loss of beta-carotene during storage and cooking, but bioavailability is still good (Sanahuja et al., 2013; De Moura et al., 2015). The most rigorous assessments of the effects of those high–beta-carotene varieties were conducted with orange-fleshed sweet potato (high in beta-carotene) in farming areas of Mozambique and Uganda. In both countries, there was increased beta-carotene intake. In Uganda, there was a positive relationship between consumption of high–beta-carotene sweet potato and positive vitamin A status (Hotz et al., 2012a). A more recent study in Mozambique found a decrease in diarrhea prevalence associated with consumption of the high–beta-carotene sweet potato (Jones and DeBrauw, 2015).
No reported experiments have tested any crop with high–beta-carotene for unintended effects. There has been concern about the potential for too high a concentration of beta-carotene in crops because of the hypervitaminosis A syndrome that can be caused by direct intake of too much vitamin A, but that is not a problem when the source is beta-carotene (Gannon et al., 2014).
Golden Rice, which was produced through genetic engineering to increase beta-carotene content, is one of the most recognized examples of the use of genetic-engineering technology to improve a crop’s nutritional value. It is based on the understanding that rice possesses the entire machinery to synthesize beta-carotene in leaves but not in the grain. The breakthrough in the development of Golden Rice was the finding that only two genes are required to synthesize beta-carotene in the endosperm of the rice grain (Ye et al., 2000). The first version of Golden Rice had a beta-carotene content of 6 µg/g. To raise the content to a point where it could alleviate vitamin A deficiency without consumption of very large amounts of rice, a second version of Golden Rice was produced by transforming the plant with the psy gene from maize. The carotene content was thereby raised above 30 µg/g (Paine et al., 2005). Varieties that yield well, have good taste and cooking qualities, and cause no adverse health effects from unintended changes in the rice could have highly important health effects (Demont and Stein, 2013; Birol et al., 2015). There have been claims that Golden Rice was ready for public release for well over a decade (Hefferon, 2015), but this is not the case.
There is a publication on a field test of the first version of Golden Rice (Datta et al., 2007), but the committee could not find information on the newer, higher–beta-carotene Golden Rice in the peer-reviewed literature. Therefore, it contacted the International Rice Research Institute (IRRI) Golden Rice project coordinator, Violeta Villegas, for an update on the status of the project. In discussions with Dr. Villegas (IRRI, personal communication, 2015), it was clear that the project is progressing with a new lead transgenic event, GR2-E, because of difficulties with the previous lead event, GR2-R. The GR2-E event has been backcrossed into varieties that have been requested by several countries including the Philippines, Bangladesh, and Indonesia. As of March 2016, Golden Rice GR2-E in PSBRc82 and BRRI dhan20 genetic backgrounds was being grown in confined field tests in the Philippines and Bangladesh, respectively. Both Golden Rice varieties underwent preliminary assessment inside the greenhouse prior to planting in confined field tests. If performance is good, the varieties will be moved to open field-testing on multiple locations. Once a food regulatory approval is received in one of the participating countries, IRRI will supply the rice with the GR2-E event to an independent third party to assess its efficacy at alleviating vitamin A deficiency.
Past issues with persons and organizations opposed to Golden Rice for a myriad of reasons may have affected IRRI’s work on the rice, but the overall project status7 points out that development of Golden Rice varieties that meet the needs of farmers and consumers and that are in full compliance with the regulatory systems of the partnering countries remains the primary objective. IRRI’s summary statement on its Golden Rice project was that “Golden Rice will only be made available broadly to farmers and consumers if it is successfully developed into rice varieties suitable for Asia, approved by national regulators, and shown to improve vitamin A status in community conditions. If Golden Rice is found to be safe and efficacious, a sustainable delivery program will ensure that Golden Rice is acceptable and accessible to those most in need.”8
Increasing concentrations of beta-carotene is only one goal of conventional crop breeding and genetic engineering. Projects for increasing iron and zinc in crops as different as wheat, pearl millet (Pennisetum glaucum), and lentil (Lens culinaris) are at varied stages of development (Saltzman et al., 2013).
FINDING: Experimental results with non-GE crop varieties that have increased concentrations of micronutrients demonstrate that both GE and non-GE crops with these traits could have favorable effects on the health of millions of people, and projects aimed at providing these crops are at various stages of completion and testing.
Altering Oil Composition
Substantial efforts have been made to increase the oxidative stability of soybean oil, a major cooking oil all over the world, as a means of avoiding trans-fats generated through the hydrogenation process and enhancing omega-3 fatty acid content of the oil for use in both food and feed applications. Soybean oil is composed principally of five fatty acids: palmitic acid (16:0, carbon number:double bond number), stearic acid (18:0), oleic acid (18:1), linoleic acid (18:2), and linolenic acid (18:3) in approximate percentages of 10, 4, 18, 55, and 13. High content of unsaturated fats creates a disadvantage in industrial processing because they are susceptible to oxidation and trans-fat generation during hydrogenation, whereas oils with a high percentage of oleic acid (about 80 percent) require less processing and offer another route to decrease concentrations
7 What is the status of the Golden Rice project coordinated by IRRI? Available at http://irri.org/golden-rice/faqs/what-is-the-status-of-the-golden-rice-project-coordinated-by-irri. Accessed October 30, 2015.
of trans-fats in food products. High-oleic acid-containing soybean was produced by downregulating expression of the fatty acid desaturating enzymes FAD2-1A and -1B to decrease the concentration of trans-fats in soybean (EFSA, 2013). In 2015, high-oleic acid soybean was commercially available in North America and was produced on a small area in the United States for specialty-product contracts (C. Hazel, DuPont Pioneer, personal communication, December 14, 2015).
Canola (Brassica napus), known in Europe as rapeseed, is the major oilseed crop in Canada. Canola was developed through conventional breeding at the University of Manitoba, Canada, by Downey and Stefansson in the early 1970s and had a good nutritional profile—58-percent oleic acid and 36-percent polyunsaturated fatty acids—in addition to low erucic acid and a moderate concentration of saturated fatty acid (6 percent). Because of demand for saturated functional oils for the trans-fat–free market, high-lauric acid GE canola was created in 1995 through an “Agrobacteriummediated transformation in which the transfer-DNA (T-DNA) contained the gene encoding the enzyme 12:0 ACP thioesterase (bay TE) from the California Bay tree (Umbellularia californica). In addition, the T-DNA contained sequences that encoded the enzyme neomycin phosphotransferase II (NPTII). The expression of NPTII activity was used as a selectable trait to screen transformed plants for the presence of the bay TE gene. No other translatable DNA sequences were incorporated into the plant genome” (Health Canada, 1999:1). The presence of lauric acid (12:0) in the oil allows it to be used as a replacement for other types of oils with lauric acid (for example, coconut and palm kernel oil) in such products as “confectionery coatings and fillings, margarines, spreads, shortenings, and commercial frying oils. It has also been used as a substitute for cocoa butter, lard, beef fats, palm oil, and partially or fully hydrogenated soybean, maize, cottonseed, peanut, safflower, and sunflower oils” (Health Canada, 1999:2). However, low yield and comparably poor agronomic traits have removed high-lauric acid canola from the commercial market. The long-term use of crops with altered oil content is uncertain.
FINDING: Crops with altered oil composition might improve human health, but this will depend on the specific alterations, how the crops yield, and how the products of the crops are used.
Genetically Engineered Foods with Lower Concentrations of Toxins
Acrylamide is produced in starchy foods when they are cooked at high temperatures. Processing of potatoes for French fries and potato chips generates acrylamide. Toasting bread also produces acrylamide. That is viewed as a problem because the U.S. National Toxicology Program (2014)
concluded that acrylamide “is reasonably anticipated to be a human carcinogen based on sufficient evidence of carcinogenicity from studies in experimental animals” and causes neurological damage at high exposure. Acrylamide is produced from a chemical reaction between asparagine and a reducing sugar, so decreasing the concentration of either is expected to decrease acrylamide. A potato line was genetically engineered to have low amounts of free asparagine and in early tests had as little as 5 percent of the acrylamide compared with non-GE potatoes when cooked at high temperatures (Rommens et al., 2008).
In 2014, USDA deregulated a low-acrylamide potato produced by Simplot Plant Sciences (USDA–APHIS, 2014c) on the basis of nonplant pest status. The company also provided information to FDA. No problems were found by FDA with respect to the company’s assessment of composition or safety (FDA, 2015). It should be noted that for many people reduced acrylamide in potatoes is expected to lower overall acrylamide intake substantially, but many foods contain acrylamide (FDA, 2000b, revised 2006). An FDA survey of commonly consumed foods showed French fries at seven McDonald’s locations had an average acrylamide concentration of 288 parts per billion (ppb), whereas Gerber Finger Foods Biter Biscuits had 130 ppb and Wheatena Toasted Wheat Cereal had 1,057 ppb, which is much more than from fast-food French fries (FDA, 2002, revised 2006).9 Any toasted bread is expected to be high in acrylamide. Therefore, how much low-acrylamide potato decreases total exposure depends on individual diets. Furthermore, EPA has established limits for exposure to acrylamide, and current actual exposures are generally below the limits.
Although the low-acrylamide potato is the only GE crop with a lower food-toxin concentration that has been deregulated in the United States, other GE crops with lower natural toxin concentrations are in the pipeline. Potatoes and other crops in the “deadly nightshade” family (Solanaceae, which includes tomato and eggplant) produce glycoalkaloids, some of which have human toxicity, as described above (see the section “Endogenous Toxins in Plants” in this chapter). Langkilde et al. (2012) conducted a compositional and toxicological analysis of the potatoes with lower solanine and higher chaconine. The study used Syrian golden hamsters instead of rats because the hamsters are very sensitive to the glycoalkaloids. There were some statistically significant differences, but they were considered not of biological relevance. At this point, the evidence is not sufficient to conclude that a low-glycoalkaloid potato would be healthier for humans.
Highly toxic chemicals (aflatoxins and fumonisins) are produced by Fusarium and Aspergillis fungi on the kernels of maize (Bowers et al.,
9 Acrylamide concentrations reported by FDA were for individual purchased food products and were not adjusted for unit-to-unit variation.
2014). Aflatoxins are considered by the U.S. National Toxicology Program (2014) to be “human carcinogens based on sufficient evidence of carcinogenicity from studies in humans.” They are also associated with many other illnesses and considered a global health problem (Wild and Gong, 2010). Fumonisins cause a number of physiological disorders and are considered possibly carcinogenic to humans (IARC, 2002). Several investigators have reported a substantial decrease in fumonisins in Bt maize compared with conventionally bred varieties (Munkvold and Desjardins, 1997; Bowers et al., 2014). However, there is no clear association between Bt maize and aflatoxin concentrations (Wiatrak et al., 2005; Abbas et al., 2007; Bowen et al., 2014).
Research continues on how to use genetic engineering to develop varieties of maize and peanut (Arachis hypogaea) that inhibit aflatoxin production, but a GE solution has so far been elusive (Bhatnagar-Mathur et al., 2015). A reduction in aflatoxin in both maize and peanut would have substantial health benefits in some developing countries (Williams et al., 2004; Wild and Gong, 2010).
FINDING: It is possible that GE crops that would result in improved health by lowering exposure of humans to plant-produced toxins in foods could be developed, but there is insufficient information to assess the possibility. However, GE plants that indirectly or directly reduce fungal-toxin production and intake would offer substantial benefits to some of the world’s poorest populations, which have the highest dietary intake of food-associated fungal toxins.
Health Effects of Farmer Exposure to Insecticides and Herbicides
Chapter 4 presents data that demonstrate substantially lower use of insecticides in some Bt crops than in conventionally bred crops. There is a logical expectation that a decrease in the number of insecticide applications would lead to lower farm-worker exposure and therefore lower health burden, especially in countries where acute poisonings due to applicator exposure are common. Racovita et al. (2015) reviewed five studies of Bt cotton in China, India, Pakistan, and South Africa that ranged from one to four growing seasons. All reported a decline in the number of insecticide applications to Bt versus non-Bt cotton. In a study in China by Huang et al. (2002), Bt cotton was treated with insecticides 6.6 times and non-Bt cotton was treated 19.8 times during the growing season. The frequency of Bt and non-Bt cotton farmers reporting poisonings were 5 percent and 22 percent, respectively in 1999, 7 percent and 29 percent in 2000, 8 percent and 12 percent in 2001. Kouser and Qaim (2011) found fewer overall insecticide treatments in a study conducted in India: 1.5 treatments
of Bt cotton and 2.2 treatments of non-Bt cotton. In this study, the farmers who used Bt cotton reported 0.19 poisonings per season while those with conventionally bred cotton reported 1.6 poisonings. Bennett et al. (2006) studied the same types of farmers in South Africa. Bt cotton was not yet widely available in the beginning of the experiment, but eventually some farmers adopted Bt cotton and decreased spraying. The study looked at overall poisonings according to hospital records over time; there were 20 poisonings in the year before common availability of Bt cotton and four in a later year, when there was 60 percent adoption of Bt cotton.
The findings of those and other studies (for example, Huang et al., 2005; Dev and Rao, 2007; Kouser and Qaim, 2013) are in line with an expectation of a decrease in poisonings when Bt cotton is grown instead of non-Bt cotton. However, Racovita et al. (2015:15), who carefully assessed each of the studies, found many shortcomings that led them to conclude that “the link between [genetically modified] crop cultivation and a reduction in number of pesticide poisonings should be considered as still circumstantial.” The shortcomings include the fact that the number of poisonings is based on farmer recall of incidents sometimes more than a year after the field season or, in the Bennett et al. (2006) study, simply based on hospital cases. Another issue was that there may have been differences in risk–avoidance behavior between farmers who did and did not plant Bt cotton. Finally, the studies focused on farmers, not farm workers, who do not control farm operations. Racovita et al. (2015) called for more rigorous studies that would address the shortcomings of previous studies, given the politicized nature of the use of Bt crops.
Farm-worker exposure to insecticides and herbicides is lower in the United States and some other developed countries than is the case for farm workers on resource-poor farms. However, there is substantial exposure, and any effects seen in the United States would be of global concern. Prospective cohort studies of health are the high benchmark of epidemiology studies, and the Agricultural Health Study (AHS) funded by the U.S. National Institute of Environmental Health Sciences used this approach to evaluate private and commercial applicators in Iowa and North Carolina. The landmark study resulted in two peer-reviewed articles on glyphosate exposure and cancer incidence (De Roos et al., 2005; Mink et al., 2012) and one on glyphosate exposure and non-cancer health outcomes (Mink et al., 2011). De Roos et al. (2005:49) concluded that “glyphosate exposure was not associated with cancer incidence overall or with most cancer subtypes we studied.” The data suggested a weak association with multiple myeloma on the basis of a small number of cases, but that association was not found in a follow-up study (DeRoos et al., 2005; Mink et al., 2012). Mink et al. (2012:440) reported on the continuation of the AHS cohort study and found “no consistent pattern of positive associations indicating a causal relationship between total
cancer (in adults or children) or any site-specific cancer and exposure to glyphosate.” Mink et al. (2011) reviewed noncancer health outcomes that included respiratory conditions, diabetes, myocardial infarction, reproductive and developmental outcomes, rheumatoid arthritis, thyroid disease, and Parkinson’s disease. They reviewed cohort, case–control, and cross-sectional studies within the AHS study and found “no evidence of a consistent pattern of positive associations indicating a causal relationship between any disease and exposure to glyphosate” (Mink et al., 2011:172).
FINDING: There is evidence that use of Bt cotton in developing countries is associated with reduced insecticide poisonings. However, there is a need for more rigorous survey data addressing the shortcomings of existing studies.
FINDING: A major government-sponsored prospective study of farm-worker health in the United States does not show any significant increases in cancer or other health problems that are due to use of glyphosate.
Increased Precision and Complexity of Genetic-Engineering Alterations
At the time that the committee wrote its report, major commercialized GE crops had been engineered by using Agrobacterium tumefaciensmediated or gene gun-mediated transformation, both of which result in semirandom insertion of the transgene into the genome. Variation in expression of the transgene was routinely observed because of the specific genomic characteristics of the insertion sites. Because of that variation, there was a need to screen large numbers of transgenic plants to identify the optimal transgenic individual. Regulations in the United States require approval of each transformation event regardless of whether the transgene itself was previously approved for release in that crop. That is at least in part because of the potential for unintended effects of each insertion.
Precision genome-editing technologies now permit insertion of single or multiple genes into one targeted location in the genome and thereby eliminate variation that is due to position effects (see Chapter 7). Such precision is expected to decrease unintended effects of gene insertion, although it will not eliminate the effects of somaclonal variation (discussed in Chapter 7).
Consider, for example, the engineering of completely new metabolic pathways into a plant for nutritional enhancement. The simplest example
would be a set of two genes, such as has been used to create Golden Rice to deliver precursors of vitamin A. A more complex example would be engineering of fish oils (very long-chain unsaturated fatty acids) to improve the health profile of plant oils; depending on the target species, this process has required introduction of at least of three and at most nine transgenes (Abbadi et al., 2004; Wu et al., 2005; Ruiz-Lopez et al., 2014). If each of those transgenes is integrated into the genome on a different chromosome on the basis of separate insertion events, it will require a number of generations of crosses to put them all together in one plant. If, instead, all the transgenes could be targeted at the same site on a chromosome either simultaneously or one after another, they would not segregate from each other as they were moved into elite varieties. From a food-safety perspective, engineering transgenes into a single target locus also ensures that expression of the whole pathway is preserved so that the correct end product accumulates. Emerging genetic-engineering technologies currently enable insertion of a few genes in one construct, but in the future that number may increase dramatically.
In the future, the scale of genetic-engineering alterations may go much further than just manipulating oil profiles. The committee heard from speakers about projects aimed at changing the entire photosynthetic pathway of the rice plant (Weber, 2014) to create an entirely novel crop (Zhu et al., 2010; Ruan et al., 2012). The committee also heard from researchers interested in developing cereal crops with nitrogen fixation. Those projects are discussed further in Chapter 8. Although the precision of future genetic-engineering alterations should decrease unintended effects of the process of engineering, the complexity of the changes in a plant may leave it not substantially equivalent to its non-GE counterpart.
It is also important to note that crops that use RNA interference (RNAi) were coming on the market when the committee was writing its report. EPA convened a science advisory panel to evaluate hazards that might arise from use of this genetic-engineering approach. The panel concluded that “dietary RNA is extensively degraded in the mammalian digestive system by a combination of ribonucleases (RNases) and acids that are likely to ensure that all structural forms of RNA are degraded throughout the digestive process. There is no convincing evidence that ingested [double-stranded] RNA is absorbed from the mammalian gut in a form that causes physiologically relevant adverse effects” (EPA, 2014c:14). When the committee was writing its report, deployment of dietary RNAi was a new technology. EPA’s panel made a number of recommendations, including investigating factors that may affect absorption and effects of dietary double-stranded RNAs and investigating the stability of double-stranded RNA in people who manifest diseases.
FINDING: The precision of emerging genetic-engineering technologies should decrease some sources of unintended changes in the plants, thus simplifying food-safety testing. However, engineering involving major changes in metabolic pathways or insertion of multiple resistance genes will complicate the determination of food safety because changes in metabolic pathways are known to have unexpected effects on plant metabolites.
Increased Diversity of Crops To Be Engineered
The most far-ranging effects of emerging genetic-engineering technologies may be the diversity of crops that will be engineered and commercialized. Commercial GE crops at the time the committee conducted its review were mainly high-production commodity crops (maize, soybean, and cotton) engineered with trans-kingdom genes, but the applications of emerging genetic-engineering technologies are much broader: these technologies can be easily applied to any plant species that can be regenerated from tissue culture. Furthermore, the emerging technologies described in Chapter 7 can focus on any gene in which an altered nucleotide sequence results in a desired trait.
As a consequence, the committee expects a sizable increase in the number of food-producing crop species that are genetically altered. Examples of new target crops include forages (grasses and legumes), beans, pulses, a wide array of vegetables, herbs, and spices, and plants grown for flavor compounds. New traits will probably include fiber content (either increased to add more fiber or decreased to improve digestibility), altered oil profiles, decreased concentrations of antinutrients, increased or more consistent concentrations of such phytochemicals as antioxidants (for example, flavonoids) and phytoestrogens (for example, isoflavones or lignans), and increased mineral concentrations. Some of these are considered further in Chapter 8.
From a food-safety perspective, the increase in crops and traits presents a number of challenges. First is the need to develop better and more detailed baseline data on the general chemical composition and probably the transcriptomic profiles of currently marketed conventionally bred varieties of the crops (see Chapter 7). Perhaps more problematic will be designing whole-food animal-testing regimens if the food from the crop cannot be used as a major component of the test animals’ diet. Maize, rice, soybean, and other grains can be added to diets at up to 30 percent without adverse effects on animal health. That is unlikely to be the case with new spices or some vegetables. It would be beneficial if new, publicly acceptable approaches for testing could be developed that do not require animal testing (NRC, 2007; Liebsch et al., 2011; Marx-Stoelting et al., 2015). Chapter 9
addresses the potential need to move to an entirely product-based approach to regulation and testing based on the novelty of a new crop or food.
FINDING: Some future GE crops will be designed to be substantially different from current crops and may not be as amenable to animal testing as currently marketed GE crops.
RECOMMENDATION: There is an urgent need for publicly funded research on novel molecular approaches for testing future products of genetic engineering so that accurate testing methods will be available when the new products are ready for commercialization.
The committee’s objective in this chapter was to examine the evidence that supports or negates specific hypotheses and claims about the risks and benefits associated with foods derived from GE crops. As acknowledged at the beginning of the chapter, understanding the health effects of any food, whether non-GE or GE, can be difficult. The properties of most plant secondary metabolites are not understood, and isolating the effects of diet on animals, including humans, is challenging. Although there are well-developed methods for assessing potential allergenicity of novel foods, these methods could miss some allergens. However, the research that has been conducted in studies with animals and on chemical composition of GE foods reveals no differences that would implicate a higher risk to human health from eating GE foods than from eating their non-GE counterparts. Long-term epidemiological studies have not directly addressed GE food consumption, but available time-series epidemiological data do not show any disease or chronic conditions in populations that correlate with consumption of GE foods. The committee could not find persuasive evidence of adverse health effects directly attributable to consumption of GE foods.
New methods to measure food composition that involve transcriptomics, proteomics, and metabolomics provide a broad, nontargeted assessment of thousands of plant RNAs, proteins, and compounds. When the methods have been used, the differences found in comparisons of GE with non-GE plants have been small relative to the naturally occurring variation found in conventionally bred crop varieties. Differences that are detected by using -omics methods do not on their own indicate a safety problem.
There is some evidence that GE insect-resistant crops have had benefits to human health by reducing insecticide poisonings and decreasing exposure to fumonisins. Several crops had been developed or were in development with GE traits designed to benefit human health; however, when the committee was writing its report, commercialized crops with health benefits
had been only recently introduced and were not widely grown, so the committee could not evaluate whether they had had their intended effects.
New crops developed with the use of emerging genetic-engineering technologies were in the process of being commercialized. The precision associated with the technologies should decrease some sources of unintended changes that occur when plants are genetically engineered and thus simplify food-safety testing. However, engineering involving major changes in metabolic pathways or insertion of multiple resistance genes will complicate the determination of food safety because changes in metabolic pathways are known to have unexpected effects on plant metabolites. Therefore, publicly funded research on novel approaches for testing future products of genetic engineering is needed so that accurate testing methods will be available when the new products are ready for commercialization.
Abbadi, A., F. Domergue, J. Bauer, J.A. Napier, R. Welti, U. Zähringer, P. Cirpus, and E. Heinz. 2004. Biosynthesis of very-long-chain polyunsaturated fatty acids in transgenic oilseeds: Constraints on their accumulation. Plant Cell 16:2734–2748.
Abbas, H.K., W.T. Shier, and R.D. Cartwright. 2007. Effect of temperature, rainfall and planting date on aflatoxin and fumonisin contamination in commercial Bt and non-Bt maize hybrids in Arkansas. Phytoprotection 88:41–50.
Abraham, T.M., K.M. Pencina, M.J. Pecina, and C.S. Fox. 2015. Trends in diabetes incidence: The Framingham heart study. Diabetes Care 38:482–487.
ADAS, 2015. Strategy support for the post-market monitoring (PMM) of GM plants: Review of existing PPM strategies developed for the safety assessment of human and animal health. EFSA supporting publication 2014:EN-739.
Ahuja, I., R. Kissen, and A.M. Bones. 2012. Phytoalexins in defense against pathogens. Trends in Plant Science 17:73–90.
American Association for the Advancement of Science. 2012. Statement by the AAAS Board of Directors on Labeling of Genetically Modified Foods. October 20. Available at http://www.aaas.org/sites/default/files/AAAS_GM_statement.pdf. Accessed October 13, 2015.
American Psychiatric Association. 2013. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Arlington, VA: American Psychiatric Publishing.
Amos, J. September 19, 2012. French GM-fed Rat Study Triggers Furore. Online. BBC News. Available at http://www.bbc.com/news/science-environment-19654825. Accessed December 13, 2015.
An, R. 2015. Educational disparity in obesity among U.S. adults, 1984–2013. Annals of Epidemiology 25:637–642.
Arjó, G., T. Capell, X. Matias-Guiu, C. Zhu, P. Christou, and C. Piñol. 2012. Mice fed on a diet enriched with genetically engineered multivitamin corn show no sub-acute toxic effects and no sub-chronic toxicity. Plant Biotechnology Journal 10:1026–1034.
Astwood, J.D., J.N. Leach, and R.L. Fuchs. 1996. Stability of food allergens to digestion in vitro. Nature Biotechnology 14:1269–1273.
Bartholomaeus, A., W. Parrott, G. Bondy, and K. Walker. 2013. The use of whole food animal studies in the safety assessment of genetically modified crops: Limitations and recommendations. Critical Reviews in Toxicology 43:1–24.
Belknap, J.K., S.R. Mitchell, L.A. O’Toole, M.L. Helms, and J.C. Crabbe. 1996. Type I and type II error rates for quantitative trait loci (QTL) mapping studies using recombinant inbred mouse strains. Behavior Genetics 26:149–160.
Bennett, R., S. Morse, and Y. Ismael. 2006. The economic impact of genetically modified cotton on South African smallholders: Yield, profit and health effects. Journal of Development Studies 42:662–677.
Berry, C. 2013. Letter to the Editor. Food and Chemical Toxicology 53:445–446.
Bhatnagar-Mathur, P., S. Sunkara, M. Bhatnagar-Panwar, F. Waliyar, and K.K. Sharma. 2015. Biotechnological advances for combating Aspergillus flavus and aflatoxin contamination in crops. Plant Science 234:119–132.
Birol, E., J.V. Meenakshi, A. Oparinde, S. Perez, and K. Tomlins. 2015. Developing country consumers’ acceptance of biofortified foods: A synthesis. Food Security 7:555–568.
Bøhn, T., M. Cuhra, T. Traavik, M. Sanden, J. Fagan, and R. Primicerio. 2014. Compositional differences in soybeans on the market: Glyphosate accumulates in Roundup Ready GM soybeans. Food Chemistry 153:207–215.
Boobis, A.R., B.C. Ossendorp, U. Banasiak, P.Y. Hamey, I. Sebestyen, and A. Moretto. 2008. Cumulative risk assessment of pesticide residues in food. Toxicology Letters 180:137–150.
Bowen, K.L., K.L. Flanders, A.K. Hagan, and B. Ortiz. 2014. Insect damage, aflatoxin content, and yield of Bt corn in Alabama. Journal of Economic Entomology 107:1818–1827.
Bowers, E., R. Hellmich, and G. Munkvold. 2014. Comparison of fumonisn contamination using HPLC and ELISA methods in Bt and near-isogenic maize hybrids infested with European corn borer or Western bean cutworm. Journal of Agricultural and Food Chemistry 62:6463–6472.
Brigulla, M., and W. Wackernagel. 2010. Molecular aspects of gene transfer and foreign DNA acquisition in prokaryotes with regard to safety issues. Applied Microbiology and Biotechnology 86:1027–1041.
Brix, A.E., A. Nyska, J.K. Haseman, D.M. Sells, M.P. Jokinen, and N.J. Walker. 2005. Incidences of selected lesions in control female Harlan Sprague–Dawley rates from two-year studies performed by the National Toxciology Program. Toxicologic Pathology 33:477–483.
Brouk, M.J., B. Cvetkovic, D.W. Rice, B.L. Smith, M.A. Hinds, F.N. Owens, C. Iiams, and T.E. Sauber. 2011. Performance of lactating dairy cows fed corn as whole plant silage and grain produced from genetically modified corn containing event DAS-59122–7 compared to a nontransgenic, near-isogenic control. Journal of Dairy Science 94:1961–1966.
Bush, R.K., S.L. Taylor, J.A. Nordlee, and W.W. Busse. 1985. Soybean oil is not allergenic to soybean-sensitive individuals. Journal of Allergy and Clinical Immunology 76:242–245.
Butler, D. October 11, 2012. Hyped GM Maize Study Faces Growing Scrutiny. Online. Nature. Available at http://www.nature.com/news/hyped-gm-maize-study-faces-growingscrutiny-1.11566. Accessed December 13, 2015.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, O. O’Donovan, E. Gelencsér, G. Ujhelyi, E. Szabó, A. Nagy, R.P. Ross, G.E. Gardiner, and P.G. Lawlor. 2012a. Effects of feeding Bt maize to sows during gestation and lactation on maternal and offspring immunity and fate of transgenic material. PLoS ONE 7:e47851.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, J.P. Cassidy, R.P. Ross, G.E. Gardiner, and P.G. Lawlor. 2012b. Effect of feeding genetically modified Bt MON810 maize to approximately 40-day-old pigs for 110 days on growth and health indicators. Animal 6:1609–1619.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, O. O’Sullivan, F. Crispie, P.D. Cotter, P.R. Ross, G.E. Gardiner, and P.G. Lawlor. 2012c. The effect of feeding Bt MON810 maize to pigs for 110 days on intestinal microbiota. PLoS One 7:e33668.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, O. O’Sullivan, F. Crispie, P.D. Cotter, P.R. Ross, G.E. Gardiner, and P.G. Lawlor. 2012d. High-throughput sequence-based analysis of the intestinal microbiota of weanling pigs fed genetically modified MON810 maize expressing Bacillus thuringiensis Cry1Ab (Bt maize) for 31 days. Applied and Environmental Microbiology 78:4217–4224.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, L. Quigley, O. O’Sullivan, P.D. Cotter, R.P. Ross, G.E. Gardiner, and P.G. Lawlor. 2013a. Sequence-based analysis of the intestinal Microbiota of sows and their offspring fed genetically modified maize expressing a truncated form of Bacillus thuringiensis Cry1Ab protein (Bt Maize). Applied and Environmental Microbiology 79:7735–7744.
Buzoianu, S.G., M.C. Walsh, M.C. Rea, J.P. Cassidy, T.P. Ryan, P.R. Ross, G.E. Gardiner, and P.G. Lawlor. 2013b. Transgenerational effects of feeding genetically modified maize to nulliparious sows and offspring on offspring growth and health. Journal of Animal Science 91:318–330.
CAC (Codex Alimentarius Commission). 2003. Guideline for the Conduct of Food Safety Assessment of Foods Using Recombinant DNA Plants. Doc CAC/GL 45-2003. Rome: World Health Organization and Food and Agriculture Organization.
CAC (Codex Alimentarius Commission). 2008. Annex 2: Food Safety Assessment of Foods Derived from Recombinant-DNA Plants Modified for Nutritional or Health Benefits in Guideline for the Conduct of Food Safety Assessment of Foods Using Recombinant DNA Plants. Doc CAC/GL 45-2003. Rome: World Health Organization and Food and Agriculture Organization.
CAC (Codex Alimentarius Commission). 2009. Foods Derived from Modern Biotechnology. Rome: World Health Organization and Food and Agriculture Organization.
Callahan, P. January 28, 2016. Court clears way for revival of worrisome weedkiller. EPA nixes approval of Enlist Duo weed killer. Online. Chicago Tribune. Available at http://www.chicagotribune.com/news/watchdog/ct-dow-enlist-duo-court-ruling-20160127-. Accessed March 21, 2016.
Calsamiglia, S., B. Hernandez, G.F. Hartnell, and R. Phipps. 2007. Effects of corn silage derived from a genetically modified variety containing two transgenes on feed intake, milk production, and composition, and the absence of detectable transgenic deoxyribonucleic acid in milk in Holstein dairy cows. Journal of Dairy Science 90:4718–4723.
Catassi, C., D. Kryszak, B. Bhatti, C. Strugeon, K. Helzlsouer, S.L. Clipp, D. Gelfond, E. Puppa, A. Sferruzza, and A. Fasano. 2010. Natural history of celiac disease autoimmunity in a USA cohort followed since 1974. Annals of Medicine 42:530–538.
CDC (Centers for Disease Control and Prevention). 2014. Prevalence of autism spectrum disorder among children aged 8 years—autism and developmental disabilities monitoring network, 11 sites, United States, 2010. Morbidity and Mortality Weekly Report 63:1–21.
Coresh, J., B.C. Astor, T. Greene, G. Eknoyan, and A.S. Levey. 2003. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third national health and nutrition examination survey. American Journal of Kidney Diseases 41:1–12.
Coresh, J., E. Selvin, L.A. Stevens, J. Manzi, J.W. Kusek, P. Eggers, F. Van Lente, and A.S. Levey. 2007. Prevalence of Chronic Kidney Disease in the United States. Journal of the American Medical Association 298:2038–2047.
Costa, J., I. Mafra, J.S. Amaral, and M.B.P.P. Oliveira. 2010a. Detection of genetically modified soybean DNA in refined vegetable oils. European Food Research and Technology 230:915–923.
Costa, J., I. Mafra, J.S. Amaral, and M.B.P.P. Oliveira. 2010b. Monitoring genetically modified soybean along the industrial soybean oil extraction and refining processes by polymerase chain reaction techniques. Food Research International 43:301–306.
Council on Science and Public Health of the American Medical Association House of Delegates. 2012. Report 2 (A-12). Labeling of Bioengineered Foods (Resolutions 508 and 509-A-11). Available at http://factsaboutgmos.org/sites/default/files/AMA%20Report.pdf. Accessed March 12, 2016.
Datta, S.K., K. Datta, V. Parkhi, M. Rai, N. Baisakh, G. Sahoo, S. Rehana, A. Bandyopadhyay, M. Alamgir, M.S. Ali, E. Abrigo, N. Oliva, and L. Torrizo. 2007. Golden Rice: Introgression, breeding, and field evaluation. Euphytica 154:271–278.
David, L.A., C.F. Maurice, R.N. Carmody, D.B. Gootenberg, J.E. Button, B.E. Wolfe, A.V. Ling, A.S. Devlin, Y. Varma, M.A. Fischbach, S.B. Biddinger, R.J. Dutton, and P.J. Turnbaugh. 2014. Diet rapidly and reproducibly alters the human gut microbiome. Nature 505:559–563.
Davis, R.K., G.T. Stevenson, and K.A. Busch. 1956. Tumor incidence in normal Sprague-Dawley female rats. Cancer Research 16:194–197.
DeFrancesco, L. 2013. How safe does transgenic food need to be? Nature Biotechnology 31:794–802.
de Luis, R., M. Lavilla, L. Sanchez, M. Calvo, and M.D. Perez. 2009. Immunochemical detection of Cry1A(b) protein in model processed foods made with transgenic maize. European Food Research and Technology 229:15–19.
De Moura, F.F., A. Miloff, and E. Boy. 2015. Retention of provitamin A carotenoids in staple crops targeted for biofortification in Africa: Cassava, maize and sweet potato. Critical Reviews in Food Science and Nutrition 55:1246–1269.
De Roos, A.J., A. Blair, J.A. Rusiecki, J.A. Hoppin, M. Svec, M. Dosemeci, D.P. Sandler, and M.C. Alavanja. 2005. Cancer incidence among glyphosate-exposed pesticide applicators in the Agricultural Health Study. Environmental Health Perspective 113:49–54.
Demont, M., and A.J. Stein. 2013. Global value of GM rice: A review of expected agronomic and consumer benefits. New Biotechnology 30:426–436.
Dethlefsen, L., and D A. Relman. 2011. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proceedings of the National Academy of Sciences of the United States of America 108: 4554–4561.
Dev, M.S., and N.C. Rao. 2007. Socio-economic Impact of Bt Cotton. Monograph No. 3. Hyderabad: Centre for Economic and Social Studies.
Diaz, C., C. Fernandez, R. McDonald, and J.M. Yeung. 2002. Determination of Cry9C protein in processed foods made with StarLink™ corn. Journal of AOAC International 85:1070–1076.
Dinse, G.E., S.D. Peddada, S.F. Harris, and S.A. Elmore. 2010. Comparison of NTP historical control tumor incidence rates in female Harlan Sprague–Dawley and Fischer 344/N rats. Toxicologic Pathology 38:765–775.
Dixon, R.A. 2001. Natural products and disease resistance. Nature 411:843–847.
Dixon, R.A. 2004. Phytoestrogens. Annual Review of Plant Biology 55:225–261.
Domingo, J.L., and J.G. Bordonaba. 2011. A literature review on the safety assessment of genetically modified plants. Environment International 37:734–742.
Dona, A., and I.S. Arvanitoyannis. 2009. Health risks of genetically modified foods. Critical Reviews in Food Science and Nutrition 49:164–175.
Duke, S.O., A.M. Rimando, P.F. Pace, K.N. Reddy, and R.J. Smeda. 2003. Isoflavone, glyphosate, and aminomethylphosphonic acid levels in seeds of glyphosate-treated, glyphosate resistant soybean. Journal of Agricultural and Food Chemistry 51:340–344.
Dung, L.T., and L.H. Ham. 2013. Comments on “Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize.” Food and Chemical Toxicology 53:443–444.
Dunn, O.J. 1961. Multiple comparisons among means. Journal of the American Statistical Association 56:52–64.
EC (European Commission). 2010a. A Decade of EU-funded GMO Research (2001–2010). Brussels: European Commission.
EC (European Commission). 2010b. Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. Official Journal of the European Union 276:33–79.
EC (European Commission). 2013. Commission implementing regulation (EU) No 503/2013 of 3 April 2013 on applications for authorisation of genetically modified food and feed in accordance with Regulation (EC) No 1829/2003 of the European Parliament and of the Council and amending Commission Regulations (EC) No 641/2004 and (EC) No 1981/2006. Official Journal of the European Union 157:1–48.
EFSA (European Food Safety Authority). 2007. Statement of the Scientific Panel on Genetically Modified Organisms on the Analysis of Data from a 90-day Rat Feeding Study with MON 863 Maize. Available at http://www.efsa.europa.eu/sites/default/files/scientific_output/files/main_documents/GMO_statement_MON863%2C0.pdf. Accessed December 13, 2015.
EFSA (European Food Safety Authority). 2010. Scientific opinion on the assessment of allergenicity of GM plants and microorganisms and derived food and feed. EFSA Journal 8:1700.
EFSA (European Food Safety Authority). 2011a. Guidance on risk assessment of food and feed from genetically modified plants. EFSA Journal 9:2150.
EFSA (European Food Safety Authority). 2011b. Scientific opinion on guidance on conducting repeated-dose 90-day oral toxicity study in rodents on whole food/feed. EFSA Journal 9:2438.
EFSA (European Food Safety Authority). 2011c. Statistical significance and biological relevance. EFSA Journal 9:2372.
EFSA (European Food Safety Authority). 2012. Review of the Séralini et al. (2012) publication on a 2-year rodent feeding study with glyphosate formulations and GM maize NK603 as published online on 19 September 2012 in Food and Chemical Toxicology. EFSA Journal 10:2910.
EFSA (European Food Safety Authority). 2013. Scientific Opinion on application EFSA-GMO-NL-2007-45 for the placing on the market of herbicide-tolerant, high-oleic acid, genetically modified soybean 305423 for food and feed uses, import and processing under Regulation (EC) No 1829/2003 from Pioneer. EFSA Journal 11:3499.
EFSA (European Food Safety Authority). 2015. Conclusion on the peer review of the pesticide risk assessment for the active substance glyphosate. EFSA Journal 13:4302.
Einspanier, R. 2013. The fate of transgenic DNA and newly expressed proteins. Pp. 130–139 in Animal Nutrition with Transgenic Plants, G. Flachowsky, ed. Oxfordshire: UK: CABI Biotechnology Series.
El Ouakfaoui, S., and B. Miki. 2005 The stability of the Arabidopsis transcriptome in transgenic plants expressing the marker genes nptII and uidA. Plant Journal 41:791–800.
Entine, J. 2014. A Science-based Look at Genetically Engineered Crops. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, September 16, Washington, DC.
EPA (U.S. Environmental Protection Agency). 1989. Good laboratory practice standards. Federal Register 54:34067.
EPA (U.S. Environmental Protection Agency). 2001a. Bacillus thuringiensis subspecies Cry1F Protein and the Genetic Material Necessary for Its Production (Plasmid Insert PHI 8999) in Corn. Available at http://ofmpub.epa.gov/apex/pesticides/f?p=chemicalsearch:3:0::no:1,3,31,7,12,25:p3_xchemical_id:1322. Accessed October 10, 2015.
EPA (U.S. Environmental Protection Agency). 2001b. EPA Releases Draft Report on Starlink corn. Online. EPA Press Release. Available at https://yosemite.epa.gov/opa/admpress.nsf/blab9f485b098972852562e7004dc686/cd9013801973259885256a0800710574?OpenDocument. Accessed March 12, 2016.
EPA (U.S. Environmental Protection Agency). 2013. Glyphosate; Pesticide tolerances. Federal Register 78:25396–25401.
EPA (U.S. Environmental Protection Agency). 2014a. Final Registration of Enlist Duo™ Herbicide. October 15. Available at http://www2.epa.gov/sites/production/files/2014-10/documents/final_registration_-_enlist_duo.pdf. Accessed October 13, 2015.
EPA (U.S. Environmental Protection Agency). 2014b. Memorandum: Response to Public Comments Received Regarding New Uses of Enlist Duo™ on Corn and Soybeans. October 14. Available at http://www2.epa.gov/sites/production/files/2014-10/documents/response_to_comments.pdf. Accessed October 10, 2015.
EPA (U.S. Environmental Protection Agency). 2014c. SAP Minutes No. 2014-02, A Set of Scientific Issues Being Considered by the Environmental Protection Agency Regarding: RNAi Technology: Problem Formulation for Human Health and Ecological Risk Assessment. Available at https://www.epa.gov/sites/production/files/2015-06/documents/012814minutes.pdf. Accessed March 13, 2016.
EPA (U.S. Environmental Protection Agency). 2015. EDSP Weight of Evidence Analysis of Potential Interaction with the Estrogen, Androgen, or Thyroid Pathway; Chemical: Glyphosate. Available at https://www.epa.gov/sites/production/files/2015-06/documents/glyphosate-417300_2015-06-29_txr0057175.pdf. Accessed March 13, 2016.
Fagan, J., M. Antoniou, and C. Robinson. 2014. GMO Myths and Truths. London: Earth Open Source.
Fasano, A., I. Berti, T. Gerarduzzi, T. Not, R.B. Colletti, S. Drago, Y. Elitsur, P.H. Green, S. Guandalini, I.D. Hill, and M. Pietzak. 2003. Prevalence of celiac disease in at-risk and not-at-risk groups in the United States: A large multicenter study. Archives of Internal Medicine 163:286–292.
FDA (U.S. Food and Drug Administration). 1979. Good Laboratory Practice Regulations Management Briefings: Post Conference Report. Rockville, MD: FDA.
FDA (U.S. Food and Drug Administration). 2000a, revised 2007. Guidance for Industry and Other Stakeholders: Toxicological Principles for the Safety Assessment of Food Ingredients (Redbook 2000). Available at http://www.fda.gov/Food/GuidanceRegulation/GuidanceDocumentsRegulatoryInformation/IngredientsAdditivesGRASPackaging/ucm2006826.htm. Accessed October 29, 2015.
FDA (U.S. Food and Drug Administration). 2000b, revised 2006. Survey Data on Acrylamide in Food: Total Diet Study Results. Available at http://www.fda.gov/Food/FoodborneIllnessContaminants/ChemicalContaminants/ucm053566.htm. Accessed October 30, 2015.
FDA (U.S. Food and Drug Administration). 2002, revised 2006. Survey Data on Acrylamide in Food: Individual Food Products. Available at http://www.fda.gov/food/foodborneillnesscontaminants/chemicalcontaminants/ucm053549.htm. Accessed December 22, 2015.
FDA (U.S. Food and Drug Administration). 2013. Biotechnology Consultation Note to the File BNF No. 000133. December 16. Available at http://www.fda.gov/Food/FoodScienceResearch/GEPlants/Submissions/ucm382207.htm. Accessed October 29, 2015.
FDA (U.S. Food and Drug Administration). 2015. Biotechnology Consultation Agency Response Letter BNF No. 000141. March 20. Available at http://www.fda.gov/Food/FoodScienceResearch/GEPlants/Submissions/ucm436169.htm. Accessed October 30, 2015.
Fernandez, A., E.N.C. Mills, M. Lovik, A. Spök, A. Germini, A. Mikalsen, and J.M. Wal. 2013. Endogenous allergens and compositional analysis in the allergenicity assessment of genetically modified plants. Food and Chemical Toxicology 62:1–6.
Feron, V.J., and J.P. Groten. 2002. Toxicological evaluation of chemical mixtures. Food and Chemical Toxicology 40:825–839.
Ferruzzi, M.G. 2010. The influence of beverage composition on delivery of phenolic compounds from coffee and tea. Physiology & Behavior 100:33–41.
Finamore, A., M. Roselli, S. Britti, G. Monastra, R. Ambra, A. Turrini, and E. Mengheri. 2008. Intestinal and peripheral immune response to MON810 maize ingestion in weaning and old mice. Journal of Agricultural and Food Chemistry 56:11533–11539.
Folta, K. 2014. Letter to the Editor. Food and Chemical Toxicology 65:392.
Fonseca, C., S. Planchon, J. Renaut, M.M. Oliveira, and R. Batista. 2012. Characterization of maize allergens—MON810 vs. its non-transgenic counterpart. Journal of Proteomics 75:2027–2037.
Forouzanfar, M.H., K.J. Foreman, A.M. Delossantos, R. Lozano, A.D. Lopez, C.J.L. Murray, and M. Naghavi. 2011. Breast and cervical cancer in 187 countries between 1980 and 2010: A systematic analysis. Lancet 378:1461–1484.
Franz J.E., M.K. Mao, and J.A. Sikorski. 1997. Glyphosate: A Unique Global Herbicide. ACS Monograph 189. Washington, DC: American Chemical Society.
Friedman, M. 2006. Potato glycoalkaloids and metabolites: Roles in the plant and in the diet. Journal of Agricultural and Food Chemistry 54:8655–8681.
Fryar, C.D., M.D. Carroll, and C.L. Ogden. 2014. Prevalence of overweight, obesity, and extreme obesity among adults: United States, 1960–1962 through 2011–2012. Available at http://www.cdc.gov/nchs/data/hestat/obesity_adult_11_12/obesity_adult_11_12.pdf. Accessed October 13, 2015.
FSANZ (Food Standards Australia New Zealand). 2013. GM Food Labelling. Available at http://www.foodstandards.gov.au/consumer/gmfood/labelling/Pages/default.aspx. Accessed December 22, 2015.
Fukushima, A., M. Kusano, R.F. Mejia, M. Iwasa, M. Kobayashi, M. Hayashi, A. Watanabe-Takahashi, T. Narisawa, T. Tohge, M. Hur, E. Sykin Wurtele, B.J. Nikolau, and K. Saito. 2014. Metabolomic characterization of knockout mutants in Arabidopsis: Development of a metabolite profiling database for knockout mutants in Arabidopsis. Plant Physiology 165:948–961.
Furgał-Dierżuk, I., J. Strzetelski, M. Twardowska, K. Kwiatek, and M. Mazur. 2015. The effect of genetically modified feeds on productivity, milk composition, serum metabolite profiles and transfer of tDNA into milk of cows. Journal of Animal and Feed Sciences 24:19–30.
Gannon, B., C. Kaliwile, S.A. Arscott, S. Schmaelzle, J. Chileshe, N. Kalungwana, M. Mosonda, K. Pixley, C. Masi, and S.A. Tanumihardjo. 2014. Biofortified orange maize is as efficacious as a vitamin A supplement in Zambian children even in the presence of high liver reserves of vitamin A: A community-based, randomized placebo-controlled trial. American Journal of Clinical Nutrition 100:1541–1550.
García-Villalba, R., C. León, G. Dinelli, A. Segura-Carretero, A. Fernández-Gutiérrez, V. Garcia-Cañas, and A. Cifuentes. 2008. Comparative metabolomic study of transgenic versus conventional soybean using capillary electrophoresis-time-of-flight mass spectrometry. Journal of Chromatography A 1195:164–173.
Gasnier, C., C. Dumont, N. Benachour, E. Clair, M.C. Chagnon, and G.-É. Séralini. 2009. Glyphosate-based herbicides are toxic and endocrine disruptors in human cell lines. Toxicology 262:184–191.
Goodman, R. 2015. Evaluating GE Food Sources for Risks of Allergy: Methods, Gaps and Perspective. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, March 5, Washington, DC.
Goodman, R.E., R. Panda, and H. Ariyarathna. 2013. Evaluation of endogenous allergens for the safety evaluation of genetically engineered food crops: Review of potential risks, test methods, examples and relevance. Journal of Agricultural and Food Chemistry 61:8317–8332.
Graf, L., H. Hayder, and U. Mueller. 2014. Endogenous allergens in the regulatory assessment of genetically engineered crops. Food and Chemical Toxicology 73:17–20.
Green, P.H., and C. Cellier. 2007. Celiac disease. New England Journal of Medicine 357:1731–1743.
Guertler, P., V. Paul, C. Albrecht, and H.H.D. Meyer. 2009. Sensitive and highly specific quantitative real-time PCR and ELISA for recording a potential transfer of novel DNA and Cry1Ab protein from feed into bovine milk. Analytical and Bioanalytical Chemistry 393:1629–1638.
Gupta. R., A. Sheikh, D.P. Strachan, and H.R. Anderson. 2007. Time trends in allergic disorders in the UK. Thorax 62:91–96.
Guyton, K.Z., D. Loomis, Y. Grosse, F.E. Ghissassi, L. Benbrahim-Tallaa, N. Guha, C. Scoccianti, H. Mattock, and K. Straif. 2015. Carcinogenicity of tetrachlorvinphos, parathion, malathion, diazinon, and glyphosate. Lancet Oncology 16:490–491.
Halle, I., and G. Flachowsky. 2014. A four-generation feeding study with genetically modified (Bt) maize in laying hens. Journal of Animal and Feed Sciences 23:58–63.
Hammond, B., R. Dudek, J. Lemen, and M. Nemeth. 2004. Results of a 13 week safety assurance study with rats fed grain from glyphosate tolerant corn. Food and Chemical Toxicology 42:1003–1014.
Hammond, B.G., R. Dudek, J.K. Lemen, and M.A. Nemeth. 2006. Results of a 90-day safety assurance study with rats fed grain from corn borer-protected corn. Food and Chemical Toxicology 44:1092–1099.
Hammond, B., D.A. Goldstein, and D. Saltmiras. 2013. Response to original research article, ‘Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize.’ Food and Chemical Toxicology 53:459–464.
Hayes, A.W. 2014. Retraction notice to “Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize.” Food and Chemical Toxicology 63:244.
He, X.Y., K.L. Huang, X. Li, W. Qin, B. Delaney, and Y.B. Luo. 2008. Comparison of grain from corn rootworm resistant transgenic DAS-59122-7 maize with non-transgenic maize grain in a 90-day feeding study in Sprague-Dawley rats. Food and Chemical Toxicology 46:1994–2002.
He, X.Y., Z. Mao, Y.B. Tang, X. Luo, S.S. Li, J.Y. Cao, B. Delaney, and L.H. Kun. 2009. A 90-day toxicology study of transgenic lysine-rich maize grain (Y642) in Sprague-Dawley rats. Food and Chemical Toxicology 47:425–432.
Health Canada. 1999. Novel Food Information—Food Biotechnology, High Lauric Acid Canola Lines 23-198, 23-18-17. Available at http://www.hc-sc.gc.ca/fn-an/alt_formats/hpfb-dgpsa/pdf/gmf-agm/ofb-096-100-a-eng.pdf. Accessed May 11, 2016.
Health Canada. 2015. Proposed Re-evaluation Decision PRVD2015-01, Glyphosate. Available at http://www.hc-sc.gc.ca/cps-spc/pest/part/consultations/_prvd2015-01/prvd201501-eng.php. Accessed March 13, 2016.
Hefferon, K.L. 2015. Nutritionally enhanced food crops; Progress and perspectives. International Journal of Molecular Sciences 16:3895–3914.
Hellenas, K.E., C. Branzell, H. Johnsson, and P. Slanina. 1995. High levels of glycoalkaloids in the established Swedish potato variety Magnum Bonum. Journal of the Science of Food and Agriculture 23:520–523.
Herman, R.A. and W.D. Price. 2013. Unintended compositional changes in genetically modified (GM) crops: 20 years of research. Journal of Agricultural and Food Chemistry 61:11695–11701.
Herman, R.A., N.P. Storer, and Y. Gao. 2006. Digestion assays in allergenicity assessment of transgenic proteins. Environmental Health Perspectives 114:1154–1157.
Hernández, A.F., T. Parrón, A.M. Tsatsakis, M. Requena, R. Alarcón, and O. López-Guarnido. 2013. Toxic effects of pesticide mixtures at a molecular level: Their relevance to human health. Toxicology 307:136–145.
Hidalgo, F.J., and R. Zamora. 2006. Peptides and proteins in edible oils: Stability, allergenicity, and new processing trends. Trends in Food Science & Technology 17:56–63.
Hilbeck, A., R. Binimelis, N. Defarge, R. Steinbrecher, A. Székács, F. Wickson, M. Antoniou, P.L. Bereano, E.A. Clark, M. Hansen, E. Novotny, J. Heinemann, H. Meyer, V. Shiva, and B. Wynne. 2015. No scientific consensus on GMO safety. Environmental Sciences Europe 27:4.
Hohlweg, U., and W. Doerfler. 2001. On the fate of plant or other foreign genes upon the uptake in food or after intramuscular injection in mice. Molecular Genetics and Genomics 265:225–233.
Hotz, C., C. Loechl, A. Lubowa, J.K. Tumwine, G. Ndeezi, A. Nandutu Masawi, R. Baingana, A. Carriquiry, A. de Brauw, J.V. Meenakshi, and D.O. Gilligan. 2012a. Introduction of beta-carotene-rich orange sweet potato in rural Uganda resulted in increased vitamin A intakes among children and women and improved vitamin A status among children. Journal of Nutrition 142:1871–1880.
Hotz, C., C. Loechl, A. de Brauw, P. Eozenou, D. Gilligan, M. Moursi, B. Munhaua, P. van Jaarsveld, A. Carriquiry, and J.V. Meenakshi. 2012b. A large-scale intervention to introduce orange sweet potato in rural Mozambique increases vitamin A intakes among children and women. British Journal of Nutrition 108:163–176.
Huang, J., R. Hu, C. Fan, C.E. Pray, and S. Rozelle. 2002. Bt cotton benefits, costs, and impacts in China. AgBioForum 5:153–166.
Huang, J., R. Hu, S. Rozelle, and C. Pray. 2005. Insect-resistant GM rice in farmers’ fields: Assessing productivity and health effects in China. Science 308:688–690.
IARC (International Agency for Research on Cancer). 2002. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 82: Some Traditional Herbal Medicines, Some Mycotoxins, Naphthalene, and Styrene. Lyon, France: IARC.
IARC (International Agency for Research on Cancer). 2006. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 88: Formaldehyde, 2-Butoxyethanol, and 1-tert-Butoxypropan-2-ol. Lyon, France: IARC.
IARC (International Agency for Research on Cancer). 2015. Glyphosate. Part of Volume 112 in International Agency for Research on Cancer Monographs on the Evaulation of Carcinogenic Risks to Humans: Some Organophosphate Insecticides and Herbicides: Diazinon, Glyphosate, Malathion, Parathion, and Tetrachlorvinphos. Available at http://monographs.iarc.fr/ENG/Monographs/vol112/mono112-09.pdf. Accessed December 20, 2015.
Ibáñez, C., C. Simó, V. García-Cañas, T. Acunha, and A. Cifuentes. 2015. The role of direct high-resolution mass spectrometry in foodomics. Analytical and Bioanalytical Chemistry 407:6275–6287.
Jackson, K.D., L.D. Howie, and L.J. Akinbami. 2013. Trends in allergenic conditions among children: United States 1997–2011. NCHS Data Brief 121:1–8.
Jaffe, G. 2014. Issues for the Committee on Genetically Engineered Crops to Consider. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, September 16, Washington, DC.
John, B. 2014. Letter to the Editor. Food and Chemical Toxicology 65:391.
Johnson, N. July 1, 2014. Retracted Roundup-fed Rat Research Republished. Online. The Grist. Available at http://grist.org/food/retracted-roundup-fed-rat-research-republished/. Accessed December 13, 2015.
Jones, Y. M., and A. De Brauw. 2015. Using Agriculture to Improve Child Health: Promoting Orange Sweet Potatoes Reduces Diarrhea. World Development 74:15–24.
Joshi, L., J.M. van Eck, K. Mayo, R.G. Silvestro, M.E. Blake, T. Ganapathi, V. Haridas, J.U. Gutterman, and C.J. Arntzen. 2002. Metabolomics of plant saponins: Bioprospecting triterpene glycoside diversity with respect to mammalian cell targets. OMICS A Journal of Integrative Biology 6:235–246.
Keese, P. 2008. Risks from GMOs due to horizontal gene transfer. Environmental Biosafety Research 7:123–149.
Kiliç, A., and M.T. Akay. 2008. A three-generation study with genetically modified Bt corn in rats: Biochemical and histopathological investigation. Food and Chemical Toxicology 46:1164–1170.
Knudsen, I., and M. Poulsen. 2007. Comparative safety testing of genetically modified foods in a 90-day rat feeding study design allowing the distinction between primary and secondary effects of the new genetic event. Regulatory Toxicology and Pharmacology 49:53–62.
Kouser, S., and M. Qaim. 2011. Impact of Bt cotton on pesticide poisoning in smallholder agriculture: A panel data analysis. Ecological Economics 70:2105–2113.
Kouser, S., and M. Qaim. 2013. Valuing financial, health, and environmental benefits of Bt cotton in Pakistan. Agricultural Economics 44:323–335.
Krimsky, S. 2015. An illusory consensus behind GMO health assessment. Science, Technology, & Human Values 40:883–914.
Krishnan, H.B., M.S. Kerley, G.L. Allee, S. Jang, W.S. Kim, and C.J. Fu. 2010. Maize 27 kDa gamma-zein is a potential allergen for early weaned pigs. Journal of Agricultural and Food Chemistry 58:7323–7328.
Kuc, J. 1982. Phytoalexins from the Solanaceae. Pp. 81–105 in Phytoalexins, J.A. Bailey and J.W. Mansfield, eds. New York: Wiley.
Kuiper, H.A., E.J. Kok, and H.V. Davies. 2013. New EU legislation for risk assessment of GM food: No scientific justification for mandatory animal feeding trials. Plant Biotechnology Journal 11:781–784.
Ladics, G.S., and M.K. Selgrade. 2009. Identifying food proteins with allergenic potential: Evolution of approaches to safety assessment and research to provide additional tools. Regulatory Toxicology and Pharmacology 54:S2-S6.
Ladics, G.S., G.J. Budziszewski, R.A. Herman, C. Herouet-Guicheney, S. Joshi, E.A. Lipscomb, S. McClain, and J.M. Ward. 2014. Measurement of endogenous allergens in genetically modified soybeans—Short communication. Regulatory Toxicology and Pharmacology 70:75–79.
Langkilde, S., M. Schrøder, T. Frank, L.V.T. Shepherd, S. Conner, H.V. Davies, O. Meyer, J. Danier, M. Rychlik, W.R. Belknap, K.F. McCue, K.-H. Engel, D. Stewart, I. Knudsen, and M Poulsen. 2012. Compositional and toxicological analysis of a GM potato line with reduced α-solanine content—A 90-day feeding study in the Syrian Golden hamster. Regulatory Toxicology and Pharmacology 64:177–185.
Lee, S.H., and B.R. Hamaker. 2006. Cys 155 of 27 kDa maize γ-zein is a key amino acid to improve its in vitro digestibility. FEBS Letters 580:5803–5806.
Ley, R.E. 2010. Obesity and the human microbiome. Current Opinion in Gastroenterology 26:5–11.
Liebsch, M., B. Grune, A. Seiler, D. Butzke, M. Oelgeschläger, R. Pirow, S. Adler, C. Riebeling, and A. Luch. 2011. Alternatives to animal testing: Current status and future perspectives. Archives of Toxicology 85:841–858.
Litten-Brown, J.C., A.M. Corson, and L. Clarke. 2010. Porcine models for the metabolic syndrome digestive and bone disorders: A general overview. Animal 4:899–920.
Liu, P., X. He, D. Chen, Y. Luo, S. Cao, H. Song, T. Liu, K. Huang, and W. Xu. 2012. A 90-day subchronic feeding study of genetically modified maize expressing Cry1Ac-M protein in Sprague–Dawley rats. Food and Chemical Toxicology 50:3215–3221.
Lividini, K., and J.L. Fielder. 2015. Assessing the promise of biofortification: A case study of high provitamin A maize in Zambia. Food Policy 54:65–77.
Livingston, M, J. Fernandez-Cornejo, J. Unger, C. Osteen, D. Schimmelpfennig, T. Park, and D. Lambert. 2015. The Economics of Glyphosate Resistance Management in Corn and Soybean Production. Washington, DC: U.S. Department of Agriculture–Economic Research Service.
Ludvigsson, J.F., A. Rubio-Tapia, C.T. van Dyke, L.J. Lelton, A.R. Zinsmeister, B.D. Lahr, and J.A. Murray. 2013. Increasing incidence of celiac disease in a North American population. American Journal of Gastroenterology 108:818–824.
Lusk, R.W. 2014. Diverse and widespread contamination evident in the unmapped depths of high throughput sequencing data. PLoS ONE 9:e110808.
MacKenzie, S.A., I. Lamb, J. Schmidt, L. Deege, M.J. Morrisey, M. Harper, R.J. Layton, L.M. Prochaska, C. Sanders, M. Locke, J.L. Mattsson, A. Fuentes, and B. Delaney. 2007. Thirteen week feeding study with transgenic maize grain containing event DAS-01507-1 in Sprague-Dawley rats. Food and Chemical Toxicology 45:551–562.
Macpherson, A.J., K.D. McCoy, F-.E. Johansen, and P. Brandtzaeg. 2008. The immune geography of IgA induction and function. Mucosal Immunology 1:11–22.
Magana-Gomez, J.A., and A.M.C. de la Barca. 2009. Risk assessment of genetically modified crops for nutrition and health. Nutrition Reviews 67:1–16.
Martin, C., Y. Zhang, C. Tonelli, and K. Petroni. 2013. Plants, diet, and health. Annual Review of Plant Biology 64:19–46.
Martín-Hernández, C., S. Bénet, and L. Obert. 2008. Determination of proteins in refined and nonrefined oils. Journal of Agricultural and Food Chemistry 56:4348–4351.
Marx-Stoelting, P., A. Braeuning, T. Buhrke, A. Lampen, L. Niemann, M. Oelgeschlaeger, S. Rieke, F. Schmidt, T. Heise, R. Pfeil, and R. Solecki. 2015. Application of omics data in regulatory toxicology: Report of an international BfR expert workshop. Archives in Toxicology 89:2177–2184.
Mazza, R., M. Soave, M. Morlacchini, G. Piva, and A. Marocco. 2005. Assessing the transfer of genetically modified DNA from feed to animal tissues. Transgenic Research 14:775–784.
Miller, E.R., and D.E. Ullrey. 1987. The pig as a model for human nutrition. Annual Review of Nutrition 7:361–382.
Miller, H.I. 1999. Substantial equivalence: Its uses and abuses. Nature Biotechnology 17:1042–1043.
Millstone, E., E. Brunner, and S. Mayer. 1999. Beyond “substantial equivalence.” Nature 401:525–526.
Mink, P.J., J.S. Mandel, J.I. Lundin, and B.K. Sceurman. 2011. Epidemiologic studies of glyphosate and non-cancer health outcomes: A review. Regulatory Toxicology and Pharmacology 61:172–184.
Mink, P.J., J.S. Mandel, B.K. Sceurman, and J.I. Lundin. 2012. Epidemiologic studies of glyphosate and cancer: A review. Regulatory Toxicology and Pharmacology 63:440–452.
Munkvold, G.P., and A.E. Desjardins. 1997. Fumonisins in maize: Can we reduce their occurrence? Plant Disease 81:556–565.
Murray, J.A., C. Van Dyke, M.F. Plevak, R.A. Dierkhising, A.R. Zinsmeister, and L.J. Melton. 2003. Trends in the identification and clinical features of celiac disease in a North American community, 1950–2001. Clinical Gastroenterology and Hepatology 1:19–27.
Murthy, H.N., M.I. Georgiev, S.-Y. Park, V.S. Dandin, and K.-Y. Paek. 2015. The safety assessment of food ingredients derived from plant cell, tissue and organ cultures: A review. Food Chemistry 176:426–432.
Muzzalupo, I., F. Pisani, F. Greco, and A. Chiappetta. 2015. Direct DNA amplification from virgin olive oil for traceability and authenticity. European Food Research and Technology 241:151–155.
Nakabayashi, R., K. Yonekura-Sakakibara, K. Urano, M. Suzuki, Y. Yamada, T. Nishizawa, F. Matsuda, M. Kojima, H. Sakakibara, K. Shinozaki, A.J. Michael, T. Tohge, M. Yamazaki, and K. Saito. 2014. Enhancement of oxidative and drought tolerance in Arabidopsis by overaccumulation of antioxidant flavonoids. Plant Journal 77:367–379.
National Toxicology Program. 2014. Report on Carcinogens, Thirteenth Edition. Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health Service.
NCI (National Cancer Institute). 2014. Surveillance, Epidemology and End Results (SEER) Program. Available at http://www.cancer.org/research/cancerfactsstatistics/cancerfactsfigures2015/index. Accessed October 29, 2015.
Nemeth, A., A. Wurz, L. Artim, S. Charlton, G. Dana, K. Glenn, P. Hunst, J. Jennings, R. Shilito, and P. Song. 2004. Sensitive PCR analysis of animal tissue samples for fragments of endogenous and transgenic plant DNA. Journal of Agricultural and Food Chemistry 52:6129–6135.
Netherwood, T., S.M. Martin-Orue, A.G. O’Donnell, S. Gockling, J. Graham, J.C. Mathers, and H.J. Gilbert. 2004. Assessing the survival of transgenic plant DNA in the human gastrointestinal tract. Nature Biotechnology 22:204–209.
Nicolia, A., A. Manzo, F. Veronesi, and D. Rosellini. 2014. An overview of the last 20 years of genetically engineered crop safety research. Critical Reviews in Biotechnology 34:77–88.
Nordlee, J.A., S.L. Taylor, J.A. Townsend, L.A. Thomas, and R.K. Bush. 1996. Identification of a Brazil-nut allergen in transgenic soybean. New England Journal of Medicine 334:688–692.
Novak, W.K., and A.G. Haslberger. 2000. Substantial equivalence of antinutritional and inherent plant toxins in genetically modified novel foods. Food and Chemical Toxicology 38:473–483.
NRC (National Research Council). 2000. Genetically Modified Pest-Protected Plants: Science and Regulation. Washington, DC: National Academy Press.
NRC (National Research Council). 2002. Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation. Washington, DC: National Academy Press.
NRC (National Research Council). 2004. Safety of Genetically Engineered Foods: Approaches to Assessing Unintended Health Effects. Washington, DC: National Academies Press.
NRC (National Research Council). 2007. Toxicity Testing in the 21st Century: A Vision and a Strategy. Washington, DC: National Academies Press.
Nwaru, B.I., L. Hickstein, S.S. Panesar, G. Roberts, A. Muraro, and A. Sheikh. 2014. Prevalence of common food allergies in Europe: A systematic review and meta-analysis. Allergy 69:992–1007.
OECD (Organisation for Economic Co-operation and Development). 1993. Safety Evaluation of Foods Derived by Modern Biotechnology: Concepts and Principles. Paris: OECD.
OECD (Organisation for Economic Co-operation and Development). 1998a. Test No. 408: Repeated Dose 90-Day Oral Toxicity Study in Rodents in OECD Guidelines for the Testing of Chemicals. Paris: OECD.
OECD (Organisation for Economic Co-Operation and Development). 1998b. Principles of Good Laboratory Practice and Compliance Monitoring. Available at http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/mc/chem(98)17&doclanguage=en. Accessed October 13, 2015.
OECD (Organisation for Economic Co-Operation and Development). 2000. Report of the Task Force for the Safety of Novel Foods and Feeds. Available at http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/mc/chem(98)17&doclanguage=enhttp://www.biosafety.be/ARGMO/Docments/report_taskforce.pdf. Accessed October 12, 2015.
OECD (Organisation for Economic Co-Operation and Development). 2006. An Introduction to the Food/Feed Safety Consensus Documents of the Task Force. Series on the Safety of Novel Foods and Feeds, No 14. Paris: OECD.
OECD (Organisation for Economic Co-Operation and Development). 2015. Safety Assessment of Foods and Feeds Derived from Transgenic Crops, Volume 2, Novel Food and Feed Safety. Paris: OECD.
Oguchi, T., M. Onishi, Y. Chikagawa, T. Kodama, E. Suzuki, M. Kasahara, H. Akiyama, R. Teshima, S. Futo, A. Hino, S. Furui, and K. Kitta. 2009. Investigation of residual DNAs in sugar from sugar beet (Beta vulgaris L.). Journal of the Food Hygenic Society of Japan 50:41–46.
Onose, J., T. Imai, M. Hasumura, M. Ueda, Y. Ozeki, and M. Hirose. 2008. Evaluation of subchronic toxicity of dietary administered Cry1Ab protein from Bacillus thuringiensis var. Kurustaki HD-1 in F344 male rats with chemically induced gastrointestinal impairment. Food and Chemical Toxicology 46:2184–2189.
Paine, J.A., C.A. Shipton, S. Chaggar, R.M. Howells, M.J. Kennedy, G. Vernon, S.Y. Wright, E. Hinchliffe, J.L. Adams, A.L. Silverstone, and R. Drake. 2005. Improving the nutritional value of Golden Rice through increased pro-vitamin A content. Nature Biotechnology 23:482–487.
Panchin, A.Y., and A.I. Tuzhikov. 2016. Published GMO studies find no evidence of harm when corrected for multiple comparisons. Critical Reviews in Biotechnology, Early Online:1–5.
Patisaul, H.B., and W. Jefferson. 2010. The pros and cons of phytoestrogens. Frontiers in Neuroendocrinology 31:400–419.
Patterson, J.K., X.G. Lei, and D.D. Miller. 2008. The pig as an experimental model for elucidating the mechanisms governing dietary influence on mineral absorption. Experimental Biology and Medicine 233:651–664.
Pecetti, L., A. Tava, A. Romani, M.G. De Benedetto, and P. Corsi. 2006. Variety and environment effects on the dynamics of saponins in lucerne (Medicago sativa L.). European Journal of Agronomy 25:187–192.
Phipps, R.H., E.R. Deaville, and B.C. Maddison. 2003. Detection of transgenic DNA and endogenous plant DNA in rumen fluid, duodenal digesta, milk, blood and feces of lactating dairy cows. Journal of Dairy Science 86:4070–4078.
Poulsen, M., M. Schrøder, A. Wilcks, S. Kroghsbo, R.H. Lindecrona, A. Miller, T. Frenzel, J. Danier, M. Rychlik, Q. Shu, K. Emami, M. Taylor, A. Gatehouse, K.H. Engel, and I. Knudsen. 2007. Safety testing of GM-rice expressing PHA-E lectin using a new animal test design. Food Chemistry and Toxicology 45:364–377.
Racovita, M., D.N. Oboryo, W. Craig, and R. Ripandelli. 2015. What are the non-food impacts of GM crop cultivation on farmers’ health. Environmental Evidence 4:17.
Ren, Y., T. Wang, Y. Peng, B. Xia, and L.J. Qu. 2009. Distinguishing transgenic from nontransgenic Arabidopsis plants by (1)H NMR-based metabolic fingerprinting. Journal of Genetics and Genomics 36:621–628.
Rhee, G.S., D.H. Cho, Y.H. Won, J.H. Seok, S.S. Kim, S.J. Kwack, R.D. Lee, S.Y. Chae, J.W. Kim, B.M. Lee, K.L. Park, and K.S. Choi. 2005. Multigenerational reproductive and developmental toxicity study of bar gene inserted into genetically modified potato on rats. Journal of Toxicology and Environmental Health, Part A: Current Issues 68:2263–2276.
Ricroch, A.E. 2013. Assessment of GE food safety using “-omics” techniques and long-term animal feeding studies. New Biotechnology 30:349–354.
Ricroch, A., J.B. Bergé, and M. Kuntz. 2011. Evaluation of genetically engineered crops using transcriptomic, proteomic and metabolomic profiling techniques. Plant Physiology 155:1752–1761.
Ricroch, A.E., A. Berheim, C. Snell, G. Pascal, A. Paris, and M. Kuntz. 2013a. Long-term and multi-generational animal feeding studies. Pp. 112–127 in Animal Nutrition with Transgenic Plants, G. Flachowsky, ed. Oxfordshire, UK: CABI Biotechnology Series.
Ricroch, A., A. Berheim, G. Pascal, A. Paris, and M. Kuntz. 2013b. Assessment of the health impact of GE plant diets in long term and multigenerational animal feeding trials. P. 234 in Animal Nutrition with Transgenic Plants, G. Flachowsky, ed. Oxfordshire, UK: CABI Biotechnology Series.
Ricroch, A.E., A. Boisron, and M. Kuntz. 2014. Looking back at safety assessment of GM food/feed: An exhaustive review of 90-day animal feeding studies. International Journal of Biotechnology 13:230–256.
Riddle, M.S., J.A. Murray, and C.K. Porter. 2012. The incidence and risk of celiac disease in a healthy US population. American Journal of Gastroenterology 107:1248–1255.
Rizzi, A., L. Brusetti, S. Arioli, K.M. Nielsen, I. Tamagnini, A. Tamburini, C. Sorlini, and D. Daffonchio. 2008. Detection of feed-derived maize DNA in goat milk and evaluation of the potential of horizontal transfer to bacteria. European Food Research and Technology 227:1699–1709.
Rizzi, A., N. Raddadi, C. Sorlini, L. Nordgård, K.M. Nielsen, and D. Daffonchio. 2012. The stability and degradation of dietary DNA in the gastrointestinal tract of mammals: Implications for horizontal gene transfer and the biosafety of GMOs. Critical Reviews in Food Science and Nutrition 52:142–161.
Roberfroid, M. 2014. Letter to the Editor. Food and Chemical Toxicology 65:390.
Rommens, C.M., H. Yan, K. Swords, C. Richael, and J. Ye. 2008. Low-acrylamide French fries and potato chips. Plant Biotechnology Journal 6:843–853.
Ruan, C.J., H.B. Shao, and J.A. Teixeira da Silva. 2012. A critical review on the improvement of photosynthetic carbon assimilation in C3 plants using genetic engineering. Critical Reviews in Biotechnology 32:1–21.
Rubio-Tapia, A., J.F. Ludvigsson, T.L. Brantner, J.A. Murray, and J.E. Everhart. 2012. The prevalence of celiac disease in the United States. American Journal of Gastroenterology 107:1538–1544.
Ruiz-Lopez, N., R.P. Haslam, J.A. Napier, and O. Sayanova. 2014. Successful high-level accumulation of fish oil omega-3 long-chain polyunsaturated fatty acids in a transgenic oilseed crop. Plant Journal 77:198–208.
Saltzman, A., E. Birol, H.E. Bouis, E. Boy, F.F. De Moura, Y. Islam, and W.H. Pfeiffer. 2013. Biofortification: Progress toward a more nourishing future. Global Food Security 2:9–17.
Sanahuja, G., G. Farré, J. Berman, U. Zorrilla-López, R.M. Twyman, T. Capell, P. Christou, and C. Zhu. 2013. A question of balance: Achieving appropriate nutrient levels in biofortified staple crops. Nutritional Research Reviews 26:235–245.
Sanders, D. 2013. Letter to the Editor. Food and Chemical Toxicology 53:450–453.
Schubbert, R., U. Hohlweg, D. Renz, and W. Doerfler. 1998. On the fate of orally ingested foreign DNA in mice: Chromosomal association and placental transmission to the fetus. Molecular Genetics and Genomics 259:569–576.
Séralini, G.E. 2014. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, September 16, Washington, DC.
Séralini, G.E., D. Cellier, and J.S. de Vendomois. 2007. New analysis of a rat feeding study with a genetically modified maize reveals signs of hepatorenal toxicity. Archives of Environmental Contamination and Toxicology 52:596–602.
Séralini, G.E., E. Clair, R. Mesnage, S. Gress, N. Defarge, M. Malatesta, D. Hennequin, and J.S. de Vendômois. 2012. Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize. Food and Chemical Toxicology 50:4221–4231.
Séralini, G.E., E. Clair, R. Mesnage, S. Gress, N. Defarge, M. Malatesta, D. Hennequin, and J.S. de Vendômois. 2014. Republished study: Long-term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize. Environmental Sciences Europe 26:14.
Shepherd, L.V.T., C.A. Hackett, C.J. Alexander, J.W. McNicol, J.A. Sungurtas, D. Stewart, K.F. McCue, W.R. Belknap, and H.V. Davies. 2015. Modifying glycoalkaloid content in transgenic potato—Metabolome impacts. Food Chemistry 187:437–443.
Simó, C., C. Ibáñez, A. Valdés, A. Cifuentes, and V. García-Cañas. 2014. Metabolomics of genetically modified crops. International Journal of Molecular Sciences 15:18941–18966.
Sinden, S.L., and R.E. Webb. 1972. Effect of variety and location on the glycoalkaloid content of potatoes. American Potato Journal 49:334–338.
Singhal, K.K., A.K. Tyagi, Y.S. Rajput, M. Singh, H. Kaur, T. Perez, and G.F. Hartnell. 2011. Feed intake, milk production and composition of crossbred cows fed with insect-protected Bollgard II® cottonseed containing Cry1Ac and Cry2Ab proteins. Animal 5:1769–1773.
Small, E. 1996. Adaptations to herbivory in alfalfa (Medicago sativa). Canadian Journal of Botany 74:807–822.
Smith, J.M. 2003. Seeds of Deception: Exposing Industry and Government Lies about the Safety of the Genetically Engineered Foods You’re Eating. Fairfield, IA: Yes! Books.
Smith, J.M. 2013. Are genetically modified foods a gut-wrenching combination? Institute for Responsible Technology. Available at http://responsibletechnology.org/glutenintroduction/. Accessed October 12, 2015.
Smith, J.M. 2014. Recommendations for the Committee on Genetically Engineered Crops. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, September 16, Washington, DC.
Snell, C., A. Bernheim, J.B. Berge, M. Kuntz, G. Pascal, A. Paris, and A.E. Ricroch. 2012. Assessment of the health impact of GM plant diets in long-term and multigenerational animal feeding trials: A literature review. Food and Chemical Toxicology 50:1134–1148.
Spisák, S., N. Solymosi, P. Ittzés, A. Bodor, D. Kondor, G. Vattay, B. Barták, F. Sipos, O. Galamb, Z. Tulassay, Z. Szállási, S. Rasmussen, T. Sicheritz-Ponten, S. Brunak, B. Molnár, and I. Csabai. 2013. Complete genes may pass from food to human blood. PLoS ONE 8:e69805.
Springob, K., and T.M. Kutchan. 2009. Introduction to the different classes of natural products. Pp. 3–50 in Plant-Derived Natural Products, A.E. Osbourn and V. Lanzotti, eds. New York: Springer-Verlag.
Steinke, K., P. Guertler, V. Paul, S. Wiedemann, T. Ettle, C. Albrecht, H.H.D. Meyer, H. Spiekers, and F.J. Schwarz. 2010. Effects of long-term feeding of genetically modified corn (event MON810) on the performance of lactating dairy cows. Journal of Animal Physiology and Animal Nutrition 94:e185–e193.
Swiatkiewicz, S., M. Swiatkiewicz, A. Arczewska-Wlosek, and D. Jozefiak. 2014. Genetically modified feeds and their effect on the metabolic parameters of food-producing animals: A review of recent studies. Animal Feed Science and Technology 198:1–19.
Taylor, A. November 25, 2015. EPA nixes approval of Enlist Duo weed killer. Online. The Des Moines Register. Available at http://www.desmoinesregister.com/story/money/agriculture/2015/11/25/epa-nixes-approval-enlist-duo-weed-killer/76386952/. Accessed December 13, 2015.
Taylor, B., H. Jick, and D. MacLaughlin. 2013. Prevalence and incidence rates of autism in the UK: Time trend from 2004–2010 in children aged 8 years. BMJ 3:e003219.
Thayer, K.A., J.J. Heindel, J.R. Bucher, and M.A. Gallo. 2012. Role of environmental chemicals in diabetes and obesity: A National Toxicology Program workshop review. Environmental Health Perspectives 120:779–789.
Trabalza-Marinucci, M., G. Brandi, C. Rondini, L. Avellini, C. Giammarini, S. Costarelli, G. Acuti, C. Orlandi, G. Filippini, E. Chiaradia, M. Malatesta, S. Crotti, C. Antonini, G. Amagliani, E. Manuali, A.R. Mastrogiacomo, L. Moscati, M.N. Haouet, A. Gaiti, and M. Magnani. 2008. A three-year longitudinal study on the effects of a diet containing genetically modified Bt176 maize on the health status and performance of sheep. Livestock Science 113:178–190.
Treutter, D. 2006. Significance of flavonoids in plant resistance: A review. Environmental Chemistry Letters 4:147–157.
Trikha, A., J.A. Baillargeon, Y.F. Kuo, A. Tan, K. Pierson, G. Sharma, G. Wilkinson, and R.S. Bonds. 2013. Development of food allergies in patients with Gastroesophageal Reflux Disease treated with gastric acid suppressive medications. Pediatric Allergy and Immunology 24:582–588.
Turnbaugh, P.J., M. Hamady, T. Yatsunenko, B.L. Cantarel, A. Duncan, R.E. Ley, M.L. Sogin, W.J. Jones, B.A. Roe, J.P. Affourtit, M. Egholm, B. Henrissat, A.C. Heath, R. Knight, and J.I. Gordon. 2009. A core gut microbiome in obese and lean twins. Nature 457:480–484.
Untersmayr, E., and E. Jensen-Jarolim. 2008. The role of protein digestibility and antacids on food allergy outcomes. Journal of Allergy and Clinical Immunology 121:1301–1308.
U.S. Census Bureau. 2014. 65+ in the United States: 2010. Washington, DC: U.S. Government Printing Office.
U.S. Pharmacopeia. 2000. Pharmacopeia, simulated gastric fluid, TS, simulated intestinal fluid, TS. United States Pharmacopeial Convention, v. 24. The National Formulary 9 (US Pharmacopiea Board of Trustees), Rockville, MD, 2235.
USDA–APHIS (U.S. Department of Agriculture–Animal and Plant Health Inspection Service). 2014a. Dow AgroSciences Petitions (09-233-01p, 09-349-01p, and 11-234-01p) for Determinations of Nonregulated Status for 2,4-D-Resistant Corn and Soybean Varieties. Final Environmental Impact Statement—August 2014. Available at https://www.regulations.gov/document?D=APHIS-2013-0042-10218. Accessed May 9, 2016.
USDA–APHIS (U.S. Department of Agriculture–Animal and Plant Health Inspection Service). 2014b. Record of Decision: Dow AgroSciences Petitions (09-233-01p, 09-349-01p, and 11-234-01p) for Determination of Nonregulated Status for 2,4-D-Resistant Corn and Soybean Varieties. Available at https://www.aphis.usda.gov/brs/aphisdocs/24d_rod.pdf. Accessed December 13, 2015.
USDA–APHIS (U.S. Department of Agriculture–Animal and Plant Health Inspection Service). 2014c. Determinations of Nonregulated Status: J.R. Simplot Co.; Potato Genetically Engineered for Low Acrylamide Potential and Reduced Black Spot Bruise. Available at http://www.regulations.gov/#!documentDetail;D=APHIS-2012-0067-0384. Accessed December 22, 2015.
USRDS (United States Renal Data System). 2014. CKD in the general population. Pp. 12–22 in 2014 USRDS Annual Data Report Volume 1. Available at http://www.usrds.org/2014/view/Default.aspx. Accessed October 13, 2015.
Valdés, L., A. Cuervo, N. Salazr, P. Ruas-Madiedo, M. Gueimondea, and S. González. 2015. The relationship between phenolic compounds from diet and microbiota: Impact on human health. Food & Function 6:2424–2439.
van den Eede, G., H. Aarts, H.J. Buhk, G. Corthier, H.J. Flint, W. Hammes, B. Jacobsen, T. Midtvedt, J. van der Vossen, A. von Wright, W. Wackernagel, and A. Wilcks. 2004. The relevance of gene transfer to the safety of food and feed derived from genetically modified (GM) plants. Food and Chemical Toxicology 42:1127–1156.
Van Eenennaam, A.L., and A.E. Young. 2014. Prevalence and impacts of genetically engineered feedstuffs on livestock populations. Journal of Animal Science 92:4255–4278.
VanEtten, H., J.W. Mansfield, J.A. Bailey, and E.E. Farmer. 1994. Two classes of plant antibiotics: Phytoalexins versus “phytoanticipins.” The Plant Cell 6:1191–1192.
Verhoeckx, K.C.M., Y.M. Vissers, J.L. Baumert, R. Faludi, M. Feys, S. Flanagan, C. Herouet-Guicheney, T. Holzhauser, R. Shimojo, N. van der Bolt, H. Wichers, and I. Kimber. 2015. Food processing and allergenicity. Food and Chemical Toxicology 80:223–240.
Wal, J.M. 2015. Assessing and managing allergenicity of genetically modified (GM) foods. Pp. 161–178 in Handbook of Food Allergen Detection and Control, S. Flanagan, ed. Cambridge, UK: Woodhead Publishing.
Walsh, M.C., S.G. Buzoianu, G.E. Gardiner, M.C. Rea, E. Gelencsér, A. Jánosi, M.M. Epstein, R.P. Ross, and P.G. Lawlor. 2011. Fate of transgenic DNA from orally administered Bt MON810 maize and effects on immune response and growth in pigs. PLoS ONE 6:e27177.
Walsh, M.C., S.G. Buzoianu, M.C. Rea, O. O’Donovan, E. Gelencsér, G. Ujhelyi, R.G. Ross, G.E. Gardiner, and P.G. Lawlor. 2012a. Effects of feeding Bt MON810 maize to pigs for 110 days on peripheral immune response and digestive fate of the Cry1Ab gene and truncated Bt toxin. PLoS ONE 7:e36141.
Walsh, M.C., S.G. Buzoianu, G.E. Gardiner, M.C. Rea, R.P. Ross, J.P. Cassidy, and P.G. Lawlor. 2012b. Effects of short-term feeding of Bt MON810 maize on growth performance, organ morphology and function in pigs. British Journal of Nutrition 107:364–371.
Walsh, M.C., S.G. Buzoianu, G.E. Gardiner, M.C. Rea, O. O’Donovan, R.P. Ross, and P.G. Lawlor. 2013. Effects of feeding Bt MON810 maize to sows during first gestation and lactation on maternal and offspring health indicators. British Journal of Nutrition 109:873–881.
Wang, X., X. Chen, J. Xu, C. Dai, and W. Shen. 2015. Degradation and detection of transgenic Bacillus thuringiensis DNA and proteins in flour of three genetically modified rice events submitted to a set of thermal processes. Food and Chemical Toxicology 84:89–98.
Weber, A. 2014. C4 Photosynthesis—A Target for Genome Engineering. Presentation to the National Academy of Sciences’ Committee on Genetically Engineered Crops: Past Experience and Future Prospects, December 10, Washington, DC.
West, J., K.M. Fleming, L.J. Tata, T.R. Card, and C.J. Crooks. 2014. Incidence and prevalence of celiac disease and dermatitis herpetiformis in the UK over two decades: Population-based study. American Journal of Gastroenterology 109:757–768.
Wiatrak, P.J., D.L. Wright, J.J. Marois, and D. Wilson. 2005. Influence of planting date on aflatoxin accumulation in Bt, non-Bt, and tropical non-Bt hybrids. Agronomy Journal 97:440–445.
Wiener, J.B., M.D. Rogers, J.K. Hammitt, and P.H. Sand, eds. 2011. The Reality of Precaution: Comparing Risk Regulation in the United States and Europe. New York: RFF Press.
Wild, C.P., and Y.Y. Gong. 2010. Mycotoxins and human disease: A largely ignored global health issue. Carcinogensis 31:71–82.
Williams, J.H., T.D. Phillips, P.E. Jolly, J.K. Stiles, C.M. Jolly, and D. Aggarwal. 2004. Human aflatoxicosis in developing countries: A review of toxicology, exposure, potential health consequences, and interventions. American Journal of Clinical Nutrition 80:1106–1122.
World Health Organization. 2014. Frequently Asked Questions on Genetically Modified Foods. Available at http://www.who.int/foodsafety/areas_work/food-technology/Frequently_asked_questions_on_gm_foods.pdf. Accessed March 12, 2016.
Wu, G., M. Truksa, N. Datla, P. Vrinten, J. Bauer, T. Zank, P. Cirpus, E. Heinz, and X. Qiu. 2005. Stepwise engineering to produce high yields of very long-chain polyunsaturated fatty acids in plants. Nature Biotechnology 23:1013–1017.
Wu, Y., D.R. Holding, and J. Messing. 2010. γ-Zeins are essential for endosperm medication in quality protein maize. Proceedings of the National Academy of Sciences of the United States of America 107:12810–12815.
Ye, X., S. Al-Babili, A. Klöti, J. Zhang, P. Lucca, P. Beyer, and I. Potrykus. 2000. Engineering the provitamin A (β-Carotene) biosynthetic pathway into (carotenoid-free) rice endosperm. Science 287:303–305.
Yudina, T.G., A.L. Brioukhanov, I.A. Zalunin, L.P. Revina, A.I. Shestakov, N.E. Voyushina, G.G. Chestukhina, and A.I. Netrusov. 2007. Antimicrobial activity of different proteins and their fragments from Bacillus thuringiensis parasporal crystals against clostridia and archaea. Anaerobe 13:6–13.
Zhang, C., A. Yin, H. Li, R. Wang, G. Wu, J. Shen, M. Zhang, L. Wang, Y. Houb, H. Ouyang, Y. Zhang, Y. Zheng, J. Wang, X. Lv, Y. Wang, F. Zhang, B. Zeng, W. Li, F. Yan, Y. Zhao, X. Pang, X. Zhang, H. Fu, F. Chen, N. Zhao, B.R. Hamaker, L.C. Bridgewater, D. Weinkove, K. Clement, J. Dore, E. Holmes, H. Xiao, G. Zhao, S. Yang, P. Bork, J.K. Nicholson, H. Wei, H. Tang, X. Zhang, and L. Zhao. 2015. Dietary modulation of gut microbiota contributes to alleviation of both genetic and simple obesity in children. EBioMedicine 2:968–984.
Zhu, X.-G., L. Shan, Y. Wang, and W.P. Quick. 2010. C4 rice—an ideal arena for systems biology research. Journal of Integrative Plant Biology 52:762–770.