Table of Contents

1. Introduction
2. Background Information - Sequential Equivalence
3. Background Information - Assessments of Nutritional Quality of Animal Feed
4. Background Information - Digestion of Food
5. Performance of Animals Given Feed Derived from Transgenic Crops and Non-Transgenic Crops
6. Comparisons of Animal Food Products, Such as Milk, Meat and Eggs Obtained from Animals Given Feed Derived from Transgenic or Non-Transgenic Crops
7. Detection of Whole Transgenes or Proteins in Food Products, Such as Milk, Meat and Eggs Obtained from Animals That Have Consumed Feed Containing Transgenic Crop Material
8. The Safety of Consuming DNA Derived from Transgenic Crops
9. Summary  

Introduction and Background Information

Transgenic crops that provide insect resistance and herbicide tolerance are currently used as feed sources by the animal production industry. Animal producers want to be certain transgenic crops are safe for animal consumption. In addition, it’s important to producers that animal feed derived from transgenic crops is as nutritious as feed from non-transgenic crops, and that animal productivity yields will not vary according to whether the source of the feed is a transgenic or non-transgenic crop.

Specific questions members of the animal production industry have asked are:

• Will the performance of animals given feed derived from transgenic crops be equal to that of animals provided with feed from non-transgenic crops?
• Are animal food products, such as milk, meat and eggs, essentially equal whether the animal has consumed feed derived from trangenic or non-transgenic crop material?
• Will the novel transgenes or the proteins they encode be present in animal food products if the feed is derived from transgenic crop material?
• Are these food products safe for human consumption if the intact transgenes or the proteins they encode occur in milk, meat or eggs?

This paper addresses these questions and provides information about how transgenic crops compare with non-transgenic crops in animal feeding and performance trials.

Before a transgenic crop is brought to market, companies must show that the transgenic variety is "substantially equivalent" to the non-transgenic variety from which it was derived. Compositional studies are required to show that the biochemical components of the plant and grain such as total protein, fiber, amino acids and fatty acids do not differ between the two versions of the variety. Therefore, the term "substantial equivalence" means that both forms of the variety are the same with the exception of the newly introduced gene and the novel protein it expresses.

Return to Table of Contents

Before a transgenic crop is brought to market, companies must show that the transgenic variety is "substantially equivalent" to the non-transgenic variety from which it was derived. Compositional studies are required to show that the biochemical components of the plant and grain such as total protein, fiber, amino acids and fatty acids do not differ between the two versions of the variety. Therefore, the term "substantial equivalence" means that both forms of the variety are the same with the exception of the newly introduced gene and the novel protein it expresses.

Return to Table of Contents

Scientists familiar with the details of food digestion and absorption predict no difference in performance, yield, composition, safety and nutritional quality of animal products derived from animals that have consumed feed that is substantially equivalent in safety, composition and nutritional quality. Understanding the processes of protein and DNA digestion in animals is helpful for objectively assessing the validity of the points discussed below.

4.1   Protein Digestion and Metabolic Fate

Proteins are broken into increasingly smaller fragments in the mammalian digestive system. Intact proteins are, for the most part, not absorbed across the gut wall per se (Gardner, 1988). One exception would be a special class of milk-borne immunoglobulins (IgA) that are specifically designed to be absorbed to provide passive immunity for newborn mammals (Gardner, 1988). A second exception is that very low levels of intact proteins or large fragments of proteins are taken up by mononuclear leukocytes, possibly through macropinocytosis, as part of the immune system surveillance of gut contents (Tsume et al., 1996). From a nutritional perspective, proteins are typically classified as either non-digestible (excreted in the feces) or digestible (absorbed as small peptides and amino acids). Digestible proteins are largely soluble in water or acid whereas the indigestible proteins are typically insoluble, being bound primarily to sugars or fiber. Within the stomach, hydrochloric acid and pepsin denature and fragment the soluble proteins (Webb, 1990). Later in the small intestine, protein-digesting enzymes secreted by the pancreas and intestinal wall further fragment the protein chains into peptides of decreasing size, typically ranging in length from a few hundred amino acids down to the component individual amino acids and extremely short peptides (two to six amino acids). The stability of proteins introduced in transgenic plants is routinely determined in an in vitro digestive fate study as part of the regulatory assessment of allergenic potential (Genetically Modified Pest-Protected Plants: Science and Regulation, 2000). This approach to assess the safety relative to food allergy is consistent with the guidance provided by the International Life Science Institute (ILSI) and which has served as primary source of guidance for regulatory agencies around the world (Metcalf et al., 1996).

Roberts et al. (1999) evaluated the oral bioavailability of three bioactive peptides of varying size: thyrotropin-releasing hormone, lutenizing hormone-releasing hormone and insulin that are 3, 10 and 51 amino acid in size, respectively. This study showed that the oral bioavailability of these peptides was inversely related to their length such that peptides longer than 10 amino acids were very poorly absorbed intact. In addition, in most cases if small peptides are absorbed by intestinal epithelial cells they are degraded intracellularly to amino acids before being absorbed into the circulation. In another study, it was shown that 0.007-0.008% of ovalbumin orally administered to humans was detectable in their circulation (Tsume et al., 1996). The authors concluded that the digestive tract provides a strong barrier to the absorption of macromolecular proteins into the body. Amino acids that enter the blood stream are used to synthesize new proteins. Any non-essential amino acids in excess of nutritional requirements are further degraded to carbon and nitrogen that are oxidized as a source of energy (carbon), or to ammonia (nitrogen), that is converted to urea by the liver and excreted. Although the indigestible proteins pass through the digestive tract, they may be partially fermented by enteric bacteria in the large intestine as a source of energy and nitrogen.

To date, the proteins introduced into biotech products approved for food and feed usage have been shown to be readily degraded in simulated gastric digestion studies (Betz et al., 2000; Metcalf et al., 1996). Hence, the introduced proteins are unlikely to be detectable in animal products consumed by humans.

4.2   DNA Digestion and Metabolic Fate

DNA, a nucleic acid, encodes the fundamental genetic information by which the vast majority of organisms convey instructions for function and survival of self to subsequent generations. DNA is an essential component of most living organisms and, as such, is present in nearly all foods and feedstuffs. In biotech crops, the introduced transgenic DNA molecules are made of the same basic chemical components as the endogenous DNA (adenine, guanine, thymine, and cytosine). Therefore, the introduction of transgenic DNA into a plant does not introduce any new chemical entities to foods or feeds. Generally, the total DNA in food contributes less than 0.02% to the total dry matter of the food (Watson and Thompson, 1988). Similarly, the amount of transgenic DNA in plants improved through biotechnology represents a small proportion of the total amount of DNA in a biotech plant (<0.0004% of the total plant DNA). For example, it has been estimated that approximately two thirds of a gram (608,000 mg) of DNA is consumed on a daily basis by a 600 kg animal such as a cow (Beever and Kemp, 2000). If 60% of the feed were from a biotech crop, the daily intake of transgenic DNA would be approximately 1.5 mg, which is approximately 0.00025% of the total amount of DNA ingested per day.

The gastrointestinal tract is constantly exposed to DNA that is released from partially or completely digested foods or feeds, ingested microbes, and DNA from intestinal microflora. Ingested food is mechanically disrupted and the released DNA is cleaved through acid hydrolysis and enzymatic digestion (especially by DNase I from salivary and pancreatic secretions) into small DNA fragments and eventually converted to single nucleotides (McAllan, 1982). The presence of various phosphatases and deaminases continue to destroy the structural integrity of any free DNA. One study with beef steers showed that plant DNA in feed is progressively degraded as it moves through their digestive tract, with over 50% being degraded in the first third of the intestine and 80% having disappeared by the time the digesta reaches the terminal ileum (McAllen, 1980). DNA given directly to steers was shown to be completely degraded into mononucleotides by the animal's digestive tract in about 4 hours (McAllen, 1982). The generated nucleotides are readily abundant in food and feed and exceed nutritional requirements of the host (Yu, 1998) and gut bacteria (McAllen, 1982). The breakdown products of DNA are absorbed for use in cellular synthetic processes as they can be found in blood and tissues (McAllen, 1982); however, as intact nucleotides they are non-essential nutrients. The nucleotides are typically deaminated before being rapidly absorbed. Once absorbed, they are further catabolized into nitrogenous bases, free bases and other metabolites including sugars and phosphates that are used in cellular biosynthetic pathways (Sonoda and Tatibana, 1978). Interestingly, intestinal epithelial cells have unique salvage pathways for using free nucleotides, owing to their high rate of cell turnover (He and Walker, 1994). Any small polynucleotide DNA fragments that might enter the body would be phagocytized by mononuclear leukocytes and further degraded by cellular enzymes and nucleases (Doerfler, 2000).

The genetic sequence for a protein introduced in a plant is only functional when the DNA (gene) is activated in the plant. The presence of DNA in the diet is so common that it is of virtually no consequence to animals and people consuming plant-derived products. A recent publication describes experiments that directly tested whether extensive feeding of DNA to mice results in detectable expression of mRNA and protein in any of several organs of the animals (Holweg and Doerfler, 2001). Approximately 50 mg of DNA was fed to the mice per day. The DNA fed to the mice encoded the green fluorescent protein (GFP) under the control of one of three strong mammalian viral promoters [human cytomegalovirus (hCMV), Rous sarcoma virus (RSV) or simian virus 40 (SV-40)]. Separate experiments used a "gene therapy" approach with intramuscular injection into mice of the GFP gene coupled with either the hCMV promoter (pEGFP-C1) or the RSV promoter. These gene therapy studies showed clearly detectable expression of the GFP protein and mRNA at the site of injection. By comparison, no GFP protein or mRNA expression was detectable in liver, spleen, blood or intestinal epithelia of 21 animals fed the exact same DNA over a three week period. Also, fragments of the GFP gene were not detectable by PCR analysis of DNA isolated from spleen, liver or tail tip samples from either this three week feeding study or a separate experiment that involved feeding 50 mg of the pEGFP-C1 DNA per day to mice over eight generations. Therefore, it can be concluded from these studies that gene/promoter constructs clearly capable of functioning in vivo when administered via a gene therapy procedure (e.g. intramuscular injection) do not lead to gene expression in somatic cells or detectable integration into the germline of animals when provided orally.

In addition to digestive processes that degrade DNA, feed-processing procedures (and food preparation methods) significantly degrades DNA, especially those that involve heating to temperatures greater than 95°C (200°F) (Forbes et al., 2000; Gawienowski et al.). For example, the stability of transgenic DNA in maize preserved as silage has been studied (Hupfer et al., 1999). The intact transgene was only detectable during the first five days of ensiling; with small fragments (about 200 bp) of DNA being identifiable using sensitive PCR methods for longer stored silage. The rapid breakdown of DNA during ensiling was not unexpected. This process creates a harsh environment that involves plant tissue being chopped which leads to cell breakage, release of cell contents including the DNA and nucleases, and mild acidic conditions from natural fermentation. Thus, feed generated by ensilage reduces an animal's dietary exposure to intact DNA, including any introduced transgenic DNA, even before ingestion and further degradation by its own digestive system.

Uptake of ingested DNA by intestinal flora causes some to ask if foreign DNA could persist in mammals. The possibility of plasmid DNA being incorporated via a normal biological process into endogenous gut bacteria is minimized due to the non-conjugative nature of typical plasmids used in recombinant DNA laboratories (Hamer, 1977) and the low frequency with which unaided transformation (uptake of naked DNA) occurs. Furthermore, beyond the difficulty of unaided transformation is the lack of stable incorporation (Behr et al., 1989) for DNA in general. Moreover, the probability of transferring such plasmids into natural bacteria in the gut environment has been calculated to be less than one in one million (Maniatis et al., 1982). Both scenarios assume the plasmid is free rather than incorporated into the plant genome, which would require that DNA be precisely removed and that it would be removed in a form that was similar enough that it could be incorporated into the animal genome.

There is also no evidence for the transfer of intact genes to humans from bacteria in the gut or from any food source (The Royal Society, 1998). The DNA remaining after digestion is small random pieces of DNA regardless of the food source. So the fundamental question is related to whether such DNA will be incorporated into the host cells in a functional way. There is no precedence for DNA being incorporated into host cells beyond the use of the basic nucleotide building blocks as nutrients. And, in fact, acid hydrolysis in the stomach is expected to depurinate most adensine and guanine nucleotides, which would result in a DNA sequence that would have very little value (Klinedinst and Drinkwater, 1992).

A series of papers from Prof. Walter Doerfler's laboratory (Doerfler, 2000; Doerfler et al., 1997; Hohlweg and Doerfler, 2001; Schubbert et al., 1994; Schubbert et al., 1997; Schubbert et al., 1998) have addressed questions on the fate of ingested DNA. In preliminary work (Schubbert et al., 1994), mice were fed circular M13 bacteriophage DNA (~ 7.2 Kb) and scientists were able to detect small DNA fragments in certain organs and tissues. These fragments were mostly 200 - 400 bp in size, although up to 1.7 Kb fragments were detected in the feces and up to 500 bp fragments were detectable in the blood. These DNA fragments were detectable within 2-7 hours after feeding. The sum of all of the DNA fragments recovered from all tissues and feces could account for 2-4% of the total M13 fed to the mice, with only 0.01% detectable in the blood. Therefore, 96-98% of the ingested DNA was likely digested quickly and completely to very small and undetectable pieces. Furthermore, in vitro incubation of intact M13 DNA in blood demonstrated complete elimination within 6 hours. The results from this pioneering work were consistent with the general understanding that DNA is a normal component in food and subjected to extensive degradation during digestion. The authors stated "The implication that a random mixture of DNA including gene fragments or intact genes of animal, plant or microbial origin should have been constantly excreted by innumerable organisms over millennia does not appear startling given the complexities of evolution. This barrage of linear DNA fragments, i.e. of recombinationally highly active DNA fragments in Nature should mitigate any concerns that one might have had in the past about biological consequences of experiments carried out with recombinant DNA over the course of the past two decennia." Earlier research⁴¹ complements these reports on studies on the fate of free DNA demonstrated complete and quick digestion in the intestine.

In follow-up studies, a re-cloning experiment with DNA from mouse splenocytes appears to have isolated one M13-positive 1.3 kb fragment clone from 109 clones screened (Einspanier et al, 2001; Schubbert et al., 1997). Moreover, this one was covalently linked to an 80-nucleotide fragment with 70% homology for the IgE receptor gene. Additional chimeric molecules of M13 DNA fragments were also found associated with bacterial DNA (authors suggest from normal gut bacteria) and with rearranged lambda phage DNA. These later findings raise serious questions about the technical conduct of these studies and suggest that the results are possibly common cloning artifacts as cautioned when creating genomic libraries by (Maniatis, 1982). That concern is independent of the question of why an 'altered' IgE sequence (or even how the IgE sequence was altered) was found in the same clone.

Finally, a recent report (Schubbert et al., 1997) from Doerfler's laboratory suggests that ingested foreign DNA can be associated with chromosomes and cross the placenta to the fetus, which led the authors to make the claim, that foreign DNA is a potential mutagen. This last report, using a plasmid containing the GFP gene for green fluorescent protein (pEGFP-C1) in addition to M13, confirmed that DNA fragments (maximum size is 0.82 kb) persist in the digestive track and can penetrate the intestinal wall. The authors extended the previous work by stating that the fragments were found in the cell nuclei by using in situ hybridization and even linked to chromatids of the parent and offspring. They also report the persistence of an M13 fragment in one fetus (a few cells) for up to three months. A conflicting study, on a separate topic (Tsukamoto, 1995), showed that DNA could not cross the placenta and is not taken up by fetuses without the addition of lipopolyamines (synthetic additive), an essential added component. Furthermore, follow-up studies from the Doerfler laboratory failed to detect fragments of the GFP gene by polymerase chain reaction (PCR) analysis of DNA isolated from spleen, liver or tail tip samples from mice fed the pEGFP-C1 DNA over eight generations, showing that the ingested DNA never became integrated into the mouse germline (Hohlweg and Doerfler, 2001).

The Doerfler work has been critically reviewed and questioned (Beever and Kemp, 2000) not only in regard to the feeding of large quantities of purified DNA to animals but also in regards to some of the methodology. In particular, a question is raised about the possibility of methodological errors given the infrequent observations of incorporated M13 DNA and that identifications were only after powerful amplification methods. Interestingly, even with high doses of foreign DNA (M13), no reports of adverse effects on the animals were presented. Further confounding the issue was the question of whether these fragments might be fully explainable by normal immuno-surveillance mechanism involving mononuclear phagocyte ingestion of foreign materials from the digestive tract.

Return to Table of Contents

Discussion

Companies and independent academic scientists have completed dozens of animal performance studies, many of which have been published, on a wide variety of species including beef and dairy cattle, broiler and layer chickens, swine, sheep and catfish. Below is a summary of the results of these published studies.

Animal performance studies are designed to demonstrate whether or not feed from transgenic crops is nutritionally equivalent to feed from non-transgenic crops by measuring growth traits such as feed intake, body weight gain and feed efficiency of the animals that consume the feed. Any unintended health effects would also be observed if they occur. The U.S. and several international regulatory agencies now require a wild or domesticated fowl growth study for regulatory approval. Additional studies with other species are conducted voluntarily to provide the livestock industry with information on the nutritional quality of transgenic crops. Unexpected changes in nutritional quality and safety are assessed through sensitive biochemical tests when the regulatory agencies determine whether or not a crop is substantially equivalent to its non-transgenic counterpart.

Transgenic and conventional corn, soybeans, canola and sugar beets have been compared in feeding studies to assess animal performance, and no differences have been found in feed derived from transgenic and non-transgenic crops. These feeding studies support the claim that feed derived from transgenic grain can be considered substantially equivalent to feed derived from non-transgenic grain.

In reviewing the animal feeding studies, the focus will be on corn and soybeans because they make up the main ingredients in most livestock rations. Specifically discussed are studies involving transgenic corn containing a gene from Bacillus thuringiensis (Bt), which protects the corn from the European corn borer and transgenic soybeans that are tolerant to the herbicide glyphosate.

5.1   Transgenic Corn

5.1a   Poultry

In a five-day study, Aulrich et al. (1998) fed Bt and non-Bt corn grain to laying hens and measured no differences in nutrient composition, body weight, digestible organic matter and protein as well as energy available for metabolism. Brake and Vlachos (1998) conducted a 38-day broiler study comparing Bt and non-Bt corn and measured no differences in mortality, body weight, or feed intake while measuring an improvement (p<0.05) in feed conversion. Carcass data did not differ between groups with one exception; breast meat yield was higher (P<0.05) in broilers fed Bt corn. However, slight differences in overall composition of the diets may have been the cause for these improvements with the Bt corn. In a third study, Halle et al. (1998) fed Bt and non-Bt corn to broilers for 35 days and found no differences in body weight gain, feed intake, feed conversion or protein digestibility. Mireles et al. (2000) conducted two studies to compare nutrient composition and availability in Bt and non-Bt corn. The first study measured true metabolizable energy (TME) and amino acid digestibility and found no difference between the two corn sources. The second study was designed to measure the performance of broiler chickens fed starter feeds. No differences in weight gain and feed efficiency were seen in chickens fed Bt or non-Bt corn. Finally, Sidhu et al. (2000) found no difference in growth, feed efficiency and fat pad weight between broilers fed diets containing glyphosate-tolerant corn and diets containing conventional corn.

5.1b   Lactating Cows

Faust and Miller (1997) fed green chop from Bt or non-Bt corn to lactating cows for 14 days. No differences were measured in feed intake, milk yield, milk composition or udder health. Mayer and Rutzmoser (1999) fed cows diets containing either Bt or non-Bt corn silage with grass silage and supplement in a five-week crossover test. They found no differences in food intake, milk production and milk composition (fat, protein, lactose, urea).

Folmer et al. (2000b) utilized a 4 X 4 Latin square design to compare the lactational performance and ruminal fermentation parameters of cows fed one of four balanced diets containing Bt or non-Bt corn silage from either early maturity or late maturity hybrids. No differences were measured between Bt and non-Bt diets at either maturity for milk production, milk composition or rumen fermentation characteristics (in-situ fiber digestibility, rumen volatile fatty acid [VFA] concentration and rumen pH). The animals fed early maturity hybrids (Bt and non-Bt) had improved (P<0.005) total rumen VFA and efficiency of production when compared to later maturity hybrids (Bt and non-Bt).

Donkin et al. (2000) fed lactating dairy cows either glyphosate tolerant and non-glyphosate tolerant corn silage and corn grain in identical mixed rations. They found no differences in dry matter intake, milk production, milk protein yield, lactose yield or milk fat yield between the two treatment groups. Likewise, no differences were found in milk composition as measured by percentage of fat, protein, lactose and non-fat solids; somatic cell count; and milk urea nitrogen.

5.1c   Beef and Sheep

A common practice is to graze cornfields after harvest with animals that have a lower nutrient requirement such as beef cows. Cows are turned out to graze the stalks and leaves that are left in the field after the combine has harvested the corn grain. Another practice is to harvest and precision chop the entire corn plant and store the chopped forage in a large silo that is relatively air-free. Over the period of a couple of weeks, the forage ferments resulting in the production of lactic acid, which preserves the forage as silage for future use.

Animal performance of beef cows grazing Bt or non-Bt corn crop residue was compared over a two-year period (Russell et al., 2000; Russell et al. 1999). There was no difference in animal performance in either year of the two-year study.

Two trials were conducted to evaluate the use of corn silage and corn residue by Folmer et al. (2000). An absence of significant European corn borer pressure resulted in similar grain yield and residue corn between Bt and non-Bt corn. In trial 1, 23 acres of Bt corn residue and 21 acres of non-Bt were divided into 3 pastures each and stocked with eight or nine steers per pasture to result in equal stocking rates. Average daily gain (avg. 0.28 kg/d) was similar between both corn sources. In addition, 16 steers were allowed access to either seven acres of Bt or non-Bt corn residue. No preference was show for grazing either field.

In trial 2, 128 steers were fed silage diets containing either Bt or non-Bt versions of two hybrids as corn silage at a 90% inclusion rate with 10% supplement. The steers were fed in a
2 x 2 factorial design and performance parameters were measured. Dry matter intake was higher (P<0 .05) for steers fed Bt than non-Bt corn silage (8.61 vs. 8.32 kg/d respectively). An interaction (P< 0.05) was observed between genotype and the Bt trait for daily gain and feed efficiency. For hybrid A, steers receiving the Bt version had improved daily gain over those receiving the non-Bt version. For hybrid B, there was no significant difference in daily gain between the steers fed the Bt vs. the non-Bt version. Steers fed both versions of hybrid A had improved (P< 0.05) feed efficiency when fed the Bt version than when fed the non-Bt version. No differences were measured in feed efficiency between steers fed the Bt and the non-Bt versions of hybrid B. Improved (P< 0.01) daily gain and feed efficiency were measured for steers fed hybrid B compared to hybrid A, although an interaction was present. The authors concluded that while hybrid genotype appeared to affect performance, there was no consistent effect on performance of growing steers due to the presence of the Bt trait.

Daenicke et al. (1999) compared the digestibility and animal performance of sheep and growing bull calves that were fed Bt corn silage or non-Bt corn silage. There were no differences in the digestibility of organic matter, fat, fiber or nitrogen free extract. Likewise, there were no differences in intake, body weight gain, feed conversion, hot carcass weight, dressing percentage and abdominal fat.

The feeding value of whole plant corn silage and crop residues over a two-year period was compared between Bt and non-Bt corn by Hendrix et al. (2000). Three studies were conducted each year: 1) performance of steer calves fed corn silage, 2) performance of beef cows grazing corn residue and 3) grazing pattern of beef cows when given a choice between Bt and non-Bt residue. There were no differences between steers fed the two corns silage sources for average daily gain or dry matter intake. Feed/gain was greater (P<0.05) for Bt vs. non-Bt corn silage. There was no difference in weight change between cows grazing the Bt and non-Bt residues. Over the entire observation period, no differences were measured in preference for one corn residue over the other between grazing cows.

In another study conducted by Russell et al. (2000), the nutritive value of the crop residues from Bt and non-Bt corn hybrids and their effects on performance of grazing beef cows was studied. No differences were measured between the residues from Bt and non-Bt residues for dry matter or organic matter composition. Over the grazing season, no differences were measured between residues for rates of change of residue composition.

In further work, Russell et al. (2001) studied the effects of grazing crop residues from Bt-corn hybrids on the performance of pregnant beef cows. Four hybrids planted on duplicate fields were utilized in the study that was conducted over 2 consecutive years. One hybrid was non-Bt while three hybrids contained the Bt gene from two different sources. Thirty Angus x Charolais x Simmental cows in midgestation were allotted between to two drylots or the eight crop residue fields to strip-graze for 126 days. Biweekly visually estimated body scores were taken with dry alfalfa hay supplemented to maintain a mean body condition score of 5 out of a 9-point scale. Crop residue yields were determined monthly from a 4 square-meter location in each grazed and ungrazed area paddock. On two consecutive days following 2 weeks of grazing, forage selected during a two-hour grazing period by one fistulated steer per field or drylot was harvested from via the rumen cannulae. Dry matter intake (DMI) was calculated from the digestibility of the forage and the fecal output in two cows per field or drylot during the same 2-day period. There were no effects on yields of harvested grain, dropped ears or grain, residue dry matter (DM) or organic matter (OM) over the 2 years. At grazing initiation, in-vitro organic matter digestibility (IVOMD) as well as acid detergent fiber (ADF) and acid detergent lignin (ADL) differed (P<0.05) by base genetics but not by Bt vs. non-Bt hybrids. Rates of change in NDF, ADF, ADL, crude protein (CP) and IVOMD over winter did not differ between hybrids. There were also no differences between hybrids for intakes of forage digestible OM, NDF and ADF. No differences were seen in the amount of hay required to maintain body condition score between hybrids.

Kerley et al. (2001) compared Bt and non-Bt corn fed to beef steers for the last 49 days of the finishing period. Thirty-six crossbred steers were allotted to six pens and fed a 75% corn diet. Growth performance and carcass parameters were measured. There were no differences in corn composition, average daily gain, feed efficiency, yield grade or quality grade between Bt and non-Bt corn hybrids.

Petty et al. (2001) compared Bt vs. isogenic non-Bt corn fed as whole plant silage (WPS) and dry rolled grain over a two-year period. Each year corn was grown under isolation and harvested as grain and WPS. A feeding study was performed each year utilizing 56 Angus and Simmental sired steers, blocked and randomly allotted by weight and breed type one month postweaning into eight pens of seven steers each. Growing diets comprised primarily of WPS were fed for 89 and 85 days in years 1 and 2 respectively and were followed by finishing diets comprised of 75% dry rolled corn, 15% WPS and 10% supplement for 101 and 84 days in years 1 and 2 respectively. During the grower phase in year one, there were no differences (P>0.05) in average daily gain or dry matter intake but feed efficiency was improved (P<0.05) for the steers fed non-Bt corn, however this difference was not measured in year 2. There were no differences (P>0.05) in average daily gain, dry matter intake or feed efficiency during the finishing phase for either year. Steers were harvested each year when they were estimated to be 75% USDA choice as a group. There were no differences (P>0.05) in carcass characteristics in either year. The investigators summarized that there were no major differences in the feeding value of the Bt-corn compared to its isogenic counterpart.

Petty et al. (2001) also evaluated herbicide resistant and non-herbicide resistant corn fed as whole plant silage (WPS) and dry rolled grain. A total of 56 Angus and Simmental sired steers were blocked and randomly allotted to treatments by weight and breed type. The grower diet of 90% WPS and 10% supplement was fed for 85 days followed by the finishing diet comprised of 75% WPS, 15% dry rolled corn and 10% supplement for 84 days. Average daily gain, dry matter intake and feed efficiency were not different (P>.05) during the grower phase, the finishing phase or over the total duration of the feeding study. Steers were harvested when it was estimated that 75% of the steers would grade USDA choice. There were no differences (P>.05) in carcass characteristics.

5.1d   Swine

Weber et al. (2000) compared grower-finisher performance and carcass characteristics from pigs fed Bt corn, the non-Bt isogenic counterpart or commodity-sourced (CS) corn. For animal performance, no differences were found in average daily gain, average feed intake, or feed efficiency between pigs fed any of the three corn varieties. Pigs fed Bt and non-Bt corn did not differ in carcass weight; however, pigs fed CS corn had heavier carcass weights and higher dressing percentages than the other two groups (P<0.05). Pigs fed the non-Bt isogenic control had smaller lean percentages and greater backfat depth at the 10th rib and P2 location than pigs fed diets containing Bt or the CS corn (P<0.05). Pigs fed the non-Bt corn had greater backfat depth at the last lumbar vertebrae than pigs fed CS corn (P<0.05). Marbling scores were highest for pigs fed Bt and non-Bt corn (P<0.05). The researchers concluded that Bt corn had no adverse effects on swine growth performance or carcass characteristics.

Herbicide tolerant and non-herbicide tolerant corn were compared in swine metabolism studies by Böhme and Aulrich (1999). The results showed no differences in protein digestibility, nitrogen free extract (NFE) digestibility or metabolizable energy (ME).

5.1e   In-Vitro

In studies designed to simulate digestion in the lower intestine, Faust (1999) compared in-vitro digestibility between corn silage derived from Bt and non-Bt corn that were ensiled at two different stages of maturity. No differences were measured in true digestibility or the digestibility of cell walls or dry matter regardless of stage of maturity.

Russell et al. (2000b) studied the nutritive value of the crop residues from Bt and non-Bt corn hybrids. No differences were measured between the residues from Bt and non-Bt residues for dry matter, organic matter composition or in-vitro digestible dry matter. Over the grazing season, no differences were measured between residues for rates of change of residue composition. The researchers also compared the performance of grazing beef cows and found no differences in the in-vitro organic matter digestibility of Bt and non-Bt residues selected by steers.

5.2   Transgenic Soybeans

Hammond et al. (1996) reported results from feeding trials comparing soybean meal derived from glyphosate-tolerant soybeans and soybean meal from the conventional counterpart in broilers, catfish and dairy cows. No differences were measured in feed intake, body weight gain, feed efficiency, breast meat composition and fat pad thickness in broilers. Catfish fed diets comparing both soybean meals exhibited no differences in weight gain, feed efficiency or meat composition. Finally, no differences were measured in feed intake, milk yield, milk composition, dry matter digestibility, and rumen fermentation end products between dairy cows fed diets containing either soybean meal.

Return to Table of Contents

Changes in the composition of a feed/forage crop, whether through conventional breeding, genetic engineering or environmental effects, may change the composition of the animal food products that humans consume (Dickinson and King, 1978; Keeler et al., 1978). The studies described in the previous section demonstrate the feed derived from transgenic crops is nutritionally comparable to feed derived from non-transgenic crops. No researcher was able to find any difference between animals fed transgenic and non-transgenic feed, irrespective of the species they studied or the variable they measured. Consequently, one would expect to see no difference in the quality of the milk, meat and eggs of animals that had eaten either transgenic or non-transgenic feed. The limited research on milk and fish fillets conducted to date supports this view.

Milk from lactating cows fed green chop Bt corn for 14 days was no different in either amount or quality of milk from cows fed green chop from the conventional counterpart (Faust and Miller, 1997). Similarly, milk from lactating cows fed complete diets containing soybean meal from glyphosate-tolerant soybeans did not differ in either amount or quality from the milk of cows fed the non-transgenic soybean variety (Hammond et al., 1996). The quality parameters they measured included the amounts of fat, protein and lactose and the number of somatic cells.

No differences were detected in catfish fed soybean meal from either the glyphosate-tolerant soybeans or the non-transgenic progenitor soybean variety (Hammond et al., 1996). They measured moisture, protein, fat and ash content in catfish fillets.

Return to Table of Contents

Digestive physiology would suggest that neither intact transgenes nor whole proteins are absorbed from the small intestine and incorporated directly into animal tissue. Scientific evidence (Beever and Kemp, 2000) and widespread opinion conclude that ingested DNA cannot be incorporated into animal tissue to any significant extent. Therefore, while small fragments of DNA may be found in meat, milk or eggs, the entire transgene will not be present in any animal product. Similarly, the proteins found in milk, meat and eggs may contain amino acids that were molecular components of the protein encoded by the transgene, but the entire protein encoded by the transgene will not be found in milk, meat or eggs.

7.1   Studies on Proteins from Transgenic Plants

A poster at the 2000 Poultry Science meeting held in Montreal presented data showing that the novel protein in glyphosate tolerant soybeans could not be detected by a double antibody sandwich ELISA specific for this protein (Strategic Diagnostic Inc.) in tissues and eggs of laying hens (Ash et al., 2000). Not surprisingly, the raw transgenic soybeans, soybean meal and complete diets were positive for the novel protein. By comparison, whole egg, egg white, liver and fecal samples were all negative for the biotech protein. The poster concluded that the digestive process of the laying hen effectively breaks down the novel protein from the soybean meal portion of the diet such that no modified protein is detectable in the liver, eggs, or feces.

Weber and Richert (2001) showed that all samples of pork loin muscle tissue from grower-finisher pigs fed Bt corn had no detectable levels of intact or immunologically reactive fragments of the novel protein, using a competitive immunoassay for the Bt protein. Their data showed that the growth performance and carcass characteristics of pigs fed Bt corn was statistically similar to that observed for pigs fed conventional corn.

7.2   Studies on DNA and Transgenes

Klotz and Einspanier (1988) published that the gene responsible for glyphosate tolerance in soybeans was not detectable by PCR followed by Southern blot in either the blood of the cow or the milk from these animals. This publication showed that the highly sensitive method of PCR followed by a Southern was able to detect a small fragment of a highly abundant endogenous chloroplast gene in blood lymphocytes but not milk. Very recently, Einspanier's laboratory has published data from a study in which dairy cows, beef steers and broiler chickens were fed either conventional maize grain or Bt corn (Einspanier et al., 2001; Flachowski et al., 2000). The investigators evaluated two DNA detection technologies [standard Polymerase Chain Reaction (PCR) and Light Cycler "real time" PCR]. Although Light Cycler PCR showed advantages for detecting Bt-maize in feed, this technique did not provide additional sensitivity beyond standard PCR methods for animal tissue samples. The presence of even a small portion of the coding region of the Bt gene was not detectable by either standard PCR or Light Cycler PCR in any samples from the cows, steers or chickens fed Bt corn. Similar to their previous report, using standard PCR technology, a small portion of the coding region of a highly abundant chloroplast gene (tRNAleu) was detectable in lymphocytes of dairy cows and in muscle, liver, spleen and kidneys of chicken, but not in dairy milk, or any tissue samples from steers. It is important to note that plastid genome copy number per cell varies depending on tissue-type, ranging from ~500 to 10,000 copies in roots and leaves, respectively (Bendich, 1987). Therefore, the copy number of plastid genes is orders of magnitude higher than a transgene in a biotech product, which typically has only one copy present per haploid genome. In addition, plastid gene sequences are also present in high numbers in the nuclear genome, with sometimes >100 copies of some sequences being observed (Ayliffe et al., 1998), such that the nuclear copies of plastid genes are an additional source of positive PCR signals. As a consequence, the high copy number of plastid genes and their subcellular localization within organelles could explain detection of these endogenous genes while transgenic DNA fragments are undetected to date.

Khumnirdpetch et al. (2001) attempted to detect transgenic DNA in broiler chickens. Broiler chickens were maintained by commercial standards and fed diets containing meal from either conventional or glyphosate tolerant soybeans from birth to seven weeks of age. Samples (meat, skin, duodenum and liver) were isolated from the birds at 1, 3, 5 and 7 weeks. Real-time PCR was used to test for the transgenic DNA in the various samples. PCR results of the broiler samples taken over this entire seven-week feeding period were all negative. The authors speculated that the negative detection results suggest that the transgenic DNA in glyphosate tolerant soybean meal has been fully degraded in the digestive tract of the broilers.

Weber and Richert (2001) also reported data on PCR studies attempting to detect both the Bt gene and an endogenous corn gene in DNA extracted from 24 pork loin samples (12 fed Bt corn and 12 fed a control conventional corn). PCR, followed by Southern blot analysis for ~200 bp fragments of the cry1Ab and shrunken-2 (sh-2) were uniformly negative. The sh-2 gene is an endogenous single-copy corn gene. By comparison, an endogenous swine gene (pre-prolactin) was readily detected in all pork loin samples, and spiking corn DNA into the extracted swine DNA also yielded positive results, indicating that the DNA quality and PCR conditions were both favorable for detection of DNA fragments, had they been present in the original samples. The PCR assay coupled with Southern blot was shown to have a limit of detection of approximately 1 to 2.5 pg of target DNA per 1 mg of input DNA, or approximately 1 genome equivalent of the target gene per PCR, the theoretical limit of assay sensitivity.

DNA degradation during the digestive process has been documented from mouse feeding studies with M13 phage DNA and recently reviewed (Doerfler, 2000). From studies feeding purified M13 phage DNA to mice, it was observed that up to approximately 0.1% of that ingested DNA could be detected in their blood (Schubbert et al., 1998; Schubbert et al., 1997). This extremely high level of DNA observed in the circulation is most likely owing to unique features of this circular, non-methylated phage DNA. Using the M13 data from mice, however, a calculation can be performed to predict the theoretical level of transgenic DNA that might be present in animal tissues, assuming uniform tissue distribution of that DNA in the farm animal. Basing uptake of transgenic DNA in farm animals on the mouse M13 phage data, it can be estimated that approximately 0.002 fg of transgenic DNA (1 femtogram equals one-trillionth of a mg, or 10-15 of a gram) might be present per mg of muscle tissue in the farm animal. No transgenic DNA has been detected in meat, milk or eggs from farm animals fed biotech products. These results are consistent with the knowledge that there are extremely small amounts of transgenic DNA in plants improved through biotechnology (<0.0004% of the total plant DNA).

However, it is important to remember that even if transgenic DNA is detected by a future study, scientific evidence and opinion concludes that ingested transgenic DNA would not be any different from ingestion of DNA already in foods, which is deemed safe. The safety of ingested DNA cannot only be derived from the long natural history of animal and human consumption of DNA, but it is also significant that, as would be expected because of digestive processes, no intact genes, only relatively small fragments, have yet been detected in animal tissues, regardless of the gene's abundance. Instead, in the published reports describing detection of DNA from ingested plants in animal tissues, only small portions of the entire coding region of these highly abundant chloroplast genes were found (Einspanier et al., 2001; Klotz and Einspanier, 1998). Furthermore, only samples from a fraction of the total number of tested animals are yielding positive detections for these highly abundant gene fragments, suggesting that most of the individual animals are degrading ingested DNA to levels below the most sensitive PCR detection limits. Therefore, the likelihood that a transgenic gene or fragments is absorbed to any significant degree following digestion remains extremely low, especially when the relatively low levels of the transgenic DNA per cell is also considered when compared to the highly abundant endogenous plastid genes.

Return to Table of Contents

Questions about the safety of transgenic DNA were originally asked in the mid-1970s, following the advent of recombinant DNA methodologies that utilized E.coli and recombinant plasmids. Maturin and Curtiss (1977) studied the intestinal fate and rate of degradation of E. coli and plasmid DNA in rats. The bacterial DNA in the small intestine degraded rapidly (initially 24 min-1) due primarily to secretion of pancreatic nucleases into the intestine and, secondarily by stomach acids. Plasmid DNA was even less stable, presumably due to its smaller size.

Both the World Heath Organization and the U.S. Food and Drug Administration concluded that there is no inherent risk in consuming DNA, including that derived from transgenic crops (WHO, 1991; US-FDA, 1992). The basis for their conclusion is that humans and other animals have always consumed DNA from a wide variety of sources including plants, animals, bacteria, parasites and viruses.

DNA is an essential component of all living things, so it is present in nearly all foods humans consume, but in relatively small amounts. As mentioned above, the total amount of DNA in food contributes only 0.02% to the total dry matter of the food (Watson and Thompson, 1998). Foods derived from transgenic crops contain only 0.001% more DNA than traditional foods. To put this into perspective, a dairy cow eating transgenic corn would consume 6.8 µg/day of transgenic DNA out of a total 'natural' dietary DNA exposure of 680 mg/day.

Both transgenic DNA and DNA from non-transgenic crops are made of the same 4 nucleotides - adenosine, guanosine, thymidine, and cytosine. The addition of transgenic DNA does not add any new chemical entity, so the digestive system is well adapted to processing transgenic DNA. The specific sequence of nucleotides in a transgene that is responsible for the production of the novel protein is destroyed once the digestive process begins. Rasche (1998) conducted a direct analysis of the stability of transgenic DNA that encodes the PAT gene in canola. In this study the transgene was completely broken down into nucleotides in actual digestive fluids taken from pork, chicken, and cow within one hour at 37oC and pH 1.5. This rate of degradation is similar to other DNA molecules studied.

The gene sequence encoding a plant protein, whether non-transgenic or transgenic, is functional only when the DNA (gene) is activated in the plant cells; just the presence of the DNA alone in the diet is so common that it is of little consequence to the host.

Return to Table of Contents

Transgenic crops have been fed to numerous species of livestock for several years. Studies comparing livestock fed transgenic and non-transgenic crops show that there is no difference between these types of crops in animal performance or composition of products from these animals. The results of these studies assure livestock producers that they can feed transgenic crops with confidence, knowing that the safety and nutritional value is comparable to the non-transgenic crops that have been fed to livestock for decades.

Return to Table of Contents