Comments on Biotechnology in the year 2000 and beyond

 Consumers Union’s comments on Docket No. 99N-4282, Biotechnology
in the Year 2000 and Beyond Public Meetings

Consumers Union*, the non-profit publisher of Consumer Reports magazine, appreciates the opportunity to comment on the Food and Drug Administrations policies on biotechnology.

Consumers Union urges the FDA to establish a mandatory process to assure the safety of genetically engineered food. That process must ensure the safety of engineered foods developed in other countries as well as in the United States. The review process should be strengthened, and be designed to achieve a standard of "reasonable certainty of no harm."

Consumers Union further urges FDA to require mandatory labeling of genetically engineered food. There is now a strong public desire for labeling, and mandatory labeling would be consistent with existing FDA policy which requires labeling of irradiated and frozen food, as well as food ingredients. In addition, new scientific data about unexpected effects in genetically engineered plants argues for labeling as a precautionary measure. In the event that some unexpected difficulty should develop with an engineered food, labeling would facilitate identification of the problem. Labeling is also essential to the health and well being of individuals with food allergies and sensitivities.

A. Scientific/Safety Issues

1. Has FDA’s consultation process achieved its intended purpose? Based on experience to date, should this regulatory approach "sunset," continue in its current state, be made mandatory or otherwise be revised?

FDA’s consultation process should be mandatory, and should be strengthened, so that the standard of safety achieved is comparable to that for food additives.

It is essential that the FDA process be mandatory so that consumers in the US and our trading partners abroad can be assured that all US genetically engineered foods have gone through a safety review. All major transnational corporations marketing genetically engineered (GE) crops in the US assert that they have brought all products forward to FDA for review. However if the review is voluntary consumers have no guarantee that all developers of GE crops will always come forward.

Of particular concern at this time is the question of assuring the safety of genetically engineered food imported from developing countries. Increasingly, developing countries are developing GE crops in their own laboratories. Many of these crops are likely to be grown for the export market–indeed this may already be the case. Among the Asian countries that have genetic engineering research programs are China, India, the Philippines and Thailand. Among the crops they are working on, according to presentations at the October, 1999 Conference of the Consultative Group on International Agricultural Research in Washington DC are rice, wheat, corn, papaya, bananas, mangos, coconut, potatoes, tomatoes, peppers, cucumbers, and tobacco. China reported 53 transgenic varieties are in commercial production or have been field tested (Zhang, 1999). Other countries with research programs include Iran, Brazil, Mexico, Egypt, South Africa and Malaysia. Systems for assuring the safety of these crops vary widely.

The FDA needs to develop a comprehensive program for assuring the safety of genetically engineered imported food. This should include requirements that all GE crops developed in other countries and sold in the US go through an FDA safety review. To achieve this, FDA will have to begin testing imported food for presence of GEFs. It will also be essential to make mandatory its current system of FDA review, for all products sold in the US market, regardless of country of origin.

The current FDA review process also needs to be strengthened and made more transparent. The "decision tree" approach outlined in the FDA 1992 policy appears to give wide latitude to companies to self-evaluate the safety of their GE products. One approach advocated by the Environmental Defense Fund and others is to require GE foods to go through a food additive petition review. From the public’s point of view, GE foods should meet the standard for food additives of "reasonable certainty of no harm." Like additives, genetically engineered varieties are not essential to the food supply, and there is no reason to introduce any GE food that in any way makes food more hazardous either to the general population, or to particular subpopulations.

FDA should have clear protocols for evaluating the known risks of GE foods, which include introduction of toxins, allergens, and nutritional changes; for catching any unexpected effects which have health implications; and for addressing any public health risks that GE foods may pose, such as exacerbation of antibiotic resistance. In addition it should establish clear benchmarks for how it defines "safety" in each case, which begin from an assumption of "reasonable certainty of no harm" from the product. FDA should give notice and request comment of its policies in this area. We recommend that FDA consider, as benchmarks for safety, that a food not be considered "safe" if the genetic engineering process has introduced a known common allergen, and if the final product includes an antibiotic marker gene.

While FDA’s review system needs to be strengthened, it is possible that a food additive review may not be the optimal system for evaluating safety, and that a review process specifically designed for genetically engineered food could yield both greater assurance of safety for consumers and more efficient use of FDA resources. FDA should develop specific proposals in this area and propose them for public comment.

We present a more detailed discussion of some of our safety concerns in our response to (2) below.

2. What newly emerging scientific information related to the safety of foods derived from bioengineered plants is there, if any? Are there specific tests which, if conducted on such foods, would provide increased assurance of safety for man or animals consuming these foods?

Toxins and Unexpected Effects

Information has appeared in the scientific literature related to the safety of foods derived from genetically engineered (GE) plants which collectively suggests that the FDA’s present regulatory approach is insufficient to ensure that GE foods will not pose health risks to those who consume it. This information relates to unexpected and unpredicted effects of gene insertions, and instability of the genetic characteristics that are introduced. This information leads to the view that FDA must scrutinize genetically engineered foods more closely than it has so far, and in particular should require long-terms (one to two year) animal feeding studies of the whole engineered food. Requiring a more detailed molecular characterization for each genetic transformation event will also help FDA evaluate the potential for risk and may provide a means for FDA to decide how much additional testing is needed.

The studies which lead to greater concern about unexpected effects can be put into two categories: unpredictability of the location and expression of transgenic DNA inserts; and differences resulting from post-translational processing (e.g. proteins from the same gene are not identical in differing organisms).

Unpredictability of the location and expression of transgenic DNA underlines need for long-term toxicity tests of engineered food.

The FDA maintains that GE is more precise than traditional breeding because just the desired gene(s) can be transferred without extra unwanted genetic material and that this increased precision "increase[s] the potential for safe, better characterized, and more predictable foods" (57 FR 22986, May 29, 1992). We disagree. Although rDNA techniques may be more precise than traditional plant breeding in terms of the identity of genetic material transferred, they are less precise in terms of where the material is transferred. Conventional plant breeding shuffles around aberrent versions (alleles) of the same genes, which basically are fixed in the chromosomal locations as a result of evolution. With GE (or rDNA techniques), one inserts genes on essentially a random basis, using a gene "gun" or other techniques (e.g. use of Ti-plasmid, chemoporation, electroporation, etc.) into a plant’s pre-existing chromosomes. Frequently the genetic material comes from living things with which the host organism(s) would never cross in nature.

The process of insertion of genetic material via GE is unpredictable with regard to a number of parameters, including: the number of inserts of transgenic DNA, their location (chromosome, chloroplast, mitochondria), their precise position (i.e. where and on which chromosome), their structure, and their functional and structural stability. While all of these parameters can have consequences, perhaps the most important is the random or semi-random nature of the physical location of the genetic insert. The inability to control where the insertion happens is of key importance. This means that each transformation event is unique and cannot be replicated because the precise location of the insertion of genetic material always will be different.

The variable insertion site can have a number of unpredictable, and potentially negative, consequences (Doerfler et al., 1997). The insertion site can affect expression of the inserted transgene itself as well as the expression of host genes (i.e. genes in the recipient organisms). The former is known as the "position effect". A classic example involved attempting to suppress the color of tobacco and petunia flowers via the transfer of a synthetically created gene designed to turn off (via anti-sense technology) a host pigment gene (van der Krol et al., 1988). The expected outcome was that all the transformed plants would have the same color flowers. However, the transformed plants varied in terms of the amount of color (or pigmentation) in their flowers as well as the pattern of color in the individual flowers. Not only that, but as the season changed (i.e. in different environments), some the flowers also changed their color or color pattern. The factors contributing to the position effect are not fully understood.

The expression of host genes can be influenced by the location of the genetic insertion as well. If the material inserts itself into "the middle" of an important gene, that gene would functionally be turned off. In one experiment, insertion of viral genetic material into a mouse chromosome lead to disruption of a gene which resulted in the death of the mouse embryos (Schnieke et al., 1983). If the "turned off" gene happened to code for a regulatory protein which prevented the expression of some toxin, the net result of the insertion would be to increase the level of that toxin.

The genetic background of the host plant can also affect the level of expression of the transferred gene, which explains the common observation that varieties of the same plant species varied widely in the ease with which they can be genetically engineered (Doerfler et al., 1997; Traavik, 1998). In some varieties, the trait can be expressed at high enough levels to have the desired impact. In others, the expression level is too low to have the desired impact. In general though, scientists do not really understand why some plant varieties yield more successful results in GE than other varieties.

To get around the common problem of an insufficient level of expression of a desired gene product, powerful regulatory elements-particularly promoters/enhancers-are inserted along with the desired transgene and used to maximize gene expression. The promoter has numerous elements that enable it to respond to signals from other genes and from the environment which tell it when and where to switch on, by how much and for how long. When inserted into another organism as part of a "genetic construct," it may also change the gene expression patterns in the recipient chromosome(s) over long distances up- and downstream from the insertion site. If the promoter (plus associated transgenes) is inserted at very different places on a given chromosome or on different chromosomes, the effects may be very different; it will depend on the nature of the genes that are near the insertion site. This uncertainty of insertion site, along with the promoter means that for all transgenic plants, there will be a fundamental unpredictability with regard to: expression level of the inserted foreign gene(s); expression of a vast number of the recipient organism’s own genes; influence of geographical, climate, chemical (i.e. xenobiotics) and ecological changes in the environment; and transfer of foreign genetic sequences within the chromosomes of the host organism, and vertical and/pr horizontal gene transfer to other organisms. Such unpredictability explains the common observations that different insertion events in the same variety can vary greatly in terms of the level of expression of the desired transgene and that the majority of transformation events do not yield useful results (i.e. the transgenic plant is defective in one way or another).

The unpredictable influence of the environment may explain what went wrong in Missouri and Texas with thousands of acres of Monsanto’s glyphosate tolerant cotton and Bt cotton, respectively. In Missouri, in the first year of approval, almost 20,000 acres of this cotton in malfunctioned. In some cases the plants dropped their cotton bolls, in others the tolerance genes were not properly expressed, so that the GE plants were killed by the herbicide (Fox, 1997). Monsanto maintained that the malfunctioning was due to "extreme climatic conditions." A number of farmers sued and Monsanto ended up paying millions of dollars in out-of-court settlements. In Texas, a number of farmers had problems with the Bt cotton in the first year of planting. In up to 50% of the acreage, the Bt cotton failed to provide complete control (a so-called "high dose") to the cotton bollworm (Helicoverpa zea). In addition, numerous farmers had problems with germination, uneven growth, lower yield and other problems. The problems were widespread enough that the farmers filed a class action against Monsanto. Just a few months ago, Monsanto settled the case out of court, again by paying the farmers a significant sum (Schanks [plaintiffs attorney], personal communication). If there could be this unexpected effect on the growing characteristics of the cotton, it is theoretically possible that their could be changes in the plant itself which affect the nutritional or safety characteristics of the plant (used as cattle feed) or the seed (the oil from which is used in a number of food products). This raises the question of whether FDA should establish procedures for assuring safety in the long term.

The unpredictability associated with the process of genetic engineering itself could lead to unexpected effects such as the production of a toxin that doesn’t normally occur in a plant or the increase in a level of a naturally occurring toxin. An example of the former occurred in an experiment with tobacco plants engineered to produce gamma-linolenic acid. Although the plants did produce this compound, another metabolic pathway ended up producing higher quantities of a toxic compound, octadecatetraenic acid, which does not exist in non-engineered plants (Reddy and Thomas, 1996).

An example of the latter occurred in an experiment involving yeast where genes from the yeast were duplicated and then reintroduced via genetic engineering (Inose and Murata, 1995). The scientists found that a three-fold increase in an enzyme in the glycolytic pathway, phoshofructokinase, resulted in a 40-fold to 200-fold increase of methylglyoxal (MG), a toxic substance which is know to be mutagenic (i.e. tests positive in an Ames test). This unexpected effect occurred even though the inserted genetic material came from the yeast itself. As the scientists themselves concluded, "Although, except for the case of microbes, we have no information as to the toxic effect of MG in foods on human beings, the results presented here indicate that, in genetically engineered yeast cells, the metabolism is significantly disturbed by the introduced genes or their gene products and the disturbance brings about the accumulation of the unwanted toxic compound MG in cells. Such accumulation of highly reactive MG may cause a damage in DNA, thus suggesting that the scientific concept of "substantially equivalent" for the safety assessment of genetically engineered food is not always applied to genetically engineered microbes, at least in the case of recombinant yeast cells. . . . Thus, the results presented may raise some questions regarding the safety and acceptability of genetically engineered food, and give some credence to the many consumers who are not yet prepared to accept food produced using gene engineering techniques" (Inose and Murata, 1995: ).

A highly controversial study is that of Ewen and Pusztai published in Lancet in late 1999 (Ewen and Pusztai, 1999). That study used potatoes that were genetically engineered to contain a chemical from the snow drop plant (a lectin, Galanthus nivalis agglutinin [GNA]) to increase resistance to insects and nematodes. Feeding experiments with rats demonstrated a number of potentially negative effects. The study found variable effects on the gastrointestinal tract, including proliferation of the gastric mucosa. Interestingly, the potent proliferative effect on the jejunum was seen only in the rats fed GE potatoes with contained the GNA gene but not in rats fed non-transgenic potatoes to which GNA had been added. Indeed, a previous feeding study utilizing GNA with a 1,000-fold higher concentration than the level expressed in the GE potatoes had found no proliferative effect (Pusztai et al., 1990). The authors proposed "that the unexpected proliferative effect was caused by either the expression of other genes of the construct or by some form of positioning effect in the potato genome caused by GNA gene insertion" (Ewen and Pusztai, 1999: 1354). Such a fine-grained feeding study, which involved utilizing young rats which were still growing and involved weighing various organs and looking very carefully for effects on various organ systems and the immune system is far more detailed than the general feeding studies done utilizing GE plants. While many criticisms have been leveled at this study, we believe it raises important questions that merit further research.

Because of the unexpected effects that are theoretically possible and which have been seen in various experiments, we feel FDA should require long-term animal feeding studies using the whole food product. Such testing should be done on growing animals, so that effects on various organ systems can be readily observed. In addition, fairly extensive data should be taken on the weights of various organs and on histopathology and immunology. In addition, there should be follow-up feeding studies if any data from the lab or field demonstrates that the genetic insert is unstable. FDA should propose its procedures for public comment so that it can get further input from the scientific community and others.

The most commonly used promoter in plant genetic engineering is one from the cauliflower mosaic virus (CaMV); all GE crops on the market contain it. A promoter has numerous elements that enable it to respond to signals from other genes and from the environment which tell it when and where to switch on, by how much and for how long. A CaMV promoter is used for a number of reasons: because it is a very powerful promoter, because it is active in all plants-monocots, dicots, algae-and inE. Coli and because it is not greatly influenced by environmental conditions or tissue types. CaMV has two promoter, 19S and 35S, but the 35S is the one most frequently used because it is the most powerful. The powerful nature of the CaMV 35S promoter means that it is not readily controlled by the host genes that surround it and often yields a high expression level of the transgene next to it. This is not unexpected as CaMV is a virus that is designed to hijack a plant cell’s genetic machinery and make many copies of itself. This also means that it is designed to overcome a plant cell’s defensive devices to prevent foreign DNA from being expressed. In the case of transgenic crops, however, the CaMV promoter is used to put the transgenes outside the normal regulatory circuits of the host organism and have them expressed a very high levels. Being placed outside of normal regulatory circuits may be one of the reasons why GEFs are known to be so unstable (Finnegan and McElroy, 1994). The questions raised by the extensive use of the CaMV 35S promoter in engineered crops should be investigated with further research (Ho et al., 1999)

Post-translational processing

Another area of study that raises serious questions about the safety of transgenic traits is the phenomenon of post-translational processing, which consists of the modification of a protein after it has been translated from the genetic message. And such post-translational processing can have a significant impact on the structure and function of a gene. Furthermore, post-translational processing can differ between organisms, so that the same gene expressed in different genetic backgrounds may have the same amino acid sequence but may differ in structure and function. Examples of such processing includes glycosylation, methylation and acetylation.

Glycosylation consists of the addition of sugar groups (usually oligosaccharides) and can dramatically affect the three-dimensional structure and thus, function of a protein. Indeed, glycosylation is thought to be connected to allergenic and immunogenic responses (Benjuoad et al., 1992). The different proteins produced from the same gene are called glycoforms. Research with recombinant human tissue plasminogen activator (rt-PA) revealed that different glycoforms were created depending on whehter the rt-PA gene was expressed in human, Chinese hamster ovary, or mouse cells (Parekh et al., 1989a,b). Different glycoforms were even produced when different human cell lines were used (Parekh et al., 1989a). The activity (or behavior) of these glycoforms differed. Further work demonstrated that when the rt-PA gene was inserted into tobacco, although it was expressed and the protein had the normal amino acid sequence, it had no physiological activity whatsoever (ref. to come). Parekh et al. (1989b) argue that recombinant glycoproteins produced in plants could be allergenic as it is known that many allergens are glycoproteins.

But perhaps the most dramatic example of how glycosylation can affect the structure and function of proteins and have negative results occurs with the prion protein, which is thought to be the causative agent for transmissible spongiform encephalopathies (Scott et al, 1999). Prion proteins are a normally found attached to the surface of cells in the nerve and immune system. Research has demonstrated that the prion proteins in people suffering nvCJD-a particularly severe form of Creutzfeldt-Jakob disease (CJD) that has been recently strongly linked to bovine spongiform encephalopathy (BSE)-have a glycosylation pattern that differs significantly from that of prion proteins from people suffering other forms of CJD and is identical to the glycosylation patterns of prion proteins from cows with BSE (Hill et al., 1997; Scott et al., 1999). This occurs despite the fact that the amino acid sequence from normal prion proteins and those suffering nvCJD is identical. In this case, the altered glycosylation pattern has had a catastrophic effect on the behavior of the prion protein.

Given that glycosylation patterns can dramatically change the structure and function of proteins and may affect anitgenicity and allergenicity, we feel that FDA should require information on the glycosylation patterns of all trans genes expressed in GE foods.

Acetylation of proteins consists of the addition of acetyl groups to certain amino acids, thereby modifying their behavior. Although incompletely understood, acetylation of the amino acid lysine has been most studied in certain groups of proteins that bind with DNA-histones and high-mobility group proteins-and such acetylation appears to be involved with the regulation of interaction of these proteins with negatively charged DNA molecules (Csordas, 1990). However, it has been discovered that some the lysine residues in rbGH are acetylated, to form epsilon-N-acetyllysine when it is produced in E. coli . Harbour et al. (1992) found this to occur at lysine residues 157, 167, 171 and 180 or rbGH, while Violand et al. (1994) found it at residues 144, 157, and 167. The creation of this mutant amino acid may be overlooked because "(T)he identification of this amino acid cannot be determined by simple amino acid analysis because the acetyl group is labile to the acidic or basic conditions normally used for hydrolysis" (Violand et al, 1994: 1089). The effect this has on the safety, structure and function of rbGH is not known as it hasn’t been actively studied.

The differences in glycosylation and acetylation that can happen when transgenes are expressed in plants or bacteria can possibly affect toxicity and therefore lend further support to the need for toxicity testing using the whole engineered food. At present, to test for acute toxicity of a given transgene, the companies invariably do not use the protein that is produced in the plant itself. Rather, in order to obtain large enough quantities of the protein for testing, the companies will put the transgene into a bacteria (invariable E. coli), isolate the expression product (i.e. the protein) and use that for the acute toxicity testing. However, the protein produced in the bacteria may be glycosylated differently than the same protein produced in the plant. Even if there are no differences in glycosylation, acetylation of lysine residue(s) could cause differences. The presence of such mutant lysine residues could easily be missed as routine amino acid analysis will remove the acetyl group; to find if there are mutant lysine residues, one must specifically look produce the transgene of interest (gene for herbicide tolerance or Bt endotoxin, for example). Thus, whenever possible, FDA should require the companies to use material derived from the transgenic plants themselves in toxicity studies rather than bacterially-derived proteins.

Methylation is the process of putting methyl groups on a molecule. Methylation of DNA, which occurs with the nucleotide bases cytosine and adenosine, is important as this appears to prevent that piece of DNA from being expressed (or "turned on"). Methylation is one of the mechanisms behind the phenomenon of "gene silencing," whereby a cell "turns off" a gene. Transgenic work has found that if you try to insert multiple copies of a gene into a plant, the plant will frequently turn off all, or all but one, of the copies of the transgene (Finnegan and McElroy, 1994). Indeed, some scientists now think that gene silencing is an important defense mechanism that plants use to prevent foreign DNA from being expressed (other mechanisms exist to try to degrade the foreign DNA before it can enter the nucleus of the cell) (Traavik, 1998; Ho, 1998). This should be combined with the recent finding that tobacco plants may contain large numbers of copies of paratetroviral-like sequences, in some cases reaching copy numbers of about 10,000 (Jakowitsch et al. 1999). This study is quite striking as it was previously thought that plant viruses rarely integrate, if at all, into host genomes. Furthermore, such integrated viral genetic material is normally silenced via methylation, so that there could be a lot of dormant viral sequences in plants. Interestingly, the cauliflower mosaic virus promoter (CaMV 35) used in virtually all transgenic plants on the market is a pararetrovirus-derived sequence (i.e. CaMV is a pararetrovirus).

With methylation, the danger exists that the CaMV 35S promoter, being a very powerful "on switch" that can have effects thousands of base pairs upstream and downstream from an insertion point, could inadvertently "turn on" a foreign gene that has previously been silent. Given the studies in the last couple of years that suggest that horizontal gene transfer may be more common than previously thought and that most such foreign DNA, if it survives and is able to incorporate itself in the host genome, is frequently "silenced" via methylation, there’s a potential risk that some nasty dormant genetic material is inadvertently turned on due to the presence of the CaMV promoter. Thus, it becomes important to know the exact insertion site of any and all genetic construct as well as knowing what the genetic sequence is for thousands of bases pairs upstream and downstream from the insertions site, and do long term toxicity tests with the whole engineered food.

What data FDA should require

Because of all the reasons stated above and because of the random nature of the genetic transformation process each random insertion of transgenic DNA will differ in location and in structure from all other inserts. It will be accompanied by a different pattern of unintended positional and pleiotropic effects due, respectively, to the location of the insert and the functional interaction of the insert with host genes. Thus, each transgenic line resulting from the same process, despite using the same vector system and plant materials under the same conditions will be distinct, and must be treated as such. Consequently, we think FDA should require the companies to submit data for each separate transgenic line. For every line, FDA should require a complete molecular characterization of each line with respect to the identity, stability and unintended positional and pleiotropic effects. And based on the results of such characterization, the agency could decide on how much toxicity data to require.

The components of a complete molecular characterization for molecular identity would include, for each transgenic or transformed line:

· Total number of inserts of transgenic DNA
· Location of each insert (organelle [chloroplast, mitochondria, etc.] or chromosomal)
· Exact chromosomal position of each insert
· Structure of each insert (whether duplicated, deleted, rearranged, etc.)
· Complete genetic map of each insert including all elements (coding region, noncoding regions, marker gene, promoters, enhancers, introns, leader sequences, terminators, T-DNA borders, plasmid sequences, linkers, etc. including any truncated, incomplete sequences)
· Complete (nucleotide) base sequence of each insert
· (Nucleotide) base sequence of at least 10kbp (10,000 base pairs) of flanking host genome DNA on either side of the insert, including changes in methylation patterns

To determine stability, the FDA needs data on both functional stability (level of expression remains constant over time and over successive generations) and structural stability (location in the genome and structural arrangement of the insert). For functional stability, FDA would need data on the level of expression of the transgene over time-throughout the lifetime of the plant as well as over a number of generations (say 3 to 5 generations). For structural stability, the FDA would need data on the physical location of the insert in the genome as well as the structure of the insert-throughout the lifetime of the plant as well as over successive generations (say 3 to 5). In addition, the FDA would require appropriate molecular probes for each insert with flanking host genome (organelle sequence) sequences in order to monitor the structural stability of the insert.

To test for unintended positional effects, the FDA could look carefully at the methylation patterns of the genes in the flanking host genome DNA (data we suggest be required under molecular identity characterization). To look for pleiotropic (as well as positional effects), each transformed line must be identified in terms of total protein profile and metabolic profiles. The total protein profiles would help to monitor for unintended changes in the pattern of gene expression while the metabolic profile would help to monitor for unintended changes in metabolism. The use of mRNA fingerprinting and protein fingerprinting as part of the protein profiles would represent a better, finer screen for detecting novel biochemical, immunological or toxicological hazards. Some such tests have been suggested by a Dutch government team and should be more carefully considered by the FDA (Kuiper et al., 1998). If any of these tests found differences, there would be more reasons to ask for more comprehensive toxicity testing.

Public Health/Antibiotic Marker Genes

In 1991-1992, when FDA was developing its policy of GE plants, the conventional wisdom in the scientific community was that DNA was a very fragile molecule that would be readily broken down in the environment and would not survive digestion in the gut. We now know that both assumptions may not always be valid (Traavik, 1998). Even though DNases (molecules that break down DNA) are widely distributed in the environment, free DNA has been found in all ecosystems (marine, fresh water, sediments) studied (Lorenz and Wackernagel, 1994). Indeed, pooled data suggest that free DNA is present in significant amounts in the environment. Larger amounts of DNA are extracted from soil than can be extracted from the cells in the soil (Steffan et al., 1988). Further studies have shown that this free DNA in the soil comes from microorganisms that no longer occur in that habitat (Spring et al., 1992) thus demonstrating that DNA can out-survive the organism it came from and still be capable of being taken up and expressed by microorganisms. Finally, yet other studies have found that pollution (i.e. xenobiotics) can affect the survivability of DNA and the possibility of its transfer to other organisms (Traavik, 1998).

These data lead to serious concerns about the antibiotic resistance marker genes that are present in virtually all engineered plants presently on the market. These genes code for proteins that confer resistance to a given antibiotic. The possibility therefore exists that these genes for antibiotic resistance could be taken up by bacteria, thus exacerbating the already very serious problem of antibiotic resistance in disease causing organisms.

In mammalian system, the question is whether foreign DNA can survive digestion, be taken up through the epithelial surfaces of the gastrointestinal or respiratory tract or not, or be excreted in feces. Studies in the 1970s (Maturin and Curtiss, 1977) and 1980s (McAllan, 1982) in rats and ruminants, respectively, suggested that nucleic acids (e.g. DNA and RNA) failed to find evidence that DNA survived digestion. Consequently, many scientists assumed that DNA was readily digested. However, the methods used to detect DNA were not very sensitive. In the mid-1990s, researchers in Germany, re-investigated the issue, using far more sensitive methods (Schubbert et al., 1994). Mice were fed DNA from the M13 bacteriophage either by pipette or by adding it to the feed pellets. Using sensitive hybridization methods and PCR (polymerase chain reaction) the authors found 2-4% of the M13 DNA in feces and 0.01-0.1% in the blood-both in serum and cell fraction. Sizeable DNA fragments (almost a quarter of the M13 genome) could be found up to 7 hours after uptake.

If free DNA is not immediately digested in the gastrointestinal tract, the possibility also exists that it can be transferred to bacteria that live there. A recent study utilizing a simulated human gut demonstrated that naked DNA had a half-life of 6 minutes, more than enough time for such DNA to transform bacteria (ref to come).

In another experiment, a genetically engineered plasmid was found to survive (6 to 25%) up to an hour of exposure to human saliva (Mercer et al., 1999). Partially degraded plasmid DNA also successfully transformed Streptococcus gordonii, a bacteria that normally lives in the human mouth and pharynx although the frequency of transformation dropped exponentially with time. Transformation occurred with either filter-sterilized human saliva or unfiltered saliva. The study also found that human saliva contains factors that increase the ability of resident bacteria to become transformed by "naked" DNA. Since transgenic DNA from food is highly unlike to be completely broken down in the mouth, it may be able to transform resident bacteria. Of particular concern would be the uptake of transgenic DNA containing antibiotic resistance marker genes, which are found the majority of GE crops presently on the market. It should be pointed out that the antibiotic marker gene present in Novartis’ Bt corn, which codes for resistance to ampicillin, is under the control of a bacterial promoter rather than a plant promoter which would further increase the possibility of expression of the ampicillin resistance gene if it were taken up by bacteria.

In September, 1998, the British Royal Society put out a report on genetic engineering that called for ending the use of antibiotic resistance marker genes in engineered food products (Anonymous, 1998). In May, 1999, the British Medical Association released a report calling for a prohibition on the use of antibiotic resistance mark genes in genetically engineered plants (BMA, 1999).

We therefore urge FDA to prohibit use of antibiotic resistance marker genes as there is no consumer benefit for the presence of such genes in engineered foods and a significant potential risk.


In the United States, about a quarter of all people say they have an adverse reaction to some food (Sloan and Powers, 1986). Studies have shown that 2 percent of adults and 8 percent of children have true food allergies, mediated by immunoglobin E (IgE) (Bock, 1987; Sampson et al., 1992). People with IgE mediated allergies have an immediate reaction to certain proteins that ranges from itching to potentially fatal anaphylactic shock. The most common allergies are to peanuts, other nuts and shellfish.

Allergens can be transferred from foods to which people know they are allergic, to foods that they think are safe, via genetic engineering. In March 1996, researchers at the University of Nebraska in the United States confirmed that an allergen from Brazil nuts had been transferred into soybeans. The Pioneer Hi-Bred International seed company had put a Brazil nut gene that codes for a seed protein into soybeans to improve their protein content for animal feed. In an in-vitro and a skin prick test, the engineered soybeans reacted with the IgE of individuals with a Brazil nut allergy in a way that indicated that the individuals would have had an adverse, potentially fatal reaction to the soybeans (Nordlee et al., 1996).

This case was resolved successfully. As Marion Nestle, the head of the Nutrition Department at New York University summarized in an editorial in the respected New England Journal of Medicine, "In the special case of transgenic soybeans, the donor species was known to be allergenic, serum samples from persons allergic to the donor species were available for testing and the product was withdrawn" (Nestle, 1996: 726). However, for virtually every food, allergists will tell you, there is someone allergic to it. Proteins are what cause allergic reactions, and virtually every gene transfer in crops results in some protein production. Genetic engineering will bring proteins into food crops not just from known sources of common allergens, like peanuts, shellfish and dairy, but from plants of all kinds, bacteria and viruses, whose potential allergenicity is largely uncommon or unknown. Furthermore, there are no fool-proof ways to determine whether a given protein will be an allergen, short of tests involving serum from individuals allergic to the given protein. This point is strongly driven home in the case of the transgenic soybean containing a Brazil nut gene, where animal tests had suggested that the transfered Brazil nut seed storage protein was not an allergen (Nordlee et al., 1996). Had the results of the animal tests been relied on and the soybeans approved, the results could have been disastrous.

Most biotechnology companies increasingly use microorganisms rather then food plants as gene donors, or are designing proteins themselves, even though the allergenic potential of these proteins is unpredictable and untestable. Consequently, Nestle continues, "The next case could be less ideal, and the public less fortunate. It is in everyone’s best interest to develop regulatory policies for transgenic foods that include premarketing notification and labeling" (Nestle, 1996: 727).

In April 1994, the EPA, FDA and USDA hosted a "Conference on Scientific Issues Related to Potential Allergenicity in Transgenic Food Crops." The conference revealed how little is actually known about the topic. Indeed, two conclusions/observations noted by the scientists at the meeting were that there are: i) no direct methods to assess potential allergenicity of proteins from sources that are not known to produce food allergy, and ii) although some assurance can be provided to minimize the likelihood that a new protein will cause an allergic reaction by evaluating its similarity with characteristics of known food allergens (i.e. whether the new protein has a similar protein sequence, is prevalent in food, is resistant to enzymatic and acid degradation, is heat stable, and is of the appropriate molecular size), no single factor is predictive. Since this meeting, FDA has appeared to have taken no significant steps to increase the scientific understanding of allergenicity or to develop a truly predictive methodology for assessing allergenicity of transgenic crops.

We urge FDA to take a leadership role in pushing for scientific research that could result in the development of a truly predictive test for allergenicity. Furthermore, a present, companies evaluate allergenicity by looking only at the similarity of the engineered proteins with characteristics of known food allergens. As pointed out at the April, 1994 Interagency conference, such a rudimentary approach is not completely predictive. We think this approach is not stringent enough. We call on FDA to develop a stringent protocol for testing for allergenicity and to publish such a protocol for comment. Furthermore, since there is no fool-proof predictive methodology for testing for allergenicity, FDA must require labeling of all GE foods to facilitate the ability to detect the appearance of new allergies.

Genetically engineered foods present a qualitatively different risk of allergenicity than do conventional foods. If a consumer develops an allergy to a new food, they will always react to that food, thereby facilitating figuring out which food is causing the allergic reaction. With a genetically engineered food, the person will react only to the genetically engineered variety and not to all varieties of the same food. Without labeling, it will be exceedingly difficult to be able to determine what the offending food is.

3. What types of food products derived from bioengineered plants are planned for the future? Will these foods raise food safety issues that would require different approaches to safety testing and agency oversight? If so, what are these approaches?

The industry is discussing plans to develop products with altered nutritional profiles and/or the insertion of biologically active substances such as are now included in dietary supplements. Psychoactive substances, vitamins and other substances may be engineered into food.

Such plants raise significant questions that we feel the FDA is not equipped to handle. The public health implications of widescale consumption of foods with significantly altered nutrient profiles could be enormous, yet FDA may not have the legal authority to address such issues. This underlines the need to label such foods as genetically engineered, so that people know it is not part of a traditional diet.

B. Public Information Issues

1. Should FDA’s policy requiring labeling for significant changes, including changes in nutrients or the introduction of allergens, be maintained or modified? Should FDA maintain or revise its policy that the name of the new food be changed when the common or usual name for the traditional counterpart no longer applies? Have these policies regarding the labeling of these foods served the public?

The FDA should require mandatory labeling of all genetically engineered food. Consumer Reports tested foods purchased in the supermarket in 1999 as to whether it was genetically engineered and found that many foods contain genetically engineered ingredients and are not labeled as such. There are both health-based, and "right-to-know" reasons for requiring labeling.

The current FDA policy does not serve the public well. A majority of consumers want labeling of all genetically engineered food, as numerous polls attest. A recent IFIC poll shows that public comfort levels with genetic engineering are declining, and support for labeling is growing. An October Gallup poll found that 68% of consumers want labeling even if it increased food costs. A summary of polling data on this topic is appended to these comments.

Recently fifty members of Congress co-signed a letter with House Minority Leader David Bonior advising FDA that in their view, FDA currently has the authority to require mandatory labeling and should do so.

All such foods should be labeled including engineered whole food, processed food containing engineered ingredients, and food produced through genetic engineering, such as milk from cows treated with a genetically engineered drug (which contain residues of the drug). It should be labeled even if the food contains no "foreign" DNA (from a source with which the crop would not cross normally in nature). In part, this is because the process of genetic engineering, regardless of what is introduced, can induce unexpected effects.

The food should be labeled regardless of whether current testing technology is capable of verifying whether the food is engineered or not. Consumer Reports found that current test technology cannot detect whether foods that are highly processed, like corn flakes, or do not contain protein, like oils, are genetically engineered. However the identity of such foods can be maintained through recordkeeping and certification procedures.

The terminology of labeling should be simple and straightforward, such as "contains genetically engineered material." The terminology should not be value laden or promotional, such as "improved through modern biotechnology."

Labeling should also be permitted which states that food does not contain genetically engineered material. We believe FDA should define "not containing" as "no detectable" where current test methodology is available. The current reliable limit of detection appears to be .01 percent; this is the benchmark Consumer Reports used in its testing. Foods containing detectable amounts of GE material should be required to label as such.

It will not be sufficient to merely allow voluntary labeling. Such a policy would lead to very few products being labeled, due to the extra effort involved in meeting standards for "no GE material," thus failing to facilitate consumer information and choice. Voluntary labeling also would not protect people with unusual food allergies or food sensitivities.

In any labeling proposals, we urge FDA not to suggest, propose or require any "contextual statement" on any products labeled as to presence or absence of genetically engineered material. All product labeling required by FDA is "shorthand" and can be misinterpreted. FDA has in the past, correctly and appropriately, generally taken the view that public education programs of various kinds, including company advertising, can dispel misconceptions. Explanatory statements along with labels have generally been deemed unnecessary and inappropriate.

The exception was with labeling of milk from cows not treated with rbGH, where the agency, in an unprecedented action, proposed a contextual statement. The FDA suggested that companies should indicate that FDA has found no significant difference between milk from cows treated with rbGH and milk from cows not treated with the drug. However there are many similar contexts in which consumers could theoretically be misled in which FDA has not required explanatory statements-indeed there is an almost endless list of such potentially misleading situations. Consumers could for example possible think that cheese "made in Wisconsin" or "made by the Amish" is somehow superior-indeed that is the goal of the labelers. Yet FDA has correctly not seen fit to suggest that labels should say that "FDA sees no significant difference between cheese made in Wisconsin or cheese made elsewhere" or sees "no significant between cheese made by the Amish and other cheese." We urge FDA to rescind the guidance suggesting a special contextual statement on milk from cows not treated with rbGH, and issue no further requirements for contextual statements on genetically engineered products.

There are both health reasons and basic "right-to-know" reasons to require labeling of generically engineered food. There are two health concerns that are relevant: allergens and unexpected effects.

Regarding allergens, while we urge FDA to prohibit introduction of known allergens into genetically engineered food, science is not yet at a point where it can successfully predict all allergens. Indeed, allergies are so varied, that for almost any food, there is probably someone who is allergic to it. Also, there are many food sensitivities (causing indigestion and the like) that do not have the same mechanism as IGE-mediated food allergies. Finally, genetic engineering is introducing many proteins into the food supply that have not previously been eaten, or which have been eaten only in small quantities. Though no one is currently allergic to them, allergies may develop as people are exposed.

The problems of uncommon allergens, food sensitivities, and unknown and new allergens can all best be addressed through mandatory labeling of all genetically engineered food. Mandatory labeling of GE food will allow individuals to identify and avoid food that cause them difficulty. Without such labeling, it will be impossible to distinguish the problem food from its conventional look-alike counterpart.

As discussed, there may also be unexpected effects from genetically engineered food. At this point, we do not know if such effects will occur frequently, rarely or hardly ever, in a way that affects health. However, as discussed earlier in these comments, there is considerable new data showing unexpected effects of genetic transformations. It therefore would be prudent to require mandatory labeling so that if any effects that affected health did occur, they could be identified and their origin determined. For example, tracking a problem such as occurred with the genetically engineered l-tryptophan dietary supplement, which led to illness and death but whose exact cause has never been determined, could be facilitated if there were mandatory labeling.

FDA also has ample precedent for requiring mandatory labeling of genetically engineered food, because it is a "material fact." FDA has generally required labeling of all processes that are of interest to consumers, including whether food is frozen, irradiated, or from concentrate. It requires ingredients and additives to be indicated. It also has many standards of identity whose sole purpose is to facilitate consumer choice. FDA should also require a label if food is genetically engineered, as this is a "material fact" to consumers. A discussion of the legal precedents is appended to these comments.

2. Should additional information be made available to the public about foods derived from bioengineered plants? If so, what information? Who should be responsible for communicating such information?

No comment.

3. How should additional information be made available to the public: e.g., on the Internet, through food information phone lines, on food labels, or by other means?

No comment


Anonymous, 1998. Call for UK genetic food watchdog. Nature online service. Sept. 3.

Bock, S.A. 1987. Prospective appraisal of complaints of adverse reactions to foods in children during the first 3 years of life. Pediatrics, 79: 683-688.

BMA (British Medical Association). 1999. The impact of genetic modification on agriculture, food and health. An interim statement. May, 1999. Board of Science and Education, British Medical Association.

Csordas, A. 1990. On the biological role of histone acetylation. Biochemistry Journal, 265: 23-38.

Doerfler, W. et al. 1997. Integration of foreign DNA and its consequences in mammalian systems. Trends in Biotechnology, 312: 401-406.

Ewen, S.W.B. and A. Pusztai. 1999. Effect of diets containing genetically modified potatoes expressing Galvanthus nivalis lectin on rat small intestine. The Lancet, 354 (9187): 1353-1354.

Finnegan, H. and McElroy. 1994. Transgene inactivation: plants fight back! Bio/Technology, 12: 883-888.

Fox, J.L. 1997. Farmers say Monsanto’s engineered cotton drops bolls. Nature Biotechnology, 15: 1233.

Harbour G.C., Garlick, R.L., Lyse, S.B., Crow, F.W., Tobins, R.H. and J.G. Hoogerheide. 1992. N-‘-acetylation can occur at lysine residues 157, 167, 171 and 180 of recombinant bovine somatotropin. Techniques in Protein Chemistry III: 487-495.

Hill, A.F., Desbruslais, M., Joiner, S., Sidle, K., Gowland, I., Collinge, J., Doey, L.J. and P. Lantos. 1997. The same prion strain causes vCJD and BSE. Nature, 389: 498-501.

Ho, M.W. 1998. Genetic Engineering: Dream or Nightmare? Gateway Books, Bath, United Kingdom.

————-, Ryan, A. and J. Cummins. 1999. The cauliflower mosaic virus promoter-a recipe for disaster? Microbial Ecology in Health and Disease, (in press).

Inose, T. and K. Murata. 1995. Enhanced accumulation of toxic compound in yeast cells having high glycolytic activity: A case study on the safety of genetically engineered yeast. International Journal of Food Science and Technology, 30: 141-146.

Jakowitsch, J., Mette, M.F., van der Winden, J., Matzke, M.A. and A.J.M. Matzke. 1999. Integrated pararetroviral sequences define a unique class of dispersed repetitive DNA in plants. Proceedings of the National Academy of Sciences, 96(23): 13241-13246.

Kuiper, H.A. et al. 1998. Food Safety Evaluation of Genetically Modified Foods as a Basis for Market Introduction. Ministry of Economic Affairs, The Hague, Netherlands.

Lorenz, M.G. and W. Wackernagel. 1994. Bacterial gene transfer by natural genetic transformation in the environment. Microbial Reviews, 58: 563-602.

Maturin, L. and R. Curtiss. 1977. Degradation of DNA by nucleases in intestinal tract of rats. Science, 196: 216-218.

McAllan, A.B. 1982. The fate of nucleic acids in ruminants. Progress in Nutrional Science, 41: 309-317.

Mercer, D.K., Scott, K.P., Bruce-Johnson, W.A., Glover, L.A. and H.J. Flint. 1999. Fate of free DNA and transformation of the oral bacterium Streptococcus gordonii DL1 by plasmid DNA in human saliva. Applied and Environmental Microbiology, 65: 6-10.

Nestle, M. 1996. Allergies to transgenic foods-Questions of policy. The New England Journal of Medicine , 334(11): 726-727.

Nordlee, J.A., Taylor, S.L., Townsend, J.A., Thomas, L.A. and R.K. Bush. 1996. Identification of a brazil-nut allergen in transgenic soybeans. The New England Journal of Medicine , 334(11): 688-692.

Parekh, R.B., Dwek, R.A., Thomas, J.R., Opdenakker, G. and others. 1989a. Cell-type-specific and site-specific N-glycosylation of type I and type II human plasminogen activator. Biochemistry, 28(19): 7644-7662.

Parekh, R.B., Dwek, R.A., Rudd, P.M., Thomas, J.R., and others. 1989b. N-glycosylation and in vitro enzymatic activity of human plasminogen activator expressed in Chinese hamster ovary cells and a murine cell line. Biochemistry, 28(19): 7670-7679.

Reddy, S.A. and T.L. Thomas. 1996. Expression of a cyanobacterial delta 6-desaturase gene results in gamma-linolenic acid production in transgenic plants. Nature Biotechnology, 14: 639-642.

Sampson, H.A., L. Mendelson, and J.P. Rosen. 1992. Fatal and near-fatal anaphylactic reactions to food in children and adolescents. New England J. Med. 327: 380-384.

Schnieke, Harbers, and Jaenisch. 1983. Embryonic lethal mutation in mice induced by retrovirus insertion into the alpha-1(I) collagen gene. Nature, 304: 315-320.

Schubbert, R., Lettmann, C. and W. Doerfler. 1994. Ingested foreign (phage M13) DNA survives transiently in the gastrointestinal tract and enters the bloodstream of mice. Molecules, Genes and Genetics, 242: 495-504.

Scott, M.R., Will, R., Ironside, J., Nguyen, H.B., Tremblay, P., DeArmond, S.J. and S.B. Prusiner. 1999. Compelling transgenetic evidence for transmission of bovine spongiform encephalopathy prions to humans. Proceedings of the National Academy of Sciences, 96(26): 15137-15142.

Spring, S. et al. 1992. Phylogenetic diversity and identification of nonculturable mangeto-tactic bacteria. Systematics and Applied Microbiology, 15: 116-122.

Steffan, R.J. et al. 1988. Recovery of DNA from soils and sediments. Applied Environmental Microbiology, 54: 2908-2915.

Traavik, T. 1998. Too early may be too late: Ecological risks associated with the use of naked DNA as a tool for research, production and therapy. Directorate for Nature Research. Trondheim, Norway.

Van der Krol;. Lenting, and Veenstra et al. 1988. An anti-sense chalcone synthase gene in transgenic plants inhibits flower pigmentation. Nature, 333: 866-869.

Violand, B.N., Schlittler, M.R., Lawson, C.Q., Kane, J.F., Siegel, N.R., Smith, C.E., Kolodziej, E.W. and K.L. Duffin. 1994. Isolation of Escherichia coli synthesized recombinant eukaryotic proteins that contain e-N-acetyllysine. Protein Science, 3: 1089-1097.

Zhang, Q. 1999. Meeting the challenges of food production: the opportunities of agricultural biotechnology in China. Paper presented at "Ensuring Food Security, Protecting the Environment, and Reducing Poverty in Developing Countries: Can Biotechnology Help? An Inernational Conference on Biotechnology" held at The World Bank, Washington, D.C. October 21-22, 1999.


* Consumers Union is a nonprofit membership organization chartered in 1936 under the laws of the State of New York to provide consumers with information, education and counsel about goods, services, health, and personal finances; and to initiate and cooperate with individual and group efforts to maintain and enhance the quality of life for consumers. Consumers Union’s income is derived solely from the sale of Consumer Reports, its other publications and from noncommercial contributions, grants and fees. In addition to reports on Consumers Union’s own product testing, Consumer Reports, with approximately 4.5 million paid circulation, regularly carries articles on health, product safety, marketplace economics and legislative, judicial and regulatory actions which affect consumer welfare. Consumers Union’s publications carry no advertising and receive no commercial support.