An example of research into this problem is available at http://bric.postech.ac.kr/science/97now/01_12now/011219c.html. An advisory committee to Health Canada has submitted proposals (http://www.hc-sc.gc.ca/vetdrugs-medsvet/amr/e_policy_dev.html#AMR Report ) for reducing the risk of overuse. "The Challenge of Antibiotic Resistance", an article by Levy (1998) in Scientific American, discusses many facets of antibiotic resistance.

At several stages of the laboratory process, developers of transgenic crops use DNA that codes for resistance to certain antibiotics, and this DNA often becomes a permanent feature of the final product although it serves no purpose beyond the laboratory stage. Will transgenic foods contribute to the existing problems with antibiotic resistance?

One of the concerns is related to the risk of horizontal gene transfer, that is, transfer of DNA from one organism to another outside of the parent-to-offspring channel. Transfer of a resistance gene from transgenic food to micro-organisms that normally inhabit our mouth, stomach, and intestines, or to bacteria that we ingest along with food, could help those micro-organisms to survive an oral dose of antibiotic medicine.

Horizontal transfer of DNA does occur under some natural circumstances. Small pieces of DNA called plasmids move from E. coli to several other organisms, and Agrobacterium tumefaciens facilitates the horizontal transfer of DNA to produce the disease known as crown gall in plants. Horizontal transfer of DNA occurs at very low frequency under laboratory conditions, for example Acinetobacter, a soil- and water-borne bacterium (Gebhard and Smalla, 1998), Streptococcus gordonii, a cause of dental cavities and heart valve infection (Mercer et al., 1999), Aspergillus niger, a fungus harnessed to produce citric acid for soft drinks (Hoffmann et al., 1994).

Whether such a transfer could happen frequently in the special environment of the human gastrointestinal tract is not known.

Several circumstances would work against a successful transfer. The acid environment of the stomach degrades DNA. Mercer et al. (1999) found that DNA was degraded within 30 seconds when it was mixed with human saliva and hydrochloric acid in a simulation of the conditions in the human stomach. And while some organisms take in foreign DNA under some circumstances, many organisms have mechanisms to destroy foreign DNA that enters unbidden.

Researchers at the University of Newcastle have reported that microorganisms in the human digestive system took up a herbicide resistance gene after the human subjects ate a meal of GM soy. The experiment was small, involving 12 people with intact digestive systems and 7 who had undergone colostomies, in which their lower intestines were removed. In people with intact digestive systems, no GM DNA was found in the stools and no microorganisms took up GM DNA. But in people who had undergone colostomies, about 4 percent of the GM DNA survived the trip through the abbreviated intestinal tract and a small number of microorganisms took up GM DNA. The GM DNA in this experiment was a herbicide resistance gene rather than an antibiotic resistance gene, but the results suggest that horizontal transfer of GM DNA can occur in the human digestive tract under some circumstances. News reports on this experiment are available at http://www.newscientist.com/news/news.jsp?id=ns99992565 and http://www.guardian.co.uk/gmdebate/Story/0,2763,756666,00.html. A report from the researchers at the University of Newcastle is available at http://www.food.gov.uk/multimedia/pdfs/gmnewcastlereport.PDF. The information pertaining to this experiment begins on page 22 of the report.

Some scientists think that a successful transfer of DNA from transgenic plants would not occur frequently enough to be detectable or to cause health problems in humans. However the actual rate of transfer under natural conditions is unknown because of a lack of research into this particular circumstance. The EU has engaged scientists to assess the risk of horizontal gene transfer from genetically modified foods (http://europa.eu.int/comm/research/quality-of-life/gmo/04-food/04-07-project.html).

A second concern about the use of antibiotic resistance genes is that some (though not all) resistance genes work by producing an enzyme that destroys or inactivates the targeted antibiotic. If such a resistance gene continues to function after it has outlived its usefulness in the laboratory, it might produce low levels of the resistance enzyme in the plant parts that we eat. While high processing temperatures would inactivate the enzyme in processed foods, ingestion of fresh or raw transgenic foods could result in the stomach containing a small amount of an enzyme that inactivates an orally administered dose of the antibiotic. Will a person taking prescribed antibiotics endanger the success of the treatment if he eats a meal of transgenic food?

The FDA faced this issue in 1993 when it considered whether to approve Calgene's FlavrSavr tomato, which contained a gene for resistance to kanamycin, neomycin, gentamicin A and gentamicin B, all of which are sometimes used to fight infections in humans. In FDA tests using a simulation of a human stomach, the enzyme was degraded by stomach acids before it could attack orally administered antibiotics. A report on the FDA's evaluation is available at http://vm.cfsan.fda.gov/~dms/opa-armg.html. Scroll down to "Appendix 1. Evaluation of the safety of the kanamycin resistance gene as a selectable marker" for a discussion of the FDA's findings.

Ciba-Geigy's Bt corn 176 also contained a gene for resistance to an antibiotic used in the treatment of human diseases. Ampicillin, a member of the penicillin family, is one of the most widely prescribed antibiotics, often used to fight infections of the ear, bladder, and kidney. In Bt corn 176 the ampicillin resistance gene was constructed so that it was active only in prokaryotic organisms, such as bacteria. This feature made it useful for early stages of development of the crop. The gene was not active in eukaryotic organisms, or "higher" plants, such as corn. Tests showed that no ampicillin-inactivating protein was produced in corn plants that contained the gene, so the presence of the resistance gene was not viewed as a problem when Bt corn 176 went through the approval process in 1995 in Canada and the United States. Details of the transgenes in Bt 176 and the factors considered in approving the use of the ampicillin resistance gene are available from the USDA (http://www.aphis.usda.gov/biotech/dec_docs/9431901p_det.HTM) for the U.S. process and from the Canadian Food Inspection Agency (http://inspection.gc.ca/english/plaveg/pbo/dd/dd9609e.shtml) for the Canadian process. (Bt 176 was discontinued in 2001 for reasons unrelated to the ampicillin resistance gene. See our Discontinued Products page for more information.)

For additional information on the potential risk from antibiotic resistance genes, see the Royal Society's 2002 report "Genetically modified plants for food use and human health--an update" (http://www.royalsoc.ac.uk/files/statfiles/document-165.pdf).

While the risks from antibiotic resistance genes in transgenic plants appear to be low, steps are being taken to reduce the risk and to phase out their use.

• The FDA recommends that developers of transgenic crops use antibiotics that are not commonly used for treatment of diseases in humans. Thus, if horizontal gene transfer does occur, micro-organisms in the body are not likely to have acquired resistance to the antibiotics that a doctor might prescribe to fight infection.
• Scientists are changing their development methods. Other marker genes, such as green fluorescent protein, or mannose (Joersbo et al., 1998), may be able to do the job that antibiotic resistance markers have done.
• Scientists are also experimenting with methods for removing the antibiotic resistance genes before the plants are released for commercial use (Dale and Ow, 1991; Ebinuma et al., 1997; Iamtham and Day, 2000; Zuo et al., 2001), so that these genes can be used during development and then eliminated from the final product.
• European scientists have developed a method for inactivating the antibiotic resistance genes in the event that they are transferred to bacteria in the digestive system of animals or humans (Libiakova et al., 2001). A special DNA sequence inserted into the antibiotic resistance gene will prevent the gene from functioning inside a bacterium. Plants are able to snip out the special sequence to let the gene function correctly.

The possibility of enhanced antibiotic resistance in the soil is also an issue of interest in connection with the use of these markers. See our discussion for more information.