Science 26 November 1999:
Vol. 286 no. 5445 pp. 1662-1666
Risks and Benefits: GM Crops in the Cross Hairs
1. Dan Ferber
1. Dan Ferber is a writer in Urbana, Illinois.
As controversy builds over the safety of genetically modified crops, the evidence so far hasn't
pinpointed any specific problems—but also can't dispel the doubts
An American entomologist publishes a study in Nature showing that pollen from genetically modified
corn kills monarch butterfly larvae; two of his colleagues denounce him in a commentary for publishing
―preliminary results‖ and imply that he is spreading rumors. A British food-safety expert writes
in Nature that a concept that underlies regulation of genetically modified food in most of the developed
world is ―pseudoscientific‖; an opponent fires back in a letter calling his commentary ―a mish-mash of
old-hat sociology and poor science.‖ A British scientist announces on television that genetically modified
potatoes stunt the growth of rats and damage their immune system; his supervisors suspend him 2 days
later. It would be hard to find a scientific debate more polarized than the one now being waged about the
safety of genetically modified (GM) crops.
But while biotech opponents talk of Frankenfoods and terminator genes and industry groups minimize
safety concerns, a small group of investigators has been taking a serious look at GM crops to see what
health and environmental risks they might pose. What they are finding is in many cases reassuring—but
not always. The plants, most of which have been modified to resist pests or weed-killing herbicides, seem
to pose minimal risks to human health, say experts. But environmental concerns such as the possibility
that the novel genes might spread to wild plants and produce new strains of weeds, although hard to
substantiate, are also proving hard to dispel.
Complicating the weighing of risk is the question of how much any potential hazards are offset by the
crops' potential benefits, such as reducing the use of chemical pesticides, lowering costs, and improving
nutritional value. Part of the problem is that, unlike drugs or pesticides, plants have never been subjected
to a risk analysis, says plant pathologist James Cook of Washington State University in Pullman, who
chaired an international panel to devise risk assessment methods for GM crops. ―We have to ask what are
the safety issues raised by plants, then apply that to crop plants with transgenes,‖ he says.
Food-safety concerns have stirred the most passionate debate among the public, prompting boycotts, bans,
and protests. But few accept the conclusions of the report that sparked the furor over GM potatoes in
Britain (Science, 22 October, p. 656). And there's little other research that might raise concerns that the
transgenic crops now on the market threaten human health. ―There's something wrong with the perception
of risk here,‖ says microbiologist Abigail Salyers of the University of Illinois, Urbana-Champaign.
For example, GM food critics worry about plant genetic engineers' practice of attaching the genes they are
trying to introduce into plants to an antibiotic-resistance gene. They can then readily select those plants
that have acquired the genes by treating them with the antibiotic, which kills any nonresistant plants. The
critics charge that the antibiotic-resistance genes, which sometimes remain in the transgenic crops, could
spread to pathogens in the body and make antibiotics less effective. But several panels of antibiotic-
resistance experts have concluded otherwise. ―Unanimously, the verdict has been that the chance of
antibiotic-resistance genes getting into intestinal bacteria is minuscule,‖ Salyers says. And if they did get in, ―the virtually unanimous verdict is that it wouldn't matter‖ because the same resistance genes are
already present in many of the bugs.
A more plausible—though still unlikely—threat to human health from transgenic foods comes from food
allergies. An allergic reaction to food can be serious, even life-threatening, if it leads to anaphylactic
shock. ―That's one you certainly want to worry about,‖ says food microbiologist Bruce Chassy of the
University of Illinois, Urbana-Champaign, a former food-safety adviser to the U.S. Food and Drug
Administration. Indeed, in a study reported in 1996 in The New England Journal of Medicine, Steve
Taylor and his colleagues at the University of Nebraska, Lincoln, showed that people allergic to Brazil
nuts are also allergic to soybeans that have been engineered to express a Brazil nut protein to make them
To Chassy, the outcome was reassuring: The results led the producer of the transgenic bean, Pioneer Hi-
Bred International, to discontinue the soy line voluntarily before it was commercialized. What's more, the
producers of GM foods screen their products for allergenicity, he says. Among other methods, they can
check to see if the amino acid sequences of the proteins made by the genes they put into crop plants
resemble those of known food allergens.
Critics say that because many proteins that trigger allergic reactions have not yet been sequenced, the
sequence comparison test will fail to detect some allergens. ―If you find a match, then you have a
problem,‖ says Rebecca Goldburg, senior scientist at the Environmental Defense Fund in New York City.
―If you don't, it doesn't say anything.‖ But Chassy notes that conventional foods already on the market,
such as peanuts and Brazil nuts, pose much higher risks of allergies than GM foods, as do plants produced
by classical breeding methods, which introduce many potential allergens into the product. ―If a zero-risk
standard prevails, we shouldn't put any new food on the market and we should get rid of a lot of old
ones,‖ he says.
Bt or not Bt?
It's the potential environmental effects of GM crops that stir deeper scientific debates, as was evident at a
recent meeting, held near Chicago on 2 November, that examined whether pollen from so-called ―Bt
corn‖—corn containing an insecticidal protein from the bacterium Bacillus thuringiensis—could harm
monarch butterflies in the field. The colorful monarch became the poster child for the anti-GM movement
last May, after entomologist John Losey and colleagues at Cornell University published a short laboratory
study in Nature showing that Bt corn pollen could kill monarch butterfly caterpillars in the laboratory.
View this table:
In this window
In a new window
Out of a group of caterpillars that had munched on milkweed leaves—the larvae's only food source—
dusted with Bt corn pollen, 44% died within 4 days, while larvae eating leaves dusted with ordinary
pollen all survived. Although perhaps not surprising—researchers had known for years that Bt bacteria,
which are themselves widely used as pesticidal sprays, could harm a variety of butterflies and moths—the
Losey study was the first one published showing that a Bt plant could directly harm a nontarget butterfly.
The study attracted widespread media attention and alarmed biotech observers throughout the world. The
resulting hubbub caused European regulators to place a moratorium on the approval of additional Bt crops
and prompted jitters among biotech investors. Even so, it was not clear whether monarchs outside the lab,
developing on milkweed plants growing near fields of Bt corn, are in fact in harm's way. To find out,
major biotechnology companies, including Monsanto Co., Novartis Seeds Inc., and AgrEvo USA, formed an unusual consortium called the Agricultural Biotechnology Stewardship Working Group (ABSWG) that
dispensed $100,000 to eight researchers at U.S. and Canadian universities to conduct further studies. The
U.S. Department of Agriculture (USDA) and individual universities funded an additional 12.
The study participants presented their early results at the Illinois meeting, where they did agree on one
point. ―The worst case scenario of a toxic pollen cloud saturating the Corn Belt and wiping out all the
Lepidoptera‖ is clearly not going to happen, says entomologist Stuart Weiss of Stanford University. But
Bt pollen might still have less dramatic harmful effects—although a press release put out by the ABSWG
early in the meeting suggested otherwise (see sidebar on p. 1663).
At issue in the new monarch studies is just how far pollen might drift from cornfields, and how toxic it
might be to any monarch larvae that eat it. The potential for harm is certainly there, says entomologist
John Obrycki of Iowa State University in Ames. Because corn pollen is relatively heavy, it is likely to
settle near cornfields. What's more, Obrycki says, ―we do find lots of milkweeds growing near cornfields,
and they are being used by monarchs.‖
At a meeting of entomologists held last March, he and graduate student Laura Hanson reported results
suggesting that enough pollen might collect on nearby milkweeds to harm the larvae. The work, which
has not yet been published, showed that about 20% of monarchs that fed for 2 days on potted milkweed
plants left at the edge of Bt cornfields died, compared to 3% of monarchs that fed on plants left near non-
Bt cornfields. At the Illinois meeting, the industry- and USDA-funded researchers presented similar
results, but their interpretation was more optimistic. In one study, for example, botanist John Pleasants of
Iowa State and entomologist Richard Hellmich of the USDA's Agricultural Research Service and Iowa
State first determined the levels of Bt pollen that are toxic to monarch larvae, then measured how much
pollen they could trap on sticky slides left near cornfields.
The results showed that even milkweeds within 1 meter of the cornfield were unlikely to be dusted with
toxic levels of Bt pollen from two of the most widely planted corn varieties, AgrEvo's CBH 351 and
Monsanto's Mon810, Hellmich said at the meeting. The researchers did find that pollen from the same
line that Obrycki tested, Novartis Seeds' 176, is sufficiently toxic to threaten monarchs feeding on
milkweeds up to 2 meters away, thus confirming Obrycki's results. But the more toxic Bt line represents
just 2.5% of the corn planted in the United States. Overall, Hellmich says, his team's results and similar
data presented by others at the meeting revealed a minimal risk to the monarch. ―A lot of the data
presented were overwhelmingly positive,‖ he concludes.
Still, monarch experts were not entirely reassured. One problem, says insect ecologist Orley Taylor of the
University of Kansas, Lawrence, who directs the conservation group Monarch Watch, is that even if Bt
exposure doesn't kill monarchs, it could make them less fit for their long migration to Mexico, where they
overwinter en masse. At the meeting, Taylor presented a model, based on current Bt corn acreage and the
butterfly's migration patterns and reproductive behavior, that predicted a worst case scenario in which 7%
of the North American monarch population would die. Although the real effect would undoubtedly be
less, he says, ―there's plenty of indication that there's going to be an impact. It's a matter of degree.‖
Bt toxins might also threaten beneficial insects indirectly, by entering the food chain. For example, in
work published in 1998 and 1999, Angelika Hilbeck and her team at the Swiss Federal Research Station
for Agroecology and Agriculture in Zurich, Switzerland, showed that green lacewing caterpillars—a
beneficial pest-eating predator—were more likely to die when they ate European corn borer caterpillars
that had fed on Bt corn than when the borers had fed on non-Bt corn. ―It's interesting science because of
what it says about the toxicology of Bt,‖ says entomologist Richard Roush of the University of Adelaide
in Adelaide, Australia. But ―a lot of us wonder whether it's really important in the field.‖ He and others note that Bt bacteria have been sprayed on farm fields for 3 decades, and that earlier studies had shown
that beneficial predator insects were unaffected.
Hilbeck argues, however, that because the toxin is expressed at high levels throughout GM crop plants,
rather than just sprayed on their surfaces, plant-eating insects could receive a much bigger dose. She has
begun field trials, and she says others should monitor the effects of Bt crops on lacewings and other
insect-eating predators before a problem develops. ―Anything is possible,‖ she says. ―There may be no
effect, but there may be a slow and chronic effect on green lacewing larvae. Then you might find,
‗Whoops, where did all the lacewings go?‘‖
Flowing genes and superweeds?
Ecologists also worry that genes such as those conferring resistance to herbicides or insect pests might
pass from the crops into wild relatives and create so-called superweeds—invasive plants with the
potential to lower crop yields and disrupt natural ecosystems. They note that a variety of crops, including
canola, squash, sunflower, and sorghum, can outcross with weedy relatives growing nearby.
Plant geneticist Val Giddings, a spokesperson for the Biotechnology Industry Organization, downplays
the risk, saying that even if such outcrossing allowed a weed to pick up a gene, it would not persist for
long in the wild. A herbicide-resistance gene, for example, would disappear from weeds outside the
confines of farm fields because there would be no herbicide to select for plants containing it. ―There is
abundant literature that demonstrates that in the absence of selection pressure, a neutral trait will be lost
over time,‖ Giddings says.
Sometimes, but not always, answers plant ecologist Allison Snow of Ohio State University, Columbus. In
a study published in April in Molecular Ecology, Snow, with Rikke Jørgensen and Bente Andersen of the
Risø National Laboratory in Roskilde, Denmark, crossed canola plants carrying the gene that encodes
resistance to the herbicide glufosinate with a weedy relative called field mustard. They found that the
gene persisted in the weed even when no herbicide was applied. What's more, the weed produced equally
fit offspring whether or not it had the herbicide-resistance gene. That means that the gene will probably
stick around, Snow says.
Another type of gene that might move to weeds are virus-resistance genes, such as those that have been
engineered into yellow squash and zucchini, says Alison Power of Cornell University. If populations of
the weedy relatives of these crops are kept in check by viruses, a virus-resistant weedy squash could
potentially outgrow ordinary plants and become more aggressive. Researchers won't know until someone
does field tests to find out, however. ―It could be a significant issue,‖ Power says, but ―we don't have
good information to go on.‖
Benefits of biotech
The backers of GM crops say that all this talk of their potential risks overlooks their benefits to
consumers, farmers, and the environment. But although the risks remain hypothetical, it's also too early to
tell whether GM crops are a proven boon, because only a few independent studies have been conducted,
and those show clear benefits for some crops but not for others, agriculture experts say.
Cotton, for example, is notorious for needing heavy doses of pesticides, so Bt cotton should offer
substantial savings and environmental benefits. Indeed, by planting modified rather than conventional
cotton on 2.3 million U.S. acres (nearly 1 million hectares) in 1998, farmers reduced chemical pesticide
use by over a million pounds (450,000 kilograms), according to a report released earlier this year by
Leonard Gianessi and Janet Carpenter of the National Center for Food and Agriculture Policy, a think tank in Washington, D.C., that is funded by industry and the USDA. What's more, cotton farmers
increased their yields by 85 million pounds (39 million kilograms) and made $92 million dollars more
than farmers who did not use the technology.
The report says, however, that not all Bt crops fared as well. Although 14 million acres (5.7 million
hectares) of U.S. cornfields—about one-fifth of the total corn acreage in the United States—were planted
with Bt corn in 1998, the increased profits from higher corn yields did not cover the extra cost of the Bt
corn seed. In addition, the Bt crop saved only 2 million of those acres (800,000 hectares) from chemical
insecticides because most farmers don't bother to spray for corn borers because spraying often doesn't
protect the corn. Researchers also worry that pest insects could develop resistance to the Bt toxins over
the next several years because the bacteria is now so widespread. That would make Bt sprays ineffective,
eliminating one of the few effective pest-control strategies available to organic farmers, who forswear
Another recent report takes a look at the pros and cons of Roundup Ready soybeans—a herbicide-
resistant line from Monsanto—and concludes that the results were mixed. On the plus side, says report
author Charles Benbrook, an independent consultant to consumer and environmental groups in Sand
Point, Idaho, and a former executive director of the National Research Council's Board on Agriculture,
Roundup Ready soybeans allow farmers to substitute Roundup for more hazardous and long-lasting
herbicides like acetochlor. And they reduce the need for farmers to till the soil to ward off weeds, which
reduces soil erosion.
But Benbrook's findings did not support industry claims that the Roundup Ready beans reduce herbicide
use by allowing farmers to kill weeds with one dose of Roundup after the soybean plants have sprouted
instead of dosing the fields with a variety of herbicides before and during the growing season. Instead, the
Benbrook reported concluded, farmers applied two to five times more herbicides of all kinds to their GM
soybean fields than to fields growing conventional soybeans. And in contrast to industry claims, a recent
study by agricultural economist Michael Duffy of Iowa State University showed that Roundup Ready
beans made Iowa soybean farmers no more money than farmers growing ordinary beans. Despite the
increased herbicide usage, applications costs were lower, but so were yields from the GM soybeans. ―You
had lower income and lower costs, so it was kind of a wash,‖ Duffy says.
Even if the technology has yielded few clear-cut benefits in the developed world, agbiotech backers say
that in the developing world, new crops in the pipeline could improve yields for farmers and make
tremendous strides toward reducing malnutrition and environmental degradation. A genetically
engineered line of rice reported earlier this year, for example, can make more vitamin A precursor and
accumulate more iron, which could prevent infections, blindness, and anemia in people in the developing
world (Science, 13 August, p. 994). Other researchers are developing plant-based vaccines to prevent
diarrheal and other diseases in the developing world, says plant biochemist Charles Arntzen, president of
the Boyce-Thompson Institute for Plant Research in Ithaca, New York.
And a Cornell group is engineering a virus-resistant papaya plant that could save crops in Brazil, Puerto
Rico, and Jamaica. A version of the plant, which resists the papaya ringspot virus, has already revived
Hawaii's papaya groves, devastated by the virus in the mid-1990s, says plant pathologist Dennis
Gonsalves, who leads the effort. ―You should go back and look now—it's beautiful,‖ he says.
But before farmers sow GM crops around the world, researchers and regulators need to do a better job
assessing the ecological risks, says Ohio State's Snow: ―We shouldn't just be waving our hands. There
really are not enough ecologists doing this research,‖ in part because research funds are scarce. And even
biotech backers acknowledge the need for better data. ―I would say that the benefits totally outweigh the
risks, but we can't ignore the risks,‖ Washington State's Cook says Science 17 December 1999:
Vol. 286 no. 5448 p. 2243
CONTROVERSY OF THE YEAR
GM Foods Under Attack
The debate over genetically modified (GM) foods exploded in 1999, becoming a worldwide public
relations disaster for the biotech industry and casting science in the role of villain. Although most fiery in
the United Kingdom, where headlines warned of ―the horrors of GM foods‖ and ―the mad forces of
genetic darkness,‖ the public fervor spread through Europe, leading the European Union to suspend the
introduction of new GM cropspending new legislation, which could be 3 years away (Science, 26
November, pp. 1662-1668).
The effects even reverberated in the United States, where the GM revolution had been proceeding all but
silently. U.S. farmers, who planted roughly half of their corn, cotton, and soy fields with
transgenic crops this year, watched with dismay as their export markets shrank. And at recent public
hearings organized by the Food and Drug Administration, many speakers voiced concerns that
the crops—often tailored to resist insects or herbicides—might be hazardous to human health or could
cross-pollinate with wild plants and create ―superweeds.‖
This eruption of public feeling w