I don't like the word "genetically modified food." Virtually all of our foods have been genetically modified. Take the apple, for example. There are literally dozens of varieties of apples. How did we get those dozens of varieties? We genetically modified the apple through conventional breeding. We crossed one kind of apple with another apple, and we produced very different apples--different color, different flavor, different functions. So almost all of our foods have been genetically modified in some way. What's different now is that we have some new techniques to do it. I usually like the idea of genetic engineering or recombinant technology. All of our foods, just about, are modified genetically.

Some people say that traditional crossbreeding is natural, and therefore, is inherently safer.

			A professor of food science and toxicology at Cornell University, Hotchkiss compares genetic engineering to traditional crossbreeding of plants, explains how science evaluates risk and how toxicity is tested. He also discusses why allergenicity is the most difficult risk issue with foods, why the current U.S. regulatory framework is adequate, and the problems that would arise if mandatory labelling is introduced. (Interview conducted September 2000.)

That's kind of what you think because they've been around a long time. But there are lots of examples, for example, the potato, which, by the way, comes from South America. It was originally purple. We bred out the purple because we like white. The potato contains a naturally occurring chemical that's quite toxic, called glycoalkaloid. Those glycoalkaloids in some potatoes, as a matter of fact, have caused severe human poisonings or near-death. When you breed potatoes, it's possible to breed high levels of that toxicant into a potato. As a matter of fact, there are a number of breeds of potatoes that have high levels. Fortunately, they did not make the marketplace--for that reason.

Another example of the risks of traditional breeding is celery. Celery naturally contains a photoactive toxicant, that is, a chemical that becomes toxic when it hits sunlight. There was a case in California where a new variety of celery was bred that, unknown to the people who bred it, had high levels of this toxicant in it. It was planted. People went along, harvested this, and the workers who harvested this came out with a very severe skin rash. Why? Because it had the high level of toxicant resulting from the commercial, normal kind of breeding. So the normal kind of breeding can produce risks, just as any other genetic or other kinds of breeding can produce risks.

The way the FDA regulates food, people haven't had to do all the research in advance and prove that it's safe before they release it.

Yes. In the examples of traditional breeding, that's correct. FDA has not required pre-market approval. In other cases, FDA does require pre-market approval--additives in foods, for example, or colors or other things that are added.

They require pre-market approval?

For conventionally bred products, no, FDA has not required pre-market approval. The examples I gave you might suggest that maybe they should, but they haven't. In other cases, FDA does require pre-market approval, particularly for things added directly to foods. In the area of genetically engineered food crops, FDA has required certain submissions, and has changed its policy as it has learned, like the rest of us, more and more about this technology and how it's applied. The FDA's responsibility is to assure safety, and if pre-market approval is required in their opinion to do that, then they will require that.

What does "food safety" mean?

The term "food safety" really, in my view, is a misnomer, a strong misnomer. Safety implies a non-event. You say something's safe, nothing bad happens. But it is impossible, particularly for science, to prove non-events. There's no way that anyone can prove a thing will not happen. So science deals in risks. Science looks at the size or the probability that a risk may occur. It's impossible for science to go around and assure people that things are safe, simply because science cannot address non-events.

So you look at risks. What is the risk of conventionally bred celery? What is the risk of breeding celery through genetic modification? That's the way science looks at it--that there is always some size of risk. Then it's really up to society to gauge whether that risk is acceptable or not.

What kinds of things might go wrong with a tomato?

Tomato is a great example, because up until the middle 1800s, a tomato was thought to be quite toxic, because it belongs to a plant family that is quite toxic in itself. People didn't eat tomatoes. So it's possible that, in breeding a tomato by whatever technique, you could introduce a gene which produces more of that toxic substance. FDA would be very interested in that.

Tomato is also very important because it's a mildly acidic food. If you make it less acidic, you change the way it must be processed and you have an increase of botulism in a tomato. That's another thing FDA would be concerned with. So there's quite a long list of things that FDA would want to know about that tomato.

No matter how that tomato was made?

However it was made, yes. As a matter of fact, I think you could logically argue that the technique of genetic engineering produces a much more specific change in the tomato, where breeding produces a whole range of unknown changes. I use the example of anybody who is a parent--you know that you cannot predict what the offspring from a mating will be. It's true of the tomato as well.

Can you test food for toxicity?

The conventional toxicology is to take a group of animals--most often rats, because we know a lot about them and they're cheap--and feed them very high levels of the chemical of interest. You can do that. You can produce 5 percent or 10 percent of an animal's diet as that single chemical, and get some idea about the toxicity. You cannot do that with whole foods, because the chemicals that make up whole foods are all in relatively small amounts.

So if a tomato had a toxic component to it, if you fed nothing but tomato to a rat, that rat would get very ill, simply from eating nothing but tomatoes, because you can't live exclusively off tomatoes. You can't concentrate a tomato enough to make it significant in a rat toxicity assay. So it is just not possible to test whole foods. The outcome, scientifically, would just have very little meaning.

So you must understand what makes up a tomato. What are the components of a tomato? What are the components of concern? Are those components toxic? And what are the levels in the whole tomato? That's really the approach that FDA and those interested in the safety of this new technology use. . . .

An early rat study with potatoes ran into some of these problems.

Exactly. You cannot test whole foods on animals. There are statistical reasons as well. In the U.S., there are 265 million residents. In order to assure safety, you'd want a very high level of probability. How many rats would it take you to assure yourself that you're going to protect all 265 million people? It would take a huge study, because the sensitivity of the study is very low.

What about allergenicity?

Allergenicity is a very difficult issue and, in my view, the most difficult risk issue with foods, whether it comes from conventional or recombinant kinds of technologies, simply because allergenicity is very, very difficult to predict. Probably the most allergenic food is peanuts. There's a protein in peanuts that is a very serious allergen for a very, very small portion of the population. It's very difficult to find out who that population is, unless they've had a very bad accident or episode with this peanut protein. Even if you did do that, you wouldn't need to test the peanut protein on them, because you already know. What does that mean in terms of allergenicity or safety? It's very difficult.

The approach is certainly that you would be very suspicious, and probably would not use peanuts as a source of genetic material for transferring into any other food source, because you would be concerned about the issue of allergenicity of peanuts.

There was a case that was picked up like that, wasn't there?

Exactly. There was a case that occurred before it demonstrated severe allergenicity in the population, as I understand it. But nonetheless, I think that case demonstrated that you would not want to use peanuts for such a source of genetic material.

From a regulatory standpoint, how safe is the U.S. food supply?

. . . Of course, we want in general to think that we have no risk with our foods, because of the special place that foods plays in our society. So we want to reduce that risk, and people are all the time trying to reduce those risks. Scientists interested in food safety typically weigh the risks, the size of the risk, and logically start at the largest risk and kind of work their way down. These days, most food safety experts think that the largest risk is due to disease-causing bacteria associated with foods.

From a regulatory standpoint, compare crops produced by traditional crossbreeding versus recombinant methods. What's the same, and what's different?

Genetically produced crops are much more closely regulated, and there are some differences. One difference is what's called the vector--the carrier for the new genetic material that comes into the new plant or animal source. That's different than conventional breeding, and the safety of that vector must be carefully regulated.

Another difference is the selection process. Typically, when you change the genetic material of a plant or animal or bacteria, whether it's by conventional crossbreeding or by recombinant DNA technology, you have to select out organisms. Typically, for recombinant DNA, that's done by putting in an antibiotic resistance marker. You select by resistance to a particular antibody. That's different, and that's a very serious issue. It's an issue that requires close scrutiny and close regulation.

So there are differences, and those differences have led the Food and Drug Administration and other agencies around the world to take a much closer look at recombinant technology compared to conventional breeding.

What about using transgenes?

That's different. In conventional breeding, you cannot cross species. In recombinant technology, it's very possible to cross species.

Some things are new with this technology. But our regulatory agencies have argued that we can handle them within the existing framework.

My own view is that the current laws and powers given to FDA, and, to some extent to USDA, are more than adequate to address the issues. Those laws basically say that foods must be safe; that it has been interpreted as low risk; that it must not be fraudulent in any way; and that certain information must be provided to consumers. I think that those laws give the regulatory agencies ample authority to ensure that in fact the new technologies meet those standards--the same standards that are applied to conventionally bred products or food additives or color additives or all the other things that are in our foods.

Critics say the agencies have been either asleep at the wheel or cheerleaders for this technology, and have rushed ahead too quickly. What do you feel?

That's a common criticism of anything people don't like--that somebody should step in. I think what people have to realize is that the Food and Drug Administration operates under a very specific law, or set of laws, that tell it exactly what it cannot and can do. It has a very long judicial history of law. FDA has to operate within the laws, and I have not seen anybody challenge them successfully to say that they are not operating under those laws.

What does that mean in relation to this? Is FDA doing all that it can? Because of the law, they can't do more?

Certainly, you can always do more, I suppose. But what it really says to me is that FDA is exercising its authority to ensure low risk or safe foods. It has done so with the genetic technology just as it has done with hundreds of other technologies, going back to pasteurization of milk. It's a new technology for them. They're likely to stumble and make a few mistakes along the way, because they don't understand the technology either. They have to learn the technology. It's expanding very rapidly. But I think their hearts are in the right place with it. I think the laws are adequate.

You don't think they're giving this technology an easy ride?

No, I don't think so. . . . And I think criticism of FDA has come from both consumers and industry.

Why can't genetically modified food be labeled?

There are two sides to that argument. One says, "Listen, people have the right to know what's in their food. And if they want to know if it's from genetically modified sources, then they have a right to know that." The other side of that argument is, "Listen, all of our food's been genetically modified. It's been processed. It's likely to have pesticides added to it. It may have come from outside the country. It may have gone through several processes. Where do you stop? Where should we put things on the label, and when should we not put things on the label?"

I think FDA basically looks back at their laws and says, "What does our law require on the label?" In my view, I'm not so sure that the current laws tell FDA that they have to put anything more than things to do with the safety of food, or things to do with economic protection, fraud, or providing nutrition information. A specific law tells FDA that foods must have nutrition information. So within their current laws, it's hard for me to see that they really have the authority to make that label mandatory.

What would be the difference between voluntary and mandatory labeling?

Voluntary labeling is certainly an option for the industry, and one they can exercise.

Can you show us the kind of labeling?

There's a lot on the label on a food product. . . . The majority of what the company puts on a label is voluntary. But some of the labeling, in terms of things like ingredients, nutritional facts, manufacturer, is all required by law. So there are things that are required by law, and there are things that are not required by law. Some things are optional, and are not required by law. FDA will require, though, that whatever is put on this label, whether it's optional or required, be truthful.

So if you want to regulate, or if you want to label and say "From a GMO," or whatever, you must have a way of assuring the consumer that this is GMO food, or from a GMO, or whatever your label says. That probably is one of the difficult issues in voluntary labeling that FDA would have to address. How can you assure the consumer that you're telling them the truth?

Would labeling cost a lot of money?

Labeling always costs a lot of money. It was argued, for example, when nutritional labeling came in, that it would cost the industry billions of dollars. That may or may not be true. On the other hand, if the consumer wants that, those prices are always passed on to the consumer. If the consumer demands that, then I say go ahead and label it, if that's really the case. The consumer will just have to bear the burden of that labeling cost.

The FDA is a medically based agency. If this technology moves from input traits to output traits--for example, oils beneficial to the heart--does the FDA have responsibility to promote that? Or is its duty just to ensure that claims are honest?

The FDA is part of the Public Health Service, and says that they are part of promoting good health. But in my view, in reality, they are more of a regulatory agency, and they have greatest responsibility for preventing unhealthy situations, to label products. But it is really not FDA's responsibility to tell people what they should and should not eat, or what's good for them and what's not good for them.

The line between foods and drugs is blurring, and has blurred over the last decade, and is likely to blur even more, and present FDA with new challenges in that area. FDA will be faced with the issue of allowing the industry to put forth foods that seem to have good health benefits, and finding a way to make those honest claims, while still allowing them access to the consumer.

There are three agencies. Do you want to say a couple of words about the others?

You have to understand that FDA does not have responsibility for environmental risks. That's primarily with EPA. There are three risks with any kind of new technology, including genetically modification. There's a food risk--that's what we mostly are talking about. There's an environmental risk. Those issues, particularly as they revolve around the monarch butterfly and ecology, lie primarily with EPA. There's also a risk to the animal when you talk about genetically altering animals. That risk lies primarily with USDA.

So there are three agencies that kind of share some of the risk regulation in these areas. You have to understand that the laws, as written in the U.S., pretty clearly divide up the pie amongst those three regulatory agencies.

Do you think people have a good idea about what these agencies do? What's your judgment of the level of trust?

I think people have, in general, a very poor idea what these agencies do, and understand them very little. I think that's not the public's fault. That's the agencies' fault, for not letting people know that they're on watch, and exactly what they're doing, and for not interacting with the public more. So I almost exclusively blame the agencies for that.

The agencies always can work better with each other. In different cabinets, there has been serious talk about combining food safety under one agency. I like that. A lot of other people don't like that. There could be better coordination. But clearly the agencies don't communicate to the public.

In terms of agencies of the government that people trust, FDA usually does pretty well in most surveys. I think that, for a small agency--the FDA is only 7,500 people, and regulates one of the largest segments of our economy--FDA usually rates pretty high in trust.

You used to work for the FDA?

I was a Public Health Service fellow, which is a little bit different than a regular government employee. It's a system of bringing people into the government for two years on short-term research projects.

Are you currently on the payroll of any major biotech company?

No, I'm not on the payroll or a consultant to any biotech companies. I have no research money from biotech companies. I actually don't do direct genetic modification of anything.

Your primary interest is food safety. Do you think this is perennial?

I think food occupies a very special place in the psyche of Americans, and for people around the world. So we're always going to be concerned about the application of technology to foods. Recent history shows that to be true. We were concerned about BST, food colors, sugar in foods; a variety of issues. The current issue happens to be something that's been called genetic modification. I don't know what the next issue will be, but I predict it'll be here.

European public opinion has removed these products from the shelf. If that happened here, what would the problems be with mandatory labeling?

If you're going to label anything in a food, you have to be able to enforce the truthfulness of that labeling. If you're going to say "GMO-free," for example, you have to first define what that means, and develop a system for enforcing it. If I say that a food does not contain genetically modified organisms, what do I mean by that? Do I mean that if I use soybean oil from a genetically modified soybean, and I cook battered fish in it and then I freeze those fish, does that fish contain a genetically modified organism or food, or doesn't it?

It's not clear. You're going to have to draw the line somewhere. What if you use an enzyme in food processing? Theoretically, tiny amounts of that enzyme could get into the food. Is that GMO or not GMO? So there are a lot of very important details that would have to be developed around any labeling initiative, in my view.

Are genetically engineered enzymes pretty widely used?

Certainly they're widely used in a variety of food manufacturing settings, in everything from cheeses to high fructose corn syrup.

So what if you were a purist?

You would ban an awfully lot of foods. As a matter of fact, if you went to the extreme with this issue, I would guess that there would be a majority of foods that would have to carry that kind of label. If you label all foods with anything, that's about the same, in my view, as labeling no foods, because people still have to eat.

Would an animal that has consumed a genetically modified crop be a GMO?

That depends on what kind of regulations they put forth. But in my view, it should not be. The genetic material from that crop does not become incorporated into the genetic material of the animal. It is simply another nutrient for that animal. But those are the kinds of issues that any labeling initiative is going to have to face, and they are not easy issues.

What would be the effect of labeling?

BST was labeled in some parts of the country. It was voluntary in other parts of the country.

What is BST?

BST is bovine somatotropin. It's the growth hormone that is given to cows to have them produce more milk. It was very controversial a decade or so ago, and there was a call for labeling of that issue as well. In my view, if labeling was implemented, depending on how it was implemented, it probably would not have a lot of effect one way or the other.

What happened in that case?

The BST labeling, as far as I know, has pretty much largely disappeared. It didn't seem to steer consumers away from milk. It didn't steer them to milk, either. It seemed to me that it had not much effect in the long term, in the big picture.

Is that your guess, if we went that way?

That would be my prediction.

What would be your prediction if some of these things came to pass? Take the worst-case scenario.

My prediction, if the worst-case labeling came in, is that it would have very little effect on the technology, the implementation of that technology, or the food products that we enjoy. I think if you look at past history, people look at labels; they read labels; but it's not necessarily the prime reason that they buy or do not buy products.

In polls, why do 80 percent of people want the information?

I think if you ask anybody, "Do you want more information?" what do they say? "No, I don't want more information?" No. We're all inclined to say, "Yes, I'd like to know more about the foods I eat." And I think that's true. But when it really comes down to it, do people really need that kind of information? Will they make buying decisions on that kind of information? Certainly, some will. But in my view, a lot will not.

wheat photograph ©h. david sewell/corbis
new content copyright ©2001 pbs online and wgbh/frontline/nova