Fighting Cancer with Phytochemicals
An interview with David E. Williams, Ph.D.
Q. How did you get interested in cancer prevention by dietary phytochemicals?
A. I've always been very interested in the effect of environmental chemicals on health, going back to my undergraduate days at Reed College in Portland. We had to do an undergraduate thesis, and it was not long after Rachel Carson's book "Silent Spring" had come out on the effect of insecticides on wildlife. That got me really interested in environmental chemicals and inspired my thesis. As my career developed I came to realize a couple of things: many of our environmental exposures are through food (what we eat and drink) and that one of the most exciting approaches is a potential for prevention of diseases related to food-borne toxic chemicals. That's what got me started in the whole area.
Q. How do you study cancer prevention by dietary phytochemicals?
A. You have to use a model. It's very difficult to design and undertake large prospective studies with humans where you provide a dietary phytochemical and then follow them for many years. That's very expensive, involves a large number of people, and sometimes you get equivocal results. So a good animal model can provide initial information on molecular mechanisms and efficacy. We've used rainbow trout as a model for a number of years, and it's turned out to be very rewarding. That was initiated by George Bailey and others, but I've adopted that model and used it to look at the effects of phytochemicals in cancer prevention. We've also used other animal models when appropriate.
Q. Why do you use trout?
A. In the early 1960s there was a large outbreak of liver cancer in rainbow trout in hatcheries on the West Coast. The researchers discovered that aflatoxin was present in to be a potent human carcinogen produced by a mold that grows on corn and peanuts, especially in hot, humid environments. It's not much of a problem in the U.S., except maybe in the Southeast, but it's a significant health problem in other parts of the world like Africa and, especially, China. Aflatoxin targets the liver. The trout is a very good model because trout metabolize aflatoxin the same way humans do, resulting in the same DNA adduct, the same oncogene mutation, and the same type of tumorhepatocellular carcinoma. So there are a lot of similarities between trout and humans. In fact, in this case, trout is a better animal model than any rodent.
Q. What's a DNA adduct?
A. A DNA adduct is a covalent bond between the carcinogen and one of the chemical bases in the double strand of DNA. There are a number of adducts formed with a carcinogen like aflatoxin, and some are particularly mutagenic and carcinogenic. Aflatoxin forms a covalent bond with guanine, one of the DNA bases, and that seems to be the adduct that's especially toxic or produces mutations very readily. Those mutations can cause the cell to commit suicide or can lead to cancer.
Q. Are there economic or statistical advantages to using trout?
A. Certainly. That's something that we exploit all the time. Typically, the per diem cost for a rat or mouse is about 25 to 50 cents, but it's about a penny a day for the rainbow trout. With trout, we are able to do very large cancer studies with high statistical power and at relatively low cost. For example, we're doing a cancer study with thousands of trout that would have cost about $7 million if we had used a mouse model and about $14 million if we'd used a rat model. Again, that allows us to address statistically very challenging questions that you just can't accomplish with a rodent model.
Q. What carcinogens have you tested in trout?
A. That's a very good question because there are some limitations and some strengths in this model. One of the limitations is that we can't study some human cancers like breast cancer, lung cancer, or prostate cancer in trout. But trout are good for studying aflatoxin, most of the other mycotoxins, and another important class of environmental carcinogens called polycyclic aromatic hydrocarbons, or PAHs, which are formed from the combustion of any organic material. When oil-derived products, coal, or anything organic are burned, these PAHs are formed. There is a significant amount of PAHs in our dietthey're just everywhere. There are a few chemical classes of carcinogens to which trout don't respond, such as heterocyclic amines, or cooked-meat mutagens, found in proteinacious food cooked at high temperatures.
Q. How do you design experiments to study cancer prevention in trout?
A. We usually do a co-exposure in which the carcinogen and presumably protective phytochemical are given together in the diet. That's possible because we use a purified diet that we make up ourselves called the Oregon Test Diet. Trout hatcheries around the country now use that diet. It's a purified diet with casein as the protein and a defined vitamin and mineral mix and fish oils. We can just mix the carcinogen and the phytochemical together in the diet and feed it to the trout, or we can give the phytochemical for a week or two before we expose them to the carcinogen.
Q. How do you decide which phytochemicals to test?
A. Sometimes it's just a guess. Sometimes we identify likely candidates from the scientific literaturewhat's worked in other modelsand then test that in our model. We're interested in phytochemicals that are a significant part of our diet. We're not interested in studying phytochemicals we wouldn't normally consume dietarily. For example, right now we're focusing on indole-3-carbinol from cruciferous vegetables, such as broccoli, Brussels sprouts, and cabbage, and chlorophyll and its derivatives, which, of course, are present in all green leafy vegetables, especially spinach.
Q. Do you give a dose to the trout that is similar to what people might consume dietarily?
A. We typically start with very large doses just to see if we are going to see an effect at all. Then we can do a dose response study to see what the degree of protection is at more realistic doses. For example, a few years ago we published a study with 10,000 trout in which we used six different concentrations of aflatoxin and six different concentrations of indole-3-carbinol. With that number of fish, we could study levels of indole-3-carbinol that are quite similar to what a human would get from supplementation. Some of the phytochemicals we investigated are not only present in food but are also available over the counter as supplements. Indole-3-carbinol is one of the best-selling phytochemical supplements. We found a protective effect at levels of about 250 parts per million, which is fairly close to what the human dose would be with supplementation.
Q. Do you find that indole-3-carbinol and chlorophyll protect against cancer by the same molecular mechanism?
A. No, and that's what really makes this research particularly interesting. Indole-3-carbinol seems to work by a number of mechanisms. One involves the induction of enzymes that are responsible for detoxification of the carcinogen.
Q. Are those enzymes the same in trout and people?
A. They're very similar, and they respond to indole-3-carbinol in a similar manner. But there are other mechanisms, too. Scientists have shown that indole-3-carbinol can affect other important steps in the cancer process and cause programmed cell death, or apoptosis, where a cell that's been mutated commits suicide. There are sensors within the cell that sense that it's undergone mutations, and if it can't repair that DNA adduct or that mutation before it divides, then it programs itself to commit suicide.
Q. Does chlorophyll protect by a different mechanism?
A. We're still studying that. Until fairly recently, it was thought that the sole mechanism of chlorophyll and its derivative, chlorophyllin, was binding to the carcinogen in the gut and preventing it from being absorbed. There was strong evidence for that. But now there's evidence for other mechanisms, too, because when we feed chlorophyll to trout, their livers turn green. So we know the chlorophyll is absorbed by the target tissue, suggesting that there could be mechanisms other than the binding mechanism.
Q. If these phytochemicals work by different mechanisms and against different carcinogens, would that suggest that it might be useful to combine indole-3-carbinol, chlorophyll, and other phytochemicals to try for more powerful prevention?
A. Yes, and that's a growing area in the chemoprevention field. If there are multiple mechanisms, then combining phytochemicals for an additive or synergistic effect would be a good strategy. For example, scientists have already shown that combining COX-2 inhibitors and phytochemicals has an additive effect in preventing colon cancer.
Q. Does cooking destroy phytochemicals in vegetables or alter them in a way that might reduce their cancer chemoprotective properties?
A. To some degree. It depends on the particular phytochemical. For example, with indole-3-carbinol, it's not such a problem because most of the derivatives we get from cooking are similar to the breakdown products you get after ingestion anyway. With chlorophyll or chlorophyllins, it's also not really a problem because some of the breakdown products in the body are similar to those formed from cooking. But there is some loss. As a general rule, I would say that these protective phytochemicals are probably higher in raw foods than in cooked foods.
Q. Once you get positive results from the trout experiments, what's the next step in assessing whether these phytochemicals may protect against cancer in humans?
A. We like to take a comparative approach, which we've done historically. If a mechanism is similar in fish, mice, and rats, it's likely that mechanism was conserved throughout evolution and probably works in humans as well.
Q. When people read about the protection afforded by phytochemicals in tea or vegetables, such as indole-3-carbinol, chlorophyll, or catechins, they're inclined to buy supplements that contain the purified extracts and take them in fairly high doses. What's the safety profile of these phytochemicals?
A. When people ask me that question I tell them it's best to get these protective phytochemicals from a balanced diet rather than from supplements. However, it may be difficult to get enough cruciferous vegetables in the diet. If you have an aversion to the taste of some of those foods, as I do, supplementation might be okay. The problem is that people tend to have the attitude that if a little is good for me, a lot must be even better, and that's not necessarily the case. And you can take too much of some of these compounds. Sometimes the purified compounds don't have the same effect as the consumption of the whole foods. For example, for years epidemiological studies found a really good correlation between blood levels of beta-carotene and a decreased risk of lung and other cancers. Scientists in the chemoprotection field were convinced that beta-carotene had to be chemoprotective, but when they finally did an intervention study, it turned out that beta-carotene was not protective. In fact, it seemed actually to enhance slightly the risk of lung cancer in smokers. In that case, beta-carotene may have just been a marker for another chemical in those foods that was having a protective effect. We still don't know that for sure, but that's one example where we were fooled by the simple correlation-type studies.
Q. Are there any chemical differences between the phytochemicals that you give to the animals and phytochemicals in the diet? Are they the same chemical compounds?
A. Yes, they are.
Q. Some of your work shows that indole-3-carbinol given after exposure to a carcinogen may actually increase the risk of developing certain tumors in trout. How relevant do you think this is to people, because, of course, we would not know if we have cells that have already been initiated by a carcinogen?
A. That's an important question. We always try to design our experiments so that we can see both beneficial as well as adverse effects or toxicity. There are risks as well as benefits. We've done a number of studies showing that feeding indole-3-carbinol long-term after initiationafter exposure to the carcinogento trout and rats produced more liver cancer. In mice, though, there was less. Further investigation revealed that these disparate effects are related to sensitivity to estrogens. In other words, indole-3-carbinol may be acting as an estrogen. Genistein, a chemical in soy, also acts as a phytoestrogen.
Q. Does the phytochemical dose affect this cancer risk?
A. Based on the studies we've done, it would probably be very difficult to get a dose high enough to be a significant risk just from dietary cruciferous vegetables. However, if you eat cruciferous vegetables and supplement with a significant number of over-the-counter indole-3-carbinol tablets, and you plan to do that long term, there could be some risk. People who supplement should ask their doctor to check their liver function periodically because there could be some potential risk to the liver associated with long-term, high-dose supplementation. The problem with indole-3-carbinol is that it's a mixture of compounds. When you take an indole-3-carbinol tablet, it reacts with the acid in your stomach and very quickly forms a mixture of as many as 24 different compounds. We know very little about the potential toxicity of some of those individual compounds.
Q. What about the safety profile of chlorophyll or its synthetic derivative chlorophyllin?
A. That seems to be a different story. We've used chlorophyllin in human medicine for years. It's been used in geriatric patients as a pretty effective deodorant. In fact, I saw an Internet site where they were selling chlorophyllin tablets to deer hunters. They were marketing it as camouflage. If you take the chlorophyllin tablets, you'll smell more like a plant and the deer won't be able to detect you. We haven't seen any adverse effects of high doses in humans, which made it really easy to move from animals to human application because chlorophyllin had already been used for other purposes in humans and found to be safe.
Q. Your recent research has focused on fetal exposure to carcinogens that may increase the risk of cancer in children. How did this hypothesis develop?
A. I became familiar with literature in cancer journals showing that a number of chemicals were capable of crossing the placenta and causing cancer in offspring born to mothers exposed to carcinogens. I got interested in the potential for altering the maternal diet and providing some protection for the fetus. If the chemical carcinogens cross the placenta, then protective phytochemicals may as well.
Q. What carcinogens commonly cross the placenta?
A. Probably the two that have been studied the most are polycyclic aromatic hydrocarbons, or PAHs, and arsenic. At least a dozen chemicals have been shown to be capable of crossing the placenta and causing cancer.
Q. What's the environmental source of arsenic? How is the mother exposed?
A. Arsenic is found in drinking water. Fortunately, in most places in the U.S., that's not much of a problem, but Artesian wells often have significant amounts of arsenic. In Bangladesh, the levels of arsenic in drinking water can be so high that there is a definite risk that babies born to those mothers will develop cancers.
Q. What's the source of the PAHs?
A. PAHs are found in the atmosphere. We breathe them all the time. They're also found in water, but the main exposure is through the diet. They get into plants, and they get into the animals that eat those plants. Even if you're a vegetarian, you're going to have some PAH exposure.
Q. Are PAHs concentrated in specific vegetables or fruits?
A. That's a good question. I haven't seen any research that has found that certain plants accumulate PAHs more than others. PAH exposure also depends on how you cook your food. Charcoal broiling typically enhances your exposure because when you burn that charcoal, the PAHs formed from combustion get deposited on the meat. And, of course, if you're a smoker, your exposure is increased from burning tobacco. One of the things that is really going to increase the exposure to PAHs in the coming years is coal burning in China. The global winds drift from west to east, so coal that burns in China gets in the atmosphere and then is deposited in the U.S., especially on the West Coast. Over the next 20 years there's going to be a huge boom in coal-fired energy production plants in China, so we may see a big increase in PAH exposure.
Q. Do some of these carcinogens also act as teratogens and cause birth defects?
A. That's not really known. Dioxins, which are chemically very similar to PAHs, do cause birth defects. For example, dioxin causes cleft palate. So the carcinogens we study probably could cause birth defects, but we haven't found many relevant studies.
Q. How common are childhood cancers and what kinds of cancers do children develop?
A. The two most common types of cancer in children are leukemias and lymphomas, which are cancers of the blood, followed by brain tumors or cancers of the nervous system. Those are the major cancers in children, but cancer in children is not common. Cancer is the leading cause of death by disease in children up to the age of about 15. The most frequent cause of death in children is accidents, followed by cancer.
Q. How do you study this problem?
A. We use a mouse model. Pregnant mice are treated with an oral dose of the carcinogen about two or three days before they would give birth so there's ample time for the carcinogen to cross the placenta. Almost all of the offspring born to that mother develop a severe T-cell lymphoma and die between three and six months of age.
Q. Do the mothers develop cancer, too, as a result of exposure?
A. No. The adult seems to be relatively resistant, although we haven't followed the mothers for a long period of time. We've focused on the offspring. What's been really interesting to me is the dramatic effect of adding indole-3-carbinol to the mother's diet. The pups born to the mothers that were fed the indole-3-carbinol, even though the pups never got any indole-3-carbinol, are substantially protected against lymphoma.
Q. Are they protected because of enhanced detoxification of the carcinogen by liver enzymes?
A. That's one theory. We're not exactly sure of the mechanism. There might be other mechanisms at work, too.
Q. Have you looked at other phytochemicals using this model?
A. We are just finishing a study with tea. We used both regular green tea and decaffeinated green tea. We also tested a major phytochemical from green tea, epigallocatechin gallate (EGCG), alone or caffeine alone. So we had four different experimental groups. It was a little disappointing in that we didn't see much protection with tea. There was a hint of some protection with the caffeinated green tea, although it needs to be analyzed statistically. Caffeine alone did provide some protection. That could be important because the FDA specifically recommends that pregnant women cut down or refrain from caffeine ingestion. Those results are pertinent to lymphoma. In the surviving mice that don't get lymphoma, 100% of them get lung cancer. The EGCG alone seems to provide some protection against lung cancer, although we're having a statistician look closely at that. I'm not giving up completely on tea, but it doesn't look to be as effective as indole-3-carbinol from cruciferous vegetables in protecting against transplacental carcinogen exposure. We're just now starting a test with chlorophyll and chlorophyllin.
Q. Is lung cancer rare in children?
A. It is. In mice, it doesn't develop until they are about middle age. The number of children that get cancer is relatively low, but certainly cancers in middle-aged people and older are not rare, and lung cancer is still the #1 cause of death from cancer in males and females in the U.S. It would be exciting if this risk could be substantially reduced by phytochemical exposure during pregnancy. We're going to do some additional experiments to test that hypothesis.
Q. You also study ethnic differences in drug-metabolizing enzymes. Why is this important?
A. This field is known as pharmacogenetics. It's becoming increasingly important in medicine because the greatest number of hospitalizations are due to adverse drug reactions. A number of adverse drug reactions are due to genetic polymorphisms. Humans are not like inbred animalswe have a lot of genetic variability. Almost all of our enzymes involved in metabolizing drugs and carcinogens have what's known as genetic polymorphisms. In other words, there's more than a single form of the enzyme. There are sometimes many different possible forms of the enzyme, depending on how many mutations the individual has in the gene coding for the enzyme. These genetic polymorphisms are often related to ethnicity. That's certainly the case with the enzyme we've been studying, flavin-containing mono-oxygenase (FMO), because the expression seems to be confined to African-Americans and Hispanics, and we don't find the enzyme in Caucasians or Asians.
Q. In what tissue is the enzyme found?
A. It's found mostly in the lung. We've been trying to determine if that makes a difference for drugs that are toxic to the lung or metabolized by the lung. Inhalation is a very good drug delivery method. Some drugs just aren't very effective when you take them orally because they don't get absorbed very well. But the absorption is almost complete if the drug is inhaled. So it's a very effective method for drug delivery.
Q. Is that because of the vascularization of the lung tissue?
A. Yes. There's only a very small distance, about a micrometer, that a drug has to travel to be absorbed in the lungs, as opposed to how far it has to go through the lining of the GI tract. There's a huge surface area in the lung, and, again, it is highly vascularized. For all those reasons, drugs are absorbed very well from the lung. We've gotten some funding from pharmaceutical companies that are developing drugs for that purpose to look at whether FMO metabolizes certain drugs in the lungs.
Q. How are drugs metabolized?
A. Almost any chemical that gets into your body is metabolized before it's excreted. Metabolism helps to make the compound more water soluble and easier to excrete in urine or feces. Often, a hydroxyl group will get attached to a molecule to make it a little more water soluble. After that, the compound may be sulfated as well. It becomes very water soluble because it's got an added polar group and a charge as a result of these chemical modifications.
Q. What drugs or chemicals have you been studying?
A. We have looked at compounds like thioureas, which are found in a number of chemicals and drugs. Lately, we've been studying ethionamide, which is a drug used to treat tuberculosis. Tuberculosis is a very significant health problem around the world that's increasing in incidence.
Q. Is the drug delivered by inhalation?
A. Yes. We want it to work in the lung. The drug is metabolized, and the sulfur group is oxygenated by bacterial enzymes, resulting in the death of the tuberculosis bacterium. We're interested in what happens if the human enzyme, instead of the bacterial enzyme, metabolizes that sulfur first. Does that lessen the effectiveness of the drug? My guess is that it probably does. There are other thiourea-containing compounds whose activation by this enzymeby oxidizing that sulfur groupmay make them toxic to the lung. That has happened in mice given a developmental drug for pain reliefthe mice died from massive lung edema. We found out that this happened because the enzyme was metabolizing sulfur to a reactive sulfenic acid, which caused the toxicity.
Q. What percentage of African-Americans and Hispanics have the active FMO enzyme?
A. It turns out that all the Caucasians and Asians we have genotyped to date have a mutation that produces an inactive enzyme. About 27% of African-Americans and about 2-7% of Hispanics, depending on whether they're from Mexico or Puerto Rico, have at least one of the genes coding for an active enzyme. We predict that about 27% of African-Americans and 5% of Hispanics would metabolize these drugs differently because they have an active enzyme, whereas Caucasians and Asians do not.
Q. Are physicians aware of these ethnic differences so that they can calibrate the amount of a drug given to patients, according to their ethnicity?
A. Not in this particular case. Physicians are becoming increasingly aware of the importance of polymorphisms of some of the major drug metabolizing enzymes. In the liver, for example, the major drug-metabolizing enzyme is cytochrome P4503A4, and physicians have learned that's an extremely important enzyme in the metabolism of drugs like cyclosporin. Cyclosporin is used in organ transplant patients to fight rejection by depressing the immune system. It's a tricky drug to work with because there's a very small difference between a therapeutic dose and a toxic dose. It's metabolized mainly by cytochrome P4503A4, and the amount of this enzyme in the liver can vary among individuals by ten- to twenty-fold. That can make a big difference in the drug dosage.
Q. Are some of these liver enzyme polymorphisms also related to ethnicity?
A. Yes, almost all of the genetic polymorphisms in these drug metabolizing enzymes have an ethnic difference. It's almost the rule. Within 10 to 20 years it will be pretty typical to measure the level of these enzymes before setting drug dosage. That will avoid a lot of the adverse drug-drug interactions that cause so many people to end up in the hospital because of an overdosethey may be deficient in an enzyme that metabolizes the drug so they can't get rid of it as fast as a person of different ethnicity.
Q. Could there be ethnic differences in the way indole-3-carbinol protects against certain cancers?
A. Yes. That's actually an under-studied area. If we could use this knowledge in medicine to design the proper dose of a therapeutic drug for an individual, then we could do the same thing with phytochemicals for protection against diseases like cancer. For example, we could genotype people once we know the mechanism by which the phytochemical works. Then we could figure out what genes are important in that mechanism and test individuals for their enzyme activity. That would allow us to better determine the optimum intake of certain phytochemicals.
Q. What dietary strategies do you think people might consider in order to minimize the risk of developing cancer?
A. I think the recommendation that has been advocated by a number of agencies like the National Cancer Institute and the Association for Cancer Research to consume five to nine servings of fruits and vegetables a day is still a good rule of thumb. Most of the really effective phytochemicals against cancer tend to be in vegetables rather than in fruit. Cutting down consumption of meat and eating more fish is also good. If you're a woman of child-bearing age, you have to choose fish carefully to avoid mercury. Food should be prepared in ways to minimize or eliminate polycyclic aromatic hydrocarbons and cooked-meat mutagens. Aside from that, I recommend some daily supplements that I take myself, such as lipoic acid, a multi-vitamin pill, 500 mg of vitamin C, and 400 IU of natural vitamin E. I think that supplements are valuable in addition to a whole food diet.
Q. Aflatoxin exposure doesn't seem to be a prevalent public health problem in this country. Do you feel there is any risk from eating peanuts, peanut butter, or bread?
A. It really isn't much of a health problem in the U.S. The FDA watches it pretty closely and their action level for aflatoxin in food products is 20 parts per billion, which rarely happens. I think the models that the FDA used to estimate cancer risk from aflatoxin are pretty conservative. There is actually much less liver cancer than would be predicted from aflatoxin exposure. I think we're pretty safe in the U.S. with respect to aflatoxin. That's not true in parts of China where mold grows on grain because it's hot and humid.
Q. Do you think that some dietary phytochemicals may be useful in treating cancer?
A. That's hard to know. One of the driving forces for drug development in the U.S. is money. Obviously you can't make a lot of money studying compounds like phytochemicals that can't be patented, so therapeutic phytochemical research won’t proceed as quickly as drug research. It's becoming a focus for researchers, but it's very difficult.
Q. Have you used trout or small rodents to study the potential therapeutic use of phytochemicals?
A. We haven't done that in trout because it takes too long for the tumor to get big enough to evaluate regression, but an athymic or nude mouse model can be used. In these mice, the thymus is removed and their immune system is completely compromised. You can implant a human tumor cell in that mouse, and the tumor will grow as it does in a human. For example, you can take human breast cancer cells and inject them under the skin of a mouse. You can then feed the mouse different phytochemicals to see if that tumor shrinks or its growth is inhibited. I'm planning to do a study with human cancer cells and indole-3-carbinol to see if we can get tumor regression. We can inhibit lymphoma transplacentally, but now we want to know if indole-3-carbinol will inhibit its growth once it's formed. So we will be continuing our cancer chemoprotection studies and also investigating potential therapeutic roles.
Last updated November, 2006