RE-EVALUATION OF DIOXIN
A Presentation by Linda Birnbaum, Director
Environmental Toxicology Division U.S. Environmental Protection Agency (EPA)
To the 102nd Meeting of the Great Lakes Water Quality Board, Chicago, Illinois
July 15, 1993
What I would like to do today is give you an overview of the state of science of dioxin and health effects. What I am going to talk about first is the effect of dioxin itself. Later on I will return to the issue that dioxin is but one of a family of chemicals and, if you really want to understand the risk out there, we need to look at the aggregate of all of them together. But the important thing again is that you have a planar molecule with halogens off to the side. I would like to stress that most of the information we have has to do with the chlorinated molecule, but the brominated molecules are just about as equally toxic, and we know almost nothing about them, and almost no environmental monitoring has been done to determine whether they are really out there or not.
Some properties of dioxin are relevant to its persistence and its bioaccumulation in the environment and why we have a problem with it. It is insoluble and is highly bioaccumulated in food chains. It's very insolubility is a real issue from a research and regulatory standpoint when the standard procedure concerns how much is in the water, whereas in fact there is almost none of it in the water itself. There is a very low vapour pressure which means that it does not volatilise very much.
How are people exposed to dioxin? Food is the major source, except for some major industrial accidents where there might be some inhalation of dioxin or some skin absorption. Essentially greater than 95% of our exposure is by the food we eat, primarily meat, fish and dairy products. Now a lot of the attention has focused on the concentrations of dioxin in fish, so that may be higher than the concentrations in meat and dairy products. But in the American diet, at least up until the present time, most Americans do not eat as much fish as they do dairy product. So we have approximately equal amounts of exposure from fish as from dairy products. There may be susceptible sub-populations who do eat a large amount of fish who will get more than the average exposure to dioxin.
Dioxin gets into the food chain by bioaccumulating in organisms in the food chain. How does it get into the food chain environment? Some dioxins are directly eliminated into water, for example, from pulp and paper mills. Most dioxin is released directly to the atmosphere and is subsequently distributed worldwide through atmospheric transport. Dioxin is very sticky, it binds to particles, is picked up on dust particles in the winds and is blown around. Nowhere in the world today is free from dioxin. This is a worldwide contaminant and can be found, with sensitive analytical techniques, even in the most remote places on earth.
When was dioxin first found in the environment, and when did it start to accumulate? Some chemical companies have been trying to convince us for a long time that dioxin has been around since the beginning of time, and that it is a product of forest fires and volcanic activity. Prior to the onset of heavy uses of chlorinated organics in industry, which really commenced about 1930, levels were extremely low, based on analysis of sediment samples. People have done analyses of Egyptian mummies from more than 2,000 years ago and frozen Eskimos from northern Canada and the levels are below the detection limit. Dioxin is a product of the modern industrialisation. Other than forest fires and volcanos, what are the other major sources? I think it is a point to remember that, unlike polychlorinated biphenyls (PCBs), dioxins and dibenzofurans have no known industrial use and they were never made for any purpose. They are contaminants of industrial processes involving certain organic compounds and chlorine. They are produced by low temperature combustion at between 300o to 400o degrees centigrade. To destroy dioxins, you have to go to over 1200o centigrade. But you can form them at combustion temperatures characteristic of wood burning in wood-burning stoves. In addition, dioxin is a product formed by the chlorine bleaching of paper and pulp products. The paper industry has been very responsible in decreasing their use of chlorine bleach, and thereby tremendously decreasing their input of dioxin and dibenzofurans into the environment. Of course, that is after 50 years of heavily contaminating the sediments, which give us a long-lasting problem.
What are the other sources of dioxins? In the U.S. Environmental Protection Agency (USEPA) dioxin reassessment, there was an attempt to use total mass balance equations to determine where the dioxin is coming into the environment. A number of environmental monitoring studies indicate that the levels do not appear to have declined since the mid-to-late 1980s. This suggests that we still have inputs as well as outputs, and that we are in a pseudo steady state. In the Great Lakes, there was a peak of dioxins in the early '70s and since then levels came down until about 1987. In the U.S., the major new sources of dioxin appears to be hospital incineration. In Germany it appears to be municipal waste incineration. It is not a major source here because we use landfills but, if we switched to municipal incineration, it may become a major source in the U.S. If anyone knows about hospital waste incineration, they are totally unregulated and they burn at very low temperatures, they have tremendous amounts of chlorinated plastic, and therefore lots of dioxin are potentially emitted from them. Diesel exhaust also appears to be a source of dioxin. Metal smelting and refining appears to be another source, but we really don't have a good mass balance. In summary, we can account for about 50% of the new input into the environment, and we are uncertain where the rest is coming from.
A typical toxic substance goes into the body and kills cells, or does one particular thing. In contrast, dioxin does a lot of things and should be considered as a hormone. I don't know what it is a hormone for, in terms of the natural sense. Hormones go into the body and they have different effects on different tissues, they can have different effect at different stages of development of the tissue, and they can have different effects on different species. I think we can say the same thing about dioxin. So you see a great deal of tissue specificity, developmental stage specificity, or age specificity, or species specificity in response.
At very high doses, in all species we have looked at, dioxin causes death. But, again, it is not your typical pesticide kind of death, where you know you give a lot to the animal and it goes four legs up in the air immediately. Death is usually preceded by a wasting syndrome, the animals can easily lose 50% of their body weight before they die and, depending on the animal, the time to death varies. So it takes guinea pigs about two weeks, mice take about three weeks, so do rats approximately, monkeys take about six weeks to die, for example. It is an inexorable process. If the level of dioxin administered to the animal, either acutely or following repeated exposure, once they reach a certain body burden, it is like a switch is thrown and the animal will, approximately x number of weeks later, die. We really don't know what they die of. You can, for example, feed the animal and prevent the wasting, and they still die. So it is not just because they used up their body glycogen. The highest levels that we know that people were exposed to, would have been lethal in guinea pigs, but they are, at least in the order of magnitude or more, lower than the levels that would kill almost any other mammalian species. So I think that the reason we have not seen wasting or death in people is we may not have bid high enough exposure, and that is not an experiment I think we are going to try.
At levels below that where you see wasting, you can see effects on the lymphoid tissues and you actually have loss of the thymus and spleen and at slightly lower levels, and in the adult male you can have atrophy of the testis. But, again, these are all relatively high dose effects. Effects on the liver -- there are some differences in different species but, in general, you see enlargement of the liver, you see accumulation of fat in the liver. In some tissue, you have hyperplasia, which is a proliferation of cells. The tissue actually gets bigger from having more cells, and this occurs in the lining of the gastro-intestinal tract, it occurs in the lining of the urinary tract, and it occurs in the bile duct, which comes from the liver.
In other kinds of cells, instead of getting hyperplasia, which is an inappropriate proliferation of cells, you get squamous metaplasia, which is an inappropriate differentiation of cells. It is not that an eye turns into an ear, but in fact one type of cell turns into another type of cell. It starts behaving like that cell. So, one of the classic kinds of symptomatology that we see, not only in monkeys for example, but, also in humans, is metaplasia of the meibomian glands of the eye. Now, meibomian glands are little glands at the base of your eyelid that secrete very small amounts of fluid. With exposure to dioxins, these actually change and start producing waxy exudates on your eye that makes vision very difficult and this is a complaint of people who have had very high levels of dioxin exposure. They also complain of problems with hearing. There are glands that line your ear canal which are call the ceruminous glands and these also undergo an inappropriate differentiation and start producing earwax. These are not the normal glands that produce normal earwax. It might be possible to monitor exposed people by collecting some of their earwax and measuring the concentrations of dioxin and related chemicals in the earwax. It will be a little hard to do mass balance equation to calculate their whole body burden, but it might be an accessible source of tissue where you could find out whether there had been exposure.
Chloracne has been the hallmark of dioxin toxicity. This is not the ordinary teenage acne but a very, severe persistent form of cystic acne. People who had been exposed over 40 years ago to dioxins in industrial accidents are still having active chloracne, not limited just to their face and back, but all over their bodies. It is characterised by hyperplasia, which is a proliferation of cells, by hyperkeratosis, which is an altered differentiation of the cells, and by changed pigmentation. It is a very severe condition. You could say it is just acne, but obviously it has changed the lives of these people.
Dioxin is a potent teratogen in a number of species in that it causes actual gross structural abnormalities, cleft pallet, and hydronephrosis in mice. In other species, it causes much more subtle developmental effects. Effects that, until the last couple of years, we were not really aware of, because they are the kind of effects that you do not see in a standard teratology study. In a standard teratology study, you dose animals during organogenesis, which is the major period of differentiation of the different organs and tissues, and then you sacrifice them just before they would normally be born. The developmental effects we are studying now in fact are things that you do not see until the animal reaches puberty, and then there are alterations in the sexual functions and reproductive behaviour of the animals. So, if you kill the animals at birth, you are not going to see what is going to happen.
I will talk a little bit more about cancer. Dioxin appears to be a carcinogen in fish, rodents and other mammals, including humans. But dioxin can also modulate the immune system resulting in an inability to fight disease. It is a very powerful immunosuppressant. It can also upregulate the immune system so that you start becoming hypersensitive, start developing autoimmunity and allergies. Depending upon the stage of the animal and the species, sometimes you observe immunosuppression and in other cases you observe upregulation.
Dioxin causes a wide variety of changes in enzyme levels and causes biochemical effects (Table 1). There has been some discussion about whether these changes represent adverse effects or just biological responses. I think many of them can be considered as biomarkers of the potential for other effects to happen. These changes in enzyme levels, including increases and decreases in the synthesis, leads to alteration in metabolism of both endogenous and foreign compounds. For example, it affects the way we handle glucose metabolism, induces cytochrome P4501A1 and P45OlA2 and regulation of other kinds of enzymatic activity.
Dioxin modulates many hormone systems and their receptors (Table 2). It affects thyroid, gastric hormones and melatonin, which is a product of the pineal gland involved in the circadian rhythm. It affects estrogen, androgens, glucocorticoids and insulin. Dioxin causes the modulation of growth factors and their receptors including, for example, Vitamin A. By affecting hormone systems, you alter the homeostasis of the animal and switch how the animal behaves. For example, dioxin can act both as an anti-estrogen by blocking estrogen activity in the breast or the uterus, and in other tissues it may act more like an estrogen. Dioxin causes decreases in circulating thyroid hormones in, for example, rats, but it causes increasing levels of circulating thyroid hormones in mice. I think we have to be very careful when we talk about these effects to realise again that hormones are very carefully modulated, and the way that we maintain homeostasis in our body is by having a balanced level. So, if you perturb the level, either by increasing it or decreasing it, you are putting yourself at risk for problems.
Many different molecules are involved in controlling the activity of cells (Table 3). But there are growth factors, such as epidermal growth factor, which tells cells to divide, and there are factors like transforming growth factor beta that tells cells to stop dividing. There are growth factors like insulin growth factor, which is involved in glucose metabolism and in many other functions in the body. Dioxin modulates the activity of many of these, leading to altered growth and differentiation. Dioxin is therefore a very potent growth dysregulator.
Mechanism of Action
How does dioxin cause these biological effects? One thing I want to impress you with is that the mechanism of action is much more complicated than we thought. A couple of years ago, we thought that TCDD came in and bound to a protein called the Ah receptor and that went on into the nucleus and interacted with DNA to lead to changes in the expression of genes and of protein. We know now that it is much more complicated. The way that the Ah receptor functions. it is never present as an isolated protein (Figure 1). Its action is controlled by other proteins, so that the Ah receptor binds to two molecules of the, proteins called heat-shock proteins. Heat-shock proteins were first identified by raising the temperature of an animal, and observing changes in these kinds of proteins. These proteins control the ability of the receptor to bind dioxins. Then thereis another protein here, which has been identified only by its molecular weight. I am going to come back to this later. We call it p5o, but we really don't know what it does. We have an hypothesis but no tests. We have this complex of four proteins binding dioxin, and then something happens. I leave this purposely vague, other than to say we lose the heat-shock proteins, we lose p50, but gain another protein which, at this point in time, is a protein called arnt. This complex -- and we don't know whether this forms in the cytoplasm or in the nucleus -- functions in the nucleus, so that the complex that actually alters genetic activity is composed of two proteins, as well as TCDD. Then the dioxin can be recycled back. Heat-shock proteins are two molecules and this is really a family of proteins, and they are developmentally regulated. You have the possibility that, at different stages, these may function differently in controlling the ability of the receptor to bind the ligand.
There is the possibility that arnt, the second protein which is actually involved with the DNA binding, is part of a family of proteins. TCDD bound to a receptor can interact with arnt 1, arnt 2, arnt 3, arnt N to form this heterodimer, the two protein complex. Each of these will only recognise one specific gene. So we have multiple levels of interaction going on here. We have multiple transcription factors, multiple DNA recognition sites, and we also know that even the Ah receptor, which at one point in time we thought was the product of a single gene, has multiple alleles of that gene so that there are Ah receptors that are slightly different sizes. We do not know yet about the specificity of the receptors and whether the receptor for the protein with a size of 106 kilodaltons will accepted the 97 kilodalton protein, in addition there is a 110 kilodalton receptor. We do not know whether or not these have any functional implications but at least there is that possibility. The point is we are not dealing with a simple system. This is really quite complicated.
To make it a little more complex, Figure 2 is a hypothetical schematic of the two ways that the dioxin Ah receptor complex functions to alter the genetic activity of the cell. First, it changes it so that you have more protein synthesis or less protein synthesis. And this is what has been demonstrated and what people have spent most of their time looking at. That, in fact, TCDD binds with this complex, which then falls apart and the dioxin receptor, plus that arnt protein, goes into the nucleus, and alters transcription and gene expression. Second, this p5O, which I told you about and we really don't know what it is, there is at least an hypothesis that this protein is a tyrosine kinase. These are proteins that put phosphate groups on tyrosine residues. Tyrosine is an amino acid in proteins, and when you phosphorylate proteins, you affect their activity dramatically. With some recent studies from a number of laboratories, including our own, we can show tyrosine kinase activity is activated minutes after dioxin exposure. Things are happening too quickly to involve changes in what is going on in the nucleus. They involve changes that are going on at the point of contact. We know also that some of the proteins that are phosphorylated by tyrosine kinase in response to dioxin are proteins that control the movement of cells through the cell cycle. Here we have a direct tie-in between a very rapid effect of dioxin, which clearly requires the receptor because it acts as a regulator of this protein. When this protein is bound to the receptor, it can't phosphorylate other protein. When you release it from activity, it now goes on and does its thing. One of the critical differences about dioxin as compared to, for example, compounds in broccoli, cauliflower, and other foods we eat that also bind to this receptor, is that those compounds are very rapidly metabolised and gotten rid of. So you have a very short-lasting effect. The problem with dioxin is that you have a persistent effect. Dioxin comes in, it binds its receptor, and it ties it up. So you get a constant signal from the nucleus to make more protein, and you have a constant phosphorylation going on. Usually phosphorylation events are very tightly regulated. They are usually very much one-off and now you locked the switch. I think the whole idea of the persistence of the dioxin activity is an important one to think about when we try and understand what the compound is doing.
So what I have shown you is that this binding to the Ah receptor is necessary for the effects of dioxin, it is the first step, but it is clearly not sufficient. Many other things have to happen subsequent to binding to the receptor. I think what you really have is a biological amplification of these responses and it occurs by a cascade of growth factors and hormones and the term that is used in the literature now when we talk about steroid hormones, the estrogens, the gluco-corticoids, the progesterones, is that you have "combinatorial complexity." What that means is that you have a complicated network -- nothing is linear -- everything is interacting, and that is exactly what is happening with dioxin. It is almost like dioxin is going into the middle of this network, and sending out signals in all directions.
There is a wide variation in acute toxicity of TCDD to adult mammals (Table 4). The guinea pig dies a week or so after being exposed to dioxin, whereas hamsters survive until you get to very high doses. You have approximately 3-5,000 fold differences in sensitivity. For many chemicals this is highly unusual. For something that functions like a hormone, this is not totally unexpected. But the point that I want to gather is that, while guinea pigs are exquisitely sensitive, and I am going to stress that these are adult guinea pigs, and adult hamsters are very resistant, most mammals tend to cluster in the neighbourhood of 100 micrograms per kilogram as approximately the lethal dose. So while the guinea pigs and the hamsters are outlyers for this response, most animals are similar. I stress the point about adult hamsters versus adult guinea pigs because if you look at hamster embryos or fetuses, they are essentially equisensitive to guinea pig embryos. If you look at rat embryos, they are essential equisensitive. There is something about the adult hamster that makes them resistant to TCDD, but the embryo responds at the same concentration as lots of other species.
With respect to dioxin, people react similarly to animal responses. Biology is inherently conservative and things tend to work the same way in many species. There is a large amount of data showing, for example, that changes in biochemical properties such as enzyme induction, in some hormonal states and in growth factors occur at similar body burdens in animals as they do in people. For example, in the ongoing occupational study conducted by National Institute of Occupational Safety and Health (NIOSH) looking at workers who were exposed to dioxin, these adult males are showing decreases in the levels of their circulating testosterone at body burdens very similar to the body burdens in adult rats. In immunotoxicity testing, human lymphocytes and cultured cells respond to the same concentration of dioxin in the media as mouse and monkey cells. in terms of developmental toxicity based on organ culture you find similar responses at similar concentrations of TCDD. For example, if you take out the embryonic palate of a rat and the embryonic palate of a human, put them in culture and expose them to the same concentration in the media, you get a similar response. Similarly, the body burden associated with chloracne in people is essentially the same as the body burden causing chloracne in monkeys, in hairless mice, or in rabbit cars. Animals with a lot of hair -- like regular mice and regular rats -- do not develop chloracne, but hairless mice do and the body burden there is essentially the same. Cancer appears to occur at similar body burdens in animals as in humans.
I will just mention some really recent data on subtle developmental effects that you would not see if you only looked when the animals were born. Dick Peterson at the University of Wisconsin is doing some very important mammalian work. A year ago he reported that if he treated pregnant rat dams towards the end of gestation with a very low level of dioxin, as low as 65 nanograms per kilogram, it resulted in demasculinisation and feminisation of the male offspring. Most of these changes were not detectable until they reached puberty. We have since repeated that study, not only with his kinds of rats, but also in another strain of rats. We have also looked at hamsters, and we get basically the same result. We see decreased sperm count, altered sexual behaviour, and shortened genitalia in these male rat pups. We have looked at both female rat and female hamster pups and we see even more dramatic changes in the females, where we see hypospadias, which is where the urethra, instead of emptying in a separate opening at the top of the clitoris, actually empties into the vagina. We see complete clefting of the clitoris, and a cleft being maintained all the way down to the vagina. In the rats, we see delayed vaginal opening and, in some cases, no vaginal opening. In the hamsters we cannot oven find an external vagina in some cases. Clearly, the animals with no vaginal opening are not going to be fertile. Although these animals appear to be cycling perfectly normally and the ovarian-pituitary axis appears to be functioning properly, we do not really understand the mechanism of what is going on and we are trying to explore it.
These are very concerning events and Dr. Guo from Taiwan, who is one of the principle investigators on the Yucheng cohort, visited my lab two days ago. This is the PCB rice oil poisoning in Taiwan in 1979, where about 2,000 peopleunfortunately cooked with rice oil that was contaminated with PCBs that were themselves contaminated with the polychlorinated dibenzofurans. The children born following this episode have been followed for the past 8-13 years, that is the age of the kids now. When they were first born they were reported to have what was called ectodermal dysplasia syndrome, which included all sorts of pigmentation problems, problems with their nails and dentition, and they were small in stature. When they did development milestones, these kids were developmentally delayed. They have continued to follow these kids. Their IQ is shifted about five points down from the rest of the population, and this has been maintained as they have grown up. It is not something they have outgrown. The children continue to be shorter in stature than matched controls and as the boys approach puberty, and some of them are now between the ages of 8-13, the ones who are 10, 11, 12 and 13 are apparently having problems with their genitalia. This is very new data, some of it will be presented this fall at the Dioxin Meeting, but it is very compatible with the data that we are seeing in the experiments.
Dioxin is a carcinogen (Table 5). There are at least 18 studies in mammals, all of which are positive. You may have heard that dioxin is a tumour promoter, and not a carcinogen, because it does not directly interact with the DNA. I think we start to dance on the heads of pins because when I am saying dioxin is a carcinogen here, if you feed animals in long-term studies without adding any known initiator, dioxin by itself still causes an increase in tumours. It does not cause only one type of tumour, it causes tumours at multiple sites. It causes it in both males and females, and it has been detected in rats, mice and hamsters. In addition, work from the U.S. EPA laboratory has indicated that dioxin causes increases of tumours in medaka, at multiple sites and short latency and at high incidence.
There are three recent human epidemiology studies (Table 6) which, I believe, deserve extra weight when we look at the dioxin epidemiology literature. Prior to these studies, there were probably equal numbers of studies that said, yes, dioxin does cause tumours in people and, no, it doesn't. The advances that these three (Fingerhut et al., 1991, Manz et al., 1991, and Zober et al., l990)* have are that they have blood levels for the cohorts. So they can actually validate their exposure in their industrial hygiene matrices with serum levels of dioxin. In that case you find an increased standardised mortality ratio related to exposure to dioxin, especially in the people who were exposed long term; people with at least 20 years of exposure. This was a generalised tumour response. I think you are all familiar with diethylstilbestrol (DES) that specifically caused vaginal adenocarcinoma in young women. The specificity of the lesion is why we were able to find out that this was a problem. If you have something that causes a generalised increase in cancer, it is very hard to pick up. There is a suggestion from two of these studies that there may be an increase in lung tumours. Well, with the background as high as it is in lung tumours, it is very hard to pick up a small number of extra cases. But in fact these studies are all very compatible with each other, showing that high levels of exposure to dioxin are associated with an increase in cancers overall. There is another study that will be published in the September issue of the American Journal of Epidemiology, based on the Seveso cohort. Seveso was a town in Italy, and in 1976 there was an explosion at a trichlorophenol plant and the area around it was highly contaminated. The serum levels in some of those people were the highest that we have ever seen. Until now there has been a suggestion in a report published in 1989 that there might be a increase in cancer, but it was just too soon, and the numbers were too small. This paper in press now actually demonstrates, based on cancer registry data in that area of Italy for 11 years since the explosion, there are very significant increases in multiple types of tumours in that population. Now that you think it is all bad, there is also a decrease in breast cancer. Remember I told you dioxin is a hormone, and it may increase some things and decrease other things. The decrease in breast cancer, by the way, has been reported in animal studies. In this report on the Seveso study to be published, the increase is in both males and females, and again at multiple sites. I really find it hard to accept the negative, or the null hypothesis. To me, the data is overwhelming that dioxin has the potential, at least at high doses, to result in cancer in people.
If we look at the dose-response relationships for dioxin (Table 7), interaction with the receptor will occur at the lowest concentration. The activation of the receptor interacting with DNA may require additional steps, and then all these other effects including enzyme induction, immunotoxicity, developmental effects occur at much lower levels than, for example, chloracne or cancer. You have to have very high levels of exposure to dioxin before you see chloracne. In cancer, in order to detect an increase in tumours, you have to have even higher doses again. With the immunotoxicity, one thing I should mention, is that in the Taiwan cohort, the children, not the adults, are reported to have elevated incidences, not only of respiratory infections, but also otitis, ear infections. In the northern Quebec Innuit population exposed to very high levels of PCB relative to the general population, children also have very high incidences of respiratory infections and otitis, and also a very decreased rate of take of vaccinations. All which would be at least compatible with the effects on the immune system.
A recent report presented at the American Society of Gynaecology has indicated that exposure to dioxin resulted in endometriosis in TCDD-exposed rhesus monkeys, many years after the cessation of exposure. These monkey were part of a cohort that was being studied at the University of Wisconsin. Seven years after the termination of exposure, one of the higher dose monkeys died after evincing severe pain and an autopsy revealed that it had fulminating endometriosis. Since then, one other monkey has died, and it died of the same cause. In monkeys endometriosis can he fatal, though it is not fatal in humans. There was a Canadian study out of Health and Welfare Canada which had reported, in abstract form, a suggestion of increased endometriosis in monkeys exposed to Arochlor 1254, which is the PCB with the highest concentration of dioxin-like PCBs in it. Because of all these findings, the Endometriosis Association paid for some veterinarians and some gynaecologists to do laproscoptic surgery on all the monkeys. There were controls, monkeys exposed to five, and monkeys that had been exposed to 25 parts per trillion of dioxin in their diet and they bid been exposed for four years, but that exposure had been terminated 10 years before this laproscoptic examination. There was a dose-related increase in both the incidence and severity of endometriosis, as compared to not only the control monkeys from this study, but to their historic controls from their monkey colony, which involves something like 300 monkeys. There are now at least two studies that are suggestive of an association, or an increased incidence of endometriosis in these monkeys and there is a possibility that it has fairly significant human application.
Dioxin is but one of a family of compounds including the naphthalenes, the dibenzofurans, the biphenyls, both the azo- and azoxybenzenes and then there were additional compounds. Remember, naphthalenes were commercially produced between World War I and World War 11, and the halowaxes were used to finish the nice wood flooring on your ships and, in fact, there were numerous incidences of chloracne occurring in some of the workers. These compounds are probably the major actors though, that when they are halogenated in the lateral positions, can interact with the Ah receptor and cause the same spectrum of biological responses.
These chemicals all act by the same mechanism, and we should be able to assign them relative potency rankings. The toxic equivalency scheme is a scheme of relative potency ranking. These compounds are considered as if they were a dilution of TCDD. So you weight them and if you assign TCDD a value of 1, you can see that the brominated dioxin is about 1/4th as toxic as TCDD itself. The tetrachlorodibenzofuran is only 1/20th as toxic despite looking like TCDD and binding very tightly to the receptor. It is much less persistent than TCDD and is readily metabolised in the mammalian systems, so the exposed animal can get rid of it, since metabolism for these chemicals is a detoxification process. The brominated-dibenzofuran is more toxic than the chlorinated, while for the dioxin it is the reverse, because the brominated dibenzofuran is harder to metabolise than the chlorinated. You can rank these compounds (Table 8) according to their relative effect. This was done for cleft palate, and it has been used in the development of toxic equivalency factors (TEFs) and many other kinds of end points have also been used. For example, receptor binding I have already mentioned. Induction of biochemical responses, like enzyme induction, and this can be done either in the animal in vivo or can be done in test tubes with cultured cells, teratogenicity, effects on the skin, dermal toxicity, immunotoxicity, or tumour promotion. All these things have been used, and looked at in total to come up with a toxic equivalency scheme. The U.S. EPA came out with interim TEFs in 1989, which I think is essentially identical to the NATO values also today being used by Scandinavian countries, and I think Canada uses the same numbers for the dioxins and furans.
The question is, what about the dioxin-like PCBs? The U.S. EPA is determining values for TEFs for the dioxin-like PCBs in fish, we have been looking at them in a long term mouse study, and Health and Welfare Canada has been looking at them in a long-term rat study. What we are finding is similar, which is that while the dioxin-like PCBs must be considered in any kind of assessment, in most cases they are not going to drive the reassessment. Congener No. 77 binds the receptor very well, but it is very rapidly metabolised in mammalian species and by many fish. So its in vivo toxicity is much less than you would predict based on something you did in culture. On the other hand, birds have very limited ability to metabolise these compounds, so you want to assess the toxicity for birds, using in vitro tests, while for species that can metabolise it, you may want to use in vitro. In other words, TEFs have to be applied with a certain degree of thought behind them.
So, at this point, I think in the general scientific community, there are three consensus points. In general I think the scientific community would say:
There are two approaches you can take to estimate dioxin risks. One is to use a biologically-based dose-response model, the other is to take any mathematical model. We prefer to put more science into the risk assessment process and develop models which are based on the mechanism of action of the compound in extrapolating from animal to human data, and from high to low dose. It is important to stress that there is no evidence that all receptor-mediated responses must be non-linear. In fact, there is no evidence that there is a threshold for responses such as interaction with the receptor and simple biochemical events. Now that says nothing about more complicated responses. I can't tell you yet whether tumour promotion or immunotoxicity or developmental toxicity may have thresholds or not. But I should stress that simple biochemical responses are occurring in the same range of body burden as developmental toxicity, and immunotoxicity. And in that range there is no evidence for a threshold or non-linear response. The other point I should make is that for those compounds we are not even starting from zero exposure. We all have these compounds in our body. There are lots of these molecules floating around in us. So we are not starting from an absence -- we are already somewhere up on the curve.
So, one approach is the modelling and the extrapolation of the experimental data. The other approach is to do some direct comparison of body burdens with responses in animals and humans. What we see are great similarities across species, and that there may be sensitive subpopulations, based on either exposure or susceptibility. Examples are subsistence fishermen who may eat much more fish and, therefore, have higher exposures, or nursing infants, since the only way we know of to reduce our body burdens from all these persistent lipophilic chemicals is, if you are a woman, have a baby. That is not a very cheerful thought, but what you are doing is partitioning the compounds out of your fats into the milk fat and eliminating them by the milk. During the short period of nursing, the infant will be exposed to much higher concentrations than for the rest of its life.
And then there is the issue of susceptibility. In the animal data, both the aquatic; data and the mammalian data suggest that the embryo/foetus is the most sensitive stage for the toxicity of these chemicals. We really don't know how sensitive the neonate is, who is getting this very large exposure. In most of the mammalian studies that I am more familiar with, the effects that are happening are occurring at the end of gestation in rodents of at the early neonatal period. But that entire range of development will all be in utero in humans.
Dioxin levels are normally expressed on a lipid or dosage basis. Because of the persistence, there is an age-associated increase in the levels. For TCDD, all of us sitting in this room have approximately 7 parts per trillion. If we look at the toxic equivalents for all the dioxins and furans, it will be about 30 parts per trillion in our bodies, and if we include the dioxin-like PCBS, that is going to be in the neighbourhood of 50 parts per trillion. These are background levels in humans in an industrial country such as the U.S. or Canada or Western Europe and Scandinavia, eating food from the grocery store. Nobody has looked at Eastern Europe and I bet there are areas where that is not going to be true. I was going to remind you that, of the 209 PCB congeners, only a very small number have dioxin-like activity, but the non-dioxin-like congeners have their own inherent biological activity and we would be mistaken if we ignored that, or if we thought that being protective of dioxin-like PCBs would protect us against the non-dioxin-like PCBS. Our current "dioxin" exposure is somewhere in the neighbourhood of maybe 30-50 parts per trillion on a toxic equivalency basis, derived primarily through the food. This kind of body burden of dioxins and furans is associated with exposure to about 1-3 picograms per kilogram per day. But nursing infants and subsistence fishermen may have higher levels of exposure.
So I think there are two views that can be looked at. The first one is, are current levels in the environment a problem? The second is that if we do not think they are a problem, should we worry about these special populations, the nursing infants and the subsistence fishermen? Do the current levels have the entire population on the brink of some kind of biological response? I am purposely vague, I am not saying adverse effect. Death is clearly an adverse effect, but is alteration of your hormonal status an adverse effect? I am not sure. It probably depends in what environment you find yourself. You know if your glucocorticoid levels are already elevated and then you are stressed, you might have a lot more problems than someone else. But are we at the level of beginning to expect responses? In fact, are there people in the population who are already experiencing subtle health effects? I should mention that male sperm count has dropped over 50% in the last 50 years, the incidence of endometriosis in the human population has increased dramatically, the age of menarche has decreased dramatically and this cannot all be accounted for by nutritional changes. I mean there are definitely things going on there. There is a recent suggestion that elevated levels of DDT are associated with increased incidences of breast cancer. Are there subtle things going on? I don't really know the answer to that but, if they are, then clearly, any increase of individual exposure would be undesirable.
Now, before I tell you what I think, I just would like to briefly mention the dioxin reassessment. About three years ago, there was a meeting on dioxin at the Banbury Center in Connecticut and, when we came back from it, I was really excited. I was also real naive. I had just joined the agency, so I wrote Erich Bretthauer a memo telling him I thought this was an opportunity for the agency to get in front of the issue, instead of always coming in at the rear, and that there was enough new information that had been gathered about the effects of dioxin that we really ought to reassess its risk. In fact, Bill Reilly, a little over two years ago, decided that we would do that and we started a multi-faceted approach including a bioaccumulation project and an aquatic toxicity project. There are many other parts of this reassessment. There has been evaluation of the literature, which has been on-going. This is a critical review of the new literature. There have been eight chapters written, and they have been done by outside experts in conjunction with U.S. EPA people as well. These were peer reviewed last September at a meeting and have been undergoing reevaluation and updating really since that time. There is also analysis of the exposure information. I should say there are actually eight to nine revised chapters that will be available shortly, the exposure scenario was also reviewed last September and that has been revised, and there are three volumes of exposure assessment. We have been looking more closely at human tissue levels, in collaboration with colleagues at the Center for Disease Control (CDC). We have been doing a lot of data collection, because what we wanted to do was try to develop extrapolation models that would be biologically based. There was some data that we felt we were missing, and we tried to identify the most sensitive responses that we could measure and obtain better estimation of the toxic equivalency factors for the coplanar PCBS. All these different kinds of bench and laboratory science have been ongoing and are feeding into a risk characterisation. We had originally hoped this would be done by now, but we have been held up by the epidemiology because all this new information which is just coming out, for example, the paper in press about cancer in Seveso, needs to be incorporated into the document. I think we are currently looking at public release of drafts of these documents probably in about October, which means that they would be taken to our external Science Advisory Board for review in probably January or February.
One thing we have tried to do in this reassessment is keep it a very open process. We have done lots of public meetings. We have solicited comments from the public. If anybody has additional comments, we are more than welcome to entertain and listen to it and incorporate it into the process. When we go out with the draft in approximately October, we will also hold a series of additional public meetings, probably in several regional offices across the country so that there will be greater access, as opposed to holding those kinds of meetings in Washington.
So, what were the results of last September's review? Well, the outside panel, at that point, said that the body burden in the general population was at or near the level where responses are expected to occur. I am going to editorialise, and I am not speaking for the agency, I am only speaking for me, but this sounds to me that this has a direct impact on a regulatory agenda.
David Villeneuve, Health and Welfare Canada: You have made reference to neurobehavioral effects in connection with the Yucheng study, I believe, and PCB and dibenzofurans. Is there any indication whatsoever that dioxins are perhaps implicated in neurobehavioral effects?
Linda Birnbaum: Yes, there is. Prenatal exposure of monkeys appears to result in altered spatial memory. You can tell monkeys that they are supposed to remember where a certain object is, and the dioxin-exposed monkeys have much more trouble with this short- term spatial memory than animals who were not prenatally exposed. Most of the Yusho and Yucheng data on responses is purely correlative, but if you try to associate them with the level of PCBs or the level of dibenzofurans they correlate with the dibenzofurans level much more than the PCB level for most of those things. So we know that certain PCBs, like some of the non-dioxin-like PCBs, are developmentally neurotoxic, but dioxin itself also appears to have a different spectrum of developmental neurotoxicity. Clearly, the sexual behaviour effects are neurotoxic effects, but they were induced developmentally.
Gerald Rees, Ontario Ministry of the Environment and Energy: Would you say that as a result of the dioxin reassessment we now have a better understanding of the implications of dioxins or are there more questions than when you started?
Linda Birnbaum: The answer is yes to both. Because if you do a good scientific study, you always raise more questions than you answer. But I think we have learned a tremendous amount. I personally feel that the identity between animals and humans is much stronger than it was couple of years ago. I mentioned the testosterone studies that have come out of the NIOSH cohort. There are also indications that highly exposed populations, not only the industrial workers but the small number of ranch-handers from Vietnam who were highly exposed, they also had the problems with testosterone and they also had problems with the glucose tolerance test and an increase in diabetes, and so did the NIOSH cohort. We don't know yet, and they are busy looking at that, whether it is Type I or Type II, Type I being auto-immune and Type II being age-associated. The ranch-handers also had increases in circulating immunoglobulin A (IgA), which would change the immune system, and increases in circulating lipids. Now those ranch-handers come back every five years for a follow up. They were back in '92. It will probably take 2-3 years before we have the results of that analysis. I think we are going to see a lot more information out of the Seveso cohort in the next couple of years. If we are going to do some more epidemiology studies, we need to look at the right population, and I don't think the right population is adult males. I think we need to be looking at adult females and we need to be looking at children born to women who were exposed and we need to follow those kids, especially for when they hit puberty.
Milton Clark, U.S. Environmental Protection Agency: One of the results that is coming out now is about the criticism of U.S. EPA's conservatism with dioxin relative to this non-threshold linear multi-stage model. Do you feel fairly comfortable we were on the right track some years ago, particularly when you look at the Fingerhut studies which show that when you extrapolate from the workers studies it models pretty closely the Kociba/Dow studies?
Linda Birnbaum: You are asking me a question I am not supposed to answer. Because until the reassessment is complete, we are not supposed to say anything. But I will say if you do the linearised multi-stage model, you come out with a certain number and if you now do a reference dose model, but look at the neurobehavioral or developmental effects, you come out with the same number, actually, even lower sometimes.
Milton Clark: But the Fingerhut extrapolation of the doses fall pretty closely on the cancer incidence.
Linda Birnbaum: If you take the human data. Well, you can do it two ways. You can take the Kociba data and predict the human response - what would be elevated risks -- and it comes out right in the ballpark of Fingerhut, Zober and Manz. You can take the Fingerhut data and predict what the animal tumour incidence would be, and it comes right out. In fact, the Fingerhut data would actually predict more tumours in animals than Kociba.
Tony Wagner, Environment Canada: Your last slide in your off-the-record comment, do some people take that as the first stop of organised brinkmanship?
Linda Birnbaum: I'm not sure I know what organised brinkmanship means.
Tony Wagner: We are at a critical stage either ban it or mount a massive cleanup.
Linda Birnbaum: The last slide was the conclusion of our outside peer panel. On the last day of the peer review, they actually sat there and they listed effects, and they listed body burdens that were associated with the effects in animals, and then they listed the body burdens associated with effects in people. I should really stress, this is a TEQ. If all you are worried about is dioxin, the levels of dioxin by itself, are probably not that high. But when you look at the sum total of what is out there, that is where the body burdens may be high enough so you might say we are having a response. Your question about ban it or not banning it I mean, what are you going to ban? We don't even know where half this stuff is coming from. My feeling, and this is purely my opinion, I have actually heard industry people say this, and this is totally off the record, is that the total level of halogenated aromatic organics in the environment is higher than they should be.
Denis Davis, Environment Canada: Do you have the historic body burden trends?
Linda Birnbaum: That is a really good question but we don't have the information. Well, if you go back to Eskimo mummies from around 1600, they were essentially non-detect. We don't have the information. There is a suggestion, Larry Needham told me about a month ago that the CDC is looking at the levels today. Some of these levels are based upon studies that were samples that were collected maybe 10 years ago, and they think the levels are a little bit lower. But we don't have enough numbers to really firm that up.
Doug Dodge, Ontario Ministry of Natural Resources: Is there any reason to look at aboriginal people in a different way than we would look at the rest of the general population, in terms of exposure and body burden? Were they lumped into that group?
Linda Birnbaum: Yes they are. We know that the Inuit's in Hudson Bay have much higher body burdens. Eric Dewailly from Quebec has actually measured PCB levels and he has done it on a congener-specific basis and they have about 10 times the TEQ.
Doug Dodge: But could that similarly be a purpose for special application in the Great Lakes, for example, where there are aboriginal people just as dependent upon natural resources as the Inuits?
Linda Birnbaum: Yes -- you people are much more familiar -- I know there is a Great Lakes Initiative where you are looking at levels, for example, for certain Indian tribes and certain subsistence fisherman. In New York State, they are looking at the Akwesasne tribe of Mohawk Indians, and trying to determine levels, but I don't think their levels are turning out to be higher than that of other people. The reason that the Inuits' levels are so high is because they eat sea mammals, and they eat blubber, and I think there have been some comparisons done with the Cree, who live in the same area, but they don't eat sea mammals. They eat caribou and stuff like that, and their levels are not elevated. But I think we are going to have questions of "what does it mean to be elevated." Is two-fold enough? I don't know the answer to that. But, if someone would say to me "what is a great big research need?" I would say that we need to develop some sort of better way to do bio-monitoring, because right now it costs $2,500 per sample to measure dioxins, furans and PCB levels. Well, you can't do a lot of samples and that is going to rapidly deplete everyone's resources. We need to develop some sort of methodology where we can get a measure which will be much more sensitive, and much more cost effective, so that we can get a better handle on "do we have populations where exposure is much higher than the general population."
Doug Dodge: Things like the neurogenesis could be racially different, could it not, so you look at different groups of people, you could get different effect levels.
Linda Birnbaum: I think that is an absolute possibility. When you deal with something like immunotoxicity, clearly it is a multi-genic factor. Many things feed into how the immune system works -- it is very easy to tickle it. I guess the Taiwanese data I find very important because they do have people who are ethnically matched, but didn't eat the rice oil. Much of the rice oil was eaten in a very limited province and at a school in that province, so they really do have pretty good match controls there, where that wouldn't really be much of a problem. I think it is a problem with the Inuit study to find the appropriate controls. The Cree, although they have a similar environment today in terms of their housing and so on, are ethnically different.
Doug Dodge: I ask this question because some policy makers in the Ontario Ministry of Natural Resources are under pressure from aboriginal people to provide them with a selected "cleaner" source of native, aboriginal-type foods and they are in special situations, because they have different reactions and we have been "using the white man" as the model of subsistence.
Linda Birnbaum: There is no doubt, for example, that many aboriginal peoples' diet is very different from the average white American or Canadian diet. At our first public meeting we had a chief from some tribe out west in Oregon, and basically said that the amount of fish that his people ate were at least 10 times what the U.S. Food and Drug Administration (U.S. FDA) estimates for fish consumption. So that obviously is going to have a major impact on what the exposure is and, very frankly, the exposure information is very limited. There has been only one market basket, and it wasn't even a complete basket study done in the entire U.S. There have been one or two done in Canada, so there are some really big holes.