Biologic Markers in Immunotoxicology (1992)

There is increasing awareness and concern within the scientific and public communities that chemical pollutants can suppress immune processes and thus cause increased development of neoplastic and infections diseases. Adverse effects on humans treated with immunosuppressive drugs, numerous studies employing experimental animals, and, to a lesser extent, isolated cases of altered immune function in humans inadvertently or occupationally exposed to xenobiotic substances support these concerns. There is no definitive evidence, as yet, that persons who live near contaminated sites or chemical-manufacturing plants have been immunologically compromised to the extent that they are at increased risk of disease. Nonetheless, there is reason to believe that chemical-induced damage to the immune system might be associated with pathologic conditions, some of which could become detectable only after a long latency. Likewise, exposure to immunotoxic xenobiotics can present additional risk to individuals with immune systems that are already fragile, for example, because of primary immunodeficiency, infancy, or old age.

Most of the experimental data on the effects of xenobiotics on immune function have been generated from animal models. The value of incorporating immunologic data for toxicologic assessment of drugs, chemicals, and biologics for evaluation of human hazard is increasingly accepted. However, as in other areas of toxicology, it is difficult to extrapolate change in a given area of immune function in experimental animals to the incidence of clinical or pathologic effects in humans.

One should not use such a term as ''chemical AIDS" in reference to chemical-induced immune dysfunction. Acquired immune deficiency syndrome (AIDS) is a well-defined disease of known viral etiology that bears no resemblance to potential chemical-induced immune-system changes. AIDS and the effects of commonly used immunomodulating drugs can be useful, however, as examples of the damage that can result from a compromised immune system in animals and humans. In addition, infection with the HIV-1 virus frequently occurs in persons concomitantly affected by other immunosuppressive agents, such as addictive drugs, malnutrition, herpesvirus-6, and Epstein-Barr virus. These agents could serve as cofactors that predispose an individual to HIV-1

infection, as well as confounding the resulting immune response.

The study of human immunodeficiency disease syndromes reveals a clear association between the suppression or absence of an immunologic function and an increased incidence of infectious or neoplastic disease. Numerous examples of such deficiency diseases have been reported and are well characterized in humans (Table 5-1). A deficiency in one or more immunologic functions can lead to severe, recurrent infections throughout life. These infections can be bacterial, viral, fungal, or protozoan, and the predominant type of infection depends on the associated immunologic lesion. Some infections can be treated with antibiotics or gammaglobulin, and in some cases the immunologic defect can be restored by bone marrow transplantation. However, other immunodeficiency diseases are much more severe. For example, children born with reticular dysgenesis have no white blood cells and usually die from infectious disease in the first year of life; children born with ataxia telangiectasia rarely survive past puberty. These diseases of genetic deficiency are more severe than those caused by environmental toxicants, because they are the result of the absence of part of the immune system. They demonstrate well-characterized consequences of immunosuppression. These same diseases would be expected to be associated with specific immunosuppression, whether the cause were genetic or environmental.

There are more than 60 inbred hybrid and mutant strains of rodents with well-defined immunodeficiencies (NRC, 1989c). Many of these animals have diseases that are comparable to the human immunodeficiency diseases listed in Table 5-1. An example is the beige mouse (the result of a recessive mutation on chromosome 13), which is a model for the human Chediak-Higashi disease syndrome. This defect results in reduced cell-mediated and natural killer immune function. Animal models have well-characterized, specific immunologic defects and known increased susceptibilities to infectious, neoplastic, and autoimmune diseases. Studies of human immunodeficiency diseases and the counterpart animal models emphasize the potentially serious consequences of immunosuppression, whether it occurs as a result of heredity, aging, or nutrition or is acquired as a result of exposure to xenobiotics.

For some time, immunosuppressive agents have been used in treating autoimmune diseases and as adjunctive therapy in organ transplantation procedures to prevent rejection by the recipient. Studies in this area have provided information on the potential clinical effects of chronic low-level immunosuppression. In addition, experimental studies with these compounds have provided comparative data between experimental animals and humans on immunosuppression that should have direct application to risk assessment.

Immunosuppressive treatments, such as x-irradiation, neonatal thymectomy, or the use of immunosuppressive drugs, result in an increased incidence of parasitic, viral, fungal, or bacterial infections. There is a well-established association between the therapeutic use of chemical immunosuppressants, such as those used in organ transplant therapy or in cancer chemotherapy, and an increased incidence of infections and neoplastic disease in humans (Ehrke and Mihich, 1985). For example, in a study of renal graft recipients undergoing immunosuppressive treatment, a 10-fold increased incidence of monoclonal gammopathies was observed (Radl et al., 1985). In another study, 50% of transplant patients developed cancer within 10 years after the operation (Penn, 1985). The tumors detected in these patients were heterogeneous and included skin and lip

tumors (21-fold increase), non-Hodgkins lymphomas (28- to 49-fold increase), Kaposi's sarcoma (400- to 500-fold increase), and carcinomas of the cervix (14-fold increase). Infections also are common in transplant patients on immunosuppressive therapy. In one study, 30% of cardiac-transplant patients treated with cyclosporin developed pulmonary infections within the first year after surgery (Austin et al., 1989). The most common pathogen was cytomegalovirus (11%), followed by Pneumocystis

carinii (10%) and Aspergillus (4%). It also has been postulated that the occurrence of some B-cell lymphomas and hepatocellular carcinomas in immunosuppressed individuals (e.g., patients with X-linked lymphoproliferative disorders) are due to their inability to control various viral infections, such as herpes simplex, human papilloma, and infections caused by Epstein-Barr virus (Purtillo and Linder, 1983). AIDS provides another example of the consequences of altered immune function in which the loss of immune responsiveness is associated with disease, most notably from Pneumocystis carinii and other opportunistic pathogens and the development of Kaposi's sarcoma, a rare form of cancer.

Infections that arise from immunosuppression will depend on the degree of suppression and potency of the infectious agent. The nature of the exposure can thus be an overriding factor in determining the nature of the infectious process. It is not surprising, for example, that the complex immune system can affect in several ways the equally complex development of cancer. Because the immune system normally provides a defense against viruses, the suppression of the immune system also can result in an increase in viral-oncogene-dependent tumors. Likewise, because tumors can generate specific nonself markers on their surfaces, similar to transplantation antigens, it is natural to expect some tumors that would normally be rejected by the host to develop under conditions of immunosuppression.

For purposes of illustration, the various effects of cyclosporin A (CsA) on the immune systems of humans and experimental animals will be compared with respect to effective doses. CsA is a fungal cyclic undecapeptide first described in 1976 (Borel et al., 1976). It is an immunosuppressive drug widely used to reduce transplant rejection, a cell-mediated immune response (Klaus and Hawrylowicz, 1984). The effects of CsA on graft survival and on the generation of an immune response have been studied in animal models of many species to predict the effects of an immunotoxicant on the human immune system. CsA has become the drug of choice in the clinical management of graft rejection, despite an increase in the risk of various infections. In humans, the daily clinical dose of CsA is 3-10 mg/kg of body weight for various indications; however, recent studies have shown that, independent of the route, the maintenance of plasma concentration of 250 nanograms/ml of CsA is needed to achieve resistance to organ rejection (Andrieu et al., 1988; Kerman et al., 1988; Martinet et al., 1988; Schmidt et al., 1988; Talal, 1988; Yocum et al., 1988; Baker et al., 1989; Sullivan and Shulman, 1989; Szer, 1989; Werner et al., 1989).

Although occasionally CsA does not alter the immune systems of patients (Chatenoud et al., 1989; Müller et al., 1989), the vast majority exhibit altered immune function. CsA reduces T-cell numbers and function during activation (Baker et al., 1989), and it inhibits the proliferative response of T cells and the secretion of interleukin-1, interleukin-2, and interferon (Kerman et al., 1988; McKenna et al., 1988; Pelletier et al., 1988). In HIV-seropositive patients, CsA reduces the number of peripheral blood CD8 (cytotoxic) T cells associated with the class I MHC (major histocompatibility complex) molecules and increases the number of CD4 (helper) T cells associated with the recognition of class II MHC molecules.

Most of the work examining CsA's in vivo effects on the immune system has been performed in rats and mice. However, in work with baboons and dogs, administration of CsA (20-40 mg/kg) inhibited the proliferative response of peripheral-blood lymphocytes to mitogen and antigen and prolonged pancreatic graft survival (Du Toit and Heydenrych, 1988; Kneteman et al., 1989; Zeitz et al., 1989). In rats, as in humans, administration of 5-30 mg/kg per day prolonged the survival of allogeneic grafts and reduced the severity of autoimmune disease (Bain et al., 1988; Tilney et al., 1988; Hall et al., 1989;

James et al., 1989; Oluwole et al., 1989; Ricordi et al., 1989; Rodrigues et al., 1989). Numerous studies show that in vivo administration of CsA can modulate the immune response in rodents. In rats and mice, administration of CsA reduced the medulla of the thymus, changed splenic morphology, and modulated T-cell proliferation, lymphokine secretion, and the accumulation of transplant-specific cytotoxic T cells (Hiramine et al., 1988, 1989; Muthukkumar and Muthukkaruppan, 1988; Orosz et al., 1988; Tanaka et al., 1988; Yoshimura et al., 1988; Armas et al., 1989; Fukuzawa and Shearer, 1989). Administration of CsA affected the susceptibility to infection with murine cytomegalovirus (MCMV) by reducing the immune response (Selgrade et al., 1989). Dean and Thurmond (1987) have compiled the results of immunotoxicity studies of CsA from a number of experimental animal species used in toxicology (Table 5-2). Comparative studies indicate that rats are slightly more susceptible to the immunosuppressive effects of CsA than are mice, but the consistency was good between the test species.

Although exceptions exist, rodent studies have been predictive for both the type of immune effects and the dose required to achieve the effects for humans given various immunosuppressive therapeutic drugs in clinical trials. The most notable exceptions are glucocorticoid steroids, which are thymolytic in rodents, but not in humans. Although more comparative immunotoxicity studies are required, rodents are now the most appropriate animal model for examining the immunotoxicity of non-species-specific compounds, according to established similarities of toxicologic profiles and ease of performing immunologic evaluations. It remains critical, however, to establish the pharmacokinetic and metabolic properties of the test compounds in humans and experimental species.

Animal data on CsA can be applied to predict immune-system effects in humans. Understanding of the mechanisms of action of immunotherapeutic agents, such as CsA,

and the relationship between the effective biologic dose and the various immunologic values, therapeutic effects, and adverse effects will provide a better basis for extrapolation of animal data to human risk. Many animal studies have shown that high-dose exposure to environmental contaminants and other nontherapeutic agents can modulate the immune system. Adverse effects after low-level chronic exposure remain to be confirmed.

Although they are not as well established as are the effects of therapeutic drugs, environmental toxicants are implicated in a number of reports that describe increased rates of neoplastic disease or infection in humans associated with immune-system changes. For example, a cluster of Hodgkin's diseases was reported in individuals from a small town in Michigan (Schwartz et al., 1978). Reduced CD4:CD8 ratios and natural-killer-cell function have been reported in asbestos workers (Lew et al., 1986). Another unconfirmed report describes a 4-year study of workers engaged in the manufacture of benzidine, a human bladder carcinogen, and suggests that individuals with depressed cell-mediated immunity (as shown by skin tests) demonstrated precancerous conditions and subsequent neoplasms (Gorodilova and Mandrik, 1978). On the other hand, no cases of neoplastic diseases were registered in workers with normal immunologic responses.

There are a number of better-substantiated clinical studies. For example, there have been observations of abnormal antibody production, prolongation of allograft rejection, and decreased resistance to disease in humans occupationally exposed to silica (Uber and McReynolds, 1982). A series of reports from Taiwan have described immunologic changes (Yu-Cheng diseases) in individuals exposed to rice oil inadvertently contaminated with polychlorinated biphenyls and dibenzofurans (Lee and Chang, 1985). These individuals primarily demonstrated altered T-cell function and increased rates of sinopulmonary infection.

Because considerable in vivo and in vitro data have been accumulated on the immunotoxic effects of lead, it can serve as an example of the way metals can affect the immune system. The adverse effects of lead vary with age. In the young, central nervous system function is the major target, whereas in adult workers, renal toxicity and anemia predominate. Because lead has a long half-life, the body burden takes considerable time to reach its maximum with a constant intake. The body burden, over time, may be a better estimate of the biologically effective dose than either the exposure concentration or the blood lead concentration. There are several reports that lead alters immunity in humans. Lead workers with elevated blood lead had decreased levels of salivary IgA and an increased incidence of colds and influenza (Ewers et al., 1982), impaired mitogen responses to phytohemagglutinin (Jaremin, 1983), and increased numbers of suppressor T cells (Cohen et al., 1989). Children with high blood lead concentrations have been reported to have diminished levels of IgM, IgG, and IgA (Wagnerová et al., 1986); others infected with Shigella enteritis had prolonged diarrhea (Sachs, 1978). There also are reports that elevated concentrations of lead do not alter immunity in humans (Reigart and Graber, 1976; Kimber et al., 1986). Nevertheless, it appears that high blood lead concentrations can damage the immune system.

There are several examples in which immunologic changes have been ascribed to exposure to various environmental pollutants, but have not been associated with any clinical changes. For example, women who have been chronically exposed at low levels to groundwater contaminated with the pesticide

aldicarb exhibited an altered number of T cells, including decreased CD4:CD8 ratios (Fiore et al., 1986). Immunologic effects also have been reported in individuals exposed to methyl isocyanate, an intermediate in the production of carbamate pesticides, after an industrial accident in 1984 in Bhopal, India (Deo et al., 1987). The effects of immune response included an increase in the number of CD4 and total T cells, but decreases in lymphocyte mitogenesis. Several persistent immune alterations, also primarily cell-mediated, have been reported in Michigan residents who ingested dairy products contaminated with polybrominated biphenyls (Bekesi et al., 1978), although Silva et al. (1979) were unable to detect any immune abnormalities in a similarly exposed cohort. Individuals occupationally exposed to 2,3,7,8-tetrachlorodibenzop-dioxin (TCDD), which is one of the subregistries established by the ATSDR (Jennings et al., 1988), also demonstrated immune changes. Jennings showed that antinuclear antibodies and immune complexes were detected significantly more frequently in the blood of dioxin-exposed workers. In addition, the number of leukocytes in peripheral blood cells was elevated in dioxin-exposed workers. Some of these changes were not confirmed in a later study (Evans et al., 1988), perhaps because of a technical error; other observations were confirmed. Benzene-induced pancytopenia, a classic symptom of chronic benzene exposure, is an immunodeficiency disease by virtue of the reduced number of immunocompetent cells that result from altered marrow function (Snyder, 1984). In fact, lymphopenia is common after exposure to organic solvents (Browning, 1965; Capurro, 1980). Alterations in the number of some cell types (decreases in CD3 and CD4) were reported to occur in solvent-exposed workers (Denkhaus et al., 1986); the effects might have some specificity.

In addition to environmental chemicals, a large number of therapeutic substances, as well as abused recreational drugs, can alter immune function in humans. Among these are diphenylhydantoin, ethanol, cocaine, and isobutyl nitrites (Newell et al., 1984; Specter et al., 1986).

Although most of the xenobiotics that alter the human immune system also can affect experimental animals, clinical studies often have been criticized for incomplete or inconsistent diagnosis of immunodeficiency, lack of clinical changes, small group size, inability to establish exposure levels, or lack of reproducibility. As might be expected, there are several studies in which immune functions were not shown to be affected after the subjects were exposed to presumably high levels of chemicals, even though the agents, such as heavy metals and TCDD, are known to be immunotoxic in animals (Reggiani, 1980; Kimber et al., 1986). Furthermore, there is no evidence that xenobiotics can influence immune systems in the general population (except through occupational or inadvertent exposure).

The question of population-wide immune system effects would be difficult to answer, because the immunologic effects one might expect in the general population would likely be subtle. The tests routinely used for clinical assessment of immunosuppressed function in humans are not very sensitive. The difficulty in identifying a recently exposed, well-defined cohort is substantial. There is considerable immunologic variability in the general population. A consensus exists that further clinical studies with better-defined cohorts and more sensitive tests will be required to assess the true potential of xenobiotics to affect human health.

A growing body of research in immunotoxicology has shown that many xenobiotic substances cause immunosuppression in laboratory animals (Table 5-3). Immunologic effects often are accompanied by increased susceptibility to challenge with infectious

agents or tumor cells. Animal studies, primarily those which use rodents, have provided a large information base about potentially immunotoxic chemicals, suggestive evidence of the mechanisms for their effects, and an appreciation that the immune system is susceptible to chemical injury. The susceptibility of the immune system is due as much to the general properties of a chemical (e.g., its reactivity to macromolecules) as it is to the complex nature of the immune system. Because the cellular events responsible for immune processes also are involved in embryogenesis, many immunosuppressive xenobiotics would be expected to be developmental toxicants.

To date, animal data have not been used to any significant extent in the assessment of human risk resulting from exposure to immunosuppressive environmental pollutants. Extrapolating short-term, high-dose animal studies to chronic low-dose human exposure

is always a problem. Moreover, this problem is compounded in that the end points (infection and neoplasia) of immune-system deficiency are secondary to other confounding factors. To use immunotoxicity data from animal studies in the quantitative assessment of risk (the prediction of the incidence of disease at a given human dose), we would have to be able to predict the disease incidence that results from a given degree of immunodeficiency. However, animal immunotoxicology is now at a point where it should be used in identification of pollutants that have the potential to induce immunodeficiency and in the estimation of the degree of hazard that environmental agents have, according to their immunosuppressive potency. Because of the potential for tragic outcomes associated with exposure to immunosuppressive environmental pollutants, prudent public health concerns indicate the need for a remedy. The following discussion concerns some of the environmental toxicants shown to alter immune function at doses at which toxicity in other organ systems is not readily apparent.

Probably the most extensively studied class of environmental pollutants are halogenated aromatic hydrocarbons (HAHs), including dibenzo-p-dioxins, dibenzofurans, polychlorinated biphenyls (PCBs), and polybrominated biphenyls (PBBs) (Vos and Luster, 1989). These compounds, many of which are widespread in the environment, are primarily used in commercial production of industrial chemicals, pesticides, flame retardants, and heat conductors. Dioxins and PCBs produce myelosuppression, immunosuppression, thymic atrophy, and inhibition of immune complement system components in almost all species tested, including primates. The most potent dioxin, TCDD, is an extremely potent immunosuppressant in mice. A dose of 1-2 µg/kg of body weight is all that is required to reduce immune function by 50%. As probably occurs with a number of immunosuppressive HAHs, the specific effects of TCDD on the immune system can vary, depending on the age of the animal at the time of chemical exposure. For example, the primary effect of perinatal TCDD exposure is persistent suppression of cellular immunity, a condition that mimics neonatal thymectomy. In contrast to perinatal exposure, TCDD exposure in adult mice, while still inducing deterioration of thymic tissue (predominantly cortical lymphoid depletion), causes a transient antiproliferative response in rapidly dividing cell populations, including hematopoietic cells and B cells. The marked and persistent suppression of T-cell function seen in neonates is not manifested in adults, although suppression of cytotoxic T-cell response and altered members of regulatory T cells have been reported.

Regardless of age at the time of exposure or the target tissue examined, immunosuppression by TCDD, as well as by PCBs, is believed to be mediated through stereospecific and irreversible binding to an intracellular receptor protein (the Ah genotype) found in the cellular targets for TCDD, including lymphoid tissue, bone marrow cells, and the thymic epithelium (Thomas and Faith, 1985). The Ah genotype was determined primarily in immune studies comparing inbred strains of Ah-responsive and nonresponsive mice in which the ability of TCDD to cause immunotoxicity correlated with the presence of the Ah locus (Vecchi et al., 1983; Tucker et al., 1986). In addition, a good correlation exists between the binding affinities of various HAHs and their ability to induce immunotoxicity (Silkworth et al., 1984; Tucker et al., 1986).

The role, if any, of microsomal enzyme induction in the cellular mechanisms responsible for immunotoxicity after TCDD binds to its receptor is unknown. The observation that the thymic epithelium contains a high concentration of receptor has led to the

suggestion that TCDD causes maturational defects in developing thymocytes via inadequate epithelial support (Greenlee et al., 1985). There is evidence, however, that alterations other than that of thymic epithelium function are responsible for immunosuppression. These alternative mechanisms include direct thymolysis events (McConkey et al., 1988), alterations in regulatory T-cell function (Kerkvliet and Brauner, 1987), alterations in B-cell differentiation (Luster et al., 1988), and stem cell inhibition (Fine et al., 1989).

Polycyclic aromatic hydrocarbons (PAHs), another well-studied class of compounds, are formed as products of incomplete combustion of fossil fuels, tobacco, and coke and in automobile exhaust. Selected PAHs, including benzo[a]pyrene (B[a]P), 7,12-dimethylbenzanthracene (DMBA), and 3-methylcholanthrene (3-MC), have been shown to produce immunosuppression and myelotoxicity; they also are carcinogenic (Dean et al., 1986). There is controversy about whether the immunosuppressive properties of PAHs are a cofactor for carcinogenicity, because they allow tumor antigen (neoantigen) to bypass normal host immune surveillance. Humoral immunity is suppressed after exposure to several PAHs, including B[a]P, DMBA, and 3-MC, although humoral immunosuppression after DMBA exposure could reside at the level of T-cell regulation (Dean et al., 1990). The finding of reduced progenitor B cells in exposed animals suggests that B cells also are directly targeted early in their maturation (Ward et al., 1984). B[a]P has been shown to impair the production of interleukin-1, implicating chemical-induced defects in accessory cell function as a contributing factor in decreased production of antibody-forming cells (Lyte and Bick, 1986). Cell-mediated immunity also is inhibited by PAHs. Cytotoxic T-cell activity in mice is suppressed after in vitro and in vivo exposure to DMBA or 3-MC (Wojdani and Alfred, 1984; Dean et al., 1985a, 1986).

Hexachlorobenzene (HCB), a fungicide and an intermediate in a variety of chemical syntheses, has been shown to have immunosuppressive properties in rodents. HCB suppresses both cellular and humoral immunity in adult mice (Loose et al., 1978) and is particularly toxic to T-cell function after in utero exposure (Barnett et al., 1987). HCB stimulates immune responses in rats after adult exposure (Vos and Luster, 1989) and perinatal exposure (Vos et al., 1983).

The immunotoxicity of benzene, for which a subregistry has been established at ATSDR, has been the subject of considerable research in humans and animals, with particular emphasis on its hematologic and leukemogenic potential (Snyder, 1984). In humans and experimental animals, the predominant hemopathy associated with benzene exposure is pancytopenia, with associated bone marrow hypoplasia (Laskin and Goldstein, 1977). Although hematopoietic progenitor cells are particularly susceptible to benzene, the mature circulating lymphocyte also responds to benzene in an antiproliferative response (Snyder et al., 1980). Chronic benzene exposure has been associated with depressed serum levels of IgA, IgG, and immune-system complement in humans. Despite the myelotoxic effect produced by benzene in experimental animals, the exact relationship between benzene-induced immunosuppression and an increase in leukemia has not been determined. Using pharmacokinetics to determine amounts of total benzene metabolized in 24 hours, Beliles and Totman (1989) have established the biologically effective dose. Using pharmacokinetics as a basis for across-species extrapolation of experimental data (benzene produces leukemogenesis at dose levels of 100-300 ppm in animals), they were able to show that the human risk of leukemia estimated from the incidence of leukemia in mice was quantitatively comparable to the risk predicted from occupational epidemiologic studies.

A large experimental data base exists on the immunosuppressive properties of metals, their inorganic salts, and organometallic compounds (Koller, 1980). Inorganic substances, including heavy metals, such as lead, cadmium, mercury, and nickel, have been studied, and the effects of exposure have been reported to range from suppression to enhancement of the immune system. Some metals are reported to induce autoimmune disease. Mercury and to a lesser extent gold and lead have been shown to induce polyclonal B-cell activation, and considerable effort has been devoted to understanding this autoimmune phenomenon (Bigazzi, 1988). Heavy metals also are sulfhydryl-alkylating agents and as such bind with high affinity to cellular sulfhydryl groups. This suggests that, at sufficient dose levels, these compounds can interfere with normal cell-to-cell communication by altering various membrane properties, thus causing immunosuppression. This observation is strengthened by the observation that the suppressive effects of lead, at least in vitro, can be reversed by the addition of exogenous thiol reagents. Organotin compounds, used as heat stabilizers, biocides, and industrial catalysts in the production of foams and rubber, also are immunotoxic in rats. These compounds target the thymus, causing severe thymic atrophy and suppression of cell-mediated immune responses (Seinen and Penninks, 1979).

Lead, in general, tends to suppress immunity in animals. It is well accepted that lead increases host susceptibility to numerous infectious agents and increases both tumor growth and development in animals (Koller, 1990). Lead also impairs cell-mediated immunity, particularly the delayed-type hypersensitivity response (Müller et al., 1977; Faith et al., 1979). Although lead affects IL-2 (interleukin-2), it does not appear to inhibit interferon (Gainer, 1974; Blakley et al., 1982; Exon et al., 1985). Furthermore, natural-killer-cell activity appears to be unaffected by lead (Talcott et al., 1985). Some macrophage activities, such as phagocytosis, oxygen metabolism, vascular clearance, and antigen presentation, are suppressed by lead. Mitogen responses are quite variable in animals exposed to lead (Lawrence, 1981; Burchiel et al., 1987; Koller, 1990). In vitro exposure to lead results in increased plaque-forming-cell responses; it stimulates the proliferation of immune cells; and it inhibits microbial killing, antigen presentation, and propagation of bone-marrow-derived macrophages. It has no effect on IL-2-receptor expression, on phagocytosis, or on IL-1 production, but it results in an increased density of Ia antigen (the murine class II MHC antigen) on B cells (Lawrence, 1981; Kowolenko et al., 1988; Koller, 1990). The differences in results obtained with lead in animal studies usually can be credited to variations in species, strains, or experimental metallothione levels, induction of which is highly variable among species and strains (Ohsawa et al., 1986). It is noteworthy that oral intake of lead by rats at a dose of 10 ppm, which does not otherwise damage their health, has been demonstrated to suppress immunity (Koller et al., 1983a,b). Thus, the immune system of rodents appears to be a sensitive target organ of lead toxicity. This may explain why rats tend to develop renal cancer after chronic administration of lead, whereas nephrotoxicity in humans does not progress to neoplasia.

Most of the animal studies discussed above involved the administration of doses of toxicants at levels that are higher than those normally found in the environment. Although many of the chemicals are immunotoxic at doses that do not produce other toxicities, there is still the question of whether the general public is at risk of immunosuppression, particularly after chronic low-level exposure to chemical mixtures. One study on the immune system of mice has

examined the effects of a complex mixture of 19 organic and six inorganic environmental toxicants commonly found in contaminated groundwater at hazardous-waste sites (Germolec et al., 1989). The authors note that although the chemical configuration does not actually occur in the environment, it is representative of a highly contaminated groundwater sample. The mixture (Table 5-4) was administered to the mice in their drinking water over a period of 14-90 days and was found, at least at the high levels, to suppress the number of granulocyte-macrophage progenitor cells in the bone marrow, the generation of specific antibody-forming cells in the spleen, and the ability of the mice to fight an infection of Plasmodium yoelii, although it did not affect other indicators of targetorgan toxicity. The authors note that although several of the components of this complex mixture had been shown to have similar immunologic and myelotoxic effects, earlier studies with the individual components suggest that none of the individual contaminants was present at sufficient concentrations to be solely responsible for the observed effects. Therefore, the effects of the complex mixture of groundwater contaminants on the immune system are likely to result from a composite of influences of the individual components. Controlled studies that examine the effects on the immune systems of laboratory animals of complexes of environmental contaminants found at environmentally relevant levels should help provide an understanding of the real immunologic risk of living in a contaminated environment.

There are many compounds for which there are no human data, but for which there are confirmatory animal data. Among these are dimethylnitrosamine (Thomas et al., 1985), phenolic food additives (Archer and Johnson, 1978), certain pesticides (O,O,S-trimethyl phosphorothioate—a contaminant in malathion), and several other organophosphate pesticides (Rodgers et al., 1985). For a complete listing of these putative immunotoxic xenobiotics and specific references, the reader is directed to more detailed books and reviews (Koller, 1990; Gibson et al., 1983; Dean et al., 1985a, 1986; Luster et al., 1987; Vos and Luster, 1989; Rodgers et al., in press).

Inhaled pollutants can enhance susceptibility to or severity of viral, bacterial, and neoplastic diseases. This could be related to, or a direct consequence of, the impaired immune function. Pulmonary immunocompetence has been assessed after exposure to phosgene and ozone. Exposure of rats to 1.0-ppm phosgene for 4 hours results in a significantly enhanced and prolonged pulmonary influenza virus infection (Ehrlich and Burleson, in press). Influenza-virus-specific cytotoxic T-lymphocyte activity in the lungs is suppressed after exposure to phosgene gas (Ehrlich et al., 1989). Phosgene inhalation also results in suppressed pulmonary natural-killer-cell (NK) activity (Burleson and Keyes, 1989), which remains suppressed through day 4 and returns to normal on day 7 after exposure. Phosgene inhalation suppresses NK activity in the spleen. It is therefore important to assess systemic immunotoxicity after inhalation exposures.

Suppressed pulmonary NK activity was observed after uninterrupted ozone exposures (23.5 hours/day) at 1.0 ppm for 1, 5, or 7 days. The suppressed immune function returned to normal 10 days after exposure stopped (Burleson et al., 1989). Van Loveren et al. (1990) report that inhalation exposure to ozone for 7 days at 0.2 or 0.04 ppm stimulates NK activity, but exposure at 0.8 ppm suppresses pulmonary NK activity.

Subchronic inhalation exposure to isobutyl nitrite results in decreased thymus weight and decreased white-blood-cell counts.

Injection or inhalation of isobutyl nitrite suppresses splenic and peripheral blood NK activity (Lotzová et al., 1984). Subchronic inhalation exposure to isobutyl nitrite also results in decreased thymus weight, decreased liver weight, decreased white-blood-cell counts, mild focal hyperplasia, and vacuolization of the epithelium lining bronchi and bronchioles (Lynch et al., 1985).

Arisine gas, used in the manufacture of semiconductor microchips, was used for inhalation exposure in mice to test for systemic immunotoxicity (Rosenthal et al., 1989). The exposure caused several hematopoietic effects: a decrease in erythropoiesis, as indicated by a reduction in erythroid-colony-forming units and femur cells; splenomegaly; and decreases in red blood cells, hematocrit, and hemoglobin with increases in white-blood-cell counts and mean corpuscular volume of red blood cells.

Alveolar macrophages provide an important

first line of defense and thus interact with inhaled xenobiotic substances and microorganisms. Alveolar macrophages exposed to xenobiotics can impair immunologic function or participate directly in disease induction. Alveolar macrophages from rabbits exposed to ozone at 0.6-0.95 ppm for 3 hours exhibit a suppressed ability to phagocytize streptococci (Coffin et al., 1968). In mice, ozone exposure at as little as 0.08 ppm for 3 hours results in an enhanced mortality of those exposed to group C streptococci (Coffin and Gardner, 1972; Miller et al., 1978). Suppressed pulmonary bactericidal activity was reported after a 4-hour exposure to ozone in mice (Goldstein et al., 1971). The ingestion of mineral dusts, such as quartz and asbestos, results in pulmonary fibrosis. Phagocytosis of asbestos by alveolar macrophages results in an increase in IgG receptor sites, enhanced ability to spread across a glass substrate, and more extensive cytoplasmic processes (Miller and Kagan, 1976). The phagocytic activity of alveolar macrophages is suppressed after inhalation of nitrogen dioxide (Gardner et al., 1969; Acton and Myrvik, 1972; Suzuki et al., 1986), which also results in impaired alveolar macrophage bactericidal activity (Valand et al., 1970) and increased susceptibility to Klebsiella pneumoniae (Ehrlich and Henry, 1968), Streptococcus pyogenes (Gardner et al., 1977), influenza virus (Henry et al., 1970; Ito et al., 1971), and cytomegalovirus (Rose et al., 1988).

Various compounds have been studied for their role in altering humoral immunocompetence in the lung by measuring the plaqueforming-cell (PFC) response to sheep red blood cells (SRBCs) as a measure of pulmonary humoral immunity. Chronic inhalation of cigarette smoke inhibits the PFC response of lung-associated lymph node cells to SRBCs (Sopori et al., 1989). Fly ash from a fluidized-bed combustor, fly ash from a pulverized-coal combustor, and quartz particles administered to Fischer-344 rats all significantly increased the number of lymphoid cells in the lung-associated lymph nodes. In contrast, quartz and fly ash from a pulverized-coal combustor suppressed the PFC response to SRBCs, but fly ash from the fluidized-bed combustor had no effect (Bice et al., 1987a). Exposure to nitrogen dioxide (Schnizlein et al., 1980), B[a]P (Schnizlein et al., 1982), and diesel exhaust (Bice et al., 1985) enhanced the PFC response to SRBCs in lung-associated lymph node cells.

Bronchoalveolar lavage fluid (BALF) contains numerous biochemical indicators of lung damage, including cells, antioxidants, proteins, enzymes, cytokines, growth factors, arachidonic acid, and metabolites. BALF analysis has been used to assess the pulmonary toxicity of many inhaled pollutants (Henderson et al., 1979a,b, 1985a,b, 1986, 1988; Henderson, 1984, 1988, 1989). BALF thus provides biologic markers that allow direct extrapolation of results from animal studies to investigations on human subjects. Most cells in BALF consist of alveolar macrophages. However, neutrophils provide an indicator of inflammatory response to infection and to inhaled xenobiotics.

Exposure of human subjects to low concentrations of ozone results in an increase in the appearance of neutrophils in BALF (Koren et al., 1988). Arachidonic acid metabolites have been measured in rat BALF after exposure by inhalation to phosgene. Phosgene exposure results in decreased levels of prostaglandin E-2 and leukotrienes in rat BALF. Production of arachidonic acid metabolites was measured by rat and human alveolar macrophages after in vitro exposure to phosgene (Madden et al., 1991). Monomethylhydrazine and dinitrogen tetroxide are toxic irritants in rocket fuel that result in pulmonary edema (DeJournette, 1977) and should be evaluated for pulmonary immunotoxicity.

The above categories of pulmonary immunologic responses can provide information to evaluate the immunotoxicity of inhaled chemicals and to compare sensitivity across

species for use in risk assessment. This cascade of pulmonary immune-system functions can now be used to obtain meaningful results for the extrapolation of data from animals to human risk assessment.

Several studies have established that dermal exposure to some chemicals, drugs, or environmental conditions can lead to selective immunosuppression in laboratory animals (Cooper et al., 1985; Furue and Katz, 1988; Halliday et al., 1988). The list includes ultraviolet light, cold, CsA, DMBA, and various tumor promoters. Immunosuppression is primarily characterized in the animals as an inability to elicit contact hypersensitivity responses to topically applied skin-reactive chemicals, although more generalized immunosuppression has been observed. The mechanisms for these events are unknown but appear to be associated with inhibition of accessory cell function of Langerhans cells at low dose levels and induction of suppressor T cells at higher dose levels. The relevance of these changes regarding actual disease, such as skin tumors or infections, is unknown.

Blood-cell lineages are derived from pluripotent cells, which in adults are primarily found in the bone marrow. Within the marrow microenvironment, these self-renewing cells mature into committed progenitor cells, which can be found in peripheral blood and tissues. The continued development of these cells is under the control of various growth factors, many of which originate in bone marrow stromal cells, which also provide a supporting matrix for development of hematopoietic cells. Hematopoietic cells involved in the immune system include monocytes, granulocytes, and lymphoid precursor cells. A variety of evidence, including the use of long-term bone marrow cultures, has demonstrated the importance of the microenvironment in regulating myeloid and lymphoid development. Bone marrow is a sensitive target for therapeutic drugs and environmental toxicants, most likely because stem cells turn over rapidly. In fact, bone marrow hematotoxicity is the dose-limiting toxicity in many patients in antineoplastic or immunosuppressive therapy.

Although not as sensitive as in vitro proliferative assays, alteration in bone marrow cellularity and differential cell counts are used as indicators of hematotoxicity. The quantitation of pluripotent hematopoietic stem cells (PHSCs) is determined in rodents by in vivo assays of spleen colony-forming units (CFU-S). More committed progenitor cells can be identified and enumerated operationally by their ability to proliferate in the presence of specific colony-stimulating factors (CSF). In vitro assays also have been developed to monitor the formation of stromal-cell-dependent colonies. These assays can establish selective effects on the stromal microenvironment, growth factor production, and direct stem cell proliferation.

A linear relationship between hematotoxicity and immunosuppression has not been established, although it is generally agreed that inhibition of normal lymphoid or myeloid stem cell development will be manifested as altered immune function, provided that the alteration is of sufficient magnitude or duration. In this respect, the toxicity of several environmental contaminants has been attributed, at least in part, to their direct effect on the hematopoietic system. For example, blood dyscrasia, including leukopenia and bone marrow hypoplasia or aplasia, and acute myelogenous leukemia have been attributed to benzene exposure (Laskin and Goldstein, 1977), and some benzene metabolites inhibit bone marrow hematopoiesis in vitro. Benzene also is carcinogenic in

animals and can cause leukemia at doses of 100-300 ppm. Heavy metals, such as arsenic, cadmium, copper, gold, iron, lead, and zinc, also cause hematologic changes in humans. Lead, for example, causes an increase in the number of reticulocytes and a concomitant increase in basophilic stippling (Albahary, 1972). Chronic lead exposure can lead to anemia.

In laboratory animals, many halogenated aromatic hydrocarbons, including chlorinated dibenzo-p-dioxins and biphenyls, produce myelotoxicity at relatively low dose levels (Vos and Luster, 1989). PAHs, such as B[a]P and DMBA, also produce myelotoxicity in laboratory animals but require higher dose levels (Legraverend et al., 1983). Other chemicals shown to suppress hematopoietic function in laboratory animals include asbestos, selected mycotoxins, glycol ethers, and organophosphate pesticides. Clinical case studies have suggested an association between occupational exposure to organophosphate pesticides and hematotoxicity in humans (Jenkyn et al., 1979).

Establishing a relationship between potential immune-function changes and actual diseases in humans has been difficult and controversial. For example, Lagakos et al. (1986) reported a high incidence of leukemia and recurrent infections in children of East Woburn, Massachusetts, exposed to drinking water contaminated with industrial solvents. Trichloroethylene (TCE) was the primary contaminant (267 ppm), but lesser amounts of tetrachloroethylene (21 ppm), 1,2-trans-dichloroethylene, and 1,1-trichloroethylene were found. Because the analysis was performed several years after the contamination occurred, the initial levels are unknown, but they are assumed to have been higher; a decrease of several orders of magnitude in 2 years in another well contaminated with TCE has been shown (Landrigan et al., 1987). Fagliano et al. (1987) have reported an increase of leukemias in a population in New Jersey exposed primarily to elevated concentrations of TCE, tetrachloroethylene, and other trihalomethanes in drinking water. Increased childhood leukemias have been observed in a population in Arizona exposed to TCE at detected levels of 8.9 and 29.0 ppb (Flood et al., 1989). However, the investigators state that available information was insufficient to conclude that there was a relationship between the environmental contaminants, the population exposure, and the observed childhood leukemia mortality. Studies by Byers et al. (1988) of family members in the East Woburn group demonstrated an increased number of individuals with altered ratios of T-cell subpopulations, autoantibodies, infection, and recurrent rashes. TCE was incriminated as the most likely contaminant, because Sanders et al. (1982) reported that TCE in the drinking water of mice suppresses both humoral and cell-mediated immunity. Dichloroacetic acid, a metabolite of TCE, was reported by Katz et al. (1983) to increase lung lesions caused by parasites in dogs. Unfortunately, as with most epidemiologic investigations, neither the concentrations at the time of exposure nor the effective biologic dose have been estimated or measured. Recently, the Agency for Toxic Substances and Disease Registry has established a registry for persons exposed to TCE (Burg, 1990). This registry could be useful in establishing a basis for the verification of the suspected relationship between animal bioassay results and the human health effects of TCE.

Some groups of individuals could be at increased risk of developing immunosuppression

as a result of exposure to environmental contaminants. Exposure in utero, when the immune system is developing, could have long-term effects on the ability of an individual to generate an immune response (Osburn and Schultz, 1973). Infants and young children also can have increased susceptibility to immune modulation by environmental toxicants, as childhood is the time when primary immunity is often developed. As a person ages, the immune system begins to decline (which leads to increased susceptibility to infection) and there is an increase in the incidence of tumors (Lange, 1978; Makinodan and Kay, 1980). Because immune function declines as an individual ages, the effect of potentially immunosuppressive chemicals can be more pronounced in adults than it is in children.

Exposure of laboratory animals to TCDD results in thymic atrophy and immunosuppression. When TCDD is administered to adult animals, suppression of the cell-mediated and humoral immune responses occurs and lasts for 1-2 months. However, the effects of TCDD on the thymus and on immune-system responses are more severe and long-lasting if TCDD is administered both before and after birth rather than only after birth (Vos and Moore, 1974; Faith and Moore, 1977; Luster et al., 1979). Thymic atrophy and cell-mediated immunosuppression also are extensive after perinatal exposure, at which time the immune system is being developed in utero (Thomas and Hinsdill, 1979; Fine et al., 1989). Upon perinatal exposure to TCDD, there is a significant reduction in early lymphopoiesis. These studies show that the developing fetal or neonate immune system could be at greater risk of suppression if it is exposed to environmental toxicants.

Other factors that can enhance the effects of immunosuppressive compounds on the immune system are smoking, diet, malnutrition, stress, and disease. For example, smoking has been shown to modulate pulmonary leukocyte function, and smokers could be more susceptible to modulation of pulmonary immune response by ozone or nitrous oxide. Further studies should be conducted to explore more fully the interactions of these risk factors with immunosuppressive environmental xenobiotics. A consensus exists that studies using better defined cohorts and more sensitive tests will be required to assess the potential of xenobiotics to impair human health.

The absorption, metabolism, distribution, and excretion of a given chemical can differ within animal species and between animals and humans. Males and females also can metabolize xenobiotics differently. Because there are sometimes variations in biotransformation of chemicals, careful attention should be given to comparing responses between species.

Metabolic rates vary as a function of body size, and smaller animals metabolize faster. Thus, mice metabolize a material more rapidly than do rats. The differences in metabolic rate can serve as a basis for scaling doses across species. However, across-species scaling is more reasonably based on compound-specific experimental results from several species rather than on arbitrary scaling factors (Davidson et al., 1986).

The route of administration also can affect pharmacokinetics and result in different disease end points. For example, toluene diisocyanate (TDI) is a potent sensitizer when it is inhaled, but it produces tumors when it is ingested. This is presumably because a carcinogenic metabolite forms in the acidic environment of the stomach. There could be marked differences between species in the metabolites formed, and this possibility needs to be explored. When marked differences in biotransformation or unexpected deviations in the metabolic rate of a compound are encountered or are not investigated, comparative-toxicity studies

must be performed in animals to provide a basis for predicting human risk.

Pharmacokinetic studies should be coupled with toxicity studies to establish which dose paradigm or marker paradigm is useful as an estimate of the biologically effective dose. The usefulness of this exercise has already been illustrated in the discussion of the desirability of maintaining a plasma CsA level of 250 ng/ml, the appropriate paradigm of biologically effective dose in this instance.

As alluded to above, one difficulty in extrapolating to humans from animals is differences between species. The differences between species may result from metabolic differences or from actual differences within the immune system (mice, for example, do not produce basophils). Therefore, care should be taken to acknowledge and minimize these differences.

The induction of disease via immunosuppression is essentially a two-stage process: suppression followed by initiation of the infectious or neoplastic response. Therefore, studies of the mechanisms by which the immune status is altered and, therefore, means by which the individual becomes more susceptible are necessary to establish the validity of markers for predicting disease and the degree of their predictability. When possible, mechanistic studies should be conducted with the appropriate consideration of the biologically effective dose and temporal susceptibility of the target site.

Most studies on the mechanisms of immunotoxicant action are performed to allow a further understanding of the effects a compound can have on the human immune system and to show potential differences between animal models and humans. That is, if the site of action of a compound can be determined, and the importance of the analogous site to the function of the human immune system can be established (as discussed with CsA), then the ability of the animal model to predict the risk to human health of exposure to a given chemical can be more fully evaluated. An additional benefit can arise by further clarifying the processes by which the immune system responds to and eliminates foreign materials. In this way, chemical toxicants can be used as tools to dissect specific parts of an immune response. For example, one study showed that O,S,S-trimethyl phosphorodithioate, a contaminant in the pesticide malathion, inhibits cytotoxic T-lymphocyte (CTL) activity, an indicator of effector T-cell function, at an early postrecognition step (Rodgers et al., 1988). The study determined that because this chemical is a potent esterase inhibitor, there is an esterase whose activity is necessary at an early post-recognition step for lysis of tumor target cells by CTLs that is distinct from the N-benzyloxycarbonyl-L-lysine thiobenzylesterase of CTL granules. This is one example of how a chemical can be used to dissect an immune response, and it demonstrates that immunotoxic chemicals can be powerful tools in basic immunologic research.

There is an increasing awareness within the scientific and public communities that toxic chemicals can suppress the immune system. These concerns are supported by numerous studies that use experimental animals and, to a lesser extent, by isolated human experience. There is not a great body of evidence that persons in the general population with putative exposure have been immunologically compromised. However, it is well established that treatment with immunosuppressive therapeutic agents and

infection with viruses, such as HIV, result in an increase in the incidence of infection and neoplasia. Exposure markers of immunosuppressive agents are blood or tissue concentrations of the parent compound or its metabolites. The variations of immune-system markers related to immunosuppression can result from various factors, such as age, lifestyle, and adjunctive exposure, as well as from exposure to immunosuppressive agents.

Disease is induced via immunosuppression in a two-stage process: In addition to the immune-system deficiency, there is an initiation of an infectious or neoplastic response. Animal bioassays based on treatment with a putative immunosuppressant followed by exposure to an antigen are therefore most useful to identify the test agent as a potential hazard. Studies of the mechanisms by which the immune status is altered and the degree to which the subject is rendered more susceptible are necessary to establish the validity of extrapolating findings to humans. When possible, mechanistic studies should be conducted with the appropriate consideration of biologically effective doses.

Animal immunologic bioassays are useful to identify possible hazards associated with human exposure to xenobiotics; to explain possible differences in susceptibility between individuals and species; and to develop a rational basis for the management of risk in medicine, in the workplace, and among members of the general public. These bioassays could be useful, because they can suggest possible mechanisms in the classification of carcinogens.

Although the hazards suggested by animal immunotoxicologic bioassays need to be further validated by comparison to humans, there is an unacceptable risk of potential adverse effects among the general public from unnecessary, continued exposure to immunosuppressive agents. Continued exposure to the pollutants at doses causing immunosuppression is unwarranted. However, retrospective studies of exposed persons could be useful, if exposure levels can be reconstructed. Studies of consistently exposed workers, for example, could be useful because of the likelihood of better monitoring of exposure concentrations that are higher than those encountered by the general public, but lower than those producting immunodeficiency in experimental animals. Modification of clinical monitoring to obtain correlations between findings from animal studies and the patient could prove to be useful, although direct application to a healthy population could be difficult.

There is no definitive evidence that individuals living in the vicinity of contaminated sites or chemical manufacturing plants have been immunologically compromised sufficiently to be at increased risk of disease. To elucidate more fully the level of human risk of increased disease after exposure to an immunosuppressive xenobiotic, persons who are occupationally exposed or exposed after an accident should be tested for possible immunotoxic effects. The higher the exposure, the greater likelihood of finding a measurable effect. Several suggestions are made to make this proposition feasible.

A profile of assays for the assessment of the immune function of human populations should be developed (a set of suggested assays can be found in Chapter 7). Within this development, the establishment of normal ranges and the validation of these assays should be emphasized so that the effects of response to immunotoxicants can be monitored. A team of scientists and technicians well versed in the assays within this profile should be appointed to respond to the need for immunotoxicity testing in case of accidental or occupational exposure. This could be handled through the federal government for the establishment of the core group. At the time of an accident, a scientist with expertise in the immunotoxicity of

the chemical in question should be included. Within this context, a questionnaire should be used to assess the increased incidence or duration of infection after exposure to an immunosuppressive chemical. Population background levels should be established for various complaints associated with infection, and the contribution of stress should be determined. Prospective, longitudinal studies of exposed human populations should be done, using the battery of tests described in Chapter 7. These studies will require funding arrangements and well-defined control populations. Ethical, controlled clinical studies that use human volunteers deliberately exposed to potential immunotoxicants should be considered.

Most immunotoxicologic evaluations are performed in germ-free, healthy, well-fed young adult animals. Although this is the standard of practice, some groups could be more susceptible to immunosuppression as a result of factors that affect the response, such as age, malnutrition, and exposure to other chemicals that weaken the immune system. Animal models should be devised to test the effects of age, malnutrition, pregnancy, and other factors on the susceptibility of an individual to immune-system suppression. This would be best accomplished through comparison of the dose-response curves for a given chemical in healthy, young adult animals.

One variable that has not been well defined in immunotoxicology is the level of immune-system suppression necessary to produce increased susceptibility to disease. Animal models are now used to determine the ability of a chemical to increase risk of disease. The end point used in these models is death, which could be caused by factors other than immunosuppression. Models should be developed that measure the effect of a chemical on the immune response and that can be used to evaluate the susceptibility of an animal to a pathogen. These data should then be compared to the effects of the chemical at the same doses on other immune-system responses. Through this type of analysis, perhaps a definition of the amount of immunosuppression necessary to increase the risk of disease could be established.

The tests routinely used for clinical assessment of immunosuppression in humans are not very sensitive. This could result from normal variations in immune function, or it could stem from the types of cells (peripheral blood cells) available for analysis. Assays that have reduced interassay variability and reduced biologic variability and are therefore more likely to detect changes at lower exposure levels should be developed. One way to reduce variability is to establish standard operating procedures.

Extrapolation from animals to humans is difficult because of metabolic differences and other factors. These differences should be taken into account and minimized wherever possible in the interpretation of data. Also, wherever possible, mechanistic studies should be incorporated into the risk-assessment profile of an immunotoxicant. Once the point of action of a xenobiotic is established, it can be determined whether the same site is present in the human immune system. Mechanistic data also can allow a determination of whether a given xenobiotic can induce increased susceptibility to human disease by illuminating the importance of the mechanism to human defense against disease.