The effects of ozone on immune function.

A review of the literature reveals that ozone (O3) exposure can either suppress or enhance immune responsiveness. These disparate effects elicited by O3 exposure depend, in large part, on the experimental design used, the immune parameters examined as well as the animal species studied. Despite the apparent contradictions, a general pattern of response to O3 exposure can be recognized. Most studies indicate that continuous O3 exposure leads to an early (days 0-3) impairment of immune responsiveness followed, with continued exposures, by a form of adaptation to O3 that results in a re-establishment of the immune response. The effects of O3 exposure on the response to antigenic stimulation also depend on the time at which O3 exposure occurred. Whereas O3 exposure prior to immunization is without effect on the response to antigen, O3 exposure subsequent to immunization suppresses the response to antigen. Although most studies have focused on immune responses in the lung, numerous investigators have provided functional and anatomical evidence to support the hypothesis that O3 exposure can have profound effects on systemic immunity.


Introduction
Broadly defined, the function of the immune system is to protect the body from damage by infectious microorganisms and neoplastic cells. The immune response is mediated by a variety of soluble factors and cells grouped according to their capacity to mediate innate (natural) or adaptive (acquired) immune responses. Innate immunity is primarily dependent on the phagocytic effector mechanisms (induding humoral factors such as complement, lysozyme, etc.), which are neither specific for particular infectious agents nor improved by repeated encounters with the same agent. Acquired immunity is specific for the inducing agent and is marked by an enhanced response upon repeated encounters with that agent, thus displaying memory. These systems do not function independently since innate and acquired immune mechanisms can interact; specific antibodies enhance the phagocytic ingestion process and cell-mediated immune responses increase the intracellular microbicidal and tumoricidal capabilities of the phagocytes. Both innate and acquired immune responses and the cells that carry out these processes are regulated by soluble factors such as cytokines and various autocoids.
The multiple interlocking mechanisms involved in the complex sequence of events that lead to the expression of the immune response can present a plethora of targets for air pollutants and xenobiotic compounds. Recognition that environmental agents can impair immune function has lead to the emerging field of immunotoxicology (1,2). As with any new area of scientific endeavor, uncertainties abound as the field is defining itself and gaining recognition. This is especially true when many questions remain unanswered in the parent discipline immunology. Against this background the literature on the effects of ozone on immune function is reviewed.
Ozone and the Immune Response A survey of the literature on the effects of 03 exposure on innate and acquired immunity reveals that the responses range from no effect to impaired and/or enhanced effects. Several factors related to the experimental systems used are responsible for this ambiguous situation. First, many studies have examined a diverse range of immunologic parameters with varying sensitivities to 03. Second, there are differences in responses to 03 among animal species.
Third, it is becoming increasingly apparent that the 03 concentrations employed and the exposure protocols utilized (i.e., continuous versus intermittent) heavily influence the responses measured. Finally, the time at which 03 exposure occurs, relative to the time at which responsiveness is measured, appears to be critical.
To unravel the variables and to determine the trends that emerge, it is helpful to group the approaches that have been used to investigate the effect of 0 on the 3 immune system into several categories. These include: a) measurement of lymphoid organ weights and cellular composition; b) determination of the functional capacity of lymphocytes in the absence of antigenic stimulation; c) measurement of the immune response following antigenic stimulation; d) assessment of the phagocytic capacity of alveolar macrophages; and e) measurement of the host response to infectious agents. Because multiple variables determine the response of the immune apparatus to experimental 03 exposure, attention is given herein to the exposure conditions utilized. Thereafter, the studies performed in our laboratories in an attempt to validate and gain insights into the mechanisms of the divergent effects of 0 on the immune system will be presented.
Environmental Health Perspectives

Effect ofOzone on Lymphoid Organ Weight and Celi Numbers
The effect of 03 exposure on lymphoid organ weights and/or cellularity have been examined in several studies. Although there is no clearcut conclusion, a general pattern of response to 03 is reflected in the changes in lymphoid tissues that can be correlated with changes in immune parameters.
Continuous 0.3 to 0.8 ppm 03 exposure for 20 to 24 hr/day elicits decreases in murine spleen, thymus, and mediastinal lymph node (MLN) weights after 1 to 3 days of exposure (3,4). The changes in MLN weight were found to be dose dependent over a range of 0.3 to 0.7 ppm 0 (4). With prolonged exposures (7-14 Jays), spleen, thymus, and MLN weights returned to control levels (3,4). Concomitant with the return of lymphoid organ weights to control levels is an increase in the number of T-cells in these organs. Thus, after an initial decrease, T-cells increase in lymphoid organs after continued 03 exposure (58).
The increases in MLN T-cell numbers were abolished by pretreatment with a monoclonal antibody for T-cells (8).
The increase in T-cell numbers is not paralleled by a change in B-cell numbers. IgG-and IgM-secreting cells in murine bronchus-associated lymphoid tissue (BALT) (9) and B-cell numbers in the lung as a whole did not change with prolonged 03 exposures (7). Likewise, exposure of rats to 0.13 ppm 03 for 1 week increased T/B cell ratios in the MLN, with the increases lasting for at least 5 days after cessation of exposure (10).
Based on the studies cited above, it is apparent that a pattern in the response to 03 can be recognized. Subsequent to an initial decrease in lymphoid organ weights and cellularity after 1 to 7 days of exposure, continued exposure of 1 to 3 weeks reveals that the organ weights not only stop decreasing, but actually return to control levels and, in some cases, exceed control levels (4). Most of these changes in lymphoid organ weights can be correlated to changes in T-cell numbers and proliferation. This apparent adaptation to 0 exposure is also mirrored in the functional studies of the immune response.
The general pattern of response to 03 exposures does not always hold, however. For example, it was observed that thymus weights continued to decrease as continuous exposures to 0.8 ppm 0 were extended to periods as long as 56 days (11). On the other hand, when mice were continuously exposed to 0.31 ppm 03 (103 hr/week for 6 months) and allowed to breathe filtered air for 5 months following exposure, an increase in the spleen weights was observed (12). Furthermore, an analysis of BALT tissue in rats showed that exposures of 2 to 3 days actually enhanced DNA synthesis and the mitotic index, both of which gradually returned to control levels with continued exposures. These changes in proliferation in BALT tissue were not reflected by changes in BALT size, which was unaffected by 03 exposure (13).
Effect of Ozone on Immune Function in the Absence of

Antigenic Stmulaion
Longitudinal studies on the effects of 03 exposure on lymphoid organ cell numbers provide information on cellular traffic and cell numbers but provide few insights into the functional capacity of the lymphocytes. A number of studies have investigated the effect of 0 exposure on the blastogenic response of lymphocytes to nonspecific mitogens. These assays measure nonspecific clonal expansion of the lymphocyte populations, a critical step during the amplification of the immune response. Thus, MLN cells stimulated with concanavalin A (ConA) during the course of continuous 0.7 ppm 03 exposure (20 hr/day, 7 days/week for 28 consecutive days) showed little change in responsiveness during the first week; however, by day 14, an enhanced reactivity was observed that continued to increase through day 28 (4,8). Short-term exposure studies have shown that the rat splenic cell responses to T-cell mitogens phytohemagglutinin (PHA) and ConA and a B-cell mitogen (Escherichia coli LPS) were significantly enhanced by exposure to 1 ppm 03 (8 hr/day for 7 days) (14). In contrast, intermittent exposure to 2 ppm 03 (8 hr/day for 4 days, 2-4 days in ambient air followed by another day of 03 exposure) was without effect while long-term exposure of mice to 0.1 ppm 03 (5 hr/day, 5 days/week for 103 days) suppressed the splenic cell responses to T-cell mitogens (PHA and ConA) but not to a B-cell mitogen [Salmonella typhosa LPS; (15)].
Natural killer (NK) cell activity has also been used as an index of 0 3-induced effects on the functional integrity of the immune surveillance system. One such study demonstrated that continuous exposure of rats to 1 ppm 03 for 10 consecutive days results in a significant decrease in pulmonary NK cell activity on days 1, 5, and 7 of exposure, with activity recovering by day 10 (16). In another study, continuous exposure of rats to 03 (0.2 ppm, 0.4 ppm, and 0.8 ppm) enhanced lung NK cell activity on day 7 at the two lower concentrations, whereas exposure at 0.8 ppm 03 had a suppressive effect (17).
The above-mentioned studies on the effects of 03 exposure on mitogen responses and NK cell activity used exposures to 03 in great excess of that found in ambient air. To investigate the effects of acute and chronic exposure to 03 at near ambient concentrations, Selgrade et al. (18) used an experimental protocol designed to mimic diurnal urban 03 exposure patterns. Rats were exposed to a background level of 0.06 ppm for a period of 13 hr, a broad exposure spike rising from 0.06 ppm to 0.25 ppm and returning to 0.06 ppm over 9 hr, and a 2-hr downtime. The exposures to this profile were 5 days/week; weekend exposures were to the background. After 1, 3, 13, 52, or 78 weeks of exposure, blood was drawn and the spleens removed. Spleen cells were assessed for NK cell activity and responses to T-cell mitogens (P1A and ConA) and a B-cell mitogen (Salmonella typhimurium glycoprotein). Peripheral blood leukocytes were also assessed for responses to T-cell mitogens. The data show that 03 exposure had no effect on NK cell activity, nor were there any 03-related changes in mitogen responses in splenic or blood leukocytes. There were also no effects of a single 3-hr exposure to 1 ppm of 03 on spleen cell responses to the mitogens immediately after exposure or at 24, 48, and 72 hr thereafter.
Several studies have also investigated the effect of 0 exposure on human peripheral blood lymphocytes. Exposure of subjects to 0.4 ppm 03 for 4 hr impaired the ability of B-cells to form rosettes with human erythrocytes, whereas the ability of T-cells to form rosettes with sheep erythrocytes was unaffected (19). Another study examined the effects of a single 2-hr exposure to 0.6 ppm 03 on the mitogenic response of peripheral lymphocytes following stimulation with PHA, ConA, pokeweed mitogen (PWM), and Candida albicans. The mitogenic response to PHA was significantly reduced at 2 and 4 weeks following 03 exposure, with normal responses observed at 2 months. In contrast, no significant changes in lymphocyte responses to ConA, PWM, and C albicans were observed (20). Finally, subjects were exposed for 2 hr to 0.12 and 0.4 ppm 03 during moderate exercise, and PHA-induced peripheral T-cell proliferation was assessed before exposure and 24 and 72 hr after exposure (21). In the absence of 03, exercise enhanced proliferation as compared Environmental Health Perspectives to resting subjects, whereas exposure to 0 with exercise appeared to reduce the capacity of the subjects to react to exercise with an enhanced proliferative response. This reduced response appeared to be more pronounced in subjects exposed to the higher concentration of 03 and appeared to be more pronounced at 24 hr than 72 hr postexposure.
Effect ofOzone on Immune Function with Antigenic Stimulation A handful of studies have examined the effect of 03 exposure on the allergic response to antigenic (ovalbumin, OA) stimulation. In general, these studies have used a Pandora's box of 03 exposure regimens and varied the time and site (aerosol, systemic, or a combination of both) of antigenic stimulation. A common thread through most of these studies is that when the antigenic stimulation was performed during 03 exposure, enhanced allergic responses were observed (22)(23)(24)(25)(26). However, when 03 exposure preceded OA stimulation, IgE antibody production was suppressed (27).
In a series of studies, Fujimaki and colleagues (3,11,28) evaluated the systemic effects of short-term 03 exposure on the humoral and cell-mediated immune response. In the first study, mice were continuously exposed to 0.8 ppm 0 for 1, 3, 7, and 14 days. After each day of exposure, the animals were tested for the primary splenic antibody response to either sheep erythrocytes (SRBCs; T-dependent antigen) or DNP-Ficoll (T-independent antigen). The antibody response to SRBCs was suppressed following all exposure periods. In contrast, no suppression was observed in the primary antibody response to DNP-Ficoll (3). When the continuous 0.8 ppm 03 exposure was extended to 56 days, suppression of the primary splenic antibody response to SRBCs, but not DNP-Ficoll, was still evident (11). Using the same exposure protocol (continuous 0.8 ppm 03 for 1, 3, 7, and 14 days), the delayed-type hypersensitivity (DTH) response to SRBCs was also examined (28). Ozone exposure gradually depressed the DTH reaction from 1 day to 7 days of exposure and returned to control levels by day 14. In the same study, the timing between 03 exposure and SRBC immunization for the DTH response was also examined. Continuous exposure to 0.8 ppm 03 for 3 days immediately before immunization (with no post-immunization exposure) had no effect on the DTH reaction, whereas continuous 0.8 ppm 03 exposure for 3 days following immunization significantly suppressed the DTH response (28). In toto, it appears that 03 exposure preferentially affects the immune repertoire of the T-cells rather than that of the B-cells.
The effect of 03 exposure on DTH responses to Listenia monocytogenes antigen has also been investigated (10). For this, rats were continuously exposed to 0.75 ppm 03 for 7 days immediately before or immediately after intratracheal infection with L. monocytogenes. The DTH response, tested at 7 and 14 days after infection (immunization), was not significantly influenced by exposure to 03 for 1 week before infection. However, if 03 exposure occurred during the Listeria infection, the DTH response was suppressed. Using the same exposure protocol, the proliferative responses of spleen cells to Listeria antigen was also examined. The effect of 03 exposure on this expression of immunity closely paralleled the effect found with DTH to Listeria antigen [i.e., whereas exposure before immunization did not decrease blastogenesis, the proliferative response was significantly suppressed in animals that were exposed to 03 immediately after immunization (10)]. In a related study, mice were continuously exposed to 0.59 ppm of 0 for 36 days. On the fifth day of exposure, lie animals were immunized with tetanus toxoid and challenged with tetanus toxin on day 27. Compared with unexposed controls, 0 -exposed mice exhibited greater mortality and morbidity following the toxin challenge (29).

Macrophage Phagocytosis
Studies on the effects of 03 exposure on innate immunity have focused on the phagocytic activity of alveolar macrophages (AMs). Acute exposures (2.5 ppm for 5 hr) decreased the in situ phagocytosis of inhaled Staphlylococcus aureus in rat lungs (30). In a similar manner, exposure to 0.4 and 0.8 ppm 0 ppm for 3 hr decreased phagocytosis ofintratracheally administered Streptococcus zooepidemicus in murine lungs in a dose-dependent manner (31).
Ex vivo studies on AM phagocytosis have found divergent effects of 03 exposure. For example, when rats were continuously exposed to 0.8 ppm 03 (for 3, 7, and 20 days), an enhancement in the ingestion of carbon-coated latex particles was observed, with the greatest increase following 3 days of exposure (32). On the other hand, rat AM phagocytosis of opsonized SRBCs was progressively suppressed from day 1 to day 3 of continuous 0.5 ppm 03 exposure, with recovery of phagocytic activity by day 6 of exposure (33). When rabbits were given a single 2-hr exposure of 0.1 ppm 03, AM phagocytosis was maximally decreased immediately after exposure and remained low throughout the ensuing day, albeit to a lesser extent; complete recovery of phagocytic activity occurred by day 7 post-exposure (34). Increasing the single 2-hr exposure dose of rabbits to 1.2 ppm 0 resulted in a continued suppression of3AM phagocytosis through postexposure day 7. Intermittent exposure of rabbits (0.1 ppm 03 for 2 hr/day for 14 days) resulted in a significant reduction of AM phagocytosis on day 3 and 7 with recovery of phagocytic potential by day 14.
To assess species differences in 0induced changes in AM phagocytosis, rats and mice were intermittently exposed to 0.4 ppm 03(12 hr/day for 7 days). After 1 day of exposure, AM phagocytosis of opsonized SRBCs was suppressed in mice and enhanced in rats. By days 3 and 7 post-exposure, rat AM phagocytosis returned to control levels, whereas mouse AM phagocytosis continued to be suppressed (35).
The effect of 03 exposure on human AM-mediated phagocytosis has also been examined. For this, subjects were exposed for 6.6 hr to either 0.08 ppm or 0.10 ppm 03 during moderate exercise (36). Bronchoalveolar lavage was performed 18 hr after exposure (36). Ozone exposure significantly decreased the ability of AMs to ingest Candida albicans yeast particles coated with serum (complement receptormediated phagocytosis). However, no such effect was observed with IgG-coated yeasts (Fc receptor-mediated phagocytosis) or unopsonized yeasts (nonspecific receptormediated phagocytosis). Finally, a human study investigated the effect of a single 0.4 ppm 03 exposure on peripheral blood leukocyte function. The capacity of polymorphonuclear neutrophils (PMNs) to phagocytize and kill Staphylococcus epidermidis was significantly suppressed at 72 hr after exposure and returned to normal levels 2 weeks after exposure (37).

The Effect ofOzone on Resistance to Infectious Agents
Recognition of the association between exposures to air pollutants and the development of acute respiratory illnesses has led to animal models that use microbiologic parameters to evaluate toxicity (38,39). Many studies on the effects of 0 on host defenses have been performed wit; a rodent model that is often referred to as the "infectivity model" (39). In this model, rodents are Volume 103, Supplement 2, March 1995 challenged by aerosol inhalation with highly virulent Klebsiella pneumoniae or Streptococcus pyogenes following various 03 exposure protocols and observed for deaths over the subsequent 14-day period. The data on the effects of 03 exposure are expressed as increases in mortality. The earliest of these studies (40) found significant increases in mortality following K pneumoniae challenge when mice were exposed to 1 ppm and 4 ppmO3 for 3 hr, 1 ppm O3 for 100 hr, and 1 ppm 03 for 4 hr/day, 5 days/week for 2 weeks. With the use of Streptococcus pyogenes, the sensitivity of the "infectivity model" increased in that significant differences in mortality were demonstrated when the infectious challenge was given simultaneously or 2 hr after exposure to 0.1 ppm 0 for 3 hr (41). Mice exposed 5 hr/day, 5 Lays/week for 103 days to 0.1 ppm 0 also exhibited increased susceptibility to death from streptococcus challenges (15). Other studies with the "infectivity model" involving 03 exposure focused on the relationship between 03 and NO2 exposure on resistance to streptococcal pneumonia (42)(43)(44). The "infectivity model" links interference with pulmonary antibacterial defenses to mortality following pulmonary challenges with virulent organisms. Another approach is to use minimally virulent organisms, such as S. aureus, that do not provoke injurious responses in the lungs (45). In this system, rodents are challenged by aerosol inhalation with staphylococci and intrapulmonary killing of the organism is assessed over a 4or 5-hr period by standard microbiologic methods. Exposure to O can occur before, after, or before and after bacterial challenge. Under normal conditions, approximately 90% of the staphylococci are killed over the initial 4-hr period of infection. The effects of 03 exposure are assessed in terms of suppression of intrapulmonary bacterial killing.
Exposure of mice to 0.1 ppm, 0.5 ppm, and 1 ppm 0 for 3 hr immediately before staphylococcaQ challenge significantly suppressed the intrapulmonary killing of S. aureus in a dose-related manner (46,47). Similarly, exposure to increasing concentrations of 0 for 17 hr before or 4 hr after bacterial challenge also impaired the intrapulmonary killing of staphylococci in a dose-dependent manner (48,49).
More recent studies with the Streptococcus zooepidemicus infectivity model found that 0.4 ppm and 0.8 ppm 0; exposure of 5-and 9-week-old CD-1 mice for 3 hr impaired intrapulmonary bacterial killing, with the most severe effect elicited by the higher dose of 03 in the younger mice (50). When the study was performed in two different mouse strains, the results showed that C3H/HeJ mice were more sensitive to the effects of 03 exposure on susceptibility to streptococcus infection than C57BI/6 mice (31).
The effect of 03 exposure on chronic respiratory infections has been assessed using Mycobacterium tuberculosis and L. monocytogenes as challenge organisms. Mice exposed to 1.5 ppm 0 (4 hr/day, 5 days/week for 2 months) exhibited no increased mortality to intravenously administered M tuberculosis (51). On the other hand, exposure of mice to 1 ppm 03 (4 hr/day, 5 days/week for up to 8 weeks) beginning at 1 or 2 weeks after inhalation challenge with M tuberculosis significantly enhanced the number of bacteria recovered from the lungs (52). With L. monocytogenes, rats were exposed continuously to 0.13 ppm, 0.25 ppm, 0.5 ppm, 0.75 ppm, 1.0 ppm, and 2.0 ppm 0 for 1 week and thereafter intratracheally infected with Listeria. Exposure to 1 ppm 03 significantly increased mortality through dysfunctions in the intrapulmonary elimination of the bacteria (10).
Only a few studies have investigated the effects of 0 exposure on the outcome of viral respiratory infection. Fairchild (53) demonstrated that exposure to 0.6 ppm 0 for 3 hr after aerosol infection of mice with mouse-adapted influenza A2/Japan 305/57 virus inhibited viral replication in the nose but did not alter influenza virus growth in the lungs of the same animals. Continuous 1.5 ppm 03 exposure also had no effect on the proliferation of virus in the lungs of mice infected by aerosol inhalation with a sublethal dose of influenza A/PR8/34 virus (54). In another study (55), mice were exposed to 1 ppm O3, 3 hr/day for 5 days and intranasally infected with mouse-adapted influenza virus A/Hong Kong/68 immediately after the first, second, third, fourth, and fifth of the five daily exposures. A 2-fold increase in the incidence of mortality was observed in mice infected after the second exposure, with no effects on percentage mortality in mice infected on the first, third, fourth, or fifth exposure. When the exposure concentration was lowered to 0.5 ppm, there were no effects on mortality in mice after the second exposure. Five daily 3-hr exposures to 1 ppm 0 had no effect on virus titers in the lungs otmice infected after either the second or fifth day ofexposure.
Wolcott et al. (56) observed a protective effect of 03 exposure to influenza virus infection. Mice were continuously exposed to 0.5 ppm OS for 2 weeks before and after aerosol infection with influenza A virus (WSN strain). Four groups were included in the study. The control mice were exposed to air for 2 weeks, virus infected and reexposed to air for 2 weeks. The experimental groups were exposed to 03 either before, after, or before and after virus infection. It was found that all animal groups exposed to 03 had a significant decrease in mortality, which correlated with a less widespread infection of the lung parenchyma. The reduced mortality was independent of peak pulmonary virus titers, pulmonary interferon titers, and pulmonary and serum-neutralizing antibodies.

Other Studies
Ozone exposure may also induce aberrant immune responses through modification of "self' determinants in the lungs by ways such as cross-linking of antigen (57), denaturation of protein (58), or by other means (59). Indeed, an autoimmune response has been attributed to 03 exposure (60).
The effect of in vitro 03 exposure on the immune responsiveness of isolated cells has also been investigated. These experiments have demonstrated that acute exposure to 0 reduces AM phagocytosis (61) in a dose-Aependent manner (62). In addition, PGE2 production by AMs was increased by 03 exposure, and superoxide release from stimulated cells decreased (62). Exposure of peripheral mononuclear cells from healthy individuals to various concentrations of 03 revealed a general decrease in immune parameters such as NK activity (63), responsiveness to mitogens, expression of activation markers on monocytes and lymphocytes (64,65), and antibody production (66). These are important observations because they demonstrate similar responses to 0 between human cells exposed in vitro and animals cells exposed in vivo. Confirming the validity of the animal models, these studies demonstrate that the response to 03 of human cells exposed in vitro is similar to that of animal cells exposed in vivo.

Studies Performed
The studies performed in our laboratories were designed to clarify some of the issues regarding the effects of 0 exposure on the innate and acquired immune system. The first goal was to investigate the modulating effect of 03 exposure on AM phagocytosis and to determine whether an adaptive phase followed prolonged exposure. The second goal was to explore whether changes Environmental Health Perspectives in the local (lung) immune responses were mirrored by changes in the systemic immune responses. The third goal was to gain some insight into the mechanisms of the systemic effects and, finally, we attempted to elucidate the mechanisms by which 03 exposure reduced the pathogenesis of experimental influenza virus infection.

Suppression and Recovery ofthe
Alvola Macropha Phagocytic System To investigate the effect of 03 exposure on AM phagocytosis and to determine whether prolonged 03 exposure results in an adaptive response, mice were exposed to 0.5 ppm 03 for 23 hr/day for 14 consecutive days. At 1, 3, 7, and 14 days of exposure, AM phagocytosis of opsonized SRBCs was assessed in vitro (67). Figure IA shows that 1 day of 03 exposure suppressed AM phagocytosis. This impairment of phagocytic activity was further exacerbated at day 3 but began to recover at 7 days and returned to control levels by 2 weeks.
Using another particle to probe the functional capacity of the AM phagocytic system in situ (45), mice were challenged by aerosol inhalation with S. aureus, and intrapulmonary bacterial killing was assessed over a 4-hr period. The data ( Figure 1B)  Days of Ozone Exposure (0.5 ppm) of opsonized SRBCs ( Figure 1A); that is, an initial suppression of intrapulmonary staphylococcal killing was followed by recovery of bactericidal activity by day 14 (67).

Modulation ofthe Immune Response
In nonimmunized mice, continuous 0.7 ppm 03 exposure results in an initial reduction of the number of cells recovered from MLN, which was followed by an increase and maintenance above basal levels as the 03 exposure continued (4,8). To determine whether the splenic events were comparable to the changes in the MLN, mice were exposed 23 hr/day for 14 consecutive days to 0.8 ppm 0 3. After 1, 3, 7, and 14 days of exposure, the number of cells in the MLN and the spleen/body weight ratios were determined. Then, to test the functional activity of these two cell populations, the proliferative response of MLN and splenic lymphocytes to PHA stimulation was measured (68). The data ( Figure 2A) demonstrate that continuous 0 exposure causes a reversible decline in M3LN cells followed by hypertrophy. Spleen weights were decreased with respect to body weight after 1 day of 03 exposure ( Figure 2B). This effect was still evident at day 3 but returned to control values following 7 days of exposure. With continuous exposure, the spleens were larger in size relative to body weight than in control animals. Cells from both MLN and splenic sources obtained after 1 day of 03 exposure had reduced responsiveness to PHA mitogen ( Figure 3A). However, this effect was ablated following prolonged exposure. In a similar manner, the NK cell activity of splenic lymphocytes was significantly reduced following 1 and 3 days exposure but was restored upon continued exposure ( Figure 3B).
Another set of experiments was designed to investigate the effect of 03 exposure on antigen-stimulated immune responsiveness. For this, an immunization protocol was adopted that maximizes the local (pulmonary) immune response. Briefly, mice were primed with an ip injection of OA and 2 weeks later boosted by aerosol inhalation of OA. One week after the aerosol boost (i.e., 3 weeks after the intraperitoneal priming injection), the MLN and splenic lymphocytes were assayed for proliferative responses to OA antigen, and the bronchoalveolar lavage fluid was assayed for specific IgG and IgA antibody to OA. Ozone exposure was incorporated into this immunization protocol to the effect that groups of animals were exposed to 0.8 ppm of 0 for 23 hr/day for 1, 3, 7, or 14 days prior to assay as shown schematically in Table 1.     Figure 4A shows that the MLN cell responses to OA antigen were not significantly different from controls for 1, 3, and 7 days of 0 exposure. However, following 14 days of exposure to 03 the MLN lymphocyte responses to OA antigen were significantly enhanced. Splenocyte responses to OA antigen were similarly unaffected by 1 and 3 days of exposure, but were significantly lower than control values at the 7-and 14-day time points.
Serum antibody titers to OA antigen were high in all animals and were unaffected by any period of 03 exposure (68). Pulmonary IgG and IgA antibody titers to OA antigen ( Figure 4B) were lower in bronchoalveolar lavage fluid from animals exposed to 03 for 7 and 14 days. At the earlier time points, IgA titers remained unchanged; IgG levels were at first suppressed (day 1) and then returned to control values (day 3) before decreasing again at the later time points.  Finally, all the experiments detailed above with exposure to 0.8 ppm 03 were also performed with continuous exposure to 0.5 ppm 0 . The results of these studies showed similar trends in responses, however, the magnitude of these responses was not as large as that observed with 0.8 ppm 03 exposure (MI Gilmour and GJ Jakab, unpublished observations).

Impairment ofPeritoneal Macrophage Phagocytosis: The Role ofProstanoids
During the course of studies on the effect of continuous 0.5 ppm 0 exposure on AM Fc-receptor-mediated phagocytosis, it was also observed that the exposure impaired peritoneal macrophage (PM) phagocytosis. Due to its high reactivity, it seemed unlikely that 03 could directly mediate systemic changes in phagocyte function. Rather, as has been suggested by others, the systemic effects of oxidant gases are likely indirect, due to mediators produced upon oxidant exposure (69). One such class of mediators may be the prostaglandins, potent immunomodulatory autacoids found in high concentrations in the airways (70,71) and serum of animals (72) and man (73) following acute 0 exposure. Furthermore, exogenously added prostaglandin E (PGE) inhibits macrophage phagocytosis (p4). To test the hypothesis that PGE plays a role in mediating the 03-induced suppression of macrophage phagocytosis, mice were continuously exposed to 0.5 ppm 0 and the effect of the cyclooxygenase inhibitor indomethacin on AM and PM Fc-receptormediated phagocytosis and PGE levels in the bronchoalveolar lavage fluid (BAL) were assessed. Figure 5A shows that the time course for 0 -induced impairment of PM phagocyto- .3 sis paralleled that of the AMs in that maximal suppression occurred at 3 days of exposure, which progressively abated as exposure continued. PGE levels in BAL fluid during the course of 03 exposure are presented in Figure 5B. indomethacin for 3 days before and during 03 exposure ablated the 03-induced increases in bronchoalveolar savage PGE levels ( Figure 6) and suppression of PM phagocytosis (Figure 7). Indomethacin also significantly reduced the effect on AM phagocytosis on the third day of exposure (Figure 7). The effect of indomethacin on 03-induced suppression of PM phagocytosis were mimicked by pretreatment of mice with the active enantiomer of the cyclooxygenase inhibitor naproxen (74). Thus, prostanoids seem to play a key role in the local and systemic decreases in phagocytic activity induced by 03. The observations that prostanoids play a significant role in the suppression of AM phagocytic function has been recently confirmed with S. zooepidemicus as shown by the reduced susceptibility of mice to 0 3-enhanced infection by indomethacin treatment (50). Reduction ofInfluenzaVi rus Pathogenesis Although 03 exposure has been shown to increase susceptibility to infectious bacterial challenges, the opposite has been observed with viral infections; ozone exposure actually diminishes the severity of influenza virus infection, as evidenced by decreased mortality and increased survival time (56).
Mortality from experimental influenza pneumonitis is due to extensive damage of alveolar tissue resulting in impaired gas exchange (75). The antiviral immune response is known to participate in influenza-induced lung injury since immunodeficient (76-80) and immunosuppressed (81-83) mice develop much less cellular infiltration and histopathology than their immunocompetent counterparts. Also, adoptive transfer of specific immunity to immunosuppressed animals during the course of influenza virus infection reestablishes lung injury and increases pneumonic deaths (84). Because the integrity of the antiviral immune response is required for the tissue destructive phase of viral pneumonitis, we tested the hypothesis that 0 exposure reduced the pathogenesis of influenza pneumonitis through suppression of the antiviral immune response.
Mice were infected with a nonlethal dose of influenza A (PR8/34 strain) and continuously exposed to 0.5 ppm 03 (23 hr/day for 14 consecutive days). On the ninth day of infection, at the height of the virus-induced pathologic changes (85,86), morphometric analysis of lung tissue, lung wet/dry weight ratios, and lavage albumin concentrations showed that 03 exposure was associated with a significant reduction in these parameters of lung injury ( Figure  8). Because viral-induced pathologic changes are infectious dose related (87,88), the effect of 03 exposure on viral proliferation and antigen accumulation in the lungs was quantitated. Figure 9 shows that neither virus proliferation nor the accumulation of viral antigen was affected by 03 exposure.
Since immunosuppression during the course of viral infection reduces the number of lymphocytes retrieved from the lungs, which, in turn, is related to reduced immunopathologic changes (86,89), the effect of 03 exposure on the recovery of the two major classes of lymphocytes from lung tissue was phenotypically quantitated. In the presence of viral infection, both the T-and B-cell populations are significantly reduced by exposure to 03 ( Figure 10) as were the antiviral serum antibody responses (85). These data support the notion of the immunosuppressive nature of 0 in that .3 immunopathologic mechanisms are involved in the virus-induced lung injury, which are mitigated by 03 exposure.   Figure 9. Comparison of infectious virus titers and total viral antigen in lung homogenates during the course of influenza virus infection in mice exposed continuously to 0.5 ppm ozone or ambient air. From Jakab et al. (85).
With lethal doses of virus, immunosuppressed mice still succumb to influenza virus infection. However, the deaths are not the early pneumonic deaths but occur later and are due to the extrapulmonary spread of the virus (90,91). The effects of continuous 0.5 ppm 0 exposure on a severe influenza A/PR/34 infection were investigated. Twenty-five percent of the unexposed control mice died between day 8 and day 12 of infection, with no deaths occurring thereafter. In contrast, in 03-exposed mice, 5% died between day 8 and day 12, with another 35% succumbing between day 12 and day 20 of the infection (GJ Jakab, unpublished observations).

Discussion
Before reaching any conclusions on the effects of in vivo exposure to 03 on the immune system, several related issues warrant discussion. These issues broadly encompass inhalation toxicology, immunotoxicology, host factors, and experimental design.  Figure 8. Morphometric analysis of lung tissue (A), lung wet/dry weight ratios and lung lavage albumin content (B) among virus-infected and -uninfected mice that were continuously exposed to 0.5 ppm ozone or allowed to breathe ambient air for 9 days. * = p< 0.05, virus-infected versus virus-infected and ozoneexposed. From Jakab et al. (85). Concepts of inhalation toxicology dictate that the animals exposed to 03 and unexposed animals are treated in an identical manner (i.e., the control animals are exposed to filtered air in an identical exposure chamber so that the only difference between the experimental and control group is the 0 exposure). Assuming that this practice is fog lowed, several issues arise with the longer term continuous 03 exposure studies that may influence the outcome of an immunologic parameter under investigation. For example, continuous 03 exposure leads to a loss of appetite in rodents as reflected by a slower weight gain in young animals (92) and, depending on the 0 concentrations, weight loss in older mice (6J Jakab, unpublished observations). Malnutrition is known to affect the integrity of the immune system (93) and, to a lesser extent, undernutrition also blunts the immune response (94). In addition to 03 exposure retarding growth due to loss of appetite, the food consumed by control and exposed animals is not the same. Ozone exposure oxidizes the food thereby altering its nutritional value (95). Variations in diet composition are known to affect the immune response (93,94,96). Because of this, animals exposed to 0 for longer periods of time have a nutritional? disadvantage that may act in concert with 03 exposure to suppress immunity. During long-term continuous 03 exposure, little can be done about the loss of appetite except to match the food intake of the control animals to that of the exposed animals. To minimize the effects of ingesting 0 -oxidized food, it may be prudent to provide 0 3-treated food for the control animals (85,86). In addition to the nutritional aspect, the health status of the experimental animals also merits consideration. Animal husbandry has greatly improved over the past few decades to the point that reputable suppliers are capable of providing animals for research purposes that are free of intercurrent diseases. However, endogenous infections of rodents are pervasive (97), and assessment of the health status of the animals requires constant vigilance by suppliers and in the research setting since intercurrent infections are known to have profound effects on toxicity and research testing (98,99). For example, one of the most common infections of rodent colonies is Sendai (parainfluenza 1) virus (100). Sendai virus infection suppresses AM phagocytosis (101,102) and the primary antibody response of splenocytes to SRBCs (103) reduces T-cell mitogenesis (104,105), increases splenic NK cell activity (106,107), alters the in vitro response to heterologous erythrocytes (108), and alters immune cell function in aging mice (109). Other endogenous rodent infectious agents of concern to toxicologists include pneumonia virus of mice, mouse hepatitis virus, rat coronavirus, minute virus of mice, Kilham rat virus, Tooland's virus, and Mycoplasma pulmonis (98,110). Because of the possibility that intercurrent infections confound studies, the validity of observed results may be at risk if the health status of animals is not defined. rat AMs seeded in 96-well microtiter plates (35); and d) visual counting of Fc-receptormediated ingestion by adherent rat AMs (33) and mouse AMs (67,74). With the use of the 3.5 pm carbon-coated latex particles, it was observed that continuous exposure to 0.8 ppm 0 for 3 days enhanced Sprague-Dawley rat AM phagocytosis (32). In contrast, the opposite was observed when the challenge particles were sensitized SRBCs in that continuous 0.5 ppm 03 exposure suppressed Wistar rat AM phagocytosis (33). These data point to vast differences in sensitivities to 03 exposure between the two assay systems, the response between the two strains of rats, or the 0.3 ppm difference in the 03 exposure concentration. In a manner similar to that exemplified with the effects of 03 exposure on AM phagocytosis, questions arise about the relative sensitivities of the assays that measure lymphocyte function. Uptake of radiolabeled thymidine following stimulation with lectins or microbial products is a common method for assessing the ability ofTand Bcells to proliferates; a step required for the amplification of the immune response. These mitogens have specific binding sites which may not be present on all cell types (111). For example, PHA binds to the Tcell receptor and CD2 antigens of most Tcells (111). Concanavlin A binds to all Tcells (CD3) and is more potent than PHA (112), while differences in the binding sites of PWM may exist between different species of animals. Although the binding of these mitogens to specific glycoproteins on the cell surface are integral to subsequent proliferation, these binding sites are many and varied on any given cell population. Thus, depending upon the mitogen employed and the species from which the cells are derived, immune defects in any population of cells may be exaggerated or underestimated during a proliferative event.
Interpretation of experimental results becomes more complex when examining specific immune responses toward antigens. Often the generation of antibody against Tcell-dependent antigens (e.g., SRBCs) are compared to antibody production against T-cell-independent antigens (e.g., DNP-Ficoll) as indicators of T-and B-cell competence respectively. Even though these regimens are optimized to provide maximal immune stimulation, the relative immunogenicity of the test antigens and the avidity and specificity of the antibodies generated may differ between assay systems. Thus, problems of antigen potency and antibody specificity may arise not only in the immunization stages in vivo but also in the detection ofantibody during immunoassay.
Insofar as there are shortfalls in conventional testing of immune function and in comparing the relative sensitivity of xenobiotic agents on the different arms of the immune response, Luster et al. (113,114) have demonstrated concordance between a chemical effect on the ability of mice to form antibody to SRBCs and its potential to cause an immunotoxic effect in any of a number of other assays measuring immune function. Changes in lymphocyte surface markers and, to a lesser extent, cell proliferation to mitogens and NK activity also showed good correlation with immunotoxicity; however, B-cell proliferation to LPS and the mixed lymphocyte reaction (graft rejection) did not.
In studying the possible immunomodulatory effects of 03 and other air pollutants, some attempts have been made to examine the effect of 03 on pulmonary immune responses such as local antibody production (22,23,68), NK activity (16), and lymphocyte function in the draining lymph nodes (68). Although these experiments are technically feasible, their interpretation can be challenging because of the constant trafficking of serum factors and immune Environmental Health Perspectives cells from the circulation to the lung and possible enhancement of these features during inflammation. For example, the observation of decreased lymphocyte sensitivity to antigen in the spleen following prolonged 03 exposure and a contrasting increase in MLN lymphocyte activity ( Figure 4A) could be explained by local inflammation and cellular recruitment to the site of oxidant damage (i.e., a simple matter of immune circulation). Until the cellular trafficking between the systemic immune compartment and that of the lung during immune stimulation are more clearly understood, these complicating factors will remain.

Considerations ofHost Factors
A number of studies have investigated the age-dependent response to 03 exposure on nonimmunologic bioparameters of the lung (115)(116)(117)(118)(119). In toto, these studies show that, depending on the end point, the age of the animals can play a significant role in the outcome of the effects of 0 exposure.
With respect to immunologic parameters, it should be kept in mind that the immune response is known to wane with age (120).
In addition to age, qualitative difference in nonimmunologic responses to 03 among animal species have also been documented (121)(122)(123)(124)(125)(126)(127). Also, it is becoming increasingly clear that pulmonary responses to 0 exposure is subject to genetic infuence (128)(129)(130)(131)(132), and it is also known that the immune response is under genetic control (133,134). While the wide diversity of receptor sites on antibody and T-cell receptors occur by gene rearrangement of the immunoglobulin superfamily, a more innate and stable genetic influence exists over host susceptibility in mice to intracellular infections such as Listeria, Mycobacteria, and Salmonellae spp. (135). In many cases, the host response, which governs protection from these organisms, is the DTH reaction, involving both T-cells and subsequent activation of macrophages. Genetic differences have also been recently observed in the ability of two different strains of mice to resist streptococcal challenge (50). It is also becoming dear that some mouse strains produce high or low titers of circulating antibody following identical immunization procedures (136). Thus, in addition to species differences, strain differences may also play a role in the immunotoxicology of 0; at this time, it is not known to what extent tie combination of variables that include age, animal species, and genetic constitution may influence the effect of 03 exposure on immune responsiveness.

Considerations ofthe Ozone Exposure Protocols
Investigations of the effects of 03 on immunologic end-points that do not involve antigen stimulation have primarily used the following protocols: a) continuous (or intermittent) exposure on consecutive days with tests performed during the exposure protocol; b) a short (hourly) exposure with tests performed immediately after exposure or at intervals thereafter; and c) continuous (or intermittent) exposure for several days or weeks with tests performed immediately after exposure or at intervals thereafter.
Using continuous 03 exposure with assays performed during a 2-to 4-week exposure period, four general response patterns emerge from the literature ( Figure 11). In the first (Figure 1 1A), an initial suppression followed by recovery has been observed with spleen weights (3), spleen/body weight ratios (68), thymus weight (4), AM phagocytosis (33,34,67,74), PM phagocytosis (74), pulmonary NK and splenic NK cell activity (16,68), MLN proliferative response (68), and AM-dependent intrapulmonary bacterial killing (67). The second response pattern (Figure 11 B), an initial suppression followed by an increased  (4,68). The third response pattern (Figure 11 C), an initial absence of a response followed by increased activity, has been observed with MLN (8) and splenic (14) proliferative responses. Finally, in Figure 1 ID, a sustained response of thymic atrophy has also been observed (11). The second exposure protocol consists of short (hourly) 03 exposures with the tests performed immediately after exposure or at various times thereafter. The data show that immediately after exposure, AM phagocytosis is suppressed (30,31,34) and then followed by reestablishment of phagocytosis as the interval between exposure and assay was lengthened. The third exposure protocol consists of continuous 0 exposure for several days (or weeks) followed by tests immediately after exposure or at various intervals thereafter. Using this exposure protocol, T/Blymphocyte ratios remained elevated for at least 5 days after exposure (10). Perhaps more studies were not performed with this 03 exposure protocol because of a lack of an effect. A nonresponse situation may be predicted by the observations that continuous 03 exposure, after a period of initial suppression, is followed by reestablishment  Figure 1lA. This is supported by our range-finding studies that showed that continuous 0.5 ppm 03 exposure for 10 consecutive days had no effect on AMdependent intrapulmonary bacterial killing of S. aureus, either immediately after cessation of exposure or at 1 and 5 days thereafter (GJ Jakab, unpublished observations).
Superimposed on the three exposure patterns are the 03 concentrations and whether the exposures are continuous or intermittent. In general, it appears that the 03 concentrations modulate the intensity and the duration of the responses. Intermittent 03 exposures can be viewed as alternating cycles of 0 -induced injury and repair that appear to blunt the intensity of the responses.
When the effects of 03 exposure on antigen stimulated responses are examined, the number of possibilities in exposure protocols is vastly increased. The simplest of these protocols is to expose animals to 0 either before or after immunization. The data show that continuous 03 exposure immediately before immunization had no effect on the DTH response to SRBCs (28), the DTH response to Listeria antigen, or the proliferative response of MLN to Listeria antigen (10), whereas 03 exposure after immunization suppressed the responses. Using primary and secondary immunizing regimens, we have investigated the timing between immunization and 03 exposure and found that as the interval was increased beyond 3 days, no effect was observed (68).

Summary
From the above discussion, it is apparent that many known and unknown factors impinge on the effect of 03 exposure on immunologic responsiveness. The extent to which nutritional factors, health status, age, species variability, genetic background, and immune assay sensitivity interact to play a role on the effect of 0 exposure on the immune system is difficult to unravel from the available data. Mechanistic studies will undoubtedly provide insights into the causes of the observed responses. Questions that beg answers include: a) Is there a level of 03 exposure below which immune responses are enhanced and above which there is a toxic effect that suppresses immune responses? b) What is the mechanism for adaptation? c) Given the fact that 03 itself never reaches the systemic immune compartment, what is the mechanism of 03induced suppression of systemic immunity? d) Where, in the sequence of events that leads to an antigen-stimulated immune response, does 0 have an effect? Is it antigen availability, antigen presentation or a step in the complex cascade of events in the afferent limb of the immune response which may result in altered effector function? As alluded to before, 03 and other oxidant gases are inflammatory when inhaled and thus cause a variety of permeability changes in the lung that could cause an increase in reactive immune tissue (BALT) or increase the availability of antigen to immunoreactive areas.
Much information is available regarding the effects of 03 exposure on immune function. However, this information has not paved major inroads toward an understanding of the mechanisms by which 03 exposure enhances/impairs/modulates immune function. The necessary database from which specific mechanistic hypotheses can be tested is available. The results of these studies will undoubtedly lead to a more complete understanding of the mechanisms by which 0 alters immune function.
Although it is expectated that future experimental studies will unravel the mechanisms by which 03 alters immune responsiveness, their relationship to human health effects will require serious consideration. The National Ambient Air Quality Standard for 0 is 0.12 ppm for a maximum of 1 hr that is not to be exceeded more than once per year. The worst case situation for human environmental exposure is in urban areas with concentrations reported as high as 0.36 ppm 0 3. (137). However, the ambient concentrations of 03 do not reflect the high 03 excursions documented in urban areas. Indeed, most experimental animal studies have had to use 03 exposures in excess of the ambient and spike urban 03 concentrations to demonstrate an immunotoxic effect. In this regard, the study of Selgrade et al. (18) found no effect on multiple immune parameters when rats were chronically exposed to 03 in a way that simulated an urban profile. Finally, in addition to the 03 concentrations, the inferences that can be drawn between human health effects and the battery of immune tests recommended in the National Toxicology Program guidelines for immunotoxicity evaluation in mice (113) need to be clarified.