Depletion of complement does not impact initiation of xenobiotic-induced autoimmune disease.

Deficiency in Daf1, a complement regulatory protein, has been shown to exacerbate development of various autoimmune diseases and recent studies have suggested that this may be explained by Daf1 acting to limit T-cell hyper-responsiveness. It has been suggested that the absence of Daf1 aggravates autoimmune disease in a complement-dependent manner, but others have shown that activation of T cells in the absence of Daf1 can be complement independent. However, the relationship between Daf1, complement components, lymphocyte activation, cytokine expression and antibody production remains to be determined in mice that are not Daf1 deficient. We have recently demonstrated, in murine mercury-induced autoimmunity (mHgIA), that an accumulation of CD44(high) Daf(low) CD4(+) T cells is associated with the development of autoimmunity. In this study we observed that complement depletion does not affect the accumulation of activated CD4(+) T cells, elevation of splenic interleukin-4 expression and autoantibody production in mHgIA. In addition, neither the accumulation of CD44(high) Daf(low) CD4(+) T cells nor the down-regulation of Daf1 expression on CD4(+) T cells was influenced by a lack of complement. In conclusion, these studies show that initiating events in xenobiotic-induced autoimmunity, including lymphocyte activation, cytokine expression and autoantibody production, are not dependent on complement.

Multiple experimental approaches of immunotoxic effects of mercury chloride in the blue mussel, Mytilus edulis, through in vivo, in tubo and in vitro exposures.

Biological impairments due to mercury discharge into the environment are now an issue of global concern. From the three forms of mercury found in aquatic ecosystems, the immunotoxic effects of mercury chloride were examined in the model animal, the blue mussel. In order to investigate the toxic potency of this chemical, three exposure regimes were carried out: chronic exposure of groups of individuals, a new protocol "in tubo" designed for sub-acute exposures of individuals, and acute exposures of target cells. Chronic exposure revealed significant immunotoxic effects after 7 days at 10(-6)M, while acute exposures showed significant inhibition of phagocytosis at 10(-4)M and 10(-3)M. In sub-acute exposures both circulating haemocytes and haemocyte mortality increased at 10(-4)M and 10(-3)M while phagocytosis and the clearance rate drew hormetic toxic effects on healthy individuals. These results suggest the use of the "in tubo" design for bivalve toxicological individual studies.

Dose and Hg species determine the T-helper cell activation in murine autoimmunity.

Inorganic mercury (mercuric chloride--HgCl(2)) induces in mice an autoimmune syndrome (HgIA) with T cell-dependent polyclonal B cell activation and hypergammaglobulinemia, dose- and H-2-dependent production of autoantibodies targeting the 34 kDa nucleolar protein fibrillarin (AFA), and systemic immune-complex deposits. The organic mercury species methylmercury (MeHg) and ethylmercury (EtHg--in the form of thimerosal) induce AFA, while the other manifestations of HgIA seen after treatment with HgCl(2) are present to varying extent. Since these organic Hg species are converted to the autoimmunogen Hg(2+) in the body, their primary autoimmunogen potential is uncertain and the subject of this study. A moderate dose of HgCl(2) (8 mg/L drinking water--internal dose 148 micro gHg/kg body weight [bw]/day) caused the fastest AFA response, while the induction was delayed after higher (25 mg/L) and lower (1.5 and 3 mg/L) doses. The lowest dose of HgCl(2) inducing AFA was 1.5 mg/L drinking water which corresponded to a renal Hg(2+) concentration of 0.53 micro g/g. Using a dose of 8 mg HgCl(2)/L this threshold concentration was reached within 24 h, and a consistent AFA response developed after 8-10 days. The time lag for the immunological part of the reaction leading to a consistent AFA response was therefore 7-9 days. A dose of thimerosal close to the threshold dose for induction of AFA (2 mg/L drinking water--internal dose 118 micro gHg/kg bw per day), caused a renal Hg(2+) concentration of 1.8 micro g/g. The autoimmunogen effect of EtHg might therefore be entirely due to Hg(2+) formed from EtHg in the body. The effect of organic and inorganic Hg species on T-helper type 1 and type 2 cells during induction of AFA was assessed as the presence and titre of AFA of the IgG1 and IgG2a isotype, respectively. EtHg induced a persistent Th1-skewed response irrespectively of the dose and time used. A low daily dose of HgCl(2) (1.5-3 mg/L) caused a Th1-skewed AFA response, while a moderate dose (8 mg/L) after 2 weeks resulted in a balanced or even Th2-skewed response. Higher daily doses of HgCl(2) (25 mg/L) caused a balanced Th2-Th1 response already from onset. In conclusion, while metabolically formed Hg(2+) might be the main AFA-inducing factor also after treatment with EtHg, the quality of the Hg-induced AFA response is modified by the species of Hg as well as the dose.

Self major histocompatibility complex class-II-specific regulatory CD4 T cells prevent both Th1- and Th2-mediated autoimmune diseases in the rat.

It is clear that functional heterogeneity of T cells may be explained by differential cytokine production. The aim of this paper was to review evidence for regulatory cells, generated after HgCl(2)-exposure. They differ from classical Th1 and Th2 cells, produce transforming growth factor-beta and interleukin-10 and exert their regulatory functions in a Th1/Th2-unrestricted fashion.

Antinucleolar autoantibody induced by mercuric chloride in mice: does sodium selenite inhibit autoantibody production?

Repeated exposure of subtoxic doses of mercuric chloride (HgCl2) can induce antinucleolar autoantibodies in susceptible mice. To study the immunopathological mechanism(s), sodium selenite (Na2SeO3), which is known to reduce the toxicity of mercury, is injected simultaneously with HgCl2 into mice. Equimolar Na2SeO3 nor the same amount of Se in Na2SeO3 as Hg in HgCl2 could not reduce antinucleolar autoantibody induction by HgCl2. These results indicate that the mechanism of autoimmunity induction by HgCl2 might be independent of its toxicity.

Blockade of OX40-ligand after initial triggering of the T helper 2 response inhibits mercuric chloride-induced autoimmunity.

Mercuric chloride (HgCl2)-induced autoimmunity in Brown Norway rats is a spontaneously resolving autoimmune response driven by the activation of T helper type 2 lymphocytes (Th2 cells). Treatment with antibody to OX40-ligand (OX40-L) from the time of the first HgCl2 injection for 12 days had little effect. Delayed treatment commenced 8 days after the first HgCl2 injection significantly suppressed immunoglobulin E production, splenomegaly, weight loss and mortality. This makes OX40/OX40-L signalling an attractive therapeutic target for Th2-driven autoimmune diseases. Intravenous administration of the murine antibody to OX-40-L elicited a vigorous anti-mouse immunoglobulin antibody response that was significantly enhanced compared to the response to control immunoglobulin. It is likely that this response significantly reduced the plasma half-life of the anti-OX40-L antibody and this observation has clear implications for the interpretation of data from experiments where anti-OX40-L is used in vivo.

Bacterial lipopolysaccharide both renders resistant mice susceptible to mercury-induced autoimmunity and exacerbates such autoimmunity in susceptible mice.

The initiation and severity of systemic autoimmune diseases are influenced by a variety of genetic and environmental factors, in particular bacterial infections and products. Here, we have employed bacterial lipopolysaccharide (LPS), which non-specifically activates the immune system, to explore the involvement of innate immunity in mercury-induced autoimmunity in mice. Following treatment of mouse strains resistant [DBA/2 (H-2(d))] or susceptible [SJL(H-2(s))] to such autoimmunity with mercuric chloride and/or LPS or with physiological saline alone (control), their immune/autoimmune responses were monitored. Resistant DBA/2 mice were rendered susceptible to mercury-induced autoimmunity by co-administration of LPS, exhibiting pronounced increases in the synthesis of IgG1 and IgE, high titres of IgG1 deposits in the kidneys and elevated circulating levels of IgG1 antibodies of different specificities. Furthermore, the percentages of the T cells isolated from the spleens of DBA/2 mice exposed to both mercury and LPS that produced pro-inflammatory cytokines were markedly increased by in vitro stimulation with phorbol myristate acetate (PMA) and ionomycin, which was not the case for splenic T cells isolated from mice receiving mercuric chloride, LPS or saline alone. In addition, exposure of susceptible SJL mice to mercury in combination with LPS aggravated the characteristic features of mercury-induced autoimmunity, including increased synthesis of IgG1 and IgE, the production of IgG1 anti-nucleolar antibodies (ANolA) and the formation of renal deposits of IgG1. In summary, our findings indicate that activation of the innate immune system plays a key role in both the induction and severity of chemically induced autoimmunity.

Exposure to mercuric chloride during the induction phase and after the onset of collagen-induced arthritis enhances immune/autoimmune responses and exacerbates the disease in DBA/1 mice.

In susceptible mice, mercuric chloride induces a systemic autoimmune response that is characterized by elevated serum levels of immunoglobulin G1 (IgG1) and immunoglobulin E (IgE), production of anti-nucleolar antibodies (ANolAs) and the formation of renal IgG deposits. We have previously shown that mercury can also enhance immune/autoimmune responses in mouse strains genetically prone to develop spontaneous autoimmune disease. Here, we investigated whether mercury can enhance the severity of murine collagen-induced arthritis (CIA), an inducible (acquired) autoimmune disease that cannot be induced by mercury itself. While mercury administered prior to the induction phase of CIA exerted little, if any, influence, administration of mercury during the induction phase and following onset aggravated the symptoms of this disease and increased the serum levels of IgE and IgG2a antibodies directed against collagen type II (CII). Furthermore, while animals injected with mercury alone exhibited a significant decrease in the ratio of the levels of interferon-gamma (IFN-gamma) to interleukin-4 (IL-4) mRNA in their spleens, this ratio was increased in mice with CIA, with or without administration of mercury. Finally, the production of anti-nuclear antibodies, a hallmark of autoimmunity in response to mercury, was observed in all mice with CIA treated with this heavy metal. Our findings suggest that exposure to mercury during the development of CIA may influence immunological factors in such a way as to synergistically promote disease development.

Lichenoid reactions of murine mucosa associated with amalgam.

BACKGROUND: In 97% of all patients with oral lichenoid reactions (OLR) associated with dental amalgam a removal of the fillings leads to a decline of the lesions, as a minimum. OBJECTIVES: The aim of this study was to determine if contact allergic or local toxic effects or both may contribute to OLR using an animal model with mercury-sensitive and non-sensitive rats. METHODS: Twenty Brown Norway rats, which have a genetic predisposition for an autoimmune syndrome after exposure to mercury and 20 Lewis rats, not mercury sensitive, were treated as follows: 10 animals of each group were sensitized with a low dose of mercuric chloride. Half of all animals received local exposure of the right buccal mucosa to amalgam (left: control), the others to amalgam alloy free of mercury. All rats were patch tested with an amalgam series. RESULTS: After 20 days of exposure 96% of all animals showed white mucosal lesions restricted to the contact zone of the alloy on the treated side, but only up to 25% had a positive patch test reaction to amalgam or inorganic mercury (INM). The lesions showed no relation to species, alloy, sensitization or patch test reaction. CONCLUSIONS: While allergic mechanisms may contribute to mucosal contact lesions in Brown Norway rats, this is less probable in Lewis rats. Mercury in general appears to be irrelevant in the development of ORL in this study. If this holds true for humans as well, patch testing with an amalgam series may be helpful in a minor fraction of all patients with OLR.

Resistance to re-challenge in the Brown Norway rat model of vasculitis is not always complete and may reveal separate effector and regulatory populations.

Administration of mercuric chloride to Brown Norway rats results in T helper type 2 (Th2)- dominated autoimmunity characterized by high immunoglobulin E (IgE) concentrations, the production of multiple IgG autoantibodies, including those to glomerular basement membrane (GBM), arthritis and caecal vasculitis. After 14 days animals immunoregulate and auto-immunity resolves even if mercuric chloride injections are continued. In a third phase, if animals are re-challenged with mercuric chloride 6 weeks later, they show only attenuated autoimmunity with lower anti-GBM antibody concentrations and arthritis scores. Resistance to the induction of anti-GBM antibodies can also be achieved following an initial challenge with low-dose (one-tenth standard dose) mercuric chloride. We have now studied this resistant phase in more detail. We have shown, first, that following an initial full-dose mercuric chloride challenge, resistance also affects susceptibility to caecal vasculitis. Second, following an initial full-dose mercuric chloride challenge, the IgE response upon re-challenge is initially accelerated but subsequently enters a resistant phase and third, following an initial challenge with low-dose mercuric chloride, resistance is also seen to the induction of caecal vasculitis but is not seen in IgE serology (where results suggest competing effector and regulatory cell populations). Studying such regulatory phases in animal models of autoimmunity may be of benefit in the future in designing new therapies for human vasculitis.

Genetic control of resistance to mercury-induced immune/autoimmune activation.

Previous studies have shown that genetic factors control the susceptibility to mercury-induced immunoglobulin (Ig)G1 antibody formation, IgE synthesis, renal IgG deposits and antinucleolar autoantibodies (ANolA) production in the susceptible mice. In this study, we examined the genetic control of resistance to these characteristics after HgCl2 injection in F1 hybrid crosses between the highly mercury resistant DBA/2 and mercury susceptible NZB (H-2d), SJL (H-2 s), A.CA (H-2f) and DBA/1 (H-2q) mice and also in backcross hybrids between (DBA/2 x SJL)F1 and SJL mice. We observed that mercury-induced immune/autoimmune manifestations were profoundly downregulated in most (if not all) of the F1 hybrids, indicating that the resistance to mercury was a dominant trait. Analysis of mercury-induced immune/autoimmune responses in the (DBA/2 x SJL) x SJL backcross hybrids suggested that only one gene or a cluster of genes determined the resistance to the ANolA production, whereas the resistance to other characteristics was controlled by two and/or three gene loci. By H-2 genotyping the backcross mice, it was found that H-2d haplotype per se could confer resistance to ANolA production. However, we did not find any significant association between the H-2d haplotype and the resistance to increase of IgG1 and IgE synthesis and the development of renal IgG1 deposits. Thus, while in DBA/2 mice, gene(s) in the H-2 loci strictly contribute to the inheritance of resistance to ANolA production; non-H-2 genes mainly govern the inheritance of unresponsiveness regarding other characteristics.

Monoclonal antibodies specific for mercuric ions.

Monoclonal antibodies (mAbs) that react with soluble mercuric ions have been produced by injection of BALB/c mice with a hapten-carrier complex designed to maximize exposure of the metal to the immune system. Three hybridomas producing antibodies that reacted with bovine serum albumin (BSA)-glutathione-HgCl, but not with BSA-glutathione, were isolated from the spleen of a mouse given multiple injections with glutathione-HgCl conjugated to keyhole limpet hemocyanin. Stable subclones were established from two of these antibodies, designated mAb 4A10 and mAb 1F10. The binding of both antibodies to immobilized BSA-glutathione-HgCl was inhibited by soluble HgCl2, and dissociation constants for mercuric chloride binding were 2.3 and 3.7 nM for mAbs 4A10 and 1F10, respectively. Both antibodies bound mercuric acetate with similar affinities, demonstrating that the antibodies were capable of binding to mercuric ions in the presence of a different counterion than the one used in the immunogen. Reactions were not observed with other metal cations by either antibody. These data demonstrate the successful induction of antibodies that react very specifically with mercuric ions in solution regardless of the presence of a carrier.

The functional affinity and specificity of autoantibodies in animal models of anti-glomerular basement membrane disease.

We studied characteristics of the anti-glomerular basement membrane (GBM) antibody response in three animal models of Goodpasture's disease: treatment of Brown Norway (BN) rats with the polyclonal activator mercuric chloride (HgCl2), and immunization of BN and Wistar Kyoto (WKY) rats with rat GBM. Serial serum samples were obtained over the time course of the models, and anti-GBM antibodies eluted from the kidneys. Functional affinity of the anti-GBM antibodies was measured in a solid-phase ELISA incorporating the mild chaotropic agent diethylamine. Evidence for shared epitope specificity with human anti-GBM antibodies was sought using competition ELISA. As with recent studies in human anti-GBM disease, there was no evidence for affinity maturation of the anti-GBM response in the serum of any of the animal models. Antibodies eluted from the kidneys were of higher functional affinity than serum antibodies only in the HgCl2-treated BN rat. There was no obvious correlation between the functional affinity of the antibodies and the severity of nephritis in the three models. Competition studies between eluted anti-GBM antibodies from the rat models and human anti-GBM antibodies did not provide any evidence for shared epitope recognition. This study provides further information on the extent to which these models reflect the human disease.

Effects of decomplementation with cobra venom factor on experimental vasculitis.

Mercuric chloride (HgCl2) induces autoimmunity in susceptible rat strains, with hyper-IgE, appearance of a number of autoantibodies, and widespread tissue injury, including necrotizing vasculitis in the gut. In the early phase of tissue injury there is granulocyte infiltration; later there is immunoglobulin deposition along basement membranes in vessels. We have analysed the role of complement in this model using cobra venom factor (CVF), which causes decomplementation lasting around 5 days. The characteristic time course when HgCl2 is given over 10 days is that tissue injury and autoantibody levels reach a peak at around day 15 (start of HgCl2 = day 0). We therefore gave CVF either early (day 0), intermediate (day 5) or late (day 10); a fourth group (controls) received HgCl2 but no CVF. At each time point, CVF caused complete decomplementation which lasted for at least 5 days. Serum IgE and autoantibody levels were similar in all four experimental groups. Tissue injury in the 'early' CVF group and in the 'late' CVF group was not significantly different from controls, but in the intermediate group tissue injury was significantly more severe than in controls. These data indicate that the complement system does not play a major role in the induction of autoantibodies by HgCl2, nor in the effector phase of tissue injury. We speculate that the exacerbation of tissue injury by CVF in the group given this agent at an intermediate stage of the model is explained by the presence of products of C3 activation which have proinflammatory effects during the phase of active granulocyte-mediated tissue injury.

Low-dose mercuric chloride induces resistance in brown norway rats to further mercuric chloride by up-regulation of interferon-gamma.

Mercuric chloride induces autoimmunity in Brown Norway rats with polyclonal B-cell activation, hyper-IgE and multiple autoantibodies. Pre-treatment with low-dose HgCl2 (one-tenth of the standard dose) induces resistance to later full-dose HgCl2; we have studied the mechanism of this resistance. Brown Norway rats given low-dose HgCl2 showed only a modest increase in serum IgE level, three logs lower than rats given standard-dose HgCl2, and no up-regulation of splenic interleukin (IL)-4 mRNA. There was up-regulation of splenic interferon (IFN)-gamma gene expression and a progressive rise in serum IFN-gamma. Neither IL-12 nor IL-18 were induced, but there was up-regulation of IL-12 receptor beta2-chain (IL-12Rbeta2) expression. IL-10 and transforming growth factor (TGF)-beta expression did not change. Serum IgE and splenic IL-4 mRNA expression remained static when these rats were rechallenged, confirming resistance. Thereafter IFN-gamma expression gradually fell, after which IL-4 expression and serum IgE rose slightly. Our observations suggest that low-dose HgCl2 confers protection in Brown Norway rats to further HgCl2 by up-regulation of IFN-gamma, associated with enhanced IL-12Rbeta2 expression. The immunological response to HgCl2 in susceptible rat strains is more complex than previously appreciated and is dose dependent, with low doses inducing a T helper '(Th)1' type of response in contrast to the 'Th2' type response associated with standard doses.

No evidence for specific in vitro lymphocyte reactivity to HgCl2 in patients with dental amalgam-related contact lesions.

Blood lymphocytes from 20 patients with oral contact lesions to dental amalgam and 10 healthy individuals were analyzed for HgCl2-induced proliferation in vitro, using both a modified assay and a conventional assay. The release of interferon-gamma (IFN-gamma) was measured in cell supernatants. Six patients displayed positive reactions in patch tests to mercuric compounds. No significant differences were recorded in HgCl2-induced proliferation in cells from patients and controls, since only few in the whole material responded to submitogenic concentrations. IFN-gamma was detectable in cell supernatants from some patients but also from controls and is not predictive of mercury allergy. Neither the phenotypes of peripheral lymphocyte subsets, the frequency of circulating cells expressing the interleukin-2 (IL-2) receptor, spontaneous lymphocyte proliferation nor concentrations of serum interleukin-6 differed between patient and control samples. In contrast to what has been claimed before, we did not find any evidence for specific in vitro lymphocyte reactivity in patients with oral contact lesions.

Prevention of experimental autoimmune uveoretinitis and experimental autoimmune pinealitis in (Lewis x Brown-Norway) F1 rats by HgCl2 injections.

Mercuric chloride (HgCl2) induces in Brown-Norway (BN) and (Lewis x Brown-Norway) F1 hybrid rats a transient autoimmune disease characterized by the production of various antibodies to self and non-self antigens and by a dramatic increase of serum IgE. Experimental autoimmune uveoretinitis (EAU) can be induced in Lewis (LEW) and (LEW x BN) F1 hybrid rats by a single immunization with retinal S-antigen (S-Ag). Besides uveoretinitis, animals immunized with S-Ag develop an autoimmune pinealitis (EAP). We demonstrate in this study that (LEW x BN) F1 hybrid rats, injected with HgCl2 7 days before S-Ag immunization, are quite efficiently protected against EAU and EAP. We also show that HgCl2-induced protection is neither due to a cytotoxic effect of HgCl2 nor to CD8+ T-cell dependent mechanisms nor to the HgCl2-induced increase of serum IgE concentration. The role of other hypothetical mechanisms, such as anti-S-Ag anti-idiotypic antibodies and/or HgCl2-induced unbalance between T-helper cell subsets, is discussed.

Resistance to xenobiotic-induced autoimmunity maps to chromosome 1.

Although evidence indicates that environmental factors play a major role in precipitating systemic autoimmunity in genetically susceptible individuals, little is known about the mechanisms involved. Certain heavy metals, such as mercury, are potent environmental immunostimulants that produce a number of immunopathologic sequelae, including lymphoproliferation, hypergammaglobulinemia, and overt systemic autoimmunity. Predisposition to such metal-induced immunopathology has been shown to be influenced by both MHC and non-MHC genes, as well as susceptibility to spontaneous lupus, in mice and other experimental animals. Among the various mouse strains examined to date, the DBA/2 appears to uniquely lack susceptibility to mercury-induced autoimmunity (HgIA), despite expressing a susceptible H-2 haplotype (H-2d). To define the genetic basis for this trait, two genome-wide scans were conducted using F2 intercrosses of the DBA/2 strain with either the SJL or NZB strains, both of which are highly susceptible to HgIA. A single major quantitative trait locus on chromosome 1, designated Hmr1, was shown to be common to both crosses and encompassed a region containing several lupus susceptibility loci. Hmr1 was linked to glomerular immune complex deposits and not autoantibody production, suggesting that DBA/2 resistance to HgIA may primarily involve the later stages of disease pathogenesis. Identification and characterization of susceptibility/resistance genes and mechanisms relevant to the immunopathogenesis of mercury-induced autoimmunity should provide important insights into the pathogenesis of autoimmunity and may reveal novel targets for intervention.

Differential regulation of expression of the MHC class II molecules RT1.B and RT1.D on rat B lymphocytes: effects of interleukin-4, interleukin-13 and interferon-gamma.

Susceptibility to induction of both T helper 1- (Th1) and Th2-mediated autoimmunity is multifactorial and involves genetic linkage to the major histocompatibility complex (MHC) class II haplotype. Brown Norway (BN) rats exposed to mercuric chloride develop a Th2-dependent systemic autoimmunity, whereas Lewis rats, which are highly susceptible to Th1-mediated autoimmunity, develop immune suppression after mercuric chloride exposure. Exposure to mercuric chloride is known to enhance B-lymphocyte expression of the MHC class II molecule RT1.B, predominantly in BN rats. We demonstrate that, in contrast, expression of RT1.D was unmodified on these B cells, whereas both RT1.B and RT1.D were up-regulated on epithelial cells. Regulation of B-cell MHC class II isotype expression was further studied in vitro, using BN rat lymph node (LN) cells. Interleukin-4 (IL-4) strongly enhanced B-cell expression of RT1.B (2.8-fold), whereas RT1.D expression was only slightly, although significantly, modified (1.2-fold). B cells from Lewis rats showed a similar IL-4-induced enhancement of RT1.B expression (2.5-fold), whereas, in contrast, RT1.D expression was unmodified. Exposure of LN cells from BN rats to interferon-gamma induced a moderate increase of B-cell MHC class II expression, predominantly of RT1.B. Strong and rapid enhancement of B-cell RT1.D expression was observed after stimulation by phorbol 12-myristate 13-acetate and ionomycin. Rat IL-13 did not modify B-cell MHC class II expression; however, it induced typical morphological changes in peritoneal macrophages. These experiments demonstrate isotype-specific and strain-dependent regulation of MHC class II expression on rat B lymphocytes, which may be of pathophysiological relevance for the strain-dependent susceptibility for Th1- or Th2-mediated autoimmunity.