Abstract
Infection with the protozoan parasite Trypanosoma cruzi leads to acute myocarditis that is accompanied by autoimmunity to cardiac myosin in susceptible strains of mice. It has been difficult to determine the contribution of autoimmunity to tissue inflammation, because other inflammatory mechanisms, such as parasite-mediated myocytolysis and parasite-specific immunity, are coincident during active infection. To begin to investigate the contribution of myosin autoimmunity to myocarditis, we selectively inhibited myosin autoimmunity by restoring myosin tolerance via injection of myosin-coupled splenocytes. This tolerization regimen suppressed the strong myosin-specific delayed-type hypersensitivity (DTH) that normally develops in infected mice, although it did not affect myosin-specific Ab production. Suppression of myosin autoimmunity had no effect on myocarditis or cardiac parasitosis. In contrast, myosin tolerization completely abrogated myocarditis in mice immunized with purified myosin, which normally causes severe autoimmune myocarditis. In this case, myosin-specific DTH and Ab production were significantly reduced. We also examined the contribution of T. cruzi-specific immunity to inflammation by injection of T. cruzi-coupled splenocytes before infection. This treatment reduced T. cruzi DTH, although there was no effect on parasite-specific Ab production. Interestingly, cardiac inflammation was decreased, cardiac parasitosis was significantly increased, and mortality occurred earlier in the parasite-tolerized animals. These results indicate that myosin-specific autoimmunity, while a potentially important inflammatory mechanism in acute and chronic T. cruzi infection, is not essential for inflammation in acute disease. They also confirm previous studies showing that parasite-specific cell-mediated immunity is important for myocarditis and survival of T. cruzi infection.
Chagas’ heart disease (CHD) 3 is a potentially fatal dilated cardiomyopathy resulting from infection by the protozoan parasite Trypanosoma cruzi. Approximately 16–18 million people are infected with this parasite, and 120 million are at risk of infection in the South and Central America (1). CHD manifests as an acute and chronic myocarditis that develops in ∼30% of T. cruzi-infected individuals. The exact mechanism(s) of disease pathogenesis, and hence treatment, are under debate (2, 3, 4, 5). One proposed mechanism is that CHD is an autoimmune disease induced by T. cruzi (6). In other words, T. cruzi infection induces autoimmune responses in the susceptible host, and these autoimmune responses cause tissue damage and inflammation. The evidence for autoimmune-mediated damage in CHD has been reviewed (7). T. cruzi has been shown to induce humoral and cellular autoimmunity to a diverse set of self Ags (reviewed in Refs. 4 and 5), including cardiac myosin (8, 9); however, only a few studies have demonstrated that autoimmunity induced by T. cruzi directly contributes to tissue inflammation. Immunization with T. cruzi proteins induces cardiac abnormalities (10, 11); transfer of CD4+ splenocytes from infected mice to naive recipients promotes rejection of cardiac grafts (12); and tolerance to a myosin-enriched heart homogenate reduces cardiac inflammation in chronically infected mice (13).
We previously showed that the level of myosin autoimmunity in acutely infected mice was similar to that in mice with autoimmune myocarditis induced by immunization with myosin in CFA (9). Myosin may be an important cardiac autoantigen because it is the most abundant protein in the heart, and immune responses to myosin are sufficient to induce myocarditis in the absence of infection (14). In addition, myosin autoimmunity is induced in response to a variety of cardiac insults, including viral infection (15), protozoal infection (9), and cardiac transplantation (16) and surgery (17). In all these instances, the precise role of myosin autoimmunity is unknown: is it an epiphenomenon or does it contribute to pathology? To address this question in our model of acute CHD myocarditis, we specifically suppressed myosin autoimmunity by Ag-coupled splenocyte tolerance and asked whether suppression of myosin autoimmunity would ameliorate acute myocarditis upon T. cruzi infection.
Ag-coupled splenocyte tolerance has been used for four decades to restore peripheral immune tolerance to haptens (18), organ homogenates (19), proteins, and peptides (20). Ag-coupled splenocyte tolerance depends on the route and timing of administration (21); MHC class II (22); the dose of Ag and number of splenocytes administered (19); and, shown most recently on CTLA-4 (23), Fas, and FasL, and cross-presentation by splenic dendritic cells (24). The mechanism of tolerization is thought to be due to induction of anergy in Ag-specific host T cells, possibly through a switch from Th1 to Th2 cell phenotype (25, 26). Ag-coupled splenocyte tolerance induction has been used to abrogate Theiler’s murine virus-induced demyelination and inflammation (27), prevent the development of lesions induced by Leishmania major infection (28), reduce mortality associated with herpes simplex virus infection (29), reduce the severity of coxsackievirus B3-induced myocarditis (30), and ameliorate several experimental autoimmune diseases (19, 31, 32, 33), including myosin-induced autoimmune myocarditis (34).
In this study, we investigated the contribution of myosin autoimmunity to inflammation and tissue damage in acute experimental CHD. We suppressed cellular myosin autoimmunity in infected and myosin-immunized mice by coupled splenocyte tolerance induction to cardiac myosin. Myosin tolerance induction did not affect myocarditis in infected mice, while it did reduce inflammation and fibrosis in myosin-immunized mice. In contrast, T. cruzi tolerance induction decreased cardiac inflammation, increased cardiac parasitosis in infected mice, and was associated with enhanced mortality. These results indicate that myosin autoimmunity does not significantly contribute to cardiac inflammation and that targeted anti-T. cruzi-specific immunity is important for myocarditis and control of parasitosis and survival in acute CHD.
Materials and Methods
Experimental animals and T. cruzi infections
Four- to six-week-old male A/J mice (The Jackson Laboratory, Bar Harbor, ME) were housed under specific pathogen-free conditions. Mice were infected by i.p. injection of 1 × 104 Brazil strain T. cruzi trypomastigotes derived from infection of tissue culture H9C2 rat myoblasts (American Type Culture Collection, Manassas, VA). Parasitemias were measured from tailbleeds by counting on a hemacytometer. Uninfected controls received an i.p. injection of Dulbecco’s PBS (Life Technologies, Grand Island, NY) of equal volume. Mice were anesthetized by a single i.p. injection of sodium pentobarbital (60 mg/kg) for each experimental manipu-lation. The use and care of mice were conducted in accordance with the guidelines of the Center for Comparative Medicine at Northwestern University.
Preparation of myosin and T. cruzi Ag
Cardiac myosin H chains were purified according to the method of Shiverick et al. (35), with modifications as described (9). T. cruzi Ag was prepared from T. cruzi epimastigotes, as described (9).
Induction of autoimmune myocarditis
Mice were immunized with purified cardiac myosin or OVA (Sigma-Aldrich, St. Louis, MO) (300 μg) in an emulsion of CFA (Difco, Detroit, MI) in a total volume of 0.1 ml. Three sites in the dorsal flank received s.c. injections. Seven days later, mice were boosted in an identical manner.
Preparation of Ag-coupled splenocytes
Splenocytes were purified from naive mice and incubated at 37°C for 10 min in Tris-NH4Cl (0.017 M Tris-free base, 0.14 M NH4Cl, pH 7.2) to lyse erythrocytes and washed twice in HBSS (Life Technologies). The splenocytes were then coupled with Ag using the cross-linking chemical, ethylene carbodiimide (ECDI; Calbiochem, La Jolla, CA), in the following manner. The splenocytes were washed in unbuffered saline (0.15 M NaCl, pH 7.2), collected by centrifugation at 1500 × g for 10 min in a 50-ml conical tube, and resuspended at a final concentration of 5 × 108 cells/ml in cold unbuffered saline containing 2 mg/ml purified cardiac myosin, 2 mg/ml BSA (Sigma-Aldrich), or 2 mg/ml T. cruzi lysate. Immediately after resuspension of splenocytes with Ag, the coupling reaction was initiated by the addition of 0.5 ml of freshly prepared ECDI (150 mg/ml in unbuffered saline) per ml of the cell suspension and incubated for 1 h on ice with periodic gentle inversion. The cells were washed three times in HBSS and maintained at 4°C until injection into mice.
Induction of peripheral tolerance
Tolerance was induced by the i.v. injection of 7.5 × 107 Ag-coupled splenocytes in 0.5 ml HBSS, as described (34). Seven days postinjection, mice were infected with T. cruzi or immunized with myosin or OVA, as described above.
Histopathology
Hearts were removed, rinsed with saline (PBS), and fixed for 24 h in 10% buffered Formalin. Fixed hearts were embedded in paraffin, sectioned, stained with H&E or Masson’s trichrome, and examined by light microscopy. Two sections were taken from each heart, one including both atria and the other both ventricles. Each section was examined for evidence of mononuclear and polynuclear cellular inflammation, necrosis and mineralization, T. cruzi pseudocysts, and fibrosis, and was assigned a histologic score of between 0 (no involvement noted) and 4 (100% involvement), with 1, 2, and 3 representing 25, 50, and 75% involvement of the histologic section (36). Independent observers obtained substantial (0.60–0.79) to almost perfect (0.80–1.00) agreement on their scoring by weighted κ statistic (37) as measured on a representative sample of 20 heart sections: overall agreement (0.82), inflammation (0.80), necrosis (0.96), pseudocysts (0.66), and fibrosis (0.92).
Serologic analysis
Levels of cardiac myosin-specific and T. cruzi-specific IgG were determined by ELISA, as described (9). Endpoint dilution titers for total IgG were defined as the highest serum dilution that resulted in an absorbance value (OD450) of 2 SDs above the mean of a negative control sample (pooled sera from uninfected mice) included on every plate.
Delayed-type hypersensitivity (DTH)
Myosin-specific and T. cruzi-specific DTH was quantitated using a standard ear-swelling assay (9). Ag-induced ear swelling was the result of mononuclear cell infiltration and exhibited typical DTH kinetics (i.e., minimal swelling at 4 h, maximal swelling at 24–48 h postinjection).
Statistical analyses
The statistical significance of DTH, parasitemia, or log-transformed (base 2) serum Ab titer was analyzed by one-way ANOVA, followed by a two-tailed t test and post hoc Bonferroni analysis. Comparison of histologic scores was analyzed by the Pearson’s χ2 test, and agreement of blinded observer histopathology scores was analyzed by weighted κ. Significant differences in mortality were analyzed by log rank test. Values of p < 0.05 were considered significant.
Results
Preservation of peripheral immune tolerance to myosin and T. cruzi decreases myosin-specific and T. cruzi-specific cellular immunity, but not humoral immunity
We were interested in investigating whether myosin autoimmunity contributes to cardiac inflammation and tissue damage in acute experimental CHD. To this end, we administered myosin-coupled splenocytes to mice before infecting them with T. cruzi to preserve peripheral immune tolerance to this self Ag (schematic shown in Fig. 1⇓). If effective, this treatment would selectively prevent the development of myosin-specific autoimmunity, while permitting other immune responses, including antiparasite and other autoimmune responses, to proceed. A BSA-tolerized group was included as a negative control, and a group tolerized with splenocytes coupled to T. cruzi lysate was added to test the effect of inhibiting parasite-specific cell-mediated immunity on acute disease. Each group was split into subgroups that were infected with T. cruzi or injected with saline. Finally, a separate set of groups was immunized with purified cardiac myosin in CFA to induce cardiac autoimmune disease and tolerized to either myosin or BSA as a positive control for tolerance induction to myosin. We have previously shown that myosin tolerization completely prevents development of the severe cardiac inflammation that develops 3 wk after immunization with myosin (34).
Protocol for the induction of Ag-specific peripheral tolerance in T. cruzi and myosin-induced myocarditis. Splenocytes from naive syngeneic donors were incubated with BSA, purified cardiac myosin (Myosin), or T. cruzi lysate (T. cruzi) in the presence of the cross-linking chemical ECDI to couple the Ag to the surface of the cells. The Ag-coupled splenocytes were suspended in HBSS, and 75 × 106 Ag-coupled splenocytes were injected into the tail vein of each naive recipient mouse. Seven days later, these tolerized mice were immunized with purified myosin or infected with T. cruzi using our standard protocols. Twenty-one days later, mice were analyzed for myocarditis by histopathology and for myosin or T. cruzi immune responses by measurement of Ag-specific DTH and Ab production.
First, we confirmed that the tolerization treatment effectively suppressed Ag-specific immunity by assaying for DTH to the tolerized Ag. There is no significant difference in myosin DTH between BSA-tolerized and untolerized myosin-immunized mice (34). Myosin tolerization led to significant inhibition of myosin-specific DTH in both myosin-immunized mice and T. cruzi-infected mice (Fig. 2⇓A). Likewise, T. cruzi tolerization effectively inhibited T. cruzi DTH (Fig. 2⇓B). To our surprise, myosin tolerization suppressed T. cruzi DTH (Fig. 2⇓B), and T. cruzi tolerization suppressed myosin DTH (A). The suppression of T. cruzi DTH by myosin tolerization was not due to a general suppressive effect of myosin, because myosin-tolerized mice immunized with OVA in CFA exhibited no suppression of OVA DTH compared with controls (data not shown).
Myosin tolerization and T. cruzi tolerization inhibit myosin-specific and T. cruzi-specific cell-mediated immunity. Myosin-specific and T. cruzi-specific DTH were measured by a 24-h ear-swelling assay. A, Myosin-specific DTH was assessed in four groups of mice: myosin immunized (Myosin Immunized), saline immunized (Saline Immunized), T. cruzi-infected (T. cruzi Infected), or injected with saline (Saline Injected). Within each group, mice were treated with the indicated tolerogen: BSA, cardiac myosin (Myosin), no tolerogen (None), or T. cruzi lysate (T. cruzi). B, T. cruzi-specific DTH was assessed in T. cruzi-infected (T. cruzi Infected) or saline-injected (Saline Injected) mice. Among the infected mice, subgroups were treated with the tolerogens described in A. n = 5 mice per group. ∗, p < 0.05 compared with both BSA-tolerized and untolerized controls. †, p < 0.05 compared with BSA-tolerized controls. Error bars indicate SEM.
Ag-coupled splenocyte tolerance sometimes decreases humoral immunity (29, 34, 38). To address whether myosin or T. cruzi tolerization affected anti-myosin or anti-T. cruzi Ab production, we measured Ag-specific Ab titers in tolerized mice. Neither myosin tolerization nor T. cruzi tolerization affected the respective specific IgG titers in infected mice compared with controls (Fig. 3⇓). There is no significant difference in myosin-specific IgG titers between BSA-tolerized and untolerized myosin-immunized mice (34). Myosin tolerization also did not affect anti-T. cruzi IgG titers, and T. cruzi tolerance induction did not affect anti-myosin IgG titers. The titers of isotype-specific Abs were also unaffected (data not shown). Myosin tolerization did inhibit anti-myosin IgG production in myosin-immunized mice, as previously reported (34).
Myosin tolerization and T. cruzi tolerization do not affect myosin-specific and T. cruzi-specific Ab production in T. cruzi-infected mice. The endpoint titers of myosin-specific and T. cruzi-specific IgG were determined by ELISA. Each circle represents the titer of one mouse. A, Mice were tolerized to either BSA (BSA) or cardiac myosin (Myosin) and immunized with purified cardiac myosin in CFA. Saline-immunized mice (Saline Immunized) were not given tolerogen. Myosin-specific IgG endpoint titers were determined 21 days later. B, Mice were tolerized to BSA (BSA), cardiac myosin (Myosin), T. cruzi lysate (T. cruzi), or nothing (None), and infected with T. cruzi. Saline-injected mice (Saline Injected) were not tolerized. Myosin-specific IgG endpoint titers were determined 21 days later. C, Mice were tolerized, as described above, infected with T. cruzi, and assayed for T. cruzi-specific serum IgG. Saline-injected mice (Saline Injected) were not given tolerogen. ∗, p < 0.05 compared with the BSA-tolerized group.
T. cruzi-induced myocarditis is not significantly affected by myosin tolerization
Next, we tested whether suppression of myosin-specific cellular immunity affected inflammation, tissue damage, or fibrosis. Infected mice tolerized to myosin exhibited no differences in inflammation, necrosis, or fibrosis compared with controls (Table I⇓ and Fig. 4⇓). In contrast, myosin-tolerized, myosin-immunized mice exhibited a significant decrease in inflammation and fibrosis, consistent with previous findings (34). There is no significant difference in myocarditis scores between BSA-tolerized and untolerized myosin-immunized mice (34). Cardiac hypertrophy induced by infection or myosin immunization was not affected by myosin tolerance induction (Table II⇓), nor were parasitemia (Fig. 5⇓) or parasite-induced mortality (Fig. 6⇓).
Cardiac histopathology of tolerized and untolerized mice. Representative cardiac sections are shown for T. cruzi-infected mice (A); BSA-tolerized, T. cruzi-infected mice (B); myosin-tolerized, T. cruzi-infected mice (C); T. cruzi-tolerized, T. cruzi-infected mice (D); BSA-tolerized, myosin-immunized mice (E); myosin-tolerized, myosin-immunized mice (F); and saline-injected mice (G). Arrowheads indicate parasite pseudocysts. More complete analysis of tissue histology and other parameters is given in Tables I⇓ and II⇓.
Neither T. cruzi tolerization nor myosin tolerization affects parasitemia. Mice were tolerized to BSA (BSA), myosin (Myosin), T. cruzi lysate (T. cruzi), or nothing (None), and infected with T. cruzi, and parasitemias were determined at different times postinfection. Saline-injected control mice (Saline Injected) were not given tolerogen. Each data point represents the mean parasitemia of the group (n = 10); error bars indicate SEM. For the T. cruzi-infected group, n = 9 on days 19–20 and n = 7 on day 21.
T. cruzi tolerization, but not myosin tolerization, increases mortality in T. cruzi-infected mice. Mice were tolerized to BSA (BSA), cardiac myosin (Myosin), T. cruzi lysate (T. cruzi), or nothing (None), and infected with T. cruzi, and the percentage of survival was charted over time (n = 30 initial mice/group). Saline-injected mice (Saline Injected, n = 15) were not given tolerogen. A Kaplan-Meier actuarial survival plot was generated, in which each data point represents the cumulative proportion of mice surviving at the particular time point, combined from three separate experiments. The T. cruzi-tolerized group had significantly (∗, p < 0.05) higher mortality compared with both BSA-tolerized and untolerized groups.
Effect of Ag-specific tolerization on myocarditis in T. cruzi-infected micea
Effect of Ag-specific tolerance on cardiac hypertrophy in T. cruzi-infected micea
T. cruzi tolerance induction decreases cardiac inflammation and enhances cardiac parasitosis and mortality in infected mice
Finally, we asked whether suppression of T. cruzi-specific cellular immune responses affected acute myocarditis. Mice tolerized with T. cruzi-coupled splenocytes exhibited a significant decrease in cardiac inflammation and an increase in cardiac parasitosis upon infection (Table I⇑, Fig. 4⇑), but had no significant effect on the numbers of circulating parasites (Fig. 5⇑). This increase in cardiac parasitosis was associated with earlier mortality (Fig. 6⇑). The T. cruzi-tolerized mice had scruffy fur and difficulty moving compared with myosin-tolerized and BSA-tolerized control animals, and, although infection generally decreased body weight, the T. cruzi-tolerized mice had an even greater decrease in body weight (Table II⇑).
Discussion
We previously showed that T. cruzi-infected mice develop strong myosin-specific autoimmunity during acute infection. In this study, we addressed the contributions of myosin autoimmunity and T. cruzi-specific immunity to acute CHD myocarditis by Ag-coupled splenocyte tolerization. In infected mice, myosin tolerance induction effectively decreased myosin DTH, while having no effect on myosin-specific IgG levels. In myosin-immunized mice, myosin tolerization decreased both myosin DTH and anti-myosin IgG, in agreement with earlier findings (34). Myosin tolerance induction did not affect myocarditis in infected mice, while it decreased inflammation and fibrosis in myosin-immunized mice. In contrast, T. cruzi tolerance induction decreased cardiac inflammation, enhanced cardiac parasitosis, and was associated with an earlier mortality in infected mice. Interestingly, T. cruzi tolerance induction suppressed myosin DTH, and myosin tolerance induction suppressed T. cruzi DTH.
Suppression of myosin autoimmunity did not affect myocarditis in infected mice, but did reduce inflammation and fibrosis in myosin-immunized mice, as previously reported (34). These results suggest that: 1) myosin tolerization was effective at reducing myosin-specific, autoimmune-mediated tissue inflammation (34), and 2) myosin autoimmunity is not essential to tissue inflammation and destruction in acutely infected mice. However, the results do not speak to whether myosin autoimmunity contributes to inflammation. Nor do the results rule out a significant role for autoimmunity to other cardiac Ags in inflammation and pathogenesis. These possibilities are currently under investigation. Still, the collective results support the notion that direct damage caused by the parasite and the immune response to parasite Ag are significant causes of tissue inflammation in acute disease, as suggested by others (2, 3).
Tolerization to parasite Ags caused a decrease in inflammation, an increase in cardiac parasitosis, and an earlier mortality in T. cruzi-infected mice. This suggests that T. cruzi-specific immunity contributes to inflammation in acutely infected mice. This conclusion is supported by experiments from Kumar and Tarleton (39), in which OVA-specific CD4+ lymphocytes accumulated in tissue infected with an OVA-expressing parasite. The decrease in cardiac inflammation in T. cruzi-tolerized, infected mice is similar to the decrease in CNS inflammation in Theiler’s virus-infected mice that had been treated with virus-coupled splenocytes (27). The enhanced cardiac parasitosis may result from the decreased ability of cellular immunity to clear infected cells from the host. In coxsackieviral infection, injection of streptococcal peptide-coupled splenocytes into mice also enhanced the viral load (30). In addition, because Ag-coupled splenocyte tolerance can potentially mediate an Ag-specific switch from Th1 to Th2 immune responses (25, 26), it is possible that the enhanced parasite load and mortality were due to suppression of Th1 responses, as reported (39, 40, 41). These results also suggest that T. cruzi-specific immunity is important in controlling cardiac parasite load and protecting mice from death. In addition, infection and lysis of infected cells may explain the necrosis and fibrosis in infected mice tolerized to T. cruzi. The inflammation that persists in parasite-tolerized, infected mice could result from incomplete tolerization, inflammation against T. cruzi Ags not present in the toleragen (e.g., Ags found in mammalian stages of the parasite), or autoimmunity, including myosin-specific autoimmunity.
Finally, tolerance to T. cruzi reduced myosin-specific DTH, and tolerance to myosin reduced T. cruzi-specific DTH. The simplest explanation for this observation is that there is immunologic cross-reactivity between a host Ag, such as myosin, and an immunodominant T. cruzi Ag (manuscript in preparation) (8, 42, 43, 44, 45). This could possibly be the myosin-B13 cross-reactive epitope described by Cunha-Neto et al. (8, 46), although this awaits further study. It is surprising that tolerization to a single protein (myosin), despite its size, could reduce the DTH response against the many Ags present in a T. cruzi lysate. The decrease in T. cruzi DTH after myosin tolerization did not lead to enhanced cardiac parasitosis or mortality, as was the case after T. cruzi tolerization. This observation supports the hypothesis that only a small number of T. cruzi Ags were tolerized by myosin, unlike T. cruzi tolerance induction, and that cell-mediated immunity to these proteins is not essential for resistance. An alternative hypothesis is that tolerization led to an altered inflammatory environment (e.g., a Th1/Th2 switch) that affected the DTH response to Ags. Although the vast majority of antigenic T. cruzi proteins characterized to date are expressed in all life cycle stages of the parasite, we are also investigating the possibility that stage-regulated proteins may contribute specifically to the autoimmune response.
In conclusion, we have shown that myosin autoimmunity is not essential for inflammation in an acute model of CHD in which the myosin-specific autoimmune response is similar to that of a purely autoimmune model in which myosin autoimmunity is pathogenic. Although these results support the parasite persistence mechanism of CHD pathogenesis, they do not rule out a role for myosin autoimmunity in pathogenesis, and we are investigating the possibility that autoimmunity to other cardiac Ags contributes to tissue damage by tolerization to total heart homogenate.
Acknowledgments
We thank Dr. A. W. Rademaker for his advice on statistical analysis and Dr. Thomas Bahk for assistance on histopathology scoring. We thank Drs. W. J. Karpus and S. D. Miller and members of their laboratories for advice and guidance throughout this project.
Footnotes
↵1 This work was supported in part by grants from the U.S. Public Health Service. J.S.L. was supported by a predoctoral fellowship from the American Heart Association, Midwest Affiliate. D.M.E. is an Established Investigator of the American Heart Association.
↵2 Address correspondence and reprint requests to Dr. David M. Engman, Northwestern University, Department of Pathology, Ward Building 6-175, 303 East Chicago Avenue, Chicago, IL 60611. E-mail address: d-engman{at}northwestern.edu
↵3 Abbreviations used in this paper: CHD, Chagas’ heart disease; DTH, delayed-type hypersensitivity; ECDI, ethylene carbodiimide.
- Received March 4, 2003.
- Accepted August 5, 2003.
- Copyright © 2003 by The American Association of Immunologists