In vivo staphylococcal superantigen-driven polyclonal Ig responses in mice: dependence upon CD4+ cells and human MHC class II

International Immunology, Vol. 13, No. 10, pp. 1291–1300

© 2001 The Japanese Society for Immunology

In vivo staphylococcal superantigen-driven polyclonal Ig responses in mice: dependence upon CD4⍣ cells and human MHC class II William Stohl, Dong Xu, Song Zang1, Kyung S. Kim, Lily Li1, Julie A. Hanson3, Stephen A. Stohlman2, Chella S. David3 and Chaim O. Jacob1 Division of Rheumatology and Immunology, 1Division of Gastrointestinal and Liver Diseases in the Department of Medicine, and 2Department of Neurology, Keck School of Medicine, University of Southern California, 2011 Zonal Avenue, HMR 711, Los Angeles, CA 90033, USA 3Department of Immunology, Mayo Clinic, Rochester, MN 55905, USA Keywords: CD8⫹ cells, staphylococcal enterotoxin B, transgenic/knockout mice

Abstract Staphylococcal enterotoxin (SE) B and seven other staphylococcal superantigens (SAg), despite promoting vigorous Ig production in human peripheral blood mononuclear cell cultures, are exceedingly poor at eliciting Ig responses in cultures of spleen cells from C57BL/10J (B10) or C3H/HeJ mice. In contrast, SEB elicits Ig responses in cultures of spleen cells from human MHC class II-transgenic mice. Whereas i.p. administration of SEB (0.2–20 µg) to non-transgenic B10 mice elicits very weak in vivo Ig responses, identical treatment of CD4⍣ cell-intact (but not CD4⍣ cell-depleted) human MHC class II-transgenic mice elicits dramatic increases in both splenic Ig-secreting cells and serum Ig levels. Over a 2-week period, the SEB-induced in vivo Ig responses peak and then plateau or fall in association with a preferential increase in splenic CD8⍣ cells. Nevertheless, in vivo depletion of CD8⍣ cells has no sustained effect on SEB-driven Ig responses. Taken together, these observations demonstrate that the effects of SAg on in vivo humoral immune responses are highly CD4⍣ cell dependent, are substantially CD8⍣ cell independent and can be successfully investigated using human MHC class II-transgenic mice. This model system may be useful in investigating the polyclonally activating effects of microbial products (prototypic environmental insults) on the development of systemic autoimmunity. Introduction Microbial superantigens (SAg) are biologically potent natural products of many infectious organisms. Their concurrent binding to Vβ elements on T cells (1) and to MHC class II molecules on SAg-presenting cells (2,3) triggers polyclonal T cell activation with preferential expansion of T cells bearing those Vβ elements (4,5). In vitro SAg stimulation of human lymphocyte populations also can promote T cell-dependent B cell differentiation and Ig production (6–13). Indeed, a compelling theoretical argument in support of a pathogenetic role for microbial SAg in development of clinical autoimmunity associated with pathogenic autoantibodies [e.g. systemic lupus erythematosus (SLE)] has been offered (14). SAg, by virtue of its potential pathogenicity, cannot be administered to humans for experimental purposes. Accord-

ingly, we sought to apply the SAg-based model to the murine system for in vivo experimentation. In this manuscript, we demonstrate that staphylococcal enterotoxin (SE) B and seven other staphylococcal SAg are exceedingly poor at eliciting Ig responses in cultures of spleen cells from C57BL/10J (B10) or C3H/HeJ (C3H) mice, but can elicit Ig responses in cultures of spleen cells from human MHC class II-transgenic mice. In addition, i.p. administration of SEB to non-transgenic B10 mice elicits very weak in vivo Ig responses, whereas identical treatment of CD4⫹ cell-intact (but not CD4⫹ cell-depleted) human MHC class II-transgenic mice elicits dramatic increases in splenic Ig-secreting cells (IgSC) and serum Ig. The rise and subsequent plateauing or decline of Ig responses in vivo is paralleled by a selective increase in splenic CD8⫹

Correspondence to: W. Stohl Transmitting editor: T. Tedder

Received 14 May 2001, accepted 11 July 2001

1292 In vivo SAg-driven Ig responses in mice cell number. Although in vivo depletion of CD8⫹ cells facilitates a heightened early IgSC response in SEB-treated mice, increased in vivo Ig responses are transient and no longer appreciated by day 14, pointing to regulation of SAg-driven Ig responses in vivo via some CD8⫹ cell-independent mechanism. Taken together, these observations provide proof of principle that staphylococcal SAg can promote in vivo Ig responses in a host bearing human MHC class II-expressing cells. This may provide a useful model system for investigating the polyclonally activating effects of microbial products (prototypic environmental insults) on the development of systemic autoimmunity (e.g. SLE).

Methods Human subjects Normal healthy adult human donors of either sex were recruited from University of Southern California personnel. Women known to be pregnant were excluded.

In vivo SAg-driven Ig responses Mice were injected i.p. with graded doses of SEB in 0.2 ml PBS or PBS alone and were sacrificed at the indicated times. The spleen cells were assayed for total IgSC by the reverse hemolytic plaque assay (21,23). Each plaque-forming cell (PFC) was taken as an IgSC. In some experiments, the mice were bled pre- and post-injection at the indicated times by tail vein puncture, and the sera were assayed for IgG and IgM levels by ELISA. Cell surface staining Murine spleen mononuclear cells were single-stained with phycoerythrin- or FITC-conjugated mAb specific for murine CD3, CD4, CD8 or CD45R (B220) or were double-stained with phycoerythrin-conjugated anti-CD4 or anti-CD8 mAb ⫹ FITC-conjugated mAb specific for Vβ8.1/8.2, Vβ8.3 or Vβ6 (PharMingen, San Diego, CA) and analyzed by flow cytometry. Cell debris, as determined by forward- and sidescatter characteristics, was electronically excluded from the analysis. At least 5000 events were analyzed for each sample.

Mice B10 and C3H mice were purchased from Jackson Laboratory (Bar Harbor, ME). B10 congenic mice transgenic for murine I-Eα (B10-A⫹E⫹), B10 congenic mice deficient for endogenous MHC class II antigens (Aβ0), and Aβ0 mice transgenic for human DR2 (Aβ0-DR2), human DR3 (Aβ0-DR3), human DQ6 (Aβ0-DQ6), human DQ8 (Aβ0-DQ8), human DR2 and DQ8 (Aβ0-DR2/DQ8), human DR3 and DQ8 (Aβ0-DR3/DQ8), human DQ8 and DQ6 (Aβ0-DR8/DQ6), and murine I-Eα (Aβ0-A–E⫹) have previously been described (15–20). These mice were initially bred at the Mayo Foundation (Rochester, MN), and were propagated and maintained at the University of Southern California. Mice of either sex were used for in vitro and in vivo studies, and were 7–13 weeks of age at the onset of any given experiment.

In vivo depletion of CD4⫹ or CD8⫹ cells To deplete CD4⫹ cells, mice were injected i.p. with 250 µg rat anti-mouse CD4 mAb GK1.5 (24) on days –3, 0 and 3 relative to SEB injection (day 0). The mice in these experiments were sacrificed on day 6. To deplete CD8⫹ cells, mice were injected i.p. with 200 µg rat anti-mouse CD8 mAb 2.43 (24) on days –3, 0, 3 and 7. The mice in these experiments were sacrificed on day 14. Control non-depleted mice were treated with equal volumes of PBS. The extent of CD4⫹ or CD8⫹ cell depletion was assessed by surface staining of spleen cells. In some CD8⫹ cell-depletion experiments, mice were sacrificed on day 3 or 7, in which case they respectively received only two (days –3 and 0) or three (days –3, 0 and 3) anti-CD8 mAb injections.

Cell populations Human peripheral blood mononuclear cells (PBMC) were isolated from venous blood by Ficoll density gradient centrifugation (21). Murine spleen mononuclear cells were isolated by mechanically teasing the spleens followed by Ficoll density gradient centrifugation. SAg SEA, C1, C2, C3, D, E and toxic-shock-syndrome toxin (TSST)-1 were purchased from Toxin Technology (Sarasota, FL). SEB was purchased from Sigma (St Louis, MO). In vitro SAg-driven Ig responses Human PBMC (2⫻106 cells/ml/well in 24-well plates) were cultured in RPMI 1640 medium supplemented with 10% FCS, glutamine and antibiotics and were stimulated with graded doses of SAg. Murine spleen mononuclear cells were cultured in an identical fashion, except that 2-mercaptoethanol (50 µM) was added to the culture medium. Supernatants were harvested at the indicated time points, and assayed for IgG and IgM levels by ELISA (22).

Statistical analysis All analyses were performed using SigmaStat software (SPSS, Chicago, IL). PFC and serum IgG and IgM results were log-transformed prior to analysis (to achieve normal distributions), whereas cell number and CD4/CD8 ratio results were analyzed in their raw forms (which routinely followed normal distributions without log transformation). Unpaired and paired t-tests were used for unmatched and matched data respectively for comparisons between two groups, and ANOVA and one-way repeated measures ANOVA tests were used respectively for comparisons among three or more groups. When the raw or log-transformed data did not follow a normal distribution or the compared populations failed the equal variance test, the non-parametric Mann–Whitney ranksum test and the Wilcoxon signed-rank test were used for unmatched and matched data respectively for comparisons between two groups, and the Kruskal–Wallis one-way ANOVA on ranks test and the Friedman one-way repeated measures ANOVA on ranks test were used respectively for comparisons among three or more groups.

In vivo SAg-driven Ig responses in mice 1293

Fig. 2. SEB-induced IgG and IgM responses in spleen cell cultures from Aβ0-DR2/DQ8 mice. Cultures were stimulated with the indicated concentrations (ng/ml) of SEB. IgG and IgM levels were determined in culture supernatants harvested at day 14.

Fig. 1. IgG and IgM responses to SAg in human PBMC and murine spleen cell cultures. Human PBMC (A) or B10 spleen cells (B) were stimulated with the indicated concentrations (ng/ml) of SEB. Culture supernatants were harvested at the indicated times, and assayed for IgG and IgM by ELISA.

Results Disparate in vitro SAg-driven Ig responses by human PBMC and murine spleen cells Human PBMC stimulated with a low dose (0.001–0.1 ng/ml) of any of eight different staphylococcal SAg, including SEB, generate vigorous IgSC responses (10). To verify that these SEB-driven outcomes are equally applicable to IgG and IgM responses, we stimulated PBMC cultures with these low doses of SEB and measured Ig isotype levels in the culture supernatants at multiple time points. Vigorous Ig responses were detected in the SEB-stimulated cultures, with peak responses occurring at days 10–14 (Fig. 1A). IgG responses usually, but not invariably, exceeded IgM responses. Strikingly, SEB over ⬎106-fold concentration range (0.0001– 100 ng/ml) was incapable of promoting Ig responses in B10 (I-A⫹, I-E–) spleen cell cultures at any time point tested. Stimulation of B10 spleen cells with any of seven additional staphylococcal SAg (SEA, SEC1, SEC2, SEC3, SED, SEE and TSST-1) also failed to increase IgG or IgM levels. Similarly, SEB did not promote Ig responses in cultures from B10 congenic mice expressing either no class II MHC antigens (Aβ0), only I-E (A–E⫹) or both I-A and I-E (A⫹E⫹) at any time point tested. Moreover, absent SEB-induced Ig responses were observed not only in cultures of spleen cells from H-2b (B10) mice but in spleen cell cultures from H-2k (C3H) mice as well (data not shown). Only by stimulating murine spleen cell cultures with extremely high SEB concentrations (艌 1000

ng/ml) could even modest Ig responses be elicited (Fig. 1B). These poor murine Ig responses induced by staphylococcal SAg parallel the previously reported poor ability of SEB to promote murine lymphocyte proliferation (25,26). SEB-driven Ig responses in cultures of spleen cells from human MHC class II-transgenic mice Cell-mixing studies demonstrated that human T cells, despite supporting SEB-driven Ig production by human B cells, could not do the same for murine B cells (data not shown). This inability may have reflected absent human MHC class II expression by the murine B cells. Although SAg can bind to surface molecules other than MHC class II (27) and SAgtriggered T cell activation and effector function do not absolutely require MHC class II⫹ SAg-presenting cells (28–32), the ability of SAg to trigger T cell activation is markedly facilitated by high-affinity binding of SAg to human MHC class II molecules (3,33,34). To assess whether murine spleen cell expression of human MHC class II molecules could restore SAg-driven Ig responses, cultures of spleen cells from Aβ0 mice (expressing no murine MHC class II molecules) transgenic for DR and/or DQ were stimulated with graded doses of SEB. Ig production was variably enhanced in cultures of spleen cells from Aβ0-DR2, Aβ0-DR3, Aβ0-DQ6, Aβ0-DQ8, Aβ0-DR2/DQ8 and Aβ0-DR3/DQ8 mice, with Aβ0-DR2/DQ8 spleen cells usually giving the best results (data not shown). Ig production in Aβ0-DR2/DQ8 cultures was routinely triggered by a SEB concentration of 1 ng/ml (Fig. 2) and higher SEB concentrations did not further enhance Ig production (data not shown). In vivo SEB-driven Ig responses in Aβ0-DR2/DQ8 mice To assess in vivo Ig responses to SEB, Aβ0-DR2/DQ8 and B10 mice were injected i.p. with 20 µg SEB or PBS. Geometric mean splenic PFC level in SEB-treated Aβ0-DR2/DQ8 mice was markedly (6.6-fold) greater than that in PBS-treated Aβ0DR2/DQ8 mice by 3 days post-injection (P ⬍ 0.001) and

1294 In vivo SAg-driven Ig responses in mice remained as elevated at day 6 (P ⬍ 0.001, Fig. 3). In contrast, identical SEB treatment of B10 mice resulted in a modest (2.1-fold) increase in geometric mean splenic PFC that was appreciated only at day 3 (P ⫽ 0.012) but was no longer detected at day 6 (P ⬎ 0.1). No increase in splenic PFC was observed in SEB-treated Aβ0 mice at any time point (data not shown). SEB-induced in vivo expansion of Vβ8.1/8.2⫹ and Vβ8.3⫹ cells in B10 and Aβ0-DR2/DQ8 mice The relatively poor in vivo Ig response in B10 mice to SEB contrasts with the reported ability of SEB to promote in ‘conventional’ mice (i.e. mice expressing murine MHC class II rather than human MHC class II) in vivo expansion of T cells expressing Vβ8, but not Vβ6, elements (35–37). To

confirm that SEB does induce a T cell response in vivo, we assessed SEB-driven expansion of Vβ8.1/8.2⫹ and Vβ8.3⫹ T cells in B10 and Aβ0-DR2/DQ8 mice. On day 3 post-SEB injection, splenic CD4⫹ cells and CD8⫹ cells expressing either Vβ8.1/8.2 or Vβ8.3 were increased in SEB-injected B10 mice compared to those in PBS-injected controls (P ⫽ 0.047 and P ⫽ 0.043 for CD4⫹ cells and P ⫽ 0.023 and P ⫽ 0.020 for CD8⫹ cells, Fig. 4A and B). In contrast, no expansion of Vβ6⫹ T cells was observed (data not shown). These results are consistent with those previously reported by others (35–37). Importantly, expansion of Vβ8.1/8.2⫹ and Vβ8.3⫹ T cells was much greater in SEB-treated Aβ0-DR2DQ8 mice (P ⫽ 0.010 and P ⫽ 0.010 for CD4⫹ cells and P ⫽ 0.030 and P ⫽ 0.003 for CD8⫹ cells, Fig. 4C and D) than that in SEBtreated B10 mice. This parallels the greater in vivo IgSC response at day 3 in SEB-treated Aβ0-DR2/DQ8 mice than that in SEB-treated B10 mice (Fig. 3). As the case for B10 mice, SEB did not induce expansion of Vβ6⫹ T cells in Aβ0DR2/DQ8 mice (data not shown). Abrogation of Ig-enhancing effects of SEB in CD4⫹ celldepleted Aβ0-DR2/DQ8 mice

Fig. 3. In vivo splenic IgSC responses to SEB in Aβ0-DR2/DQ8 and B10 mice. Aβ0-DR2/DQ8 (Tg) or B10 mice were injected i.p. with PBS or SEB (20 µg). Mice were sacrificed at the indicated times and the spleens were assayed for total IgSC. Each symbol represents an individual mouse. The asterisks indicate the geometric means.

SEB-induced Ig production by human B cells in vitro is CD4⫹ cell-dependent (10,38). To determine whether this is also the case for SEB-induced Ig production in vivo, Aβ0-DR2/DQ8 mice were depleted of CD4⫹ cells and were injected with SEB. CD4⫹ cell depletion in anti-CD4-treated mice was virtually complete (Fig. 5A). Importantly, SEB had no discernible effect on splenic PFC responses in these CD4⫹ cell-depleted mice (P ⬎ 0.1, Fig. 5B), results that are in sharp contrast to the stimulatory effects of SEB in CD4⫹ cell-intact controls (P ⫽ 0.006). Thus, in complete agreement with results from

Fig. 4. SEB-induced in vivo expansion of Vβ8.1/8.2⫹ and Vβ8.3⫹ CD4⫹ and CD8⫹ cells in B10 and Aβ0-DR2/DQ8 mice. B10 (A and B) and Aβ0-DR2/DQ8 (C and D) mice were injected i.p. with 20 µg SEB (⫹) or PBS (–). On day 3, the mice were sacrificed and the numbers of spleen CD4⫹ cells and CD8⫹ cells expressing Vβ8.1/8.2 (A and C) or Vβ8.3 (B and D) for each mice were determined. Note that the ordinate scales differ among the individual panels. The asterisks indicate the arithmetic means.

In vivo SAg-driven Ig responses in mice 1295 our human in vitro model system (10,38), SEB-induced Ig responses in vivo also are highly CD4⫹ cell-dependent. Transient effect of in vivo depletion of CD8⫹ cells on in vivo SEB-driven Ig responses The plateauing of splenic PFC responses in (CD4⫹ cell-intact) SEB-treated Aβ0-DR2/DQ8 mice at day 6 and their decline at the later time points (Fig. 3 and data not shown) suggested the development of important in vivo homeostatic down-regulatory events. As early as day 3 post-SEB injection, there was an increase (P ⫽ 0.056) in total spleen mononuclear cell number in Aβ0-DR2/DQ8 mice (Fig. 6A). This increase was predominantly due to increased T (CD3⫹) cells (P ⫽ 0.008, Fig. 6C) without any discernible increase in B (B220⫹) cells (P ⬎ 0.1, Fig. 6B). Strikingly, the increase in T cells was largely due to a dramatic expansion of CD8⫹ cells (P ⫽ 0.008,

Fig. 6E) with variable increases in CD4⫹ cells (P ⬎ 0.1, Fig. 6D), resulting in a marked drop in CD4/CD8 ratios (P ⬍ 0.001, Fig. 6F). Preferential expansion of CD8⫹ cells persisted for at least an additional 11 days (data not shown; see Fig. 7 below). No differences in any of the measured parameters were appreciated in SEB-injected B10 or Aβ0 mice at any time point (P ⬎ 0.1 for each comparison, Fig. 6 and data not shown). Despite the dramatic and preferential expansion of CD8⫹ cells, in vivo depletion of CD8⫹ had only a transient effect on SEB-induced Ig responses. Treatment of mice with anti-CD8 mAb was very effective in depleting CD8⫹ cells at any time point tested (Fig. 7A and C). At day 3 post-injection, SEB induced vigorous splenic PFC responses in both CD8⫹ cellintact and CD8⫹ cell-depleted mice (P ⬍ 0.001 for each comparison, Fig. 7B), with geometric mean splenic PFC of

Fig. 5. SEB-induced in vivo IgSC response in CD4⫹ cell-depleted Aβ0-DR2/DQ8 mice. Aβ0-DR2/DQ8 mice were depleted (dep) of CD4⫹ cells by injection i.p. of 250 µg anti-CD4 mAb on days –3, 0 and 3. Control CD4⫹ cell-intact (int) mice were injected with an equal volume of PBS at the same time points. On day 0, mice were injected i.p. with PBS (–) or 20 µg SEB (⫹). On day 6, the mice were sacrificed and spleen CD4⫹ cell number (A) and PFC (B) for each mouse were determined. The asterisks indicate the arithmetic means (A) or the geometric means (B).

Fig. 6. Spleen cell composition in Aβ0-DR2/DQ8 and B10 mice in response to SEB. Aβ0-DR2/DQ8 (Tg) and B10 mice were injected i.p. with PBS (–) or 20 µg SEB (⫹). On day 3, the mice were sacrificed, and total spleen mononuclear cell number (A), B (B220⫹) cell number (B), T (CD3⫹) cell number (C), CD4⫹ cell number (D), CD8⫹ cell number (E) and CD4/CD8 ratio (F) for each mouse was determined. The asterisks indicate the arithmetic means.

1296 In vivo SAg-driven Ig responses in mice

Fig. 7. SEB-induced in vivo IgSC response in CD8⫹ cell-depleted Aβ0-DR2/DQ8 mice. Aβ0-DR2/DQ8 mice were depleted (dep) of CD8⫹ cells by injection i.p. of 200 µg anti-CD8 mAb on days –3 and 0 (A and B) or on days –3, 0, 3 and 7 (C and D). Control CD8⫹ cell-intact (int) mice were injected with an equal volume of PBS at the same time points. On day 0, mice were injected i.p. with PBS (–) or 20 µg SEB (⫹). On day 3 (A and B) or day 14 (C and D), the mice were sacrificed and spleen CD8⫹ cell numbers (A and C) and PFC (B and D) for each mouse were determined. The asterisks indicate the arithmetic means (A and C) or the geometric means (B and D).

the latter mice being ~3-fold greater than that of the former (P ⫽ 0.001). However, this enhancing effect of CD8⫹ cell depletion on SEB-driven IgSC responses was no longer present at day 14. Whereas geometric mean splenic PFC in SEB-injected CD8⫹ cell-intact and CD8⫹ cell-depleted mice remained ~2-fold greater than those in their corresponding PBS-injected controls (P ⫽ 0.02 and P ⫽ 0.003 respectively, Fig. 7D), differences in splenic PFC between CD8⫹ cell-intact and CD8⫹ cell-depleted mice were no longer detectable (P ⬎ 0.1). A similar lack of effect of CD8⫹ cell depletion on SEB-driven splenic PFC responses was also observed as early as day 7 (data not shown). In vivo SEB-driven IgG and IgM responses in various strains of Aβ0 mice transgenic for human MHC class II To determine whether increased splenic IgSC induced by SEB actually resulted in increased circulating Ig, additional Aβ0-DR2/DQ8 and B10 mice were injected i.p. with graded doses of SEB or PBS. SEB reproducibly promoted increased serum IgM in Aβ0-DR2/DQ8 mice (Fig. 8). At a dose of 20 µg, SEB uniformly induced serum IgM to rise in Aβ0-DR2/DQ8 mice, with peak responses at day 7 and only modest (statistically insignificant) waning by day 12. In these mice, even SEB doses as low as 0.2 µg often induced substantial elevations in serum IgM. In contrast, no tested dose of SEB (0.2–20 µg) promoted increased serum IgM in B10 mice. PBS injection had no significant effects on serum IgM at any time point in either Aβ0-DR2/DQ8 or B10 mice (P ⬎ 0.1 for each comparison). Although treatment of Aβ0-DR2/DQ8 mice with either SEB or TSST-1 enhanced serum IgM levels, treatment with neither SAg reproducibly enhanced serum IgG levels (data not

shown). To determine whether the failure to mount SAgdriven IgG responses was unique to Aβ0-DR2/DQ8 mice, we tested other human MHC class II-transgenic mice for in vivo responsiveness to SEB. The response patterns differed considerably among different human MHC class II-transgenic strains (Fig. 9). Similar to Aβ0-DR2/DQ8 mice, Aβ0-DQ8 mice demonstrated a substantial increase in serum IgM following SEB injection, whereas the IgG response was variable and usually feeble. In contrast, SEB induced not just a substantial IgM response in Aβ0-DQ6 mice but induced a large IgG response as well. At day 7, geometric mean serum IgG concentration in SEB-injected Aβ0-DQ6 mice was ⬎10-fold greater than that in corresponding PBS-injected mice (P ⫽ 0.018) and remained ⬎7-fold greater at day 14 (P ⫽ 0.016). In the limited number of animals studied, no differences in total, CD3⫹, CD4⫹, CD8⫹ or B220⫹ numbers of spleen cells or in spleen CD4/CD8 ratios were detected between Aβ0DQ6 and Aβ0-DQ8 mice (P ⬎ 0.1 for each comparison, data not shown). Of note, IgG responses of Aβ0-DQ8/DQ6 mice were intermediate between those of Aβ0-DQ8 and Aβ0-DQ6. As expected, IgM and IgG levels remained unchanged in PBS-injected controls of all strains tested (Fig. 9). Discussion Microbial SAg, products of organisms ubiquitous in nature, can dramatically affect the in vivo immune status of the host. Administration of staphylococcal SAg to mice leads to expansion and, subsequently, to physical deletion of and/or development of anergy by T cells expressing specific Vβ elements (35–37,39–41). In Rhesus monkeys challenged with SEB, specific Vβ⫹ T cells undergo a triphasic response,

In vivo SAg-driven Ig responses in mice 1297

Fig. 8. Serum IgM responses to SEB in Aβ0-DR2/DQ8 and B10 mice. (Left) Aβ0-DR2/DQ8 mice were injected i.p. with PBS, or 0.2, 2 or 20 µg SEB. The mice were pre-bled on the day prior to injection, and were serially bled on days 3, 7 and 12 post-injection. Sera were assayed for IgM. Each set of connecting points represents an individual mouse. (Right) Groups of six to nine Aβ0-DR2/DQ8 (Tg) or B10 mice were injected i.p. with PBS or 20 µg SEB. The mice were serially bled as above. The sets of connecting points represent the geometric means ⫾ SEM of the individual groups. SEB induced ⬎90% increase in geometric mean serum IgM levels in Aβ0-DR2/DQ8 mice at days 7 and 12 post-injection (P ⬍ 0.001 and P ⫽ 0.002 respectively), but had no significant effect on serum IgM levels in B10 mice at either time point (P ⬎ 0.1 for each comparison).

Fig. 9. Serum Ig responses to SEB in Aβ0 mice transgenic for different human MHC class II genes. Aβ0-DQ8 (left), Aβ0-DQ6 (center) and Aβ0-DQ8/DQ6 (right) mice were injected i.p. with PBS or 20 µg SEB. The mice were pre-bled on the day prior to injection, and were serially bled on days 7 and 14 post-injection. Sera were assayed for IgG (top) and IgM (bottom) concentrations. Each set of connecting points represents an individual mouse.

characterized by a hyperacute decline, followed by a brief dominant expansion, followed by a prolonged drop (42). Despite the considerable effort devoted to the study of SAg-mediated effects on in vivo T cell biology, the effects of SAg on in vivo Ig responses has received relatively little attention. We had previously demonstrated (10) and confirmed in this study (Fig. 1) that in vitro stimulation of human PBMC with low concentrations (0.001–0.1 ng/ml) of SEB induces

vigorous Ig responses. Nevertheless, these in vitro results could not be replicated with spleen cells from B10 or C3H mice. A wide range of SEB concentrations had no Igpromoting effects, and only at extremely high concentrations (1000–5000 ng/ml) could even small incremental Ig production be detected (Fig. 1). These poor SEB-driven murine Ig responses in vitro are consistent with the poor ability of SEB to promote murine lymphocyte proliferation in vitro (25,26).

1298 In vivo SAg-driven Ig responses in mice In contrast, we were able to induce Ig production in cultures of spleen cells from Aβ0-DR2/DQ8 mice (Fig. 2). Moreover, Aβ0-DR2/DQ8, but not B10 or Aβ0, mice mounted substantial in vivo Ig responses to single i.p. injections of SEB (Figs 3 and 8 and data not shown). Although certain staphylococcal SAg can enhance proliferation and Ig production of preactivated B cells (7,43) and can protect certain B cell subpopulations from apoptosis (44), we know of no reports describing activation of resting B cells in the absence of T cells or T cell factors with SEB or other staphylococcal SAg. Indeed, SEB-driven in vivo Ig responses in Aβ0-DR2/DQ8 mice were abrogated when the mice were depleted of CD4⫹ cells (Fig. 5), demonstrating a CD4⫹ cell-dependent process. The failure of B10 mice to mount vigorous in vivo Ig responses to single i.p. SEB injections does not indicate an absolute inability by ‘conventional’ mice to respond in vivo to SEB or other staphylococcal SAg and generate strong Ig responses. Rather, it likely reflects the greater ability of such SAg to drive T cell proliferation (activation) in the presence of cells bearing human MHC class II compared to that in the presence of cells bearing murine MHC class II (25,26). Indeed, in vivo expansion of Vβ8.1/8.2⫹ and Vβ8.3⫹ T cells in B10 mice in response to SEB was significant, albeit quantitatively less than that in Aβ0-DR2/DQ8 mice (Fig. 4). Moreover, very high concentrations of SEB (1000–5000 ng/ml) promote in vitro Ig production in B10 spleen cell cultures, so considerable in vivo Ig responses in these mice might in principle be induced with single i.p. injections of very large quantities of SEB (even if not practical). Nevertheless, the very poor SEB-driven Ig responses in ‘conventional’ mice highlight a great advantage to human MHC class II-transgenic mice in that vigorous Ig responses can be induced with very modest (experimentally achievable) amounts of SAg in the transgenic mice. The in vivo Ig responses following single injections of SEB in Aβ0-DR2/DQ8 mice plateaued and/or declined after ~6–7 days (Figs 3 and 8, and data not shown). CD8⫹ cells increased in disproportionate number to any increases in CD4⫹ cells in the spleens of SEB-injected Aβ0-DR2/DQ8, but not B10, mice, and this CD8⫹ cell expansion persisted through day 14 (Figs 6 and 7). Nevertheless, depletion of CD8⫹ cells had only a transient effect on SEB-driven Ig responses in vivo (Fig. 7). This result suggests either that the few remaining CD8⫹ cells were sufficient to effect ‘normal’ down-regulation or, more likely, that down-regulation can be achieved via a CD8⫹ cellindependent mechanism. Since CD4⫹ cells, in the absence of CD8⫹ cells, can effectively regulate in vitro SEB-driven Ig responses via a CD95-based pathway (43), it may be that CD4⫹ cells can also effectively regulate in vivo SEB-driven Ig responses via a CD95-based pathway in the absence of CD8⫹ cells. Although SEB did not reproducibly enhance serum IgG levels in Aβ0-DR2/DQ8 or Aβ0-DQ8 mice, SEB did promote robust IgG (and IgM) responses in Aβ0-DQ6 mice (Fig. 9 and data not shown). Why Aβ0-DQ6 mice responded to SEB with such vigorous IgG production while the other tested human MHC class II-transgenic mice responded more poorly is uncertain at present. Aβ0-DQ6 and Aβ0-DQ8 mice have similar transgene copy number and expression of MHC class II (C. David, unpublished observations), so quantitative

differences in presentation of SEB to T cells in vivo are unlikely to account for the differences in IgG responses. There is heterogeneity in binding of staphylococcal SAg to different human MHC class II molecules (45,46), so subtle differences in SEB binding to DQ6 compared to its binding to DQ8 and/ or DR2 may contribute to differential IgG responses to SEB. It also remains to be determined whether the low baseline serum IgG levels in Aβ0-DQ6 mice are related to the vigorous SEB-induced IgG responses and whether differential cytokine profiles and/or expression of various co-stimulatory molecules (e.g. CD154, ICOS) may be involved. In any case, the ability of SEB to promote in vivo IgG and IgM responses in certain human MHC class II-transgenic mice (e.g. Aβ0-DQ6) provides proof of principle that SAg can promote in vivo Ig responses in a physiologic setting in which human MHC class II-expressing cells are present. This should permit us to analyze by multiple approaches the contributory effects of a prototypic environmental factor (microbial SAg) to development of serological and/or clinical autoimmunity. First, the appropriate human MHC class II genes can be introduced into autoimmune-prone strains of mice (e.g. NZM). These mice can be challenged with SAg and acceleration of autoimmunity (or lack of same) can be monitored. Second, the interplay between SAg stimulation and discrete genetic abnormalities (e.g. lpr, gld) that likely affect immune homeostatic recovery from a polyclonal insult can be assessed by breeding the genetic defect(s) into human MHC class IItransgenic mice and challenging the mice with SAg. Third, based on a report of SAg inducing in vivo Ig responses in SCID mice reconstituted with human T cells ⫹ human B cells (47), we can reconstitute SCID mice with T cells and B cells from SLE patients and monitor the effects of SAg on development and/or acceleration of autoimmunity.

Acknowledgements The authors thank Hal Soucier for his flow cytometry operation. This work was supported in part by NIH grants AR41006 (W. S.), NS18146 (S. A. S.) and AI14764 (C. S. D.), and by an Arthritis Foundation Biomedical Science Grant (W. S.).

Abbreviations Aβ0 B10 C3H IgSC PBMC PFC SAg SE SLE TSST

B10 congenic deficient for endogenous MHC class II C57BL/10J C3H/HeJ Ig-secreting cell peripheral blood mononuclear cell plaque-forming cell microbial superantigen staphylococcal enterotoxin systemic lupus erythematosus toxin-shock-syndrome toxin

References 1 Choi, Y., Herman, A., DiGiusto, D., Wade, T., Marrack, P. and Kappler, J. 1990. Residues of the variable region of the T-cellreceptor β-chain that interact with S. aureus toxin superantigens. Nature 346:471. 2 Fraser, J. D. 1989. High-affinity binding of staphylococcal enterotoxins A and B to HLA-DR. Nature 339:221.

In vivo SAg-driven Ig responses in mice 1299 3 Mollick, J. A., Cook, R. G. and Rich, R. R. 1989. Class II MHC molecules are specific receptors for staphylococcus enterotoxin A. Science 244:817. 4 Choi, Y., Kotzin, B., Herron, L., Callahan, J., Marrack, P. and Kappler, J. 1989. Interaction of Staphylococcus aureus toxin ‘superantigens’ with human T cells. Proc. Natl Acad. Sci. USA 86:8941. 5 Kappler, J., Kotzin, B., Herron, L., Gelfand, E. W., Bigler, R. D., Boylston, A., Carrel, S., Posnett, D. N., Choi, Y. and Marrack, P. 1989. Vβ-specific stimulation of human T cells by staphylococcal toxins. Science 244:811. 6 Mourad, W., Scholl, P., Diaz, A., Geha, R. and Chatila, T. 1989. The staphylococcal toxic shock syndrome toxin 1 triggers B cell proliferation and differentiation via major histocompatibility complex-unrestricted cognate T/B cell interaction. J. Exp. Med. 170:2011. 7 Fuleihan, R., Mourad, W., Geha, R. S. and Chatila, T. 1991. Engagement of MHC-class II molecules by staphylococcal exotoxins delivers a comitogenic signal to human B cells. J. Immunol. 146:1661. 8 Moseley, A. B. and Huston, D. P. 1991. Mechanism of Staphylococcus aureus exotoxin A inhibition of Ig production by human B cells. J. Immunol. 146:826. 9 Crow, M. K., Zagon, G., Chu, Z., Ravina, B., Tumang, J. R., Cole, B. C. and Friedman, S. M. 1992. Human B cell differentiation induced by microbial superantigens: unselected peripheral blood lymphocytes secrete polyclonal immunoglobulin in response to Mycoplasma arthritidis mitogen. Autoimmunity 14:23. 10 Stohl, W., Elliott, J. E. and Linsley, P. S. 1994. Human T celldependent B cell differentiation induced by staphylococcal superantigens. J. Immunol. 153:117. 11 Martinez-Arends, A., Astoul, E., Lafage, M. and Lafon, M. 1995. Activation of human tonsil lymphocytes by rabies virus nucleocapsid superantigen. Clin. Immunol. Immunopathol. 77:177. 12 Armerding, D., Hren, A., Callard, R. E., Fu, S. M. and Mudde, G. C. 1996. Induction of cognate and non-cognate T-cell help for B-cell IgE production in relation to CD40 ligand expression. Int. Arch. Allergy Immunol. 111:376. 13 Hofer, M. F., Newell, K., Duke, R. C., Schlievert, P. M., Freed, J. H. and Leung, D. Y. M. 1996. Differential effects of staphylococcal toxic shock syndrome toxin-1 on B cell apoptosis. Proc. Natl Acad. Sci. USA 93:5425. 14 Friedman, S. M., Posnett, D. N., Tumang, J. R., Cole, B. C. and Crow, M. K. 1991. A potential role for microbial superantigens in the pathogenesis of systemic autoimmune disease. Arthritis Rheum. 34:468. 15 Le Meur, M., Gerlinger, P., Benoist, C. and Mathis, D. 1985. Correcting an immune-response deficiency by creating Eα gene transgenic mice. Nature 316:38. 16 Gosgrove, D., Gray, D., Dierich, A., Kaufman, J., Lemeur, M., Benoist, C. and Mathis, D. 1991. Mice lacking MHC class II molecules. Cell 66:1051. 17 Kong, Y. M., Lomo, L. C., Motte, R. W., Giraldo, A. A., Baisch, J., Strauss, G., Ha¨ mmerling, G. J. and David, C. S. 1996. HLA-DRB1 polymorphism determines susceptibility to autoimmune thyroiditis in transgenic mice: definitive association with HLA-DRB1*0301 (DR3) gene. J. Exp. Med. 184:1167. 18 Nabozny, G. H., Baisch, J. M., Cheng, S., Cosgrove, D., Griffiths, M. M., Luthra, H. S. and David, C. S. 1996. HLA-DQ8 transgenic mice are highly susceptible to collagen-induced arthritis: a novel model for human polyarthritis. J. Exp. Med. 183:27. 19 Bradley, D. S., Nabozny, G. H., Cheng, S., Zhou, P., Griffiths, M. M., Luthra, H. S. and David, C. S. 1997. HLA-DQB1 polymorphism determines incidence, onset, and severity of collagen-induced arthritis in transgenic mice: implications in human rheumatoid arthritis. J. Clin. Invest. 100:2227. 20 Taneja, V., Griffiths, M. M., Luthra, H. and David, C. S. 1998. Modulation of HLA-DQ-restricted collagen-induced arthritis by HLA-DRB1 polymorphism. Int. Immunol. 10:1449. 21 Stohl, W., Posnett, D. N. and Chiorazzi, N. 1987. Induction of T cell-dependent B cell differentiation by anti-CD3 monoclonal antibodies. J. Immunol. 138:1667.

22 Abo, W., Gray, J. D., Bakke, A. C. and Horwitz, D. A. 1987. Studies on human blood lymphocytes with iC3b (type 3) complement receptors. II. Characterization of subsets which regulate pokeweed mitogen-induced lymphocyte proliferation and immunoglobulin synthesis. Clin. Exp. Immunol. 67:544. 23 Gronowicz, E., Coutinho, A. and Melchers, F. 1976. A plaque assay for all cells secreting Ig of a given type or class. Eur. J. Immunol. 6:588. 24 Williamson, J. S. P. and Stohlman, S. A. 1990. Effective clearance of mouse hepatitis virus from the central nervous system requires both CD4⫹ and CD8⫹ T cells. J. Virol. 64:4589. 25 Cole, B. C., Sawitzke, A. D., Ahmed, E. A., Atkin, C. L. and David, C. S. 1997. Allelic polymorphisms at the H-2A and HLA-DQ loci influence the response of murine lymphocytes to the Mycoplasma arthritidis superantigen MAM. Infect. Immun. 65:4190. 26 Proft, T., Moffatt, S. L., Berkahn, C. J. and Fraser, J. D. 1999. Identification and characterization of novel superantigens from Streptococcus pyogenes. J. Exp. Med. 189:89. 27 Rogers, T. J., Guan, L. and Zhang, L. 1995. Characterization of an alternative superantigen binding site expressed on a renal fibroblast cell line. Int. Immunol. 7:1721. 28 Dohlsten, M., Hedlund, G., Segren, S., Lando, P. A., Herrmann, T., Kelly, A. P. and Kalland, T. 1991. Human major histocompatibility complex class II-negative colon carcinoma cells present staphylococcal superantigens to cytotoxic T lymphocytes: evidence for a novel enterotoxin receptor. Eur. J. Immunol. 21:1229. 29 Herrmann, T., Romero, P., Sartoris, S., Paiola, F., Accolla, R. S., Maryanski, J. L. and MacDonald, H. R. 1991. Staphylococcal enterotoxin-dependent lysis of MHC class II negative target cells by cytolytic T lymphocytes. J. Immunol. 146:2504. 30 Chapes, S. K., Hoynowski, S. M., Woods, K. M., Armstrong, J. W., Beharka, A. A. and Iandolo, J. J. 1993. Staphylococcus-mediated T-cell activation and spontaneous natural killer cell activity in the absence of major histocompatibility complex class II molecules. Infect. Immun. 61:4013. 31 Avery, A. C., Markowitz, J. S., Grusby, M. J., Glimcher, L. H. and Cantor, H. 1994. Activation of T cells by superantigen in class IInegative mice. J. Immunol. 153:4853. 32 Lamphear, J. G., Stevens, K. R. and Rich, R. R. 1998. Intercellular adhesion molecule-1 and leukocyte function-associated antigen3 provide costimulation for superantigen-induced T lymphocyte proliferation in the absence of a specific presenting molecule. J. Immunol. 160:615. 33 Fleischer, B. and Schrezenmeier, H. 1988. T cell stimulation by Staphylococcal enterotoxins: clonally variable response and requirement for major histocompatibility complex class II molecules on accessory or target cells. J. Exp. Med. 167:1697. 34 Hedlund, G., Dohlsten, M., Lando, P. A. and Kalland, T. 1990. Staphylococcal enterotoxins direct and trigger CTL killing of autologous HLA-DR⫹ mononuclear leukocytes and freshly prepared leukemia cells. Cell. Immunol. 129:426. 35 White, J., Herman, A., Pullen, A. M., Kubo, R., Kappler, J. W. and Marrack, P. 1989. The Vβ-specific superantigen staphylococcal enterotoxin B: stimulation of mature T cells and clonal deletion in neonatal mice. Cell 56:27. 36 Kawabe, Y. and Ochi, A. 1991. Programmed cell death and extrathymic reduction of Vβ8⫹ CD4⫹ T cells in mice tolerant to Staphylococcus aureus enterotoxin B. Nature 349:245. 37 MacDonald, H. R., Baschieri, S. and Lees, R. K. 1991. Clonal expansion precedes anergy and death of Vβ8⫹ peripheral T cells responding to staphylococcal enterotoxin B in vivo. Eur. J. Immunol. 21:1963. 38 Stohl, W. and Elliott, J. E. 1995. Differential human T celldependent B cell differentiation induced by staphylococcal superantigens (SAg): regulatory role for SAg-dependent B cell cytolysis. J. Immunol. 155:1838. 39 Kawabe, Y. and Ochi, A. 1990. Selective anergy of Vβ8⫹,CD4⫹ T cells in staphylococcus enterotoxin B-primed mice. J. Exp. Med. 172:1065. 40 Rellahan, B. L., Jones, L. A., Kruisbeek, A. M., Fry, A. M. and Matis, L. A. 1990. In vivo induction of anergy in peripheral Vβ8⫹ T cells by staphylococcal enterotoxin B. J. Exp. Med. 172:1091.

1300 In vivo SAg-driven Ig responses in mice 41 McCormack, J. E., Callahan, J. E., Kappler, J. and Marrack, P. C. 1993. Profound deletion of mature T cells in vivo by chronic exposure to exogenous superantigen. J. Immunol. 150:3785. 42 Kou, Z.-C., Halloran, M., Lee-Parritz, D., Shen, L., Simon, M., Sehgal, P., Shen, Y. and Chen, Z. W. 1998. In vivo effects of a bacterial superantigen on macaque TCR repertoires. J. Immunol. 160:5170. 43 Stohl, W., Elliott, J. E., Lynch, D. H. and Kiener, P. A. 1998. CD95 (Fas)-based, superantigen-dependent, CD4⫹ T cell-mediated down-regulation of human in vitro immunoglobulin responses. J. Immunol. 160:5231.

44 Domiati-Saad, R. and Lipsky, P. E. 1998. Staphylococcal enterotoxin A induces survival of VH3-expressing human B cells by binding to the VH region with low affinity. J. Immunol. 161:1257. 45 Herman, A., Croteau, G., Sekaly, R.-P., Kappler, J. and Marrack, P. 1990. HLA-DR alleles differ in their ability to present staphylococcal enterotoxins to T cells. J. Exp. Med. 172:709. 46 Mollick, J. A., Chintagumpala, M., Cook, R. G. and Rich, R. R. 1991. Staphylococcal exotoxin activation of T cells: role of exotoxin–MHC class II binding affinity and class II isotype. J. Immunol. 146:463. 47 Martensson, C., Ifversen, P., Borrebaeck, C. A. K. and Carlsson, R. 1995. Enhancement of specific immunoglobulin production in SCID-hu-PBL mice after in vitro priming of human B cells with superantigen. Immunology 86:224.