Endogenous Superantigens Shape Response to Exogenous Superantigens

Abstract

Endogenous superantigen-mediated thymic negative selection resulted in a paucity of mature T cells bearing T-cell receptor (TCR) Vβ8 in the periphery. Consequently, the magnitude of immune response to exogenous superantigen staphylococcal enterotoxin B, which activates TCR Vβ8⁺ T cells, was significantly reduced and conferred protection from superantigen-induced mortality.

Superantigens are a family of microbial proteins that bind directly to major histocompatibility complex (MHC) class II molecules outside of the peptide-binding groove and activate CD4⁺ and CD8⁺ T cells expressing certain T-cell receptor (TCR) Vβ families (10). Superantigens encoded by retroviruses integrated into the genome are called “endogenous superantigens,” and those of bacterial origin are called “exogenous superantigens” (3). The outcome of superantigen-mediated activation differs for mature T cells and immature thymocytes. While the former proliferate vigorously, the latter rapidly undergo apoptosis. Therefore, thymic expression of endogenous superantigens encoded by integrated retrovirus transmitted through the germ line causes deletion of thymocytes bearing certain TCR Vβ families, resulting in a paucity of those TCR Vβ families in the mature peripheral T-cell pool (13).

Endogenous superantigen-mediated TCR Vβ deletions are reported in humans (8). HERV-K18, a human endogenous retrovirus-encoded superantigen, can cause deletion of TCR Vβ7⁺ thymocytes (7). Incidentally, TCR Vβ7-bearing T cells are strongly stimulated by the exogenous superantigen staphylococcal enterotoxin A (SEA) in humans (10). Individuals expressing HERV-K18 in the thymus would theoretically mount a poor immune response to SEA due to a dearth of mature TCR Vβ7⁺ T cells in the periphery and thus could be protected from SEA-induced toxic shock syndrome. Due to the extreme toxicity of bacterial superantigens to humans, we tested the hypothesis that endogenous superantigen-mediated TCR Vβ deletions can modulate the immune response elicited by an exogenous bacterial superantigen using the well-established HLA class II transgenic mouse model (2, 11, 14, 15).

We have generated two independent lines of HLA class II transgenic mice expressing HLA-DR3 (DRB1*0301) (12) and HLA-DR2 (DRB1*1501) (1). These two HLA-DR alleles share the same DRα chain, DRA*0103. Previous studies have shown that expression of transgenic class II molecules in these mice is comparable with murine MHC class II molecules. However, it should be noted that these mice completely lack all endogenous MHC class II genes (6), and the T-cell responses are restricted only by the transgenic HLA class II (1). While the levels of expression of transgenic class II (Fig. 1A) and distribution of CD4⁺ and CD8⁺ T cells (Fig. 1B) are comparable between these lines, the presence of some endogenous superantigen in the DR2 lines deleted CD4⁺ as well as CD8⁺ T cells bearing TCR Vβ8 in thymus, resulting in negligible representation of this TCR in the periphery (Fig. 2). Deletion of TCR Vβ8 in thymus is dependent on the expression of a functional class II molecule, as mice expressing the DRα chain alone without the HLA-DRB1*1501 β chain do not delete TCR Vβ8 (Fig. 3). Since this deletion is MHC class II dependent and occurs in both CD4⁺ and CD8⁺ T cells during transition from the double-positive (DP) to the single-positive stage in thymus, we conclude that this deletion is mediated by endogenous superantigen.

FIG. 1.

MHC class II expression and T-cell development in HLA-DR transgenic mice. (A) Splenocytes from age-matched HLA-DR3 (DRB*0301) and HLA-DR2 (DRB*1501) transgenic mice (n ≥ 4 mice/group) expressing common DRA*0103 were stained with L227 (anti-DR antibody) to check expression of transgenic HLA-DR by flow cytometry. Shown are percentages of cells expressing DR. (B) CD4⁺ and CD8⁺ T-cell development in DR3 and DR2 mice (n ≥ 4 mice/group) as determined by flow cytometry. Shown are percentages of cells positive for CD4 or CD8.

FIG. 2.

Thymic selection in HLA-DR transgenic mice. Thymocytes from age-matched HLA-DR3 mice (n ≥ 4 mice/group) expressing both DRα and DRβ chains and HLA-DR2 transgenic mice expressing the DRα chain but expressing or lacking the DRβ chain were analyzed by flow cytometry for CD4 and CD8 coreceptors. Note the significant presence of TCR Vβ8⁺ T cells among the single- positive (SP) subsets in DRβ chain-negative DR2 littermates and the absence of TCR Vβ8⁺ T cells in DRβ chain-positive DR2 littermates, indicating the requirement of functional HLA class II molecules for thymic negative selection. Only CD8 SP cells are analyzed, as CD4 SP cells would be absent in the absence of functional class II molecules.

FIG. 3.

TCR Vβ repertoire in HLA-DR transgenic mice. Splenocytes from age-matched HLA-DR3 and -DR2 transgenic mice (n = 3 to 6 mice/group) were stained for CD4, CD8, and TCR Vβ-specific fluorochrome-conjugated antibodies and analyzed by flow cytometry. Shown are percentages of cells expressing a specific TCR Vβ family within the respective gated population.

It is very well established that staphylococcal enterotoxin B (SEB) activates CD4 and CD8 T cells bearing TCR Vβ8 in mice (5, 11, 12). Since HLA-DR2 mice but not HLA-DR3 mice delete CD4⁺ and CD8⁺ T cells bearing TCR Vβ8, this would be an excellent model to study the in vivo effects of endogenous superantigen-mediated TCR Vβ deletion on the immune response to an exogenous superantigen.

Contrary to the expectation that the absence of TCR Vβ8-bearing T cells in HLA-DR2 mice would result in substantial reduction in SEB-induced immune activation, we observed that splenocytes from HLA-DR2 mice showed a strong in vitro T-cell response to SEB, albeit lesser than DR3 mice (Fig. 4). To gain a better insight into this process, HLA-DR3 and HLA-DR2 mice were challenged in vivo with purified SEB and the TCR Vβ usage was determined by flow cytometry 3 days later. In HLA-DR3 transgenic mice, as shown by us earlier, TCR Vβ8⁺ T cells increased in both CD4⁺ and CD8⁺ subsets (Fig. 5). However, in HLA-DR2 transgenic mice, since TCR Vβ8-bearing T cells are deleted, they did not contribute much to the peripheral T-cell pool. However, T cells bearing TCR Vβ7 expanded by nearly 10-fold in both CD8⁺ and CD4⁺ T-cell populations. In addition, T cells bearing TCR Vβ14, which did not respond in DR3 mice, expanded significantly in response to SEB in DR2 mice. Together, TCR Vβ7 and Vβ14 contributed to 55 to 60% of the T-cell pool following SEB injection, whereas they contributed to less than 25% of the CD4⁺ and CD8⁺ T-cell pool in naive mice. Similar results were obtained with another DR2 transgenic line expressing the same DRα chain but a different DRβ chain (DRB*1502), which also shows TCR Vβ8 family deletion (data not shown).

FIG. 4.

In vitro T-cell response to SEB. Splenocytes from age-matched HLA-DR3 and -DR2 transgenic mice (n = 3 to 6 mice/group) were cultured in vitro with the indicated concentrations of SEB for 42 h. Cell proliferation was determined by thymidine incorporation assay.

FIG. 5.

TCR Vβ repertoire in superantigen-challenged HLA-DR transgenic mice. Splenocytes from age-matched HLA-DR3 and -DR2 transgenic mice (n = 3 to 10 mice/group) challenged with 10 μg of SEB 3 days earlier were stained for CD4, CD8, and TCR Vβ-specific fluorochrome-conjugated antibodies and analyzed by flow cytometry. Shown are percentages of cells expressing specific a TCR Vβ family within the respective gated population.

We next studied the extent of thymocyte deletion in SEB-treated mice. Whereas the peripheral T cells undergo expansion, the CD4 CD8 DP thymocytes undergo deletion following in vivo challenge with SEB (4). While there was a dramatic reduction in DP thymocytes in SEB-treated DR3 mice, DR2 mice had only minimal reduction in DP thymocyte subsets (Fig. 6). We also measured serum cytokine levels in these mice following SEB challenge (Table 1). While SEB-challenged HLA-DR3 mice had a dramatic increase in various proinflammatory cytokines in the serum as early as 4 h, neither DRB*1501 nor DRB*1502 mice showed such a change in serum cytokine profile after SEB challenge. In addition, while about 80% (8/10) of the DR3 mice succumbed to SEB-induced mortality during the study, none of the DR2 (DR*1501 or DR*1502) mice died (0/3 in each group).

FIG. 6.

SEB-induced thymic deletion in HLA-DR transgenic mice. Thymocytes from age-matched HLA-DR3 and -DR2 transgenic mice (n = 3 to 10 mice/group) challenged with 10 μg of SEB 3 days earlier were stained for CD4 as well as CD8 and analyzed by flow cytometry. Shown are the reductions in single-positive (SP) and DP thymocyte subsets.

TABLE 1.

Serum cytokine levels in different HLA-DR transgenic mice^a

HLA expressed	Mean level of indicated cytokine (pg/ml) ± SE
	IL-12p70	TNF-α	IFN-γ	IL-6	MCP-1
DR3 (0301)	596 ± 138	2,212 ± 523	7,466 ± 3,203	30,061 ± 6,400	39,710 ± 5,350
DR2 (1501)	67 ± 17	216 ± 7	263 ± 35	47 ± 202	1,143 ± 8
DR2 (1502)	67 ± 17	151 ± 6	291 ± 70	130 ± 237	861 ± 61

^aHLA-DR3; HLA-DR2 (1501), and HLA-DR2 (1502) mice (n ≥ 3/group) were challenged with either 10 μg of SEB or phosphate-buffered saline (PBS) alone by intraperitoneal injection. Mice were bled 4 h later, and levels of indicated cytokines in sera were determined by cytometric bead assay (BD Biosciences, San Diego, CA) as per manufacturer's instructions. PBS-challenged mice had negligible amounts of cytokines in the serum and hence are not shown. IL-12p70, interleukin-12p70; TNF-α, tumor necrosis factor alpha; IFN-γ, gamma interferon; MCP-1, monocyte chemoattractant protein-1.

In conclusion, we have shown for the first time that the immune response to exogenous superantigen that preferentially activates T cells bearing TCR Vβ families deleted by endogenous superantigen is significantly reduced to the extent of conferring protection from exogenous superantigen-mediated mortality. Endogenous superantigens could thus have clinical implications other than autoimmunity (9). Further, in the absence of T cells bearing the high-affinity TCR Vβ specific for a given exogenous superantigen (Vβ8 in this case), T cells bearing low-affinity TCR Vβ families (Vβ7 and Vβ14 in this case) can compensate to a certain extent.