Activation of antigen-presenting cells by microbial products breaks self tolerance and induces autoimmune disease (original) (raw)
Low incidence of spontaneous EAE in 5B6 transgenic B10.S mice. The 5B6 TCR was derived from an encephalitogenic CD4+ T cell clone generated from lymph node cells of an SJL mouse immunized with PLP139–151 (14). Mice transgenic for the 5B6 TCR were generated on the FVB background, backcrossed onto the EAE-susceptible SJL and EAE-resistant B10.S backgrounds and were monitored routinely for the development of spontaneous EAE. The incidence of spontaneous EAE was strikingly different in the two strains studied. While a mean of 40% 5B6 transgenic mice on the SJL strain developed spontaneous EAE, only 4% transgenic mice on the B10.S background developed disease when monitored over the same backcross generations and time period (Table 1). The 5B6 transgenic B10.S mice that did develop spontaneous EAE showed mild signs of clinical disease with a mean disease severity of 1.3 (data not shown). The significantly lower incidence of spontaneous EAE in transgenic B10.S mice could not be explained by exposure to different environmental antigen, as the SJL and B10.S colonies were housed under identical conditions in the same room of our animal facility, which is maintained under sterile SPF conditions. Taken together, these results show that in contrast to 5B6 transgenic SJL mice, B10.S transgenic mice were relatively resistant to development of spontaneous EAE, as overall only a small percentage of transgenic mice developed mild clinical signs of EAE spontaneously.
Low incidence of spontaneous EAE in 5B6 TCR transgenic B10.S mice
T cell development in 5B6 transgenic B10.S mice. To determine whether the relative resistance to spontaneous EAE observed in the 5B6 transgenic B10.S mice could be attributed to the deletion of autoreactive T cells either in the thymus or periphery, we examined the T cell populations in these mice by flow cytometry. Thymocytes from 5B6 transgenic B10.S mice were skewed toward the single-positive CD4 population at the expense of CD8+ T cells (Figure 1A). A pronounced skewing toward the CD4+ T cell population was expected, as the transgenic TCR genes were isolated from a MHC class II–restricted CD4+ T cell clone. Consistent with earlier results with TCR transgenic systems (15), this pattern is a strong indication that the transgenic T cells were positively selected. Essentially more than 95% CD4+ splenic T cells expressed the transgene-encoded TCR Vβ6 chain in 5B6 transgenic B10.S mice (Figure 1B). The transgenic TCR Vβ6 chain on CD4+ splenocytes was expressed at a physiological level, as the expression level of endogenous Vβ6 in nontransgenic littermates was very similar (Figure 1B). The total number of peripheral T cells in 5B6 transgenic mice was similar to that of nontransgenic controls (approximately 38 × 106 vs. 34 × 106 cells), indicating that there was no gross deletion of self-reactive T cells in the 5B6 transgenic B10.S mice.
Flow cytometry analysis of thymocytes and peripheral T cells in 5B6 transgenic B10.S mice. (A and B) T cells from thymi (A) and spleens (B) of mice were stained with the indicated antibodies (PE- or allophycocyanin-conjugated anti-CD4, FITC-conjugated anti-CD8, and PE-conjugated anti–TCR Vβ6). Dot plots representing two-color flow cytometry analysis of 5B6 transgenic mice with or without RAG-2 deficiency (TG/RAG-2–/– or TG, respectively) and nontransgenic littermates (NLM) are shown. Numbers in quadrants refer to percentages of gated cell populations. Total numbers of spleen cells are shown above dot plots.
Allelic exclusion of the TCRα chain is less tightly controlled than that of the TCRβ chain, resulting in the expression of nontransgenic TCRα chains paired with transgenic TCRβ chains. We could not directly examine the extent of endogenous TCR Vα expression in the 5B6 transgenic B10.S mice due to the lack of a specific antibody. However, we assessed the percentage of CD4+ T cells in peripheral blood expressing nontransgenic TCR Vα chains by staining with available antibodies against TCR Vα 2, Vα 3.2, Vα 8, and Vα 11. Assuming that there was no preferential pairing of 5B6 transgenic TCRβ chains with a nontransgenic TCRα chains, we estimated that 90% of peripheral CD4+ T cells in B10.S mice bear the transgenic TCR complex (data not shown). To further substantiate the finding that the majority of CD4+ T cells expressed the transgenic TCR and to conclusively determine whether the transgenic T cells were efficiently selected on the autoimmune-resistant B10.S background, we examined T cell populations in thymi and spleens of 5B6 transgenic B10.S mice that were deficient in RAG-2. In thymi of 5B6 transgenic RAG-2–/– mice, which expressed the transgenic TCR exclusively, T cells were not deleted but were efficiently selected to the CD4+ compartment (Figure 1A). The efficient positive selection by the transgenic TCR seen in the thymus was reflected in the periphery, where 80% of splenocytes were CD4+ and expressed the transgenic 5B6 TCR. The absolute numbers of CD4+Vβ6+ T cells in RAG-2–/– transgenic mice and normal transgenic mice were comparable, confirming the above estimations that expression of endogenous TCRs in the transgenic B10.S mice was minor (Figure 1B).
In summary, the transgenic 5B6 TCR-expressing T cells were not deleted but were efficiently selected on the autoimmune-resistant B10.S background. Thus, neither central nor peripheral deletion of transgenic T cells was the mechanism responsible for resistance to spontaneous EAE in these mice.
T cells from 5B6 transgenic B10.S mice are not anergic but respond to PLP139–151. To determine whether T cells from 5B6 transgenic B10.S mice responded as efficiently as T cells isolated from 5B6 transgenic SJL mice, we stimulated T cells purified from spleen cells of the transgenic mice with PLP139–151 presented by DAS cells (I-As-transfected fibroblasts) as artificial APCs. T cells from both 5B6 transgenic SJL and B10.S mice showed a strong and comparable proliferative response to PLP139–151 (Figure 2A). These data demonstrate that T cells from 5B6 transgenic SJL and B10.S mice have a comparable capacity to proliferate in response to PLP139–151 presented by neutral APCs.
Responses of 5B6 transgenic T cells to PLP139–151. (A) Proliferative response of 5B6 transgenic T cells to PLP139–151 presented by DAS cells. CD4-enriched T cell samples from 5B6 transgenic SJL and B10.S mice were stimulated with the indicated numbers of PLP139–151–pulsed DAS cells. T cell proliferation was assessed by [3H]thymidine incorporation assay. The mean cpm ± SD of triplicate cultures are shown. (B) Proliferative response of T cells from 5B6 transgenic B10.S or SJL mice to PLP139–151 presented by syngeneic APCs. Splenocyte samples from unimmunized 5B6 transgenic and nontransgenic littermate (NLM) B10.S or SJL mice were enriched for CD4+ T cells and were cultured with irradiated splenocytes from nontransgenic littermates in the presence of the indicated concentrations of PLP139–151 or control peptide PLP178–191. Proliferative responses were determined by [3H]thymidine incorporation assay. The mean cpm ± SD of triplicate cultures of one experiment representative of three are shown. (C) IL-2 response to PLP139–151 of T cells from 5B6 transgenic B10.S or SJL mice. Supernatants from cultures in B were assayed in duplicate by ELISA for cytokine production. Representative data from one of three experiments are shown. (D) INF-γ response of spleen cells from 5B6 transgenic B10.S or SJL mice to PLP139–151 stimulation. Culture supernatants from 5B6 transgenic SJL or B10.S whole spleen cells stimulated with the indicated concentrations of PLP139–151 were assayed in duplicate by ELISA for INF-γ production.
We next asked whether T cells from 5B6 transgenic B10.S mice were functional when stimulated by antigen presented by syngeneic APCs. Thus, we determined the proliferative and cytokine responses of CD4+ T cells from 5B6 transgenic B10.S mice after stimulation by PLP139–151 in vitro. Splenocyte samples enriched for CD4+ T cells from unimmunized 5B6 transgenic B10.S mice proliferated vigorously when stimulated with PLP139–151 presented by syngeneic nontransgenic splenocytes as APCs (Figure 2B). This response was highly antigen specific, since the T cells failed to proliferate to the control I-As binding peptide PLP178–191. In contrast, PLP139–151–specific proliferation was not detected in CD4+ T cells isolated from naive nontransgenic littermates (Figure 2B). In comparison with B10.S T cells, CD4-enriched T cell samples from 5B6 transgenic SJL mice demonstrated a similar proliferative response following antigen stimulation (Figure 2B). To determine whether immune deviation to an anti-inflammatory cytokine profile could explain the relative tolerance to spontaneous EAE in the 5B6 transgenic B10.S mice, we examined the cytokine response of transgenic T cells following stimulation with PLP139–151 in vitro. Supernatants of PLP139–151–stimulated CD4+-enriched T cell cultures were assayed for secretion of IL-2, IL-4, IL-10, TNF-α, and IFN-γ by ELISA. We consistently found that 5B6 transgenic CD4+ T cells stimulated with PLP139–151 produced IL-2 in significant amounts, in contrast to T cells from nontransgenic littermates (Figure 2C and data not shown). The amounts of IL-2 detected in the culture supernatants of CD4+ T cells from 5B6 transgenic SJL mice were generally lower than those in cultures from transgenic B10.S mice. The amounts of IL-4, IL-10, INF-γ, and TNF-α were generally below the detection limit in supernatants from both the transgenic and negative littermate cultures, indicating that T cells from 5B6 transgenic B10.S mice had a naive phenotype. Because we did not detect INF-γ in culture supernatants of purified CD4 T cells from 5B6 transgenic B10.S and SJL mice, we examined the production of this cytokine by whole spleen cells after stimulation with PLP139–151. Significant and comparable amounts of INF-γ production in PLP139–151–stimulated spleen cell cultures of 5B6 B10.S or SJL mice were observed (Figure 2D). This observation indicates that spleen cells from 5B6 transgenic B10.S mice can mount antigen-specific INF-γ responses. Since purified transgenic CD4+ T cells did not produce INF-γ upon antigen-specific stimulation, the production of INF-γ in spleen cell cultures suggests that this cytokine is produced by an alternative source such as APCs, CD8 T cells, or NK cells. Taken together, these data demonstrate that T cells from 5B6 transgenic B10.S mice were not anergic but were highly responsive to PLP139–151 presented by either artificial or syngeneic APCs and showed antigen-specific responses comparable to those of T cells from 5B6 transgenic SJL mice.
APCs in B10.S mice show a lower basal state of activation and stimulate PLP-specific T cells less efficiently than do SJL APCs. Based on our findings above, we concluded that the relative resistance to spontaneous EAE in the 5B6 transgenic B10.S mice could not be explained as being due to thymic deletion or peripheral tolerance of transgenic T cells. To determine whether the failure of 5B6 transgenic T cells in B10.S mice to mediate spontaneous EAE was due to a defect in APCs, we examined the activation/maturation state and responsiveness of APCs in wild-type B10.S mice.
One of the hallmarks of activation and maturation of APCs such as dendritic cells (DCs) is upregulation of MHC class II and costimulatory molecules (16). Thus, we examined the activation/maturation state of APCs in wild-type B10.S mice by determining the expression level of I-As on subsets of APCs, as defined by CD19+ (B cells), CD11b+, CD11c+ (DCs), and F4/80+ (macrophages). The activation/maturation state of APCs from B10.S mice was compared with that of APCs from EAE-susceptible SJL mice. As expected in unimmunized mice, the examined subsets of APCs expressed low to moderate surface levels of MHC class II. Within each APC subset, the number of MHC class II+ cells was lower in B10.S than in SJL wild-type mice (Figure 3A). Furthermore, APCs from B10.S mice consistently expressed lower levels of MHC class II than did those from SJL mice, as determined by the mean fluorescence intensity (Figure 3, B and C). We also examined the activation state of BM-derived APCs by culturing BM from wild-type SJL and B10.S mice with GM-CSF and assessing surface expression of MHC class II on CD11b+, CD11c+, or F4/80+ cells. The generation of APCs in vitro yielded lower numbers of MHC class II–expressing APC subsets from B10.S BM than from SJL BM (Figure 3D). In agreement with the data from ex vivo APCs, the activation/maturation state, as assessed by MHC class II expression, of BM-derived CD11b+, CD11c+ and F4/80+ APCs from B10.S mice was lower than that of APCs from SJL mice (Figure 3E).
Activation/maturation state and T cell-activating capacity of APCs. (A–E) Spleen cells (A–C) or BM-derived APCs (D and E) from wild-type B10.S and SJL mice were stained with PE-conjugated antibodies against the indicated subsets of APCs and biotinylated anti–I-Aq/s. The level of MHC class II expression was determined on gated APC subsets as indicated (%) (A and D) by flow cytometry and is shown as mean of fluorescence intensity (MFI) (B, C, and E). Data from one of three experiments with similar data are shown. (F) The indicated numbers of irradiated spleen cells from wild-type SJL or B10.S mice as APCs in presence of PLP139–151 were incubated with 5B6 transgenic SJL T cells and were subsequently pulsed with [3H]thymidine. T cell proliferation was measured by [3H]thymidine incorporation assay. Values are shown as mean cpm ± SD of triplicate wells. One experiment representative of three is shown.
To determine the functional significance of the low activation/maturation state of APCs from B10.S mice compared with SJL APCs, we examined their ability to stimulate 5B6 transgenic T cells in proliferation assays. Purified T cells from 5B6 transgenic SJL mice were stimulated with irradiated spleen cells as APCs from nontransgenic SJL or B10.S mice in presence of PLP139–151. APCs from B10.S mice were less efficient in stimulating 5B6 transgenic T cells to proliferate than were APCs from SJL mice at constant concentrations of PLP139–151 (Figure 3F). This difference in T cell–activating capacity between the APCs from the two strains was seen at different APC/T cell ratios, indicating that it was APC dependent. However, it was not dependent on the source of the transgenic T cells, as similar results were obtained in proliferation assays using T cells from 5B6 transgenic B10.S mice (data not shown). These data indicate that antigen-pulsed APCs from B10.S mice are less efficient in inducing T cell proliferation than are APCs from SJL mice. In summary, splenic APCs from wild-type B10.S mice show lower activation/maturation states and induce a lower proliferative response in 5B6 transgenic T cells than do wild-type SJL APCs.
APCs from B10.S mice are activated by CpG ODNs. It has recently been shown that synthetic ODNs that contain unmethylated CpG motifs (CpG ODNs) trigger activation and maturation of murine APCs such as dendritic cells, macrophages, and B cells through the Toll-like receptor TLR9 (17–19). To exclude the possibility that APCs from B10.S mice have a general defect in activation, we determined whether they could be activated in vitro by sequence-specific CpG ODNs. CpG ODN induced upregulation of MHC class II and CD86 on CD19+, CD11b+, and CD11c+ subsets of spleen cells from B10.S mice and expanded the examined APC subsets compared with stimulation with non–CpG ODN (Figure 4). Taken together, these data demonstrate that CpG ODNs can expand and activate B10.S APC populations in vitro, as determined by upregulation of MHC class II and the costimulatory molecule CD86.
CpG ODN induces expansion of APC subsets and upregulation of MHC class II and CD86. Spleen cells from B10.S mice were stimulated with CpG ODN or non–CpG ODN. Spleen cells were subsequently stained with FITC-conjugated anti-CD11b, anti-CD11c, and anti-CD19 and PE-conjugated anti-CD86 or biotinylated anti–I-Aq/s. Expression of CD86 and MHC class II was determined on CD11c+-, CD11b+-, or CD19+-gated cells. Dotted lines indicate isotype controls. Numbers refer to mean of fluorescence intensity of specific staining (upper number) shown on gated APC subsets (%).
In vivo activation of APCs induces EAE in 5B6 transgenic mice on the resistant B10.S background. To examine the role of APCs in EAE, we next asked whether CpG ODN–mediated activation of APCs could break self tolerance and induce EAE in the 5B6 transgenic B10.S mice. Administration of CpG ODN alone, which activates APCs through TLR9 (19), was sufficient to break T cell tolerance and induce EAE in 33% of 5B6 transgenic B10.S mice (Table 2). However, administration of CpG ODN in presence of low amounts of PLP139–151/IFA induced severe EAE in the majority of 5B6 transgenic B10.S mice (75%). In striking contrast, when non–CpG ODN plus PLP139–151 was administered, that is, in absence of TLR9-mediated APC activation, none of the 5B6 transgenic mice developed any clinical signs of EAE. These findings indicate that even in presence of exogenous autoantigen, activation of APCs is critical for development of autoimmune disease in this TCR transgenic model.
Immunization of 5B6 transgenic B10.S mice with CpG ODN induces EAE
To determine whether abrogation of self tolerance in 5B6 transgenic B10.S mice was specific to the activation of APCs via TLR9, we examined whether systemic administration of microbial agents such as LPS and PTX also resulted in EAE in these mice. LPS activates DCs to release IL-12 by signaling through TLR4 and induces INF-γ-secreting CD4+ T cells in vivo (20, 21). 5B6 transgenic B10.S mice injected with LPS showed a low incidence rate of 13%, resulting in an earlier onset of very mild EAE (Table 3). None of the nontransgenic littermate control mice developed any signs of clinical EAE (data not shown). PTX, an inhibitor of Gαi proteins with diverse physiological effects such as breaching the blood-brain barrier and promoting lymphocytosis (22, 23), has recently been shown to induce INF-γ responses in T cells in absence of IL-12 via activation of APCs (24, 25). Administration of PTX to 5B6 transgenic B10.S mice resulted in EAE (mean peak severity, 2.6 ± 0.3) in all of the treated animals with clinical signs starting at 6 days (Table 3). The PTX-induced EAE was specific for the 5B6 TCR transgenic mice, as none of the nontransgenic littermate controls showed any signs of clinical EAE. IL-12 is thought to play a crucial role in induction of INF-γ production by T cells (26). To directly elucidate the role of IL-12 in breaking T cell tolerance in our transgenic model, we treated 5B6 transgenic B10.S mice every other day with rmIL-12 in PBS for a total of 11 administrations. Over the monitoring period of 44 days, only one IL-12–treated mouse out of six developed clinical EAE, with delayed onset compared with PTX- or LPS-induced EAE. In contrast, none of the 5B6 transgenic B10.S mice that were injected with PBS as controls developed any signs of clinical EAE (data not shown).
Induction of EAE in 5B6 transgenic B10.S mice via TLR9-independent pathways
Taken together, our results indicate that in vivo activation of APCs via the innate immune receptor TLR9 could break self tolerance and induce EAE in the 5B6 transgenic B10.S mice. Activation of APCs via TLR4 or administration of IL-12 to induce INF-γ production by T cells was less effective in inducing EAE in these mice than was TLR9-mediated APC activation. In contrast, PTX, which has pleiotropic effects on the immune system inducing APC activation, was more effective in mediating EAE in 5B6 transgenic B10.S mice than was CpG ODN.