Mycoplasma superantigen initiates a TLR4-dependent Th17 cascade that enhances arthritis after blocking B7-1 in Mycoplasma arthritidis-infected mice

Authors


Summary

Mycoplasma arthritidis is a natural pathogen of rodents causing arthritis, toxic shock and necrotizing fasciitis. It secretes a potent superantigen (SAg), MAM, that differentially affects the immune system depending upon presence or absence of TLR4, thus potentially influencing disease outcomes. Here, we establish that antibody to co-stimulatory molecule B7-1(CD80) enhances arthritis in wild-type C3H/HeSnJ (TLR2+4+) mice but suppresses arthritis in TLR4-defect C3H/HeJ (TLR2+4−) mice. Also, blockade of the B7-1/CD28 co-stimulatory pathway in C3H/HeSnJ mice resulted in a marked increase in an alternative co-stimulatory pathway ICOS/ICOSL that was associated with elevation of the IL-17/Th17cascade with enhanced IL-23, IL-6, and the RORγt and STAT3 transcriptional factors on CD4+ T cells. Anti- B7-1 also increased inflammatory chemokines and the stress protein HMGB1 that promotes cellular infiltration to joints. Using a MAM-deficient strain of M. arthritidis, a monoclonal antibody to TLR4 and a TLR4-defective mouse strain, we established that both MAM and TLR4 are required for the systemic and local joint triggering of the Th17/IL-17 cascade in mice treated with anti-B7-1 antibody. Importantly, blocking of IL-17 with anti-IL-17 antibody suppressed the elevated arthritis in M. arthritidis-infected mice treated with anti-B7-1 antibody. Thus, this unique model of arthritis illustrates how microbial agonists can bridgeinnate and adaptive immune responses to redirect signalling pathways, thus promoting chronic inflammatory and autoimmune disease.

Introduction

Mycoplasma arthritidis is a natural pathogen of rodents causing acute to chronic arthritis, resembling human rheumatoid arthritis (RA), a lethal toxicity, and a necrotizing fasciitis-like syndrome (Cole et al., 2000). The M. arthritidis model of inflammatory disease has been extensively studied in our laboratory in regards to both host and microorganism components, resulting in the identification of animal strains that differ in their genetic susceptibility to disease as well as the identification and derivation of organisms exhibiting differing degrees of virulence. M. arthritidis secretes a phylogenetically distinct superantigen (SAg), M. arthritidis mitogen (MAM) (Cole et al., 1981; 2000), which we have proposed might play a key role in these inflammatory diseases, although by itself it does not cause overt clinical outcomes. Unlike other SAgs and other microbial agonists with the possible exception of certain bacterial heat shock proteins, MAM alone can directly interact with both TLR2 and TLR4, two important innate immune recognition molecules (Mu et al., 2005). As part of its superantigenicity, it also interacts with class II MHC molecules and the Vβ chains of the T-cell receptors (TCRs) for antigens thus bridging the innate and adaptive immune systems (Cole and Atkin, 1991; Cole et al., 2000).

We have previously established that mouse strains expressing both TLR2 and TLR4 (i.e. wild-type C3H/HeSnJ or C3H/HeN) develop a Th2-type cytokine profile after administration of MAM, but that TLR4-defective mutant mice (i.e. C3H/HeJ) develop a Th1-type immune response (Mu et al., 2005). We also reported that manipulation of B7-1 (CD80) but not B7-2 (CD86) function can modulate Th1/Th2 immune responses depending upon presence or absence of TLR4 in mice injected with MAM SAg or in mice infected with live mycoplasma (Mu et al., 2006). Previously, we demonstrated that blocking of B7-1 could enhance arthritis in mice infected with M. arthritidis (Mu et al., 2006). More recently, we show that MAM/TLR4 interaction can also lead to an enhancement of the IL-17/Th17 cascade that is not seen in a mouse strain lacking a functional TLR4 (Mu et al., 2011).

Recent studies examining T-cell differentiation have established two new subsets of Th cells. The Th17 cells produce IL-21, IL-22, CCR6 and IL-23 receptor (IL-23R), and express higher levels of RORγt and STAT3 transcriptional factor (Weaver et al., 2006; Korn et al., 2009), and the regulatory T cells (Treg) express the FOXP3 transcription factor (Hori et al., 2003; Sakaguchi et al., 2010). These two newly defined T-cell subsets, as well as previous well described Th1 and Th2 subsets, play a crucial role in host immune responses and disease. Although there is controversy over the specific signals definitively required for the initiation of Th17 differentiation (i.e. IL-6 and TGFβ), it is well established that Th17 cells clearly require IL-23 for their proliferation, maintenance, stabilization and pathogenicity (Korn et al., 2009). Also, there is a growing appreciation that Th cell activation, differentiation and effector T-cell function are regulated by co-stimulatory molecules. CD28 is an important co-stimulatory molecule expressed by T cells. It binds to B7-1 (CD80) and B7-2 (CD86) on activated antigen-presenting cells (APCs). Engagement of CD28 with B7 molecules prevents T-cell anergy that leads to T-cell activation, proliferation and cytokine production (Bluestone, 1995). It is well documented that the functional outcome of B7-1 and B7-2-mediated signalling appears to be distinct (Freeman et al., 1995; Kuchroo et al., 1995; Pentcheva-Hoang et al., 2004). Furthermore, the progression and outcome of some infectious/autoimmune diseases are suppressed in a different manner by blocking B7-1 or/and B7-2 molecules (Kuchroo et al., 1995; Miller et al., 1995; Odobasic et al., 2008).

There is increasing evidence that IL-17/Th17 response is associated with various human autoimmune disorders (Korn et al., 2009) such as rheumatoid arthritis (RA) (Cascao et al., 2010), multiple sclerosis (MS) (Matusevicius et al., 1999), systemic lupus erythematosus (SLE) (Crispín et al., 2010) and psoriasis (Kryczek et al., 2008). Recently, attention has been directed towards to the role played by TLRs, microbial agonists and new inflammatory cascades involved in the IL-17/Th17 pathway (Fischer and Ehlers, 2008; Mills, 2011). However, the precise interaction between all these components which result in different clinical expressions largely remain to be defined. Our recent findings have shown that there is a profound difference in in vivo development of Th17 response as the consequence of divergent MAM interactions with different TLRs that control Th cell differentiation in the MAM SAg model system (Mu et al., 2011). Since MAM alone fundamentally alters cytokine profiles in vivo, which promotes the Th17 cascade by TLR4-dependent pathway, using M. arthritidis disease model we are now in a position to more definitively determine how these components interact to direct the immune events responsible.

Thus, the present study was undertaken to define the role played by the MAM SAg in arthritis caused by M. arthritidis and to elucidate the resulting cellular and molecular pathways that lead to the control of the inflammatory response in this natural model of inflammatory arthritis. In this study, we examined the mechanisms involved in the enhanced arthritis in M. arthritidis-infected mice that were pre-treated with anti-B7-1 antibody. Here we show that blocking of B7-1 greatly upregulates the IL-17/Th17 inflammatory response and enhances arthritis in wild-type C3H mouse strains that express both TLR2 and TLR4. Importantly, we also demonstrate that the IL-17/Th17 cascade and its associated cytokines, chemokines and transcription factors seen systemically and in the joints of mice infected with M. arthritidis, are all dependent upon a unique interaction of MAM with the TLR4 molecule. Thus, MAM plays a pivotal role in the inflammatory response of the host to the M. arthritidis infection.

Results

Effect of anti-B7-1 treatment on arthritis induced by M. arthritidis in C3H/HeSnJ versus C3H/HeJ mice

In the first experiments we established that treatment of mice with anti-B7-1 antibody profoundly alters their susceptibility to the arthritogenic effects of M. arthritidis depending upon presence (C3H/HeSnJ, TLR2+/4+) or absence (C3H/HeJ, TLR2+/4−) of a functional TLR4 molecule. Anti-B7-1 or control antibody (100 μg per mouse) were given twice to mice prior to the injection of M. arthritidis strain 158p10p9 (5 × 108 cfu per mouse). Anti-B7-1 antibody markedly enhanced arthritis in C3H/HeSnJ mice at all time periods through 28 days as compared with control antibody (P <0.01, Fig. 1) and arthritis was of earlier onset as compared with mice receiving control antibody being apparent after only 24 h versus 5 days in control mice. Anti-B7-1 antibody did not enhance arthritis in C3H/HeJ mice and, in fact, slightly but significantly suppressed the arthritis versus that seen in mice receiving control antibody (P < 0.05, Fig. 1). Arthritis in C3H/HeJ mice was largely resolved by 21 to 28 days.

Figure 1.

Effect of anti-B7-1 antibody on arthritis induced by M. arthritidis in two different C3H mouse strains. C3H/HeSnJ (TLR2+/4+) and C3H/HeJ (TLR2+/4−) mice were injected with anti-B7-1 monoclonal or isotype-matched control antibodies (100 μg each mAb/mouse respectively) twice at day −2 and 0. At day 0 mice were injected i.v. with 5 × 108 cfu M. arthritidis 158p10p9 and scored for arthritis through 28 days. Mice injected with normal saline (NS) instead of mycoplasma were included as controls. Antibody against B7-1 does not cause disease in NS-injected mice. The data are pooled from three separate experiments and mean scores for anti-B7-1-injected (close square) and control antibody-treated (open square) mice ± SEM are shown.

IL-17 response to infection with M. arthritidis

Previous studies had established that the cytokine profiles in C3H mouse substrains induced by the MAM SAg depended upon presence or absence of a functional TLR4 molecule (Mu et al., 2001; 2005) and that B7-1, but not B7-2, played a key role in this response (Mu et al., 2006). In addition, we recently demonstrated that mice injected with MAM induced a very early IL-17 response in C3H/HeSnJ mice versus a significantly later and lower IL-17 response in C3H/HeJ mice (Mu et al., 2011). Inasmuch as IL-17 can be associated with inflammatory and autoimmune diseases (Korn et al., 2009; Marwaha et al., 2012), we investigated further whether the enhanced arthritis resulting from anti-B7-1 treatment might be associated with an increase in the IL-17 response to infection by live M. arthritidis. Initial experiments here used splenocytes and sera but subsequently we employed the joint cells and CD4+ T cells from peripheral lymph node cells (PLNs) of M. arthritidis-infected mice as being more relevant for understanding the role of Th17 in the diseases process.

In Fig. 2A, we show that M. arthritidis induces elevated IL-17 in sera and supernatants from cultured activated splenocytes taken from C3H/HeSnJ mice that spontaneously released cytokines. The levels of IL-17 in sera and supernatant from cultured splenocytes were greatly enhanced by two- to threefold in the presence of anti-B7-1 antibody at all time periods between 5 and 28 days (P <0.01). The results were confirmed in that a significantly higher upregulation of mRNA expression of IL-17 and RORγt were seen in CD4+ T cells from PLNs taken from mice receiving anti-B7-1 antibody treatment (Fig. 2A). The cytokines that are known to be associated with Th17 differentiation and IL-17 induction i.e. IL-6, IL-23(p19) and TGFβ, were all also elevated by up to 2- to 10-fold in sera collected from mycoplasma-injected mice in the presence of anti-B7-1 antibody (Fig. 2B). The IL-12p40 cytokine, a dimmer between IL-12 and IL-23 (Lyakh et al., 2008), which is commonly present during inflammatory processes, was similarly expressed in mice given either anti-B7-1 or control antibody (P > 0.1).

Figure 2.

Effect of anti-B7-1 treatment on IL-17 and Th17-associated cytokines in vivo in M. arthritidis-injected mice. C3H/HeSnJ mice were injected with 5 × 108 cfu M. arthritidis 158p10p9 following the treatments with an anti-B7-1 or control antibodies as described in Fig. 1 legend. Mice injected with NS ± anti-B7-1 antibody were used as controls.

A. (i) The kinetics of serum IL-17 levels were determined. Mice were sacrificed at intervals through 28 days and the sera were collected and assayed for IL-17 content. Sera from five mice of each group were assayed in each experiment for each specific time point. Data were pooled from three experiments and are expressed as means ± SEM. (ii) Effect of anti-B7-1 antibody on constitutive cytokine production by splenic cells isolated from M. arthritidis-infected mice. Splenic cells were collected at varying days after anti-B7-1 treatment and an injection with live organisms. Cells were cultured with fresh medium for an additional 24 h in the presence of gentamicin and culture supernatants were harvested and assayed for IL-17. To determine if CD4+ T cells were affected by anti-B7-1 treatment in response to M. arthritidis, CD4+ T cells were isolated and purified from peripheral lymph nodes and total RNA were extracted, levels of IL-17 mRNA (iii) and RORγt (iv) expression were measured by quantitative RT-PCR (qRT-PCR). The transcripts were analysed and normalized to β-actin. Data are expressed as fold change, with values from mice injected with NS ± anti-B7-1 antibody. Data were pooled from three experiments and are expressed as means ± SEM.

B. Serum levels of Th17-associated cytokines IL-6, IL-23, IL-12p40 and TGFβ from C3H/HeSnJ mice were analysed at varying time points to determine the effect of anti-B7-1 treatment on C3H/HeSnJ mice infected with M. arthritidis 158p10p9. Sera from five mice of each group were analysed in each experiment for each specific time point. Data are pooled from three experiments and are expressed as means ± SEM.

Next, to determine if above findings could be duplicated in vitro, we stimulated naïve splenocytes from C3H/HeSnJ mice with live M. arthritidis 158p10p9 (1 × 105 cfu per 1 × 107 cells per culture). In in vitro experiments, IL-17 and IL-23(p19) in culture supernatants were all significantly enhanced by anti-B7-1 antibody (data not shown). Also, the levels of IL-21 and IL-22 mRNA were significantly elevated in CD4+ T cells from peripheral lymph nodes following anti-B7-1 treatment. The expression of the IL-23R transcripts, which is a signature marker expressed on pathogenic Th17 cells was also greatly elevated (data not shown). Thus, in vitro data are consistent with those shown in Fig. 2.

Effect of anti-B7-1 antibody on Th1 and Th2 cytokines

Returning now to the in vivo model, we examined the effect of anti-B7-1 on cytokine profiles thought to be associated with Th1 and Th2 cells (Fig. 3). Anti-B7-1 markedly suppressed serum IFNγ in mice given live M. arthritidis in comparison with the level of those seen with control antibody. Serum IL-4, a signature Th2 cytokine, was not affected significantly by anti-B7-1 antibody whereas the levels of IL-6 and IL-10, two cytokines that are considered to share the features of both Th17 and Treg/type 1 regulatory T cells (Tr1), were distinctly increased by treatment with anti-B7-1. All of these results were confirmed in that expression of mRNAs on CD4+ T cells from PLNs were suppressed for IFNγ, unchanged for IL-4, and elevated for IL-6 and IL-10.

Figure 3.

Effect of anti-B7-1 treatment on serum Th1 and Th2 cytokine profiles induced in M. arthritidis-infected mice.

A. Serum samples obtained from the experiments described in Fig. 2 were analysed for IFNγ, IL-4, IL-6 and IL-10. Data are pooled from three experiments and are expressed as means ± SEM.

B. Peripheral lymph nodes were collected from the mice as described in Fig. 2. CD4+ T cells were isolated and total RNA was extracted; levels of IFNγ, IL-4, IL-6 and IL-10 expressions were determined by qRT-PCR. Data are expressed as fold change, compared with values from mice injected with NS ± anti-B7-1 antibody. Data are pooled from two to three experiments and are expressed as means ± SEM.

Effect of anti-B7-1 antibody on expression of molecules expressed by Th17 cells

Next, we measured the expression of CCR6, IL-23R and ICOS molecules which are hallmarks of Th17 cells (Korn et al., 2009) on CD4+ T cells in PLNs from mice injected with M. arthritidis that had received control or anti-B7-1 antibody. Flow cytometric analyses were used to determine the percentage of cells that expressed each of the latter markers (Fig. 4A–C). The results show a two- to threefold increase in the % of CD4+ T cells expressing CCR6, IL-23R and ICOS in mice treated with anti-B7-1 antibody versus those receiving control antibody (P <0.01 and P <0.05 respectively). In confirmation of these results, levels of CCR6, IL-23R and ICOS mRNA expressions by CD4+ T cells of PLNs (Fig. 4D–F) were likewise increased in mice receiving anti-B7-1 antibody. The significance of this finding is the fact that increase in these markers demonstrates that anti-B7-1 treatment causes proliferation and stabilization of the Th17 subset.

Figure 4.

Anti-B7-1 treatment increases expression of Th17-associated surface molecules.

A–C. Th17-associated molecules CCR6, ICOS and IL-23R expressed by CD4+ T cells at varying time points after treatment with an anti-B7-1 mAb or a control antibody in mice injected with 5 × 108 cfu M. arthritidis 158p10p9. Peripheral lymph nodes were isolated and single cells were prepared for flow cytometric analysis. The percentage of IL-17+/CD4+ cells were gated and recorded. Data are pooled from three experiments and are expressed as means ± SEM

D–F. Concurrent measurement on the levels of CCR6, ICOS and IL-23 receptor expressions by CD4+ T cells in PLN were conducted and these transcripts were assayed by qRT-PCR. Mice injected with NS ± anti-B7-1 antibody were also used as controls. Data are expressed as fold change, compared with values from mice injected with NS ± anti-B7-1, and are pooled from three experiments and are expressed as means ± SEM.

Can similar effects be seen in the joints of arthritic mice treated with anti-B7-1 antibody? C3H/HeSnJ mice given antibodies and M. arthritidis (158p10p9) were sacrificed at intervals through 10 days post injection; ankle joints were removed and joint tissues were extracted for total RNA. As shown in Fig. 5, anti-B7-1 antibody markedly increased the levels of IL-17, CXCL2, IL-6 and high-mobility group box protein 1 (HMGB1) in comparison with control antibody. CXCL2 is a chemokine produced mainly by neutrophils and promotes T-cell infiltration, whereas HMGB1 is a member of the inflammatory stress protein family and is produced by many cell types and has been shown to amplify TLR4-mediated inflammation in inflammatory arthritis (Harris et al., 2012). Thus, the elements of the IL-17 cascade are significantly higher in the affected joints of mice treated with anti-B7-1 antibody.

Figure 5.

Effect of anti-B7-1 treatment on joint inflammation in M. arthritidis-infected mice. Mice treated with anti-B7-1 or control antibodies were given M. arthritidis 158p10p9 as before and RT-PCR analysis of inflammatory cytokine transcripts in the ankle joint tissues at varying time points as indicated was performed. The following transcripts were analysed and then normalized to β-actin: IL-17A, CXCL2, IL-6 and HMGB1. Data are expressed as fold change, compared with values from mice injected with NS ± anti-B7-1. The results are pooled from three separate experiments and are expressed as means ± SEM.

Role of MAM in promotion of the IL-17 cascade

The results above have established the similarity between the levels of IL-17 and its associated cytokines shown here by M. arthritidis and those seen previously in mice given the pure native MAM superantigen (Mu et al., 2011). Since M. arthritidis possesses inflammatory components other than MAM, a more direct approach was taken to establish whether MAM is a key player in the events following infection of mice with live organisms. To address this issue, we used the wild-type M. arthritidis 158KD strain and its MAM knockout derivative, MAMKO (courtesy of Dr Kevin Dybvig, University of Alabama at B irmingham) and compared the course of disease in C3H/HeSnJ mice. Although peak arthritis scores were similar for both M. arthritidis strains, wild-type M. arthritidis 158KD exhibited a very early onset of disease (maximum scores were seen by 3 days post injection) whereas the MAMKO derivative strain exhibited a delayed onset of disease (data not shown). Thus, although MAM can alter the kinetics of disease development, clearly factors other than MAM must contribute to disease expression and outcomes.

Next, we examined the role of MAM in the production of IL-17 and its transcriptional factor STAT3, which along with RORγt and IL-23 promote development of Th17 cells (Korn et al., 2009). Total peripheral lymph node cells from arthritic C3H/HeSnJ mice were collected at intervals through 21 days post infection and CD4+ T cells were extracted and assayed for levels of IL-17 mRNA expression and tested for the phosphorylated STAT3 activity (Fig. 6). Mice given M. arthritidis MAMKO exhibited minimal levels of both IL-17 mRNA and STAT3 versus those seen in the wild-type 158KD strain (P <0.01). The results strongly suggest that IL-17 and the activated STAT3 require the presence of MAM.

Figure 6.

Effect of wild-type versus MAMKO M. arthritidis on the Th17-associated immune response in mice. Mice were injected i.v. with 5 × 108 cfu wild-type M. arthritidis KD or its MAMKO derivative and were scored for arthritis through 21 days. The wild-type M. arthritidis with functional MAM gene promotes IL-17 response and the levels of Th17-associated transcriptional factor STAT3. Mice were injected with either wild-type or MAMKO M. arthritidis, and CD4+ T cells were isolated and purified from peripheral lymph nodes and total RNA and protein extracts were prepared, levels of IL-17 mRNA (A) and STAT3 (B) expression were determined. Data were pooled from three experiments and are expressed as means ± SEM.

In a supplemental experiment, we next examined events occurring in the joints of arthritic C3H/HeSnJ mice infected with 158KD or MAMKO organisms at intervals through 10 days (Fig. 7). Neutrophils were predominant in mice injected with either 158KD or MAMKO organisms, but they were significantly higher (P <0.05) in mice infected with the wild-type 158KD versus the MAMKO-treated mice (Fig. 7); this corresponded with higher levels of CXCL2 expression. In addition, the numbers of CD3+ T cells and macrophages were lower in the joints of mice receiving the MAMKO organisms. mRNA levels of IL-17 and HMGB1 expression were correspondingly lower on cells of mice receiving the MAMKO mycoplasmas. Thus, MAM also has strong influence on the migration of inflammatory cells to the arthritic joints as well as enhanced expression of inflammatory markers.

Figure 7.

Effect of wild-type versus MAMKO M. arthritidis on immune cell infiltration in mouse ankle joint tissues.

A. Ankles from wild-type or MAMKO M. arthritidis-infected C3H/HeSnJ mice were removed at different time points as indicated. Synovial membranes were scraped from articular surfaces and digested with collagenase. Cell numbers of CD3+ T cells, CD11b+c neutrophils and F4/80+ macrophages were measured by flow cytometry. The total number of each cell population were acquired and calculated. Groups consisted of six wild-type M. arthritidis- and six MAMKO-infected mice. Data shown are pooled from three separate experiments and are expressed as means ± SEM.

B. mRNA expression of IL-17, CXCL2 and HMGB1 in mouse ankle joints at different time points post wild-type and MAMKO M. arthritidis infection. Ankle RNA isolation and qRT-PCR were performed. The tested transcripts were analysed and normalized to β-actin. Data shown are pooled from three separate experiments and are expressed as means ± SEM.

The role of TLR4 in regulation of the IL-17/Th17 cascade in mice infected with M. arthritidis

We previously established that TLR4 plays an important role in regulating the immune pathways following administration of the MAM SAg (Mu et al., 2005; 2006; 2011). The present studies described above indicate that MAM also plays a key role in regulating the immune responses to live M. arthritidis. Using blocking antibody to TLR4 or control antibody we next examined whether TLR4 plays a similar role in immune outcomes mediated by wild-type 158KD versus its MAMKO derivative M. arthritidis. Mice were treated with antibodies prior to the administration of organisms and serum samples were collected at intervals through 21 days and peripheral lymph node cells collected through 10 days. In Fig. 8A we show that C3H/HeSnJ mice receiving control antibody and wild-type M. arthritidis strain 158KD induced elevated levels of serum IL-17 that peaked at 1600 pg ml−1 between 10 and 14 days post injection and stayed high through 21 days.

Figure 8.

Role of TLR4 and B7-1 on IL-17 response and the Th17 elaboration in vivo in wild-type versus MAMKO M. arthritidis-infected mice.

A. Differential effect of anti-TLR4 on serum IL-17 response in wild-type versus MAMKO M. arthritidis-infected mice. C3H/HeSnJ mice were infected with wild-type M. arthritidis, KD or the MAMKO derivative strain after treatment with an anti-TLR4 or control. At varying time points, mice were sacrificed and the sera were collected for cytokine assays for IL-17. Groups consisted of five wild-type M. arthritidis- and five MAMKO-infected mice. Data are pooled from three experiments and are expressed as means ± SEM.

B. Increase in the percentage of IL-17 + CD4+ cells is dependent upon the presence of TLR4 in wild-type M. arthritidis-infected mice. As described in (A). At varying time points after M. arthritidis injection, CD4+ cells were isolated and purified from peripheral lymph nodes of M. arthritidis, 158 KD-infected mice pre-treated with anti-TLR4 or control antibody. The percentage of CD4+ cells expressing IL-17 was determined. Data shown are pooled from three separate experiments and are expressed as means ± SEM.

C. Differential effect of anti-TLR4, anti-B7-1 and a combination of antibodies on constitutive IL-17 in M. arthritidis KD-infected mice. Mice were injected with wild-type live mycoplasma after treatment with anti-TLR4, anti-B7-1, or a combination of both antibodies, or an isotype-matched anti-mouse antibody (100 μg each mAb per mouse respectively) on two separate occasions. At varying time points after M. arthritidis injection, mice were sacrificed and peripheral lymph node cells (107 cells ml−1) were isolated and cultured with fresh plain RPMI 1640 medium for 24 h. Culture supernatants were assayed for IL-17 by capture ELISA.

In striking contrast, mice receiving anti-TLR4 antibody and M. arthritidis 158KD exhibited reduced peak serum IL-17 levels of only 400–500 pg ml−1, a 75–80% reduction (P <0.01). As we predicted, MAM-deficient M. arthritidis, MAMKO failed to elicit IL-17 levels higher than 250 pg ml−1 irrespective of the antibody treatment. To confirm a role for TLR4, we measured the per cent of cells expressing IL-17 among the CD4+ T-cell population from peripheral lymph node cells of C3H/HeSnJ mice infected with M. arthritidis 158KD (Fig. 8B). After 7 days post injection, IL-17+CD4+ cells were present in 9% of the lymph node population in mice which received control antibody in contrast to only 1.5% in mice receiving anti-TLR4 antibody.

Since anti-B7-1 elevates IL-17 and anti-TLR4 decreases IL-17 in mice given wild-type M. arthritidis, we questioned which of these two processes are dominant. Mice were pre-treated with either control antibody, anti-B7-1, anti-TLR4 or a mixture of anti-TLR4 and anti-B7-1 prior to receiving M. arthritidis 158KD. The CD4+ T cells from peripheral lymph node cells were removed at intervals through 28 days post injection and cultured in vitro in the plain medium for 24 h and supernatants then assayed for IL-17 content. As shown in Fig. 8C, anti-B7-1 antibody significantly elevated constitutive IL-17 levels to approximately two- to threefold (1900 pg ml−1) as compared with mice receiving control antibody (800 pg ml−1) (P <0.01). Anti-TLR4 antibody exhibited a marked inhibitory effect on IL-17 production (300 pg ml−1, P <0.01) which was lower than mice receiving control antibody. When both antibodies were given anti-B7-1 did not elevate IL-17 in mice that had been treated with anti-TLR4. The suppressive effects of anti-TLR4 lasted at least through 28 days. It thus appears that the upstream events of an initial interaction of TLR4 with MAM is dominant over subsequent administration of anti-B7-1.

Next, we examined the effect of anti-TLR4 antibody on the cytokine/chemokine responses in synovial ankle joint tissues from C3H/HeSnJ mice following administration of M. arthritidis 158KD. Arthritic mice were sacrificed 10 days post injection and total synovial tissues were collected, extracted and measured for mRNA expression of selected cytokines (Fig. 9A) and chemokines (Fig. 9B). Mice receiving anti-TLR4 antibody exhibited decreased levels of IL-17 and IL-23p19 mRNAs from the arthritic joints compared the values seen with control antibody. In contrast, anti-TLR4 significantly elevated levels of the Th1 cytokines, IL-12p35, IL-1β and TNFα mRNAs as compared with control antibody (P <0.05 and P <0.01 respectively). In addition, anti-TLR4 treatment significantly inhibited the chemokine mRNA levels of CXCL1, CXCL2, CXCL5, CXCL12 and CCL2 as compared with mice given control antibody (Fig. 9B). It is generally considered that these chemokines are produced by a variety of inflammatory cells including neutrophils, macrophages and synovial fibroblasts that are crucial for the recruitment of Th17 cells to inflammatory joints.

Figure 9.

Differential effect of anti-TLR4 treatment on the profiles of cytokines and chemokines expressed in the ankle joints of wild-type M. arthritidis-infected mice. Mice were pre-treated with anti-TLR4 or control Ab and followed by M. arthritidis 158KD. Synovial tissues were extracted for mRNA and the mRNA transcripts were measured for the expression of selected cytokines (A) and chemokines (B) by qRT-PCR analysis at varying time points as indicated. All transcripts analysed were normalized to β-actin. Data are pooled from three separate experiments and are expressed as means ± SEM.

Finally, we asked whether IL-17 might play a more direct role in arthritis. Using the protocol for Fig. 1, mice were injected twice with control antibody, anti-B7-1, anti-IL-17 or a combination of anti-B7-1 plus anti-IL-17 prior to an injection of live M. arthritidis. Mice were scored for arthritis for 14 days and thereafter sacrificed for later immunological studies. The results show that anti-B7-1 treatment markedly enhanced arthritis as seen previously (Fig. 1). Treatment of mice with anti-IL-17 alone had no effect on arthritis, but anti-IL-17 antibody suppressed the increased arthritis mediated by anti-B7-1 antibody. Thus, the increased arthritis following an anti-B7-1 treatment was solely due to the blockade of the B7-1(CD80) co-stimulatory molecule.

Discussion

This study has resulted in four major findings in understanding how M. arthritidis and its SAg MAM contribute to the unique immunological and pathological changes that occur in our mouse model of inflammatory disease. First, anti-B7-1 antibody enhances the severity of arthritis in M. arthritidis-infected mice by promoting IL-17 and Th17-associated cytokines and the ICOS/ICOSL co-stimulatory pathway. Second, the MAM SAg, which is secreted by live M. arthritidis, plays a critical role in this process since IL-17 and its transcription factors were suppressed in mice injected with the MAMKO derivative of M. arthritidis. Third, a TLR4 signal is required for development of the IL-17/Th17 cascade since enhancement of the Th17 response could be blocked with anti-TLR4 antibody. Fourth, anti-TLR4 antibody inhibits the IL-17 cascade in the joints of mice with the wild-type M. arthritidis strain and mycoplasmas lacking MAM failed to develop an IL-17 response in joints. Thus, IL-17 indeed contributes to the disease pathogenesis of arthritis. Our results indicate for the first time that, there is a direct link between blocking of B7-1 in elevated arthritis as well as upregulation of IL-17/Th17 cascade and an associated increase of the ICOS/ICOSL co-stimulatory pathway in mice infected with M. arthritidis, all of which are dependent upon the presence of both the MAM SAg and TLR4.

Earlier studies using the MAM SAg had established that cytokine profiles in vivo were influenced by presence (C3H/HeSnJ) or absence (C3H/HeJ) of a functional TLR4 molecule, resulting in a Th2 versus a Th1 profile respectively (Mu et al., 2001). Although both B7-1 and B7-2 were required for full T-cell activation, only B7-1 participated in T-cell differentiation. Additionally, blocking of B7-1 reversed the cytokine profiles in response to MAM (Mu et al., 2006). C3H/HeSnJ (TLR2+/4+) mice exhibited an early IL-17 response, whereas C3H/HeJ mice which lacked a functional TLR4 showed a delayed and lesser production of IL-17 in response to MAM (Mu et al., 2011). The present studies were undertaken to determine how the interactions of MAM, TLR4 and B7-1 might influence the immune outcomes and clinical disease mediated in wild-type C3H/HeSnJ mice to infection with live M. arthritidis.

There is considerable evidence that that the B7-1 co-stimulatory molecule is a key player in the manifestations of infectious and autoimmune disease and that blocking the B7-1 molecule in vivo shows therapeutic potential in treating autoimmune and infectious diseases in animals (Kuchroo et al., 1995; Miller et al., 1995; Racke et al., 1995), although in one study NOD mice treated with anti-B7-1 exhibited a more severe disease that associated with an enhancement of Th1 response (Lenschow et al., 1995).

In contrast to most of these disease models, the first unique observation in the present study was that anti-B7-1 antibody drastically enhanced arthritis in wild-type C3H/HeSnJ mice infected with M. arthritidis whereas, while anti-B7-1 caused a significant decrease in arthritis in TLR4-mutant C3H/HeJ mice. Interestingly, C3H/HeSnJ mice exhibited a very early onset of disease as well as a chronic phase that lasted for the 28-day duration of the study. To further understand the mechanisms involved in the elevated arthritis in anti-B7-1-treated C3H/HeSnJ mice we therefore examined more closely the role of B7-1, TLR4 and MAM in the immunological events taking place in these animals infected with live M. arthritidis.

The ability of anti-B7-1 antibody to enhance arthritis in C3H/HeSnJ mice was found to be associated with a marked prolonged elevation of the IL-17/Th17 cascade with both serum and constitutive levels of IL-17 as well as the upregulation of the mRNA of the Th17-specific transcriptional factor RORγt expression. Serum levels of IL-6 and TGFβ which together are critical for Th17 differentiation, were also elevated in anti-B7-1-treated arthritic mice. It is noteworthy that we showed previously (Mu et al., 2011) that the increased IL-17 by MAM SAg was also associated with elevated IL-6 and IL-23 as IL-6 is require for the initiation of Th17 differentiation and IL-23 is critical for maintaining the Th17 stability and the induction of pathogenicity in Th17 cells. Blocking antibody to IL-6 and IL-23 both drastically reduced IL-17 production (Korn et al., 2009; Mu et al., 2011). Using M. arthritidis, we demonstrated (Fig. 3) that anti-B7-1 also elevated the levels of Th2-type cytokines IL-6 and IL-10 but reduced Th1 cytokine IFNγ presumably reflecting the association of B7-1 with Th1 cytokine promotion (Mu et al., 2006).

In the M. arthritidis model, the percentage of IL-17+CD4+ T cells present in the draining peripheral lymph nodes of arthritic mice receiving anti-B7-1 antibody were elevated as was expression of the signature Th17 cell surface markers CCR6 (Hirota et al., 2007) and the IL-23 receptor. In addition, we made the novel observation that anti-B7-1 markedly increased the expression of inducible co-stimulatory molecule (ICOS) in CD4+ T cells (Fig. 4). This co-stimulatory molecule is also known to be associated with the enhanced IL-17/Th17 response (Park et al., 2005) and has also been shown to potentiate Th2 cytokines IL-4 and IL-10, and can augment Th1 cytokine IFNγ (Simpson et al., 2010). It has been reported that ICOS/ICOSL co-stimulatory signalling plays a key role in the development of autoimmunity in animal models of disease, such as CIA, EAE and myasthenia gravis (Korn et al., 2009). It was also reported by Odobasic et al. that blockade of B7-2 co-stimulatory signalling can suppress IL-17 response in arthritic C57BL/6 mice sensitized to bovine serum albumin (BSA) indicating B7-2 might also play a role in the Th17 differentiation (Odobasic et al., 2008). In the present study, as B7-2 does not control MAM-induced T-cell differentiation (Mu et al., 2006), it is likely that blocking B7-1 would require an alternative pathway other than B7-2. It is possible that CTLA-4 or PD-1/PD-L1/2 might also be involved, and, in fact, a preliminary experiment by us showed that anti-ICOSL inhibited arthritis by 50% and partially suppressed the IL-17 cascade (data not shown). Additional work is needed to clarify the role played by each of these pathways.

Previous in vivo studies on M. arthritidis and MAM SAg have focused on the resulting systemic immunological changes. Here, we demonstrate that at the site of inflammation, the arthritic joint, the same changes are seen in regards to the development of the IL-17/Th17 cascade. Treatment of mice with anti-B7-1 resulted in elevated levels of ankle joint tissue mRNAs for IL-17, IL-6, CXCL2, a chemokine known to be a recruiter of autoreactive T cells (Ha et al., 2010) and HMGB1, a member of the heat-shock protein family (HSPs) which is also known to amplify TLR4-mediated inflammation (Hreggvidsdottir et al., 2009) and can be associated with autoimmune disease (Andersson and Harris, 2010).

Our present work also establishes that MAM plays a definitive role in upregulation of the IL-17/Th17 cascade in C3H/HeSnJ mice given M. arthritidis after the pre-treatment of anti-B7-1 antibody. Thus, experiments using the wild-type 158KD strain versus its MAMKO derivative showed that arthritis onset was markedly delayed in the MAMKO strain versus the wild-type 158KD M. arthritidis. Particularly striking was that levels of IL-17 and another Th17-associated transcriptional factor, STAT3, in the lymph nodes of arthritic mice given wild-type M. arthritidis were virtually absent in animals given organisms lacking MAM (Fig. 6). Furthermore, the numbers of infiltrating CD3+ T cells, neutrophils and macrophages and the corresponding mRNA levels of IL-17, CXCL2 and HMGB1, were all significantly higher in mouse joint tissues receiving wild-type 158KD versus the MAMKO M. arthritidis (Fig. 7).

We have now confirmed that TLR4 is instrumental in initiation of the IL-17/Th17 cascade since antibody to TLR4 given to C3H/HeSnJ mice prior to infection with wild-type M. arthritidis reduced serum levels of IL-17 to just 25% of those seen in mice given control antibody and levels were no higher than those in sera from mice injected with the MAMKO strain (Fig. 8). Also, there was a parallel decrease of IL-17+CD4+ T cells (Th17) in draining peripheral lymph nodes in mice given anti-TLR4 antibody and constitutive IL-17 levels in response to mycoplasma infection were virtually eliminated by anti-TLR4. Interestingly, anti-TLR4 antibody abolished the rise in IL-17 induced by anti-B7-1 suggesting that the innate immune response to pathogen agonist, MAM is dominant over the subsequent adaptive immune response. Anti-TLR4 antibody had a similar effect on cytokine profiles present in ankle joint tissues in that mRNA levels of IL-17, IL-23p19, IL-6 and the chemokines CXCL1, 2, 5 and 12 as well as CCL2 were all decreased by anti-TLR4 treatment (Fig. 9). These chemokines are responsible for the influx of ‘inflammatory cells’ to the joint and appear to be associated with TLR4. Thus, the IL-17/Th17 response is suppressed not only in the absence of MAM, but also by blocking TLR4, asserting that both MAM and TLR4 are required. Another key question is whether IL-17 has a ‘pathogenic’ role in the development of arthritis, especially the elevated disease mediated by the administration of anti-B7-1 antibody. The results presented here clearly establish that antibody to IL-17 blocks the elevated arthritis due to injection of anti-B7-1 antibody (Fig. 10).

Figure 10.

Role of IL-17/Th17 in arthritis severity. C3H/HeSnJ (TLR2+/4+) mice were pre-injected with control antibody, anti-B7-1, anti-IL-17 or a combination of anti-B7-1 and anti-IL-17 prior to the administration of live M. arthritidis. The animals were scored for arthritis through 14 days. The determination of arthritis were similar as described in Fig. 1. The data are pooled from two separate experiments and arthritis scores for different antibody treatments ± SEM are shown.

Although the ability of anti-B7-1 to increase arthritis and the IL-17/Th17 cascade in C3H/HeSnJ mice appears solely due to the interaction of MAM with TLR4, some arthritis and lower levels of the IL-17 response were still present in mice receiving control antibody or normal saline when infected with M. arthritidis strain 158KD. Arthritis was still present in mice receiving the MAMKO strain of M. arthritidis although disease onset was remarkably delayed. Blocking of TLR4 did not consistently reduce arthritis in wild-type M. arthritidis-injected mice (data not shown) although all of the Th17-associated cytokines were markedly reduced. Clearly other signalling pathways and components of live mycoplasmas are playing a role in immune responses and disease outcomes in mice infected with M. arthritidis. In this context, we note that M. arthritidis possesses a number of pathogenic components such as adhesins (MAA1 and MAA2) (Washburn et al., 1998; 2000) which aid the organisms to attach to host cells and a bacteriophage that has variously been associated with organism virulence (Voelker et al., 1995; Clapper et al., 2004). In addition, a family of lipoproteins have been identified by us which possess TLR2-dependent macrophage-activating properties and which induce a pro-inflammatory cytokine response after in vivo administration (Hasebe et al., 2006). It is this latter finding that likely plays a role in the overall immune response of the host to M. arthritidis. Clearly there are at least three distinct pathways which might contribute to arthritis development; first, the IL-17/Th17 cascade that is TLR4 and MAM-dependent; second, an inflammatory Th1 cascade mediated by a MAM/TLR2 interaction in the absence of TLR4 (Mu et al., 2001; 2005), and third, the mycoplasma lipoprotein/TLR2 interaction which also promotes a Th1 response (Hasebe et al., 2006). The second and third pathways are apparently responsible for the presence of inflammatory cytokines IL-12, TNFα and IL-1β that are seen in ankle joints in infected-mice following pre-injection of anti-TLR4 antibody (Fig. 9).

An important issue that relates to the work presented here is that TLR4-defect C3H/HeJ mice are exceedingly susceptible to a lethal toxic shock syndrome by M. arthritidis (Mu et al., 2001). These mice produce a strong Th1 cytokine response to MAM as well as to M. arthritidis and can develop a very early onset of arthritis as shown in Fig. 1 and has previously been documented (Mu et al., 2001). We show here that whereas anti-B7-1 elevates arthritis in TLR4+ C3H/HeSnJ mice, it suppresses arthritis in TLR4-defective C3H/HeJ mice and alleviates the toxic shock syndrome in C3H/HeJ mice (Mu et al., 2006). Thus, the rapid onset of lethal toxic shock in C3H/HeJ mice given M. arthritidis is likely due to an intense Th1 cytokine release by the combined interaction of MAM and lipoproteins with TLR2, which when TLR4 is present suppresses this response leading to a Th2/Th17 less toxic cytokine milieu. Our previous work has already documented that MAM/TLR4 interactions can suppress the Th1 response induced by the engagement of MAM with TLR2 (Mu et al., 2005).

Do our observations relate to the development of autoimmune disease? The IL-17/Th17 cascade has increasingly been shown to associate with human autoimmune disease such as rheumatoid arthritis, multiple sclerosis, SLE and etc. (Korn et al., 2009). Experimental models such as CIA, EAE, murine type 1 diabetes in NOD mice suggest that the IL-17/Th17 response may, in fact, play an instrumental role in the inflammatory process. Although it remains to be established whether M. arthritidis infections are associated with inflammatory/autoimmune disorders, it is known that the MAM SAg can trigger or cause flares in autoimmune CIA in mice. In fact, recent studies have indicated that the elevated arthritis seen in CIA is due to a MAM-dependent upregulation of the IL-17 cascade (H.H. Mu, M.M. Nourian and B.C. Cole, in preparation).

In conclusion, we have shown that the MAM SAg plays an instrumental role in the immune response and inflammation induced by M. arthritidis. MAM links the innate immune system with the adaptive immune responses via TLR 4. Blocking B7-1 promotes an alternative pathway of IL-17/Th17 differentiation which directly contributes to the arthritic response. Another contributing factor to this scenario is that MAM, as a SAg, can rapidly bind to the MHC class II molecules present on the APCs (Cole and Atkin, 1991). In fact, we hypothesized the interaction of MAM with TLRs is also present on APCs (Mu et al., 2005). Our findings have provided new insights into how the products of microbial agents can modify host immune and pathological responses to promote the development of inflammatory and autoimmune diseases. Identifying the key signal transduction molecules responsible for these interactions is likely to provide key targets for clinical intervention which might also have relevance for human inflammatory and autoimmune diseases.

Experimental procedures

Mice

Female C3H/HeJ (TLR2+/TLR4−, H-2Eα+, tlr4Def) and C3H/HeSnJ (TLR2+/TLR4+, H-2Eα+, tlr4Pos) mice were purchased from Jackson Laboratory (Bar Harbor, ME). All mouse strains used in this study are H-2Eα+ and highly MAM-reactive. Mice were maintained in specific pathogen-free conditions at the Comparative Medicine Center (CMC) at the University of Utah Health Science Center and were used at 12–16 weeks of age. The CMC guarantees strict compliance with regulations established by the Animal Welfare Act.

Induction of arthritis by M. arthritidis

Wild-type M. arthritidis 158p10p9 and 158KD strains were used in the study respectively. The 158p10p9 strain was described previously and maintained in our laboratory (Cole et al., 2000; Mu et al., 2000); the 158KD strain was the gift from Dr Kevin Dybvig at the University of Alabama at Birmingham. A strain of M. arthritidis with MAM gene functionally inactivated (MAMKO) developed by Dybvig et al. as described was also used in this study to determine the role of MAM in the pathogenesis of arthritis (Luo et al., 2008). Mycoplasmas were harvested by centrifugation at 27 000 g, washed once in serum-free medium, suspended in normal saline (NS). Mice were injected i.v. in groups of five to eight mice with 5 × 108 cfu of M. arthritidis. Mice were examined and scored for arthritis severity at day 1, 2, 3, 5, 7, 10, 14, 21 and 28 after injection as described previously (Mu et al., 2001).

Blocking antibodies used in vivo studies

Neutralizing anti-mouse antibodies against B7-1 (clone 3A12), TLR4 (clone MTS510) and functional grade purified isotype-matched control antibodies were purchased from eBioscience, San Diego, CA. Anti-mouse IL-17 blocking antibody was obtained from R&D System (Minneapolis, MN, Cat# MAB421).

Animal procedures and in vivo mAb treatment

For in vivo priming studies, groups of five to eight mice were injected i.v. with 5 × 108 cfu live M. arthritidis at day 0. Control mice were given normal saline (NS) injections. At different time points after injections, mice were bled by cardiac puncture, and the sera were collected for analysis of cytokines. Spleen and draining peripheral lymph nodes (PLNs) were removed and dissected.

For antibody blocking studies, prior to M. arthritidis injection, groups of five to eight mice were injected i.p. twice with either anti-B7-1, anti-IL-17, anti-TLR4, or a combination of anti-B7-1 and anti-IL-17 (Fig. 10). Ig isotype-matched control antibody(ies) were injected at day −2 and day 0. Each mouse received 100 μg of each antibody at each injection.

Preparation of single lymphoid cells and purified CD4+ T-cell subsets

Single cell suspensions of spleens and PLNs were prepared by pressing lymphoid tissues through a nylon mesh filter. Erythrocytes were lysed in ACK lysing buffer and cells washed and resuspended in phosphate-buffered saline (PBS) with 2% fetal bovine serum.

For the preparation of CD4+ T lymphocytes, CD4+ T cells were isolated from splenic or PLNs using the MagCellect Mouse CD4+ T Cell Isolation Kit (R&D Systems, Minneapolis, MN). The resulting populations were evaluated by flow cytometry and were greater than 93% of each cell population.

Isolation of cells from ankle joint tissue

Cells from ankle joint tissue were prepared as described by Lochhead (Lochhead et al., 2012). Briefly, single-cell suspensions were prepared from the rear ankle joint tissue, following removal of skin. The joint tissues were partially removed from bone using 20-gauge syringe needles to facilitate digestion by incubation in RPMI 1640 containing 0.2 mg ml−1 endotoxin-free Liberase™ (Roche) and 100 μg ml−1 DNase I (Sigma-Aldrich) for 1 h at 37°C. After incubation, gentle pipetting further disrupted tissue and intact tissue was broken apart using the end of a 5 ml syringe. The single-cell suspension was filtered through a 100 μm cell strainer and centrifuged.

Quantitative cytokine analysis by capture ELISA

Single cell suspensions were cultured at 107 cells ml−1 in freshly prepared serum-free medium consisting of RPMI 1640, 1% Nutridoma-NS (Boehringer Mannheim, Indianapolis, IN), 200 mM l-glutamine, antibiotics and 5 × 10−5 M 2-mercaptoethanol. M. arthritidis-injected mice were sacrificed at various time points after injection. Single cell suspensions were prepared and cultured without stimulation for constitutive production of cytokine responses (Mu et al., 2001; 2005). All medium used in their preparations and for cytokine release contained 50 μg ml−1 gentamicin to inhibit mycoplasma growth. Supernatants were harvested, and assayed for specific cytokine content by Ready-SET-Go™ mouse cytokine ELISA sets (eBioscience, San Diego, CA) according to the manufacturer's instructions.

ELISA for measurement of STAT3 activation

The detection of phosphorylated STAT3 activity was performed using the STAT3 [pY705] phosphoELISA™ kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions (Mu et al., 2011). STAT 3 was detected in the protein lysate samples.

RNA and quantitative real-time PCR

Total RNA was isolated from 1 × 107 cells using RNeasy Mini Kits (QIAGEN, Valencia, CA). RNA was precipitated in isopropanol and washed in 75% ethanol at room temperature. Following DNase treatment to eliminate contaminating genomic DNA (Turbo DNase, Ambion), total RNA was further purified using the RT2 qPCR-Grade RNA Isolation kit (QIANGEN-SABiosciences, Frederick, MD) and quantified using a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington DE). For PCR array analysis 0.5 μg of total RNA was used to reverse transcribe to cDNA using the RT2 First Strand kit as per manufacturer's instructions (SABiosciences). The PCR was performed on an ABI 7900 (Applied Biosystems) as described previously (Osborne-Hereford et al., 2008; Mu et al., 2011). For each 96-well plate, 25 μl of master-mix containing cDNA and a mouse specific-signalling pathway RT2 Profiler PCR plate was prepared. The studies employed four sets of PCR arrays, namely ‘Toll-Like Receptor Signalling Pathway’, ‘Th17 for autoimmunity & inflammation array’, ‘Th1-Th2-Th3 array’ and ‘Inflammatory response and autoimmunity’ from QIANGEN-SABiosciences. Amplification was performed in accordance with the manufacturer's guidelines. The amount of mRNA was measured as fold change relative to the amount in control untreated mice. To determine fold change, each sample was normalized against the housekeeping gene β-actin by obtaining the difference in threshold cycle number between β-actin and cytokines or RORγt. This number was then subtracted from the mean normalized value of control untreated mice to generate the uncorrected difference. The fold change of each sample was then determined by raising the amplification efficiency (generated by the PCR machine) to the power of the uncorrected difference.

Flow cytometric analysis

Monoclonal antibodies against the following Ags were used for immunofluorescence analysis: CD4 (clone RM4-5), CCR6 (CD196, clone 29-2L17), IL-23 receptor (IL-23R, clone 753317), ICOS (clone 7E.17G9), CD11b (clone M1/70) and Ly6G/Gr-1 (clone RB6-8C5). All FITC-, PE- or PerCP-conjugated mAbs and control Ig were obtained from eBiosciences (San Diego, CA), R&D Systems (Minneapolis, MN) and BioLegend (San Diego, CA). Cells (1 × 106 cells in 100 μl) were stained with the fluorescence-conjugated antibodies or control mAbs on ice for 30 min, then washed thoroughly with PBS to remove unbound antibody. Cells were then centrifuged, washed and resuspended in 0.2 ml of PBS, for storage of cells, 1% paraformaldehyde in PBS was added to test vials. For each sample, 20 000 cells were analysed on a FACScan (Becton Dickinson, Mountain View, CA). Cells were gated according to size and scatter to eliminate dead cells and debris. Immunofluorescence, flow cytometry and data analysis were performed by using CELLQuest™ software (Becton Dickinson). The percentage of cell surface expressions on cells was calculated.

Statistical analysis

The method for statistics (the two-tailed Student's t-test) using Statview 5.0.1 (SAS Institute, Cary, NC, http://www.statview.com). anova was used to determine the significance between mouse strain susceptibility and severity to M. arthritidis-induced arthritis. Significant differences among groups were determined by implementing a variant of the F test that is sensitive to differences among groups with large variances. A value of P <0.05 was considered statistically significant (*P <0.05, **P <0.01).

Acknowledgements

This research was supported by a grant from the Nora Eccles Treadwell Foundation (to H.-H. Mu). We thank Kimberly Wright and Neil Xia for excellent technical assistance. We also thank Dr Kevin Dybvig (University of Alabama at Birmingham, Birmingham, AL) for the generous gift of wild-type M. arthritidis 158KD and MAM knockout mutant, MAMKO. All experiments using real-time qRT-PCR were performed in the DNA Core Facility at the University of Utah Health Science Center.

Ancillary