T Cell and Dendritic Cell Abnormalities Synergize to Expand Pro-Inflammatory T Cell Subsets Leading to Fatal Autoimmunity in B6.NZBc1 Lupus-Prone Mice

We have previously shown that B6 congenic mice with a New Zealand Black chromosome 1 (c1) 96-100 cM interval produce anti-nuclear Abs and that at least two additional genetic loci are required to convert this subclinical disease to fatal glomerulonephritis in mice with a c1 70-100 cM interval (c1(70-100)). Here we show that the number of T follicular helper and IL-21-, IFN-γ-, and IL-17-secreting CD4+ T cells parallels disease severity and the number of susceptibility loci in these mice. Immunization of pre-autoimmune mice with OVA recapitulated these differences. Differentiation of naïve T cells in-vitro under polarizing conditions and in-vivo following adoptive transfer of OVA-specific TCR transgenic cells into c1(70-100) or B6 recipient mice, revealed T cell functional defects leading to increased differentiation of IFN-γ- and IL-17-producing cells in the 96-100 cM and 88-96 cM intervals, respectively. However, in-vivo enhanced differentiation of pro-inflammatory T cell subsets was predominantly restricted to c1(70-100) recipient mice, which demonstrated altered dendritic cell function, with increased production of IL-6 and IL-12. The data provide support for the role of pro-inflammatory T cells in the conversion of subclinical disease to fatal autoimmunity and highlight the importance of synergistic interactions between individual susceptibility loci in this process.


Introduction
Systemic Lupus Erythematosus (SLE) is a generalized autoimmune disease characterized by the production of autoantibodies, particularly those directed against nuclear antigens, which form immune complexes that deposit in tissues. Studies of SLE in humans and lupus-prone mice indicate that multiple genetic polymorphisms affecting diverse immune populations interact with each other to produce the lupus phenotype. Among these populations are T helper (Th) cells. Although early studies demonstrated a predominant role for Th1 cells in lupus, several recent studies suggest that two other pro-inflammatory Th cell subsets, T follicular helper (Tfh) and Th17 cells, are also pathogenic [1].
Tfh cells are a distinct subset of Th cells that provide help for antigen specific B cell responses in the context of germinal centers (GC) and produce high levels of IL-21 [2,3]. A potential role for this population in the pathogenesis of lupus was first suggested by the observation that lupus-prone mice with a homozygous point mutation in the Roquin gene, demonstrated expansion of their Tfh population, and subsequently supported by demonstration of similar expansions in MRL lpr and BXSB/ Yaa lupus-prone mice [4].
Although Th17 cells are defined by their IL-17 production, they produce a variety of other cytokines including IL-21, IL-22, TNF-α, IL-6 and IL-9 [5]. Expansion of this population has been demonstrated in several lupus-prone mouse strains, including (New Zealand Black (NZB) x SWR) F 1 , TNF receptor 1 and 2 gene-deleted New Zealand Mixed 2328, and BXD2 mice [6,7,8]. Notably, introduction of a null gene for the IL-17A receptor onto the BXD2 background significantly attenuated production of IgG autoantibodies and nephritis [8]. Despite compelling evidence that Tfh and Th17 cells play a central role in lupus pathogenesis, the genetic basis leading to the aberrant activation of these cell populations remains unknown.
To characterize the immunologic abnormalities that promote lupus, our laboratory has produced a series of congenic mouse strains with homozygous NZB chromosomal intervals crossed onto the non-autoimmune C57BL/6 (B6) background. In previous experiments we showed that mice with a NZB c1 interval extending from 70-100 cM (c1(70-100)) develop a severe lupus phenotype, with high titers of anti-dsDNA Abs and glomerulonephritis (GN), leading to death of ~40% of the mice by 8 months of age. This phenotype appeared to result from at least 3 genetic loci, as indicated by progressively attenuated disease in mice with NZB c1 intervals extending from 88-or 96-100 cM [9]. Here we show that the disease severity in these mice parallels the expansion of pro-inflammatory T cell subsets, specifically Th1, Th17, and Tfh cells. We further demonstrate that this expansion can be recapitulated following immunization of pre-autoimmune mice with an exogenous antigen. This T cell skewing results from a combination of immune cell functional abnormalities in congenic mice that localize to different regions within the c1 70-100 interval. Naïve T cell functional abnormalities that lead to expansion of IFN-γand IL-17-producing cells localized to the 96-100 and 88-96 intervals, respectively, whereas dendritic cell (DC) functional abnormalities that promote expansion of all the proinflammatory T cell subsets localized to the 88-96 and 70-88 intervals. Notably, altered DC function appeared to play a critical role in this expansion, because in the absence of DC abnormalities, minimal expansion of pro-inflammatory T cell subsets was seen. Our findings provide insight into how individual susceptibility loci, which alone produce modest changes in immune function, interact synergistically to profoundly alter immune function leading to severe clinically relevant autoimmune disease.

Results
Expansion of pro-inflammatory CD4 + T cell subsets in NZB c1 congenic mice B6 congenic mice with NZB c1 intervals extending from 96-100 cM (172. .0 Mb; c1(96-100)), 88-100 cM (170.3-183.0 Mb; c1(88-100)) or 70-100 cM (126.6-183.0 Mb; c1(70-100)) demonstrate progressively more severe disease with increasing length of the c1 interval ( Figure 1). Since increases in the number and size of GC paralleled disease severity in these mice, we postulated that changes in Th cell number/function were producing these differences. To address this possibility, Th cell subsets were examined in 4-mo-old B6 and congenic mice, using flow cytometry. As shown in Figure  2A& B, the proportion and number of Tfh cells (gated as CD4 + CD44 hi CD62L lo CXCR5 hi PD1 hi ) was significantly increased in c1(88-100) and c1(70-100) mice, whereas the level of these cells in c1(96-100) mice was similar to B6 mice. Consistent with the increases in Tfh, and our previous findings, there was a trend to increased proportions of GC B cells in all three congenic mouse strains with the greatest increase seen in c1(70-100) mice (Figure S1A& B). The expansion of Tfh was further confirmed by immunofluorescence microscopy where increased numbers of Tfh cells were seen in the GC of c1(88-100) and c1(70-100) mice (Figure S1C& D). Dashed lines indicate regions of undefined origin. Polymorphic microsatellite markers and single nucleotide polymorphism (SNP) markers were used to discriminate between NZB and B6 DNA at the termini of the regions according to the NCBI 2007 (m37 release) mouse genome assembly (www.ensembl.org). Potential candidate genes within the interval are indicated above the chromosomal map. Phenotypic features of NZB c1 congenic mouse strains are shown to the right of the c1 congenic mice genetic map [9]. To examine the other Th subsets, cytokine-producing CD4 + T cells were quantified by flow cytometry following stimulation of freshly isolated splenocytes for 4 hrs with PMA and ionomycin, and intracellular staining for IL-4, IFN-γ, and IL-17 ( Figure 2C). There were no differences between strains in the proportion of IL-4 producing cells, but a trend to a progressive increase in IFN-γ and IL-17 producing T cells with increasing size of the NZB c1 interval was seen ( Figure 2D). Similar findings were obtained when CD4 + T cells were stimulated in-vitro with anti-CD3 and -CD28 Abs, and secretion of various cytokines was quantified in the supernatants ( Figure S2). While there was no significant difference between the mouse strains in the production of IL-2 and IL-4, there was a progressive increase in the secretion of IFN-γ, IL-17, and IL-21 that correlated with increasing length of the c1 interval.
To further define the CD4 + T cell populations secreting these cytokines, intra-cellular cytokine levels were examined in cells stained with anti-CD3, -CD4, -CXCR5, and -PD1 to permit discrimination between Tfh (CD4 + CXCR5 hi PD1 hi ) and conventional CD4 + T cells (including Th17 and extrafollicular T cells). This revealed that the increase in IL-21 and IFN-γ secreting cells observed in c1 mice results from increases in the numbers of both Tfh and conventional CD4 + T cells that secrete these cytokines (Figure S3A& B), which positively correlated with the number of NZB genetic loci. A significant proportion of the IL-21 secreting cells also secreted IFN-γ (20-40% Tfh, 40-60% conventional), and conversely IFN-γ secreting cells also secreted IL-21 (40-60% Tfh, 30-50% conventional), with the proportion of co-secretors paralleling the length of NZB interval in congenic mice ( Figure S3A and data not shown).
The majority of IL-17 secreting cells were seen in the conventional CD4 + T cell population ( Figure S3B), with ~5% of the cells also secreting IFN-γ and 80% of cells secreting IL-21 in all mouse strains examined ( Figure S3A and data not shown). Consistent with the flow cytometry findings, the majority of IL-17 secreting cells were seen within the T cell zone and the number of these cells was increased in c1(88-100) and c1(70-100) mice (Figure S3C& D).

Intrinsic skewing of the immune system towards increased generation of Tfh, Th17 and Th1 cell subsets in c1 congenic mice
To determine whether the increased production of IL-21, IL-17, and IFN-γ in c1 congenic mice was a consequence of the breakdown in tolerance to nuclear antigens, or resulted from intrinsically altered immune function leading to skewed Tfh, Th17 and Th1 development, we investigated the immune response to OVA as a representative exogenous antigen. Young pre-autoimmune 8-wk-old B6 and c1 congenic mice were immunized i.p. with OVA emulsified in CFA, using PBS emulsified in CFA as a control. The mice were sacrificed 14 days later and the proportions of various T cell subsets and GC B cells were examined. Consistent with our previous results ( Figure 2 and Figure S1), there was a progressive increase in the proportion of Tfh and GC B cells corresponding to increasing size of the NZB c1 interval, following OVA-CFA immunization ( Figure 3A). No significant differences were observed with PBS-CFA immunization. To assess the cytokine profile of the OVA-specific T cells, splenocytes isolated from OVA-primed mice were re-stimulated in-vitro with OVA for 72 h. Cytokine levels were measured in tissue culture supernatants and the amount of OVA-specific cytokine production was determined by subtracting cytokine production in the absence of OVA. As seen in 4-mo-old unimmunized mice, there were progressive increases in IFN-γ, IL-17, and IL-21 production with increasing length of the NZB c1 interval ( Figure 3B). Thus, the immune system in c1 congenic mice appears to be intrinsically skewed toward increased production of Th1, Th17, and Tfh cytokines, regardless of the specificity of the antigen.

Altered T cell differentiation in c1 congenic mice results from defects affecting T and non-T cell function
To determine the immune defects that lead to the increased differentiation of CD4 + T cells into Th1, Th17 and Tfh cells in c1 congenic mice, several approaches were used. In the first approach, naïve T cells from the spleens of 8-wk-old preautoimmune mice were isolated and induced to differentiate into various T cell subsets using cocktails of cytokines and mAbs (see Materials and Methods). Under Th0 conditions there was minimal differentiation of either B6 or c1 congenic T cells into IL-21-(< 0.21%), IL-17-(<0.12%), IFN-γ-(<0.82%), or IL-4-(<0.41%) secreting cells with similar levels seen for all mouse strains ( Figure 4A and data not shown). In contrast, under Th1-inducing conditions, all c1 congenic mice demonstrated increased differentiation to IFN-γ-secreting cells compared to B6 mice ( Figure 4B), suggesting that a genetic locus in the NZB c1 96-100 cM interval promotes differentiation of this cell subset. Using Th17-inducing conditions, both c1(88-100) and c1(70-100) naïve T cells demonstrated equivalently increased differentiation to IL-17-producing cells compared to B6 or c1(96-100) T cells. Thus, a genetic locus located within the NZB c1 88-96 interval alters T cell function to promote IL-17 secretion. In contrast, similar proportions of Th2 and IL-21-producing cells were seen for all mouse strains tested under their respective cytokine inducing conditions, suggesting that the increased proportions of IL-21 producing cells seen in-vivo in c1 congenic mice do not arise from a T cell functional defect.
The second approach used to examine the altered T cell differentiation in c1 congenic mice was adoptive transfer of B6 or congenic T cells into B6 or congenic recipients in a reciprocal fashion. To facilitate these investigations, an OT-II TCR transgene (Tg) with specificity for OVA/A b , was crossed onto the various mouse backgrounds. Naïve CD4 + T cells were then purified from the spleens of young 8-wk-old OT-II TCR Tg B6 and c1 congenic mice, and injected into the tail vein of 8wk-old B6.Thy1.1 or c1(70-100). Thy1.1 mice. Mice were then immunized with OVA and the differentiation of naïve OT-II T cells into various Th cell subsets determined by flow cytometry, gating on the transferred Thy1.2 + population ( Figure 5A). These results confirmed the in-vitro Th differentiation results, showing that the enhanced IFN-γ-and IL-17-, but not IL-21-, secreting cell differentiation arises in part from intrinsic T cell defects localizing to the NZB c1 96-100 and 88-96 intervals, respectively (Figure 5B& C). However, there was also an important role for the environment in the increased differentiation that was observed, because OT-II T cells from all of the mouse strains demonstrated enhanced differentiation to Tfh, Th1, Th17, and IL-21-secreting populations when transferred into c1(70-100). Thy1.1 mice. Indeed, only minimal non-significant increases in the proportion of IFN-γ-and IL-17secreting cells for the relevant c1 congenic T cells were seen upon adoptive transfer into B6 mice. This finding suggests that the increased differentiation of these T cell subsets in c1 congenic mice is critically dependent upon cellular and/or cytokine cues that are not provided by the B6 environment.

DC from c1(88-100) and c1(70-100) mice demonstrate altered function that promotes differentiation of proinflammatory T cell subsets
Cues from DC play an important role in directing the differentiation of T cells following Ag challenge. We therefore contrasted the ability of DC from the various strains of mice to direct the differentiation of OT-II T cells when cultured with low concentrations of OVA in-vitro. To this end, bone marrow was isolated from 8-12-wk-old B6 and c1 congenic mice and cultured with FLT3L for 7 days to expand DC. This yielded bone marrow DC (BMDC) that were ~25% plasmacytoid DC (pDC) and ~30% myeloid DC (mDC) with the remaining cells having an indeterminate phenotype. Similar proportions and numbers of DC were seen for all strains (data not shown). The BMDC were then co-cultured with OVA 323-339 peptide and OT-II T cells from B6 or c1 congenic mice for 4 days in the presence of GM-CSF. As shown in Figure 6A, BMDC from c1(70-100) mice demonstrated a significantly enhanced ability to induce differentiation of Th1 cells compared to those from B6 and c1(96-100) mice, and similar non-significant trends were seen for c1(88-100) mice and for Th17 and Tfh cell differentiation. For Th1 cells this increased induction was only seen for OT-II cells from the congenic mouse strains, indicating that T cell and DC defects must interact with each other to induce this phenotype. Similar findings were observed for Th17 cells, where differences between induction of Th17 differentiation by B6 and c1(88-100) or c1(70-100) DC were most pronounced for c1(88-100) and c1(70-100) T cells. Thus, BMDC from congenic mice appear to be able to direct differentiation of T cells in a way that is compatible with the altered differentiation that is observed in-vivo. Experiments using BMDC expanded with GM-CSF (data not shown) or whole splenocytes ( Figure S4A) as antigen-presenting cells yielded very similar results for comparison of B6 and c1(70-100) cells. To further explore the mechanisms by which BMDC from c1(88-100) and c1(70-100) mice promote differentiation of proinflammatory T cell subsets, mDC activation and cytokine production was examined in the co-culture system. Consistent with their ability to enhance differentiation of Th1 and to a lesser extent Th17 and Tfh cells, BMDC from c1(88-100) and c1(70-100) mice secreted elevated levels of IL-12 and IL-6 which achieved statistical significance for c1(70-100) mice ( Figure 6B). Similar findings were seen for c1(70-100) splenic mDC when whole splenoctyes were used as antigenpresenting cells ( Figure S4B). A trend to elevated levels of MHC class II and B7.2 were also seen on c1(88-100) mDC, which were further increased on c1(70-100) mDC ( Figure 6B). Notably, these changes were independent of the strain of T cells with which the DC were co-cultured (data not shown). In contrast to the data observed for mDC, no differences in cytokine secretion or activation were seen between strains for pDC in the culture ( Figure 6C).
In lupus, the immune response is focused on nuclear antigens contained in apoptotic debris. We have previously shown that in NZB c1 congenic mice there is a breach of tolerance to these antigens, resulting in spontaneous priming of histone-reactive T cells [10]. This observation suggests that the DC in these mice may have processed and presented nuclear antigens. Since these nuclear antigens can activate TLRs, enhancing DC activation and presentation, we investigated whether the BMDC abnormality in c1 congenic mice leads to altered TLR responses. Consistent with the results of our coculture experiments, mDC from c1(88-100) and c1(70-100) mice demonstrated significantly increased intracellular levels of IL-12 and a trend to increased intracellular levels of IL-6 in response to CpG stimulation (Figure 7 A& B). Increased intracellular levels of IL-6 were also observed for c1(70-100) derived mDC following stimulation with Poly(I:C). No differences were seen for the secretion of IFN-α, IL-23 or TNFα ( Figure 7C& D), nor were differences seen for MHC-II or B7.2 expression following TLR stimulation (data not shown). These  findings indicate that the altered DC function in c1(88-100) and c1(70-100) mice also affects the response to certain TLR signals.

Discussion
In this paper, we show that differences in the severity of renal disease that we have previously published for a series of NZB c1 congenic mouse strains correlate with the expansion of pro-inflammatory T cell subsets including, Th1, Th17, and Tfh cells. These findings are compatible with previous work suggesting that these cells populations drive pathogenic autoantibody production and/or inflammatory changes in the kidney [1,2,8,11].
Mice with the shortest NZB c1 interval, c1(96-100), showed no evidence of Tfh cell expansion. Consistent with the lack of Tfh cell expansion, the major increase in cytokine production in these mice appeared to arise from the conventional T cell subset, where slight increases in the number of IFN-γ, IL-17, and IL-21 secreting cells were seen. In contrast, c1(88-100)  mice demonstrated significant increases in Tfh and IFN-γ-, IL-17-, and IL-21-producing T cells. While our experiments do not allow us to definitively conclude which expanded cell populations are driving the increased disease severity in these mice, we have previously shown that IgG2a Ab and complement are deposited in their kidneys [9], implicating IFNγ-producing T cells in this process. However, it is likely that Tfh also play a role, since we have shown that CD40L is necessary for production of GC and nephritis in NZB and c1 mice ( [12] and unpublished observations). Notably, the Tfh cells in c1(88-100) mice do not produce significant amounts of IL-17. This finding contrasts with those observed in BXD2 lupusprone mice, where substantial numbers of IL-17-producing Tfh cells were seen and introduction of an IL-17R knockout attenuated disease [8].
Although c1(70-100) mice showed trends to increased numbers and/or proportions of IFN-γ-, IL-17-, and IL-21secreting cells compared to c1(88-100) mice, the most marked differences were for IL-21-secreting cells, particularly those that also secreted IFN-γ (data not shown). Since the severity of renal disease in our mice is closely associated with IgG deposition in the kidney [9], it is likely that these changes augment kidney disease through enhanced selection of pathogenic IgG in the GC. Nevertheless, we cannot exclude a possible role for the IFN-γ or IL-17-producing cells in either providing extra-follicular T cell help or directly impacting inflammation in the kidney in these mice. In keeping with the latter possibility, both IFN-γ-and IL-17-secreting cells have been found infiltrating the kidneys of c1(70-100) mice (unpublished observations).
Experiments in young pre-autoimmune mice immunized with a representative foreign antigen recapitulated the same types of pro-inflammatory cell expansions as seen in older mice. Using various approaches it was demonstrated that alterations in both T cell and DC function contribute to the changes observed. Enhanced differentiation of naïve T cells to IFN-γproducing cells was localized to the NZB c1 96-100 interval. Although the genetic polymorphism leading to this altered differentiation has not been definitively identified, it is likely that it arises from polymorphisms in the Slam family. The NZB c1 96-100 interval overlaps with the region containing the Sle1b lupus susceptibility allele identified in NZM2410 mice. NZB, NZM2410, and a variety of other autoimmune mouse strains share the same Slam allele which differs from that of B6 mice at genetic loci for multiple Slam family members. The top candidate gene in this interval is Ly108, which encodes a selfligating membrane glycoprotein that has at least three alternatively spliced isoforms differing in their cytoplasmic domains [13,14,15]. In autoimmune mouse strains, expression of Ly108-1 is increased, whereas Ly108-2 expression is decreased and Ly108-H1 is absent [14,16]. Since stimulation with anti-Ly108 antibody induces T cell IFN-γ secretion, it is possible that increased expression of the Ly108-1 variant in c1(96-100) T cells leads to enhanced differentiation to IFN-γproducing cells [17]. Alternatively, the absence of Ly108-H1 expression could produce this phenotype, since introduction of a BAC Ly108-H1 transgene onto mice lacking this isoform was associated with reduced numbers of IFN-γ-producing CD4 + T cells in-vivo [16]. Recently, B6.Sle1b mice were found to have an expansion of Tfh cells that was first detectable at 6-8 months of age (a time point 2 months later than the mice examined in our study) and that could be corrected with the Ly108-H1 transgene [18]. Since c1(96-100) also lack the Ly108-H1 isoform (our own unpublished observations), the findings reported herein suggest that this expansion does not arise from an intrinsic T cell functional defect.
A second T cell functional change, leading to increased generation of IL-17-producing T cells, appeared to require the NZB c1 88-96 interval. It is currently unknown whether this functional change arises solely from genetic polymorphisms located within the c1 88-96 interval or results from interaction between polymorphisms in the c1 96-100 and 88-96 intervals. Candidate genes within the 88-96 interval include: the retinoid X receptor gamma (Rxrg), a member of the RXR family of nuclear receptors which have been shown to modify the balance between Th1 and Th2 cells [19,20,21]; and Pre-B cell leukemia homeobox 1 (Pbx1) that has also been shown to influence T cell differentiation. Recently, increased T cell expression of a novel splice isoform of this gene, Pbx1-d, was found in B6 congenic mice with the Sle1a lupus susceptibility locus [22].
Despite the presence of intrinsic T cell functional abnormalities in c1 congenic mice, this does not appear to be sufficient to induce altered spontaneous T cell differentiation in vivo, pointing to a critical role of environmental cues in the induction of the abnormal differentiation of T cells in c1(88-100) and c1(70-100) mice. Nevertheless, T cell abnormalities also appear to be essential, as B6 T cells did not differentiate efficiently to IFN-γ-or IL-17-producing cells following adoptive transfer into c1(70-100) mice in-vivo. Our experiments suggest that altered DC function provides one of the environmental cues that enhances pro-inflammatory T cell differentiation in c1(88-100) and c1(70-100) mice. DC from c1(70-100), and to a lesser extent c1(88-100), mice demonstrated enhanced production of IL-12 which has been shown to promote differentiation of naïve T cells to Th1 and Tfh phenotypes [23,24,25] and increased levels of IL-6 which has been shown to promote differentiation of naïve T cells to Th17 and Tfh phenotypes [26,27,28,29]. Following these initial interactions, IL-21 production by activated Th17 and Tfh cells could act in an autocrine manner to further direct DC-primed CD4 + T cells to become Th17 and Tfh cells [30,31,32], resulting in a positive feedback loop. It is likely that the enhanced ability of mDC from c1(70-100) and, to a lesser extent, c1(88-100) mice to upregulate MHC class II and B7.2 in response to T-dC interaction further augments the differentiation of proinflammatory T cell subsets in these mice.
In summary, we demonstrate that T cell and DC defects, derived from several genetic loci, synergize to convert preclinical disease to fatal GN by leading to expansion of proinflammatory T cells. This data joins an increasing body of data from the study of congenic mouse strains demonstrating that impact of individual genetic loci on immune function and autoimmunity is highly dependent upon their genetic/ immunologic context [33,34,35,36]. These studies have important implications for the study of human autoimmune disease, in that they provide an explanation for how the presence of a susceptibility locus in the family members of a patient with autoimmune disease can be compatible with relatively normal immune function, whereas the same locus in the patient leads to profoundly altered immune function. Thus, the identification of individuals with an increased likelihood of developing autoimmune disease must necessarily involve characterization of multiple interacting genetic loci.

Ethics statement
Mice were housed in a Canadian Council on Animal Care approved facility at the Toronto Western Research Institute, part of the University Health Network. All mice used and experiments performed in this study were approved by the Animal Care Committee of the University Health Network (Animal Use Protocol 123).

Mice
B6 and NZB mice were purchased from Taconic (Germantown, NY) and Harlan Sprague Dawley (Blackthorn, England), respectively. B6.OT-II TCR Tg and B6.Thy1 a IgH a mice were originally obtained from Taconic and The Jackson Laboratory (Bar Harbor, ME), respectively, and bred in our facility. Congenic mice were generated as previously described [9]. OT-II TCR Tg and Thy1 a IgH a (termed Thy1.1 for simplicity) congenic mice were produced by polymorphic marker assisted backcrossing. Only female mice were examined and all mice were specific-pathogen free.

Flow cytometry
Half a million RBC-depleted splenocytes were incubated with mouse IgG (Sigma-Aldrich) for 15 min prior to staining with various combinations of directly-conjugated mAbs. Allophycocyanin-or PerCP-Cy5.5-conjugated streptavidin (SA) (BD Biosciences) were used to reveal biotin-conjugated Ab staining. Dead cells were excluded by staining with 0.6 µg/ml PI (Sigma-Aldrich). Events were acquired using a LSR II instrument (BD Biosciences) and analyzed using FlowJo software (Tree Star Inc., Ashland, OR). The following directly conjugated mAbs were purchased from BD Biosciences: Biotin-

Statistical analysis
Comparisons of differences between groups of mice for continuous data were performed using one-way ANOVA followed by Dunns' post-test for multiple comparisons. All statistical analyses were performed using GraphPad software (La Jolla, CA, USA). Horizontal lines indicate the mean of each group examined. Significance levels were determined by one-way ANOVA with Dunns' post-test. The p values for significant differences between B6 and congenic mouse strains are shown with *p<0.05, **p<0.01, ***p<0.001. Bars with p values above denote significant differences between congenic strains. (TIF) Figure S2. c1 congenic mice exhibit increased production of cytokines secreted by Tfh, Th1 and Th17 populations. Splenic CD4 + T cells were purified from 4-mo-old B6, c1(96-100), c1(88-100), and c1(70-100) mice using negative selection and were cultured with plate-bound anti-CD3 antibody in the presence of anti-CD28 for 48 h. Culture supernatants were assayed for cytokine production in triplicate with the levels of IL-2, IL-4, IL-17, and IFN-γ being determined using a cytokine bead array, and for IL-21 by ELISA. Each point represents the determination from an individual mouse. Horizontal lines indicate the mean for each population examined. Significance levels were determined by one-way ANOVA with Dunns' post-test. The p values for significant differences between B6 and congenic mouse strains are shown with *p<0.05, **p<0.01, ***p<0.001. Bars with p values above denote significant differences between congenic strains. (TIF) Figure S3. Identification of cytokine-producing T cell subsets in c1 congenic mice. Freshly isolated splenocytes from 4-mo-old mice were stained with anti-CD3, -CD4, -CXCR5, and -PD1, permeabilized and then stained for intracellular IL-17 and IFN-γ production (as described in Figure  2) with the addition of IL-21R/Fc chimera to detect IL-21 production.