PPARγ Negatively Regulates T Cell Activation to Prevent Follicular Helper T Cells and Germinal Center Formation

Peroxisome proliferator-activated receptor gamma (PPARγ) is a transcription factor that regulates lipid and glucose metabolism. Although studies of PPARγ ligands have demonstrated its regulatory functions in inflammation and adaptive immunity, its intrinsic role in T cells and autoimmunity has yet to be fully elucidated. Here we used CD4-PPARγKO mice to investigate PPARγ-deficient T cells, which were hyper-reactive to produce higher levels of cytokines and exhibited greater proliferation than wild type T cells with increased ERK and AKT phosphorylation. Diminished expression of IκBα, Sirt1, and Foxo1, which are inhibitors of NF-κB, was observed in PPARγ-deficient T cells that were prone to produce all the signature cytokines under Th1, Th2, Th17, and Th9 skewing condition. Interestingly, 1-year-old CD4-PPARγKO mice spontaneously developed moderate autoimmune phenotype by increased activated T cells, follicular helper T cells (TFH cells) and germinal center B cells with glomerular inflammation and enhanced autoantibody production. Sheep red blood cell immunization more induced TFH cells and germinal centers in CD4-PPARγKO mice and the T cells showed increased of Bcl-6 and IL-21 expression suggesting its regulatory role in germinal center reaction. Collectively, these results suggest that PPARγ has a regulatory role for TFH cells and germinal center reaction to prevent autoimmunity.


Introduction
Nuclear receptors constitute a superfamily of ligand-dependent transcription factors that regulate diverse aspects of metabolism and homeostasis. This family is further subdivided into three subclasses based on their DNA-binding and ligand-binding properties, including classical steroid hormone receptors, adopted orphan receptors, and orphan receptors [1][2][3][4]. Peroxisome Proliferator-activated Receptors (PPAR) are emerging adopted orphan nuclear receptors that regulate lipid and glucose metabolism, as well as cell growth, differentiation, apoptosis, and immunity, in a wide variety of cells. The PPAR subfamily consists of three subtypes including PPARa, PPARb/d, and PPARc. Each member binds multiple ligands, regulates target genes, and plays different biological roles. The ligands for PPARc have been used for therapeutic applications to treat atherosclerosis, obesityinduced insulin resistance, and inflammatory bowel disease [5][6][7]. For example, synthetic PPARc agonists called thiazolidinediones (TZD) improved insulin sensitivity in metabolic syndrome [5,8]. The PPARc agonist 15-deoxy-g 12,14 -prostaglandin J 2  has been shown to downregulate T cell proliferation and IL-2 production [9]. Another PPARc agonist, ciglitazone, has been shown to increase regulatory T (Treg) cell differentiation from naïve T cells [10], while pioglitazone has been demonstrated to ameliorate EAE severity through selective inhibition of Th17 differentiation [11] and inhibit human alloresponses in a humanized model of graft arteriosclerosis [12]. More recently, PPARc has been shown to regulate Treg accumulation in visceral adipose tissue, and pioglitazone treatment was required for complete recovery from insulin sensitivity in obese mice [13].
Autoimmune disease occurs when the immune system recognizes self-molecules and responds by inducing autoantibody production, germinal center (GC) formation, and chronic inflammation [14][15][16]. Follicular helper T (T FH ) cells are involved in the antibody response by providing selection signals to GC B cells and promoting plasma and memory cell differentiation [17]. Bcl-6 mediates the development of T FH cells, which are characterized by expression of the chemokine receptor CXCR5, ICOS, and PD-1 [18][19][20][21]. Recent studies have suggested T FH cells as a potential therapeutic target for treating autoimmune diseases [15,22,23].
Although previous studies focused on PPARc agonists revealed its regulatory function in T cells, the mechanism and intrinsic roles of this nuclear receptor in T FH cells and autoimmunity is largely uncharacterized. Therefore, we investigated the role of PPARc in T cells using CD4 T cell-specific PPARc knockout mice. Interestingly, we found that PPARc-deficient T cells exhibit enhanced proliferation following T cell receptor (TCR) stimulation and increased cytokine production under Th1, Th2, Th17, and Th9 differentiation conditions. Interestingly moderate autoimmune phenotypes developed spontaneously in 1-year-old CD4-PPARc KO mice with increased activated T cells, T FH cells, and GCs. Furthermore, PPARc regulated T cell activation by modulating the stability of IkBa, Foxo1, and Sirt1, which are negative regulators of NF-kB and T FH cells. These results suggest the important regulatory roles of PPARc in T FH cells and GC formation in the prevention of autoimmunity.

PPARc is highly expressed in CD4 T cells upon TCR stimulation
To determine its expression in various lymphocyte subsets, the level of PPARc mRNA was examined in FACS-purified CD4 + T, CD8 + T, and CD19 + B cells by quantitative real-time PCR analysis. PPARc expression was significantly higher in CD4 + T cells compared to CD8 + T cells and CD19 + B cells ( Figure 1A), suggesting that this nuclear receptor may play important roles in CD4 + T cells. In addition, PPARc mRNA expression was significantly increased following TCR stimulation ( Figure 1B), further implying its regulatory role in controlling activated CD4 + T cell functions.
Next, we generated mice in which PPARc was specifically deleted in CD4 + T cells by breeding PPARc fl/fl with CD4-Cre transgenic mice ( Figure 1C&D). PPARc-deficiency did not affect the frequency of CD4 + , CD8 + , and Foxp3 + natural regulatory T cell (nTreg) in the thymus of 6-to 8-week-old mice (Figure S1A-D), the proportion of lymphocyte population including B cells, NK cells, and naïve or memory T cells in the spleen (Figure S1E-L). Altogether these data suggest that PPARc deletion in CD4 + T cells does not affect thymic and splenic cell populations that PPARc is dispensable for T cell development.

PPARc-deficient T cells are hyper-reactive to TCR stimulation
Although PPARc is dispensable for lymphocyte development, increased PPARc expression following TCR stimulation led us to examine its function during T cell activation. To address this, we examined cytokine production from purified CD4 + CD25 2 T cells stimulated with anti-CD3 and anti-CD28 antibodies. Various cytokines including IFN-c, IL-4, IL-17, and IL-2 were increased in PPARc-deficient T cells compared to littermate control T cells, (Figure 2A-D) suggesting that these cells are hyper-reactive to TCR stimulation. In addition, PPARc-deficient T cells proliferated significantly more than control CD4 + T cells ( Figure 2E), suggesting that PPARc plays a role as a negative regulator of T cell activation and proliferation.
Next, we investigated cytokine production after CD4 + CD25 2 T cells were differentiated into various helper T cell subsets including Th1, Th2, Th17, and Th9. Interestingly, production of all the signature cytokines secreted by each helper T cell subset, namely IFN-c, IL-13, IL-17, and IL-9, respectively, were significantly increased in PPARc-deficient T cells ( Figure 2F) without affecting differentiation efficiency ( Figure 2G). These results suggest that PPARc has a negative regulatory role for T cell activation and proliferation.
PPARc maintains the stability of IkBa, Foxo1, and Sirt1 to regulate TCR signaling A previous study reported that PPARc inhibits NF-kB in macrophages [24]. However, the molecular basis of how PPARc regulates T cell activation has not been fully elucidated. To determine the mechanism, we assessed ERK and AKT phosphorylation in splenic naïve CD4 + T cells (CD4 + CD25 2 CD62L high ) stimulated with anti-CD3/CD28-coated beads. Following TCR stimulation for 20 min, the PPARcdeficient CD4 + T cells showed increased ERK and AKT phosphorylation compared to littermate control T cells ( Figure 3A), suggesting that PPARc inhibits events of TCR downstream signaling pathway.
We performed microarray analysis to identify possible target genes (e.g., Foxo1) regulated by PPARc (data not shown). Interestingly, TCR stimulation of PPARc-deficient CD4 + T cells resulted in reduced expression of IkBa, Foxo1, and Sirt1 ( Figure 3B), which have been recently identified as negative regulators of T cell activation [25,26]. In addition, phosphorylated p65 level was increased in PPARc-deficient CD4 + T cells suggesting that PPARc contributes to the stability of IkBa, Foxo1, and Sirt1 to regulate NF-kB in activated T cells.

PPARc deficiency in T cells spontaneously enhances activated T cells in vivo
The TCR-stimulated hyper-reactivity observed in PPARcdeficient T cells from 6-to 8-week-old CD4-PPARc KO mice prompted us to investigate spontaneous changes in the T cell phenotype of 1-year-old CD4-PPARc KO mice. CD62L is a cell adhesion molecule, which is highly expressed on naïve T lymphocytes. The frequency of CD4 + CD62L low cells was increased significantly in the mesenteric lymph nodes (MLN) of CD4-PPARc KO mice compared to age-matched control mice ( Figure 4A&B). The proportion of CD4 + CD62L low cells of CD4-PPARc KO mice was augmented compared to littermate control. These results are supported by an increase of the early activation marker CD69 expressing CD4 + T cells ( Figure 4C&D). Additionally, we investigated the expression of IL-7Ra, which has been shown to be decreased in activated T cells [27]. PPARc deficiency in T cells from 1-year-old CD4-PPARc KO mice showed significantly reduced proportion of CD4 + IL-7Ra high cells in the MLN ( Figure 4E&F). Collectively, these data demonstrate that PPARc deficiency in T cells contributes to their hyper-reactivity, which induces spontaneous T cell activation in vivo.

PPARc-deficient T cells in mice develop a spontaneous autoimmune phenotype
Because the number of activated T cells was increased in 1-yearold CD4-PPARc KO mice, we next investigated autoimmune phenotype in these animals. Flow cytometric analysis of splenocytes from 6-month-old CD4-PPARc KO mice demonstrated that the proportion of CD19 + B cells was increased while CD4 + T cells were slightly decreased compared to littermate control mice ( Figure 5A). Based on this observation, we hypothesized that the hyper-reactive PPARc-deficient T cells promote greater B cell activation, thereby leading to autoantibody production. Our data reveal that the proportion of CD19 + and CD138 + plasma cells was increased in the spleen of CD4-PPARc KO mice ( Figure 5B&C). Moreover, the level of anti-nuclear antibodies (ANA) in serum from 6-month-old mice ( Figure 5D) and the level of anti-doublestranded DNA (dsDNA) antibodies in serum from 1-year-old mice ( Figure 5E) were higher than age-matched control mice.
To further examine the autoimmune phenotype, inflammation in the kidneys was examined and scored by histologic analysis of hematoxylin and eosin-stained sections ( Figure 5F&G). Kidney from 1-year-old CD4-PPARc KO mice exhibited an expanded double layer around the glomerulus with thicker tubules compared to age-matched control mice. These results suggest that PPARc deficiency in T cells leads to autoantibody production and spontaneous autoimmune disease in aged mice.

PPARc-deficient T cells contribute to T FH cells and GC development
T FH cells are crucial for providing help to B cells within the germinal center, where somatic hyper-mutation, class-switch recombination, and differentiation of memory B cells and longlived plasma cells occur [28][29][30][31]. Due to increased B cell numbers, including plasma cells, in 6-month-old CD4-PPARc KO mice, we hypothesized that PPARc deficiency in T cells lead to increased T FH cells in vivo. Previous report has shown that PSGL-1 low population enriched more T FH cells than PSGL-1 high [31] so that PSGL-1 low gating strategy was applied to analyze T FH cells.
Flow cytometric analysis revealed an increased proportion of CXCR5 + PD-1 + T FH cells in 1-year-old CD4-PPARc KO mice ( Figure 6A -C). The number of GCs found in the spleen of these animals was also increased compared to age-matched control mice ( Figure 6D&E). In addition, proportion of germinal center B cells (CD95 + GL7 + from B220 + cells) were increased in the spleen with elevated amount of IgG1 and reduced IgM from the sera in 1year-old CD4-PPARc KO mice ( Figure 6F&G). These results suggest that the spontaneous autoimmune phenotypes observed were related to increased T FH cells and enhanced GC reaction.
To further confirm that PPARc regulates induction of T FH cells and GC reaction, 8-week-old mice were immunized with sheep red blood cells (SRBC) and the T FH cells and GC number were analyzed on day 7 ( Figure 7A). Following SRBC immunization, the proportion of CXCR5 + PD-1 + T FH cells ( Figure 7B&C) and total number of spleen cells were significantly increased ( Figure 7D) which is correlated to enhanced GC number ( Figure 7E&F) in CD4-PPARc KO mice than in control animals. Next, we examined IL-21 and Bcl-6 expression levels, which are important molecules for induction of T FH cells and GC reaction. CD4-PPARc KO T

Discussion
Based on its ligand-dependent functions, PPARc has been demonstrated to be an important negative regulator in T cells, macrophages, and dendritic cells [9,24]. Previous studies have also suggested that PPARc is an important therapeutic target for treating autoimmune disease [5][6][7]. Here, for the first time, we characterized CD4-PPARc KO mice to investigate the intrinsic role of PPARc in T cells and autoimmunity related to T FH cells and GC reaction.
We found that, among lymphocytes, PPARc was expressed predominantly in CD4 + T cells. The expression level of PPARc, however, increased following TCR stimulation. This is consistent with a previous study that reported an increase in PPARc expression following anti-CD3 antibody stimulation and PPARc ligand (15d-PGJ2 and ciglitazone) treatment, which it also contributed to the inhibition of T cell proliferation and IL-2 production [9]. Taken together, these results imply that PPARc expression and TCR stimulation function together to regulate T cell response.
One recent study reported that pioglitazone, a PPARc ligand, inhibited PPARc-meditated differentiation of Th17, but not Th1, cells in an autoimmune EAE model, which suggested the possibility of selective regulation of Th17 cells by PPARc [11]. However, we found significantly increase of all the cytokines, including IFN-c, IL-13, IL-17, and IL-9 in each T cell differentiation skewing media without affecting differentiation efficiency possibly due to hyper-proliferative property of PPARcdeficient T cells. This increased T cell sensitivity seems to be correlated with higher AKT and ERK phosphorylation as well as rapid IkBa degradation following TCR stimulation.
Previously PPARc was studied to inhibit the NF-kB and AP-1 transcription factors in macrophages and T cells [24,25,32] and PPARc agonists were shown to inhibit inflammatory cytokine production in monocytes [33], as well as proliferation or IL-2 production in T cells following TCR stimulation [9]. Our results demonstrated a loss of Sirt1 and Foxo1, two negative regulators of T cell activation, in PPARc-deficient T cells following TCR stimulation. Sirt1, a histone deacetylase, was shown to inhibit AP-1 and is required for T cell tolerance [26]. Foxo1 is also required for inhibiting T cell activation and acts as a T cell-intrinsic regulator of tolerance [34]. The levels of Sirt1 and Foxo1 expression were significantly down-regulated in PPARc-deficient T cells following TCR stimulation, suggesting that PPARc controls the stability or expression of Sirt1 and Foxo1 to inhibit NF-kB or AP-1 activation. In addition, Foxo1 is a positive transcriptional regulator of Sirt1 and is deacetylated by Sirt1 [35]. Analysis of ChIP-seq data from BIOBASE (http://www.biobase-international.com) identified various PPARc response elements within the Foxo1 promoter region based on the study of genome wide analysis of PPARc binding region in 3T3-L1 cells [36]. Foxo1 signaling through NF-kB has been shown to mediate proinflammatory cytokine production by interacting together in a synergistic manner in macrophages [37]. Another study also revealed the role of Foxo1 in T cell tolerance by demonstrating that T cell-specific deletion of Foxo1 caused hyper-reactivity of T cells [34]. In addition, Foxo1-deleted CD4 T cells also spontaneously increased T FH cells and GC formation as well as autoantibody production in 8-week-old mice [34]. ChIP analysis examining the interaction of PPARc on Foxo1 promoter region will be performed in a future study to elucidate more detail mechanism.
Previous studies demonstrated treatment of T cells with a PPARc agonist increases Foxp3 expression and iTreg generation [10,38]. However, our results demonstrated that PPARc is not required for Treg generation suggesting PPARc agonists possibly increase Foxp3 expression PPARc independently. One recent study reported that PPARc was dispensable for Foxp3 expression in Treg cells yet critical for inhibiting colitis [39] while PPARc was recently shown to influence the migration of Treg cells to visceral adipose tissues by regulating chemokine receptor expression without affecting Treg differentiation [13]. Collectively, these results indicate that PPARc is required for the in vivo suppressive function related to the migration of Treg cells into inflammatory regions, but not Foxp3 expression itself.
This study is the first to characterize CD4-PPARc KO mice to demonstrate spontaneous autoimmune phenotypes along with T FH cells and GC formation. T FH cells have recently been emerging as a T helper subset expressing Bcl-6 and IL-21 that interacts with B cells during GC formation. CD28-or ICOSdeficient mice exhibit defective GC formation, affinity maturation, and T FH cell development [21,[40][41][42][43]. CD28 or ICOS costimulatory signals and TCR signals are essential for the induction of T FH cells [40,41]. Reports have also shown that stronger binding to major histocompatibility complex (MHC) class II and TCR is important for T FH cell development [44]. Therefore, increased AKT phosphorylation and reduced expression of negative regulators to NF-kB in PPARc-deficient T cells following TCR stimulation could contribute to increased T FH cell generation in vivo. We also demonstrate an increased proportion of T FH cells and number of GCs following SRBC immunization in CD4-PPARc KO mice. In addition, PPARc deficient Th2 and Th17 cells produced higher IL-21 and Bcl-6 suggesting more T FH cells possibly are generated and contributed GC reaction during antigen stimulation in vivo. One previous study suggested that PPARc sustains Foxo1 expression [45] and another study demonstrated that mice lacking Foxo1 in T cells exhibited a large population of T FH cells and a decreased number of Treg cells [25], suggesting a molecular mechanism for increase of T FH cell differentiation may involving relationship of PPARc and Foxo1.
In conclusion, we have revealed the importance of PPARc in regulating T cell activation related to T FH cells and autoimmunity. Modulating PPARc to dampen abnormal T cell activation and T FH cell mediated GC reaction can be exploited for treating autoimmune diseases.

Materials and Methods
Mice B6.129-Pparg tm2Rev /J (PPARc fl/fl ) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). To generate CD4-specific PPARc knockout mice, PPARc fl/fl mice were crossed with CD4-Cre +/2 transgenic mice. All mice were maintained at the Yale University and Hanyang University mouse facilities under pathogen-free conditions.

Ethics statement
All animal protocols used in this study were approved by the Yale Institutional Animal Care and Use Committee and Hanyang Animal Care and Use Committee.

Western blot analysis
CD4 + T cells were enriched from isolated splenocytes using antibody-coated magnetic beads. The negatively selected CD4 + T cells were labeled with CD62L MACS bead (Miltenyi Biotec, Bergisch Gladbach, Germany) to purify CD4 + CD62L + naïve T cells. The purity of naïve T cells was 95,96% (data not shown). The isolated naïve T cells were stimulated with anti-CD3 and anti-CD28 antibody coated Dynabeads (Life Technologies, Carlsbad, CA) at 37uC for 0, 5, 10, and 20 min. The cells were then lysed with RIPA buffer (Cell Signaling Technology, Danvers, MA) on ice for 20 min and the protein content was determined using the Pierce BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA). The proteins were resolved on 9% SDS polyacrylamide gels and transferred onto PVDF membrane (Millipore, Temecula, CA). After blocking with 5% skim milk, the membrane was incubated with specific antibodies against P-ERK, P-AKT, P-P65, IkBa, Foxo1, Sirt1 and b-actin (Cell Signaling Technology) overnight at 4uC. HRP-conjugated secondary antibody followed by incubation with Supersignal West Femto ECL solution (Thermo Fisher scientific) were used to detect primary antibody binding.

Quantitative real-time PCR
RNA from FACS-sorted CD4 + , CD8 + , and CD19 + cells was isolated using the RNeasy Mini kit (Qiagen, Venlo, Limburg) according to the manufacturer's protocol. cDNA synthesis was performed using the iScript kit (Bio-Rad, Hercules, CA). Real-time PCR was performed on a Bio-Rad CFX Connect real-time PCR detection system using iQ SYBR Green Supermix (Bio-Rad). The level of PPARc expression was assessed in triplicate with specific primers and normalized to b-actin.

ELISA
Cytokine levels in the supernatants of polarized Th1, Th2, Th17, and Th9 cultures were determined by ELISA using antibodies against mouse IL-2 (Biolegend, San Diego, CA), IFNc, IL-13, IL-17 and IL-9 (eBioscience) according to the manufacturer's instructions. Anti-dsDNA antibody in mouse serum was determined by ELISA (Alpha Diagnostic International Inc, San Antonio, TX).

Confocal microscopy
Mice aged 6 to 8 weeks old were immunized with sheep red blood cells (SRBC) and spleens were harvested 9 days later. Spleens from 1-year-old mice were also collected to analyze T FH cells and GC formation. The tissues were frozen in OCT reagent and sectioned into 7-mm slices. The frozen sections were stained with anti-PNA-FITC (Sigma-Aldrich, St. Louis, MO), IgD-PE, and CD4-APC (eBioscience) overnight at 4uC. After washing, Anti-Fade reagent (Invitrogen, Life Technologies, Carlsbad, CA) was added to the slides, which were visualized using a Leica DM IRE2 confocal microscope.

Histology
Kidney tissues from 1-year-old mice were isolated, fixed, embedded in paraffin, and stained with hematoxylin and eosin. For each mouse, the level of kidney inflammation and more than 15 glomerular, tubular, or interstitial areas were evaluated and scored blindly for glomerular cellularity, leukocyte infiltration, severity of tubular lesions, mesangial matrix expansion, crescent formation, and interstitial mononuclear cell infiltration in the medulla and cortex. The severity of kidney lesions was determined by scoring each feature (from 0 to 3) and then calculating the mean of each set of scores. For example, glomerular inflammation was scored as follows: 0 = normal or no inflammatory cells; 1 = few inflammatory cells; 2 = moderate inflammation; and 3 = severe lymphocyte infiltration.

Statistical analysis
All data were analyzed statistically analyzed with the Student's t-test and Mann-Whitney test using Prism5 (GraphPad, San Diego, CA). p-values (P) less than 0.05 were considered statistically significant.