STAT4 Regulates the CD8⁺ Regulatory T Cell/T Follicular Helper Cell Axis and Promotes Atherogenesis in Insulin-Resistant Ldlr^−/− Mice

Atherosclerosis is a multifactorial chronic inflammatory disease that is characterized by the accumulation of modified lipoproteins and immune cells in the aortic wall, vascular dysfunction, and low-grade chronic inflammation. Atherosclerosis is the prominent cause of cardiovascular diseases and mortality in many countries (1, 2). Inflammation impacts all stages of atherogenesis and is a common denominator among atherosclerosis, obesity, insulin resistance (IR), and type 2 diabetes (3). Although the role of IR in atherosclerosis is recognized, there are limited numbers of murine models that can be used to investigate the simultaneous effects of IR and inflammation on the induction of atherosclerosis (4). Low-density lipoprotein receptor–deficient (Ldlr^−/−) mice develop obesity, IR, and atherosclerosis when fed a diabetogenic diet with added cholesterol (DDC) (5). Importantly, this study showed that male mice fed a DDC diet have increased macrophage (MΦ) infiltration and inflammation in adipose tissues, as well as significantly increased MΦ content in the aorta within atherosclerotic plaques and accelerated atherogenesis (5). Therefore, we chose to use this model of diet-induced obesity and IR to investigate a role for STAT4 in IR-accelerated atherosclerosis.

STAT4 is an important regulator of inflammatory immunity and autoimmune-related inflammation (6). The activation of STAT4 is regulated by the proinflammatory cytokines IL-12 and IL-23 released primarily by APCs (6, 7). STAT4 and the T-box transcription factor T-bet are key transcription factors that support the differentiation of Th1 cells. Importantly, the expression of STAT4 also suppresses the expansion of Foxp3⁺CD4⁺ regulatory T cells (Tregs) (8). Thus, STAT4 is involved in the regulation of a delicate balance between the proinflammatory and suppression arms of the immune response. Recent work on the biology of STAT4 has revealed that it is not expressed exclusively by T and NK cells; it can also be detected in IFN-α– or LPS-stimulated monocytes and dendritic cells (7, 9, 10). Although the role of STAT4 in myeloid cells is not well understood, there is some evidence suggesting a key role for it in MΦ biology. Attenuated microbicidal activity and reduced NO and IFN-γ levels were detected in STAT4-deficient MΦs (9–11). Our data also demonstrated reduced cytokine production, low activation, and diminished expression of CCR2 in STAT4-deficient M1 or M2 MΦs (12).

Although several subsets of T cells and MΦs play an important role in atherosclerosis, additional leukocyte populations, including T follicular helper (Tfh) cells and CD8⁺ Tregs, are now reported to have a role in this disease (13). Tfh cells assist in the induction of germinal centers and support the activation and differentiation of memory B cells and plasma cells regulating the generation of Ag-specific Abs (13). Importantly, Clement et al. (13) demonstrated that Tfh cells support atherogenesis via the production of pathological Abs and generation of highly active germinal centers. The investigators also showed that CD8⁺ Tregs control Tfh cell development and the formation of follicular helper–germinal center B cells during atherogenesis. The IL-12/STAT4 pathway is involved in the gene expression of Il21 and Bcl6, both of which are necessary for the generation of Tfh cells (14–17). In line with this observation, it is reported that the in vitro differentiation of human Tfh cells is supported by STAT3/STAT4 signaling (18). However, the role of STAT4 in the generation of Tfh cells under atherosclerosis-prone conditions has not been examined.

Mounting evidence has demonstrated that a population of CD8⁺CD122⁺ Tregs controls the generation of autoreactive CD4⁺ T cells, as well as the formation of Tfh cells (19, 20), suppressing autoimmune and alloimmune responses. Importantly, in atherosclerosis-prone conditions, CD8⁺ Tregs suppress the development of Tfh cells and the formation of germinal centers in Apoe^−/− mice (13). Although the functions of CD8⁺ Tregs are under active investigation, the transcriptional network that controls the differentiation of CD8⁺ Tregs is unknown. In this study, we demonstrate that STAT4 suppresses CD8⁺ Treg functions and affects a well-known ability of CD8⁺ Tregs to defeat generation of Tfh cell and germinal B cells in vivo. Additionally, STAT4 supports MΦ activation and modulation of the proinflammatory immune composition within the aorta. The results obtained in this study could lead to novel drug therapy using inhibitors against STAT4 to regulate the immune response and IR-related inflammation to provide a duel strategy to combat IR-associated atherogenesis.

Stat4^−/− mice (21) were crossbred with Ldlr^−/− mice (The Jackson Laboratory, Bar Harbor, ME) to generate Stat4^−/−Ldlr^−/− mice. For some experiments, C57BL/6 and Stat4^−/− mice were used. Beginning at 8 wk of age, male Stat4^−/−Ldlr^−/− and Ldlr^−/− mice were fed a DDC diet (protein 20.5%, fat 36.0%, carbohydrates 35.7%, cholesterol 0.15%, product number S6524; Bio-Serv) for 11, 16, or 24 wk. All animals were kept in specific pathogen–free conditions, and animal experiments were approved by the Eastern Virginia Medical School Animal Care and Use Committee.

The aortas of Stat4^−/−Ldlr^−/− and Ldlr^−/− mice were collected, stained with Oil Red O (ORO), microdissected longitudinally, and pinned as described earlier. Images were scanned, and the surface area percentage occupied by lesions was determined by two independent investigators using ImageJ (National Institutes of Health). Hearts were harvested and fixed with 4% paraformaldehyde via cardiac puncture. From the point of the appearance of aortic valve leaflets, sequential 5-μm-thick sections were cut, and six sections over a 300-μm distance were collected and analyzed by Russell-modified Movat staining, as previously described (12). Total cholesterol and triglyceride levels were determined according to the manufacturer’s instructions.

Single-cell suspensions from the aorta were prepared as previously described (12, 22). Briefly, mice were anesthetized using CO₂, and blood was collected via cardiac puncture. Next, the heart was perfused with PBS containing 20 U/ml heparin by cardiac puncture. Aortas were microdissected and enzymatically digested for 1 h at 37°C with 125 U/ml Collagenase type XI, 60 U/ml hyaluronidase type I-s, 60 U/ml DNase I, and 450 U/ml Collagenase type I (Sigma-Aldrich, St. Louis, MO) in PBS, as described previously (12, 22). Aortas, spleens, para-aortic lymph node (LNs), and peripheral LN (PLNs) were rubbed in a 70-μm cell sieve (Corning Incorporated Life Sciences, Tewksbury, MA). Erythrocytes in spleens were lysed using ACK lysis buffer (8.29 mg/ml NH₄Cl, 1 mg/ml KHCO₃, 0.372 mg/ml EDTA; all from Sigma-Aldrich). Cell numbers were determined using trypan blue (MP Biomedicals, Solon, OH) and a hemocytometer. Intracellular staining for T-bet, Foxp3, CD68, and Bcl6 was performed using a FIX & PERM Cell Fixation & Permeabilization Kit (BD Biosciences, San Jose, CA). A Cytek DXP8 Color (Cytek Development) upgraded FACSCalibur (BD Biosciences) was used to collect samples, and data analysis was conducted with FlowJo (TreeStar, Ashland, OR). In all flow cytometry experiments, isotype controls and fluorescent-minus-one controls were used to set appropriate gating for the samples. To exclude doublets from analysis, a forward scatter area against forward scatter linear gate was used.

The following Abs were used: CD19-PECy7 (1D3), CD8-allophycocyanin (53-6.7), Ly6C-FITC (AL-21), CD68-PE (FA11), CD44-FITC (IM7), CD45-PerCP (30-F11), CD69-PE (H1.2F3), CXCR5-ef450 (SPRCL5), CD138-allophycocyanin (281-2), CD21-PE (7G6), CD275-PE (HK5.3), CD122-allophycocyanin (TM-b1), CD11b-PB (M1/70.15), IFN-γ–eFluor 450 (XMG1.2), CD4-PerCP (L3T4), F4/80–allophycocyanin–eFluor 780 (BM8), Foxp3-PE (MF23), CD3–allophycocyanin–CY7/eFluor 780 (17A2), Ly-6G–PE (1A8) (all from eBioscience, San Diego, CA), and anti-mouse CD16/CD32 (The Lymphocyte Culture Center, University of Virginia, Charlottesville, VA). For some staining, we used PerCP-CD4, biotinylated-CXCR5, followed by allophycocyanin-streptavidin, or FITC–GL-7, PE-FAS, and PerCP-B220 staining (all from BioLegend). To distinguish between live and dead cells, Viability Dye eFluor 650 (eBioscience) was used. Anti-CD3 and anti-CD28 were used for in vitro assays (eBioscience). Recombinant proteins mouse TGF-β and IL-2 were purchased from PeproTech (Rocky Hill, NJ). LPS was purchased from Sigma-Aldrich.

Total RNA was extracted from splenic cells and peritoneal MΦs using TRIzol Reagent (Invitrogen, Life Technologies, Grand Island, NY). DNase I treatment was used to remove contaminating genomic DNA (QIAGEN, Germantown, MD). Approximately 1 μg of total RNA was reverse transcribed to cDNA by synthesis reaction using a Promega Reverse Transcription System containing Reverse Transcription 10× Buffer without MgCl₂, MgCl₂, Random Primers, 10 mM dNTP Mix, AMV Reverse Transcriptase (HC), and Recombinant RNasin Ribonuclease Inhibitor. Real-time PCR was performed using TaqMan probes from Applied Biosystems (Carlsbad, CA), 10 mM dNTPs, 10× PCR buffer without MgCl₂, MgCl₂, and Jumpstart Taq polymerase (Sigma-Aldrich) for iNos, Mrc1, Il21, and Il27 for splenic cells and Ifnγ, Il6, Tgfβ, and Il27 for LPS-stimulated peritoneal MΦs. Ct values for cDNA were determined using a CFX96 Real-Time System with a C1000 Thermal Cycler (Bio-Rad). The results were normalized to housekeeping gene 18S.

Briefly, CD8⁺ T cells from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice were isolated using a stem cell kit (EasySep Mouse CD8⁺ T Cell Isolation Kit; STEMCELL Technologies) and plated with recombinant human (rh)TGF-β (2 ng/ml), rhIL-2 (100 U/ml), plate-bound anti-CD3 (1 μg/ml), and soluble anti-CD28 (1 μg/ml) in complete RPMI 1640 medium supplemented with 10% FBS, 1% penicillin/streptomycin, 1% GlutaMAX, 1% HEPES, 0.5% nonessential amino acids, 0.5% sodium pyruvate, and 50 μM 2-ME) for 3 d. Subsequently, FACS analysis was used to identify differentiated CD8⁺ Tregs as CD8⁺CD122⁺ cells (23).

Briefly, peritoneal MΦs from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice were plated in complete RPMI 1640 medium supplemented with 10% FBS, 1% penicillin/streptomycin, 1% GlutaMAX, 1% HEPES, 0.5% nonessential amino acids, 0.5% sodium pyruvate, and 50 μM 2-ME for 2 h. Thioglycollate was not used for the isolation of peritoneal MΦs. Lymphocytes were removed by washing with complete media. Next, LPS (500 ng/ml)–supplemented complete media was used to activate peritoneal MΦs for 48 h, and the conditioned media and the peritoneal MΦs were collected. Ldlr^−/− splenic CD4⁺ or CD8⁺ T cells were isolated using a stem cell kit (STEMCELL Technologies) and plated with immobilized anti-CD3, soluble anti-CD28, and the conditioned media from the Stat4^−/−Ldlr^−/− and Ldlr^−/− LPS-activated peritoneal MΦs. Differentiation into Th cell types or CD8⁺CD122⁺ Tregs was determined 48 h later using flow cytometry.

To perform glucose tolerance tests (GTT), mice that were fasted overnight were injected i.p. with filter-sterilized glucose (2 g/kg) in 0.9% NaCl. A tail vein blood sample was taken before the injection and at 10, 20, 30, 60, 90, and 120 min after injection for measurements of blood glucose levels. The insulin tolerance test (ITT) was conducted by i.p. injection of insulin (0.75 U/kg) in 0.9% NaCl. A tail vein blood sample was taken immediately before and at 15, 30, 45, and 60 min after the injection for analysis of blood glucose levels. GTTs and ITTs were performed on the same set of mice that were used for en face analysis.

CD8⁺CD122⁺ STAT4-sufficient or STAT4-deficient Tregs (CD45.2⁺) were transferred i.v. into C57BL/6 (CD45.1⁺) mice (0.5 × 10⁶ cells per recipient, n = 4 per group). A group of C57BL/6 mice that did not receive T cells was used as a control. All mice were immunized s.c. with 1000 μg of keyhole limpet hemocyanin (KLH) emulsified in CFA as we described previously (14). Seven days after immunization, draining LNs from the recipients were collected, and cell suspensions were prepared and stained with Abs (PE-Bcl6, PerCP-CD4, biotinylated-CXCR5, followed by allophycocyanin-streptavidin or FITC–GL-7, PE-Fas, and PerCP-B220) and analyzed by flow cytometry. Ag-specific IgM, IgG, and IgA Abs in sera from immunized mice were measured with ELISA. In brief, serum samples were added in a 3-fold serial dilution to plates coated with 10 μg/ml KLH. Ag-specific Abs were detected with biotinylated goat anti-mouse IgM, IgA, or rat anti-mouse IgG Abs (SouthernBiotech).

Sorted CD8⁺CD122⁺ Stat4^−/−Ldlr^−/− or CD8⁺CD122⁺ Ldlr^−/− cells were transferred i.v. into 8-wk-old Ldlr^−/− mice (0.9 × 10⁶ cells per recipient, n = 5 per group). For basal levels of plaque burden, a group of Ldlr^−/− mice (n = 6) was injected with PBS. After 11 wk of DDC feeding, aortas were analyzed for plaque burden using ORO. Additionally, aortas, spleens, blood, para-aortic LNs, and PLNs were collected from Ldlr^−/− recipients that received CD8⁺CD122⁺ Stat4^−/−Ldlr^−/− or CD8⁺CD122⁺ Ldlr^−/− donor cells. Cell suspensions from PLNs, para-aortic LNs, and spleen were prepared, stained for Tfh cell and germinal center B cells, and analyzed by flow cytometry. Obtained plasma was analyzed for the presence of IgM, IgG1, IgG2a, IgG2b, IgG k, IgG λ, IgG3, and IgA using a Mouse Immunoglobulin Isotyping ELISA Kit (BD Biosciences).

Data were analyzed by GraphPad Prism 6, and comparisons were made using the Student or Mann–Whitney U test with the data expressed as mean ± SEM. Comparisons of three or more groups were conducted using ANOVA, and multiple comparisons used the Tukey test. Statistical significance was set at p < 0.05.

We recently demonstrated a proatherogenic role for STAT4 in an Apoe^−/− mouse model of atherosclerosis (12); however, the role of STAT4 in IR-accelerated atherogenesis remains undefined. Schreyer et al. (24) reported that the consumption of a DDC results in obesity, IR, and accelerated atherogenesis in Ldlr^−/− mice (5). Therefore, to investigate the role of STAT4 in IR-accelerated atherosclerosis, we generated Stat4-deficient Ldlr^−/− (Stat4^−/−Ldlr^−/−) mice and used a DDC for the course of this study. In line with the study by Schreyer et al. (24), we detected impaired glucose tolerance in the GTTs (Fig. 1A) and diminished insulin sensitivity in the ITTs (Fig. 1B) in Ldlr^−/− mice fed a DDC in comparison with Ldlr^−/− mice fed a chow diet (CD). These results using the area under the curve of DDC-fed versus CD-fed mice indicate that DDC feeding results in glucose intolerance and IR in Ldlr^−/− mice. Next, we investigated the impact of STAT4 deficiency on the development of IR in DDC-fed Stat4^−/−Ldlr^−/− mice. These mice displayed reduced area under the curve in the ITTs and GTTs, indicating improved insulin sensitivity and glucose homeostasis (Fig. 1A, 1B) compared with diet-matched Ldlr^−/− mice.

To directly assess the impact of STAT4 deficiency on atherogenesis, we examined the plaque burden in Stat4^−/−Ldlr^−/− and Ldlr^−/− male mice fed a DDC for 16 wk. Male Stat4^−/−Ldlr^−/− mice had a 36% reduction in aortic lesions in comparison with age- and diet-matched Ldlr^−/− control mice (5.8 ± 0.6 and 9.1 ± 1.2%, respectively, p < 0.02, Fig. 1C). Moreover, there was a reduction in cross-sectional plaque area in the aortic roots of Stat4^−/−Ldlr^−/− mice compared with Ldlr^−/− mice (Fig. 1D). Diet- and age-matched Stat4^−/−Ldlr^−/− and Ldlr^−/− mice display no differences in body weight, plasma cholesterol, or triglyceride when fed a DDC for 16 wk (data not shown). To further examine the role of STAT4 in the regulation of atherosclerosis at the advanced stage of atherogenesis, we next analyzed Stat4^−/−Ldlr^−/− and Ldlr^−/− mice that were fed a DDC for 24 wk. STAT4 deletion resulted in ∼45% reduction in plaque burden in Stat4^−/−Ldlr^−/− mice in comparison with Ldlr^−/− controls (9.6 ± 2.0 and 17.4 ± 1.2%, respectively, p < 0.005, Fig. 1E). Altogether, these results clearly demonstrate a prominent proatherogenic role for STAT4 at the different stages of IR-accelerated atherogenesis in DDC-fed Ldlr^−/− mice.

Atherosclerosis-prone conditions induce differentiation of Tfh cells and increased germinal center formation, with a concomitant elevation in plasma levels of Igs in aged Apoe^−/− mice (13). To gain insight into a role for STAT4 in the regulation of Tfh cell content in atherogenesis, we first analyzed the cellularity and immune composition of secondary lymphoid organs in Stat4^−/−Ldlr^−/− and Ldlr^−/− mice fed a DDC diet. We found a reduced number of CD4⁺ cells in spleens of Stat4^−/−Ldlr^−/− mice versus Ldlr^−/− mice (Fig. 2A). Interestingly, reduced expression of CD69, as an early activation marker, and CD44, as a marker of a long-term memory phenotype, was detected in splenic CD4⁺ T cells from Stat4^−/−Ldlr^−/− mice in comparison with Ldlr^−/− mice (Fig. 2B), suggesting at least some defective T cell activation under the conditions of STAT4 deficiency. Importantly, Stat4^−/− mice have no differences in splenic T and B cell composition in comparison with C57BL/6 mice under normal/noninflamed conditions (21, 25). In agreement with these data, we also found no difference in the percentage of splenic CD4⁺ and CD8⁺ T cells between Stat4^−/− and C57BL/6 mice (CD4⁺ 21.4 ± 0.8 and 21.7 ± 0.9%, CD8⁺ 15.1 ± 0.8 and 13.7 ± 0.6% for Stat4^−/− and C57BL/6, respectively). Thus, the reduction in splenic CD4⁺ cells in Stat4^−/−Ldlr^−/− mice is likely due to specific effects of STAT4 deficiency under inflammatory conditions of atherosclerosis.

Tfh cells are detected in spleens and tertiary lymphoid structures at the advanced stages of atherosclerosis (13). To determine whether STAT4 plays a role in Tfh cell content under DDC-induced atherosclerotic conditions, we examined the relative proportion and number of Tfh cells in the spleen using the expression of the Tfh cell–defining chemokine receptor CXCR5. STAT4 deficiency caused a significant reduction in the numbers of splenic CD4⁺CXCR5⁺ Tfh cells in Stat4^−/−Ldlr^−/− mice in comparison with Ldlr^−/− mice (1.1 ± 0.2 × 10⁶ versus 2.2 ± 0.6 × 10⁶, respectively, p < 0.05, Fig. 2C). One of the major roles of Tfh cells is to regulate the differentiation of B cells to plasma B cells, resulting in the generation of specific Ab responses (26). Next, we sought to determine whether reduced numbers of Tfh cells in the spleen of Stat4^−/−Ldlr^−/− mice would affect plasma B cell content under the conditions of STAT4 deficiency. There was a significant reduction in the number of CD19⁺CXCR5⁻CD138⁺CD21⁻ plasma B cells in Stat4^−/−Ldlr^−/− mice compared with Ldlr^−/− controls (Fig. 2E, 2F).

CD8⁺ Tregs are important suppressors of T cells that control the development of autoreactive CD4⁺ T cells. CD8⁺ Tregs also target Tfh cells because of their high basal expression of the Qa-1 molecule; thus, they indirectly regulate the formation of germinal centers, H-chain class switching, and affinity maturation of Abs (20). CD8⁺ Tregs are typically identified as CD8⁺CXCR5⁺CD275⁺CD122⁺ cells and are found in a small numbers in secondary lymphoid tissues (20). An example of the gating scheme used for analysis is depicted in Fig. 3A. Interestingly, a recent study demonstrated a protective role for splenic CD8⁺ Tregs in atherosclerosis via the downregulation of Tfh cell number and reduced generation of germinal center formation and plasma B cells (13). Although the transcriptional network that is responsible for the induction of CD4⁺ Tregs is well defined, transcriptional factors that are responsible for the induction/maintenance of CD8⁺ Tregs remain elusive. Because we found reduced numbers of Tfh cells in spleens of Stat4^−/−Ldlr^−/− mice, and CD8⁺ Tregs are known to be an important factor in the process of Tfh cell regulation, we hypothesized that STAT4 might be an essential factor in CD8⁺ Treg differentiation. To test this hypothesis, we examined the cellularity of spleens isolated from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice. We detected a significant increase in the number of CD8⁺CXCR5⁺CD275⁺CD122⁺ T cells in Stat4^−/−Ldlr^−/− mice (p < 0.05, Fig. 3B), despite a reduction in total CD8⁺ T cells (Fig. 3C). Interestingly, there was no significant difference in Tfh cell or CD8⁺ Treg content between spleens of Stat4^−/− and C57BL/6 mice (data not shown). Thus, our results suggest that STAT4 might regulate components of the humoral response in atherogenesis via the fine-tuning of a CD8⁺ Treg/Tfh cell/B cell axis.

It is well-known that STAT4 is a transcription factor involved in the differentiation of Th1 cells (6) and negative regulation of CD4⁺ Treg generation (8, 27). In this study, we hypothesized that STAT4 might also be important as a negative regulator for the differentiation of CD8⁺ Tregs upon Treg-stimulating conditions in vitro. To test this concept, we isolated CD8⁺ splenic T cells from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice and cultured them with plate-bound anti-CD3 in complete RPMI 1640 media supplemented with soluble anti-CD28, rhIL-2, and rhTGF-β for 3 d, as described (23). We detected an increased percentage of STAT4-deficient CD8⁺ Tregs compared with STAT4-sufficient CD8⁺ Tregs in analyzed cell cultures (Fig. 3D), suggesting a unique cell-intrinsic role for STAT4 in regulating CD8⁺ Treg development.

To further clarify the potential mechanism through which STAT4 deficiency reduces plaque formation, we examined the aortic immune composition, with a specific focus on CD8⁺ Tregs, Tfh cells, and MΦs. STAT4 deficiency was associated with an increase in aortic CD8⁺ Tregs (CD8⁺CXCR5⁺CD275⁺CD122⁺ T cells) in Stat4^−/−Ldlr^−/− mice fed a DDC diet for 16 wk in comparison with Ldlr^−/− controls (Fig. 4A). Furthermore, STAT4 deficiency led to decreased number of Tfh cells within the aorta of Stat4^−/−Ldlr^−/− mice versus Ldlr^−/− mice (Fig. 4B). These results indicated a potential STAT4-dependent fine-tuning of the balance of CD8⁺ Tregs and Tfh cells in the aortic wall during atherogenesis.

MΦ numbers strongly correlate with the severity of arterial inflammation and atherogenesis in mice and humans. Therefore, we sought to determine whether the presence of MΦs was also affected by STAT4 deficiency in Stat4^−/−Ldlr^−/− aortas. We detected a significant reduction in the percentage of CD11b⁺F4/80⁺ cells in the aorta of Stat4^−/−Ldlr^−/− mice versus Ldlr^−/− mice (p < 0.01, Fig. 4C), with a trending decrease in the number of MΦs (CD11b⁺F4/80⁺) in Stat4^−/−Ldlr^−/− aortas (Fig. 4C). Altogether, STAT4 deficiency reduces the numbers of Tfh cells and MΦs within the aortic wall of Stat4^−/−Ldlr^−/− mice, but it supports the presence of CD8⁺ Tregs within the aorta.

The initiation and progression of atherosclerosis are accompanied by a complex immune response, and MΦs play a key role in the modulation of arterial and systemic inflammatory processes. In this study, we sought to evaluate MΦ content and activation status in the IR Ldlr^−/− mice model and examine how MΦs might affect CD8⁺ Treg and Tfh cell differentiation. We detected a similar reduction in the percentages (data not shown) and numbers of CD11b⁺CD68⁺Ly6C⁺ (Fig. 5A) and CD11b⁺CD68⁺MHC-II⁺ proinflammatory MΦs (Fig. 5B) in the spleens of Stat4^−/−Ldlr^−/− mice in comparison with Ldlr^−/− mice. To further investigate the potential impact of STAT4 deficiency on MΦs, we examined the gene expression of phenotypic markers for M1 MΦs and M2 MΦs (Nos2 and Mrc1, respectively). The gene expression by Ldlr^−/− MΦs was set at 1, and gene expression by Stat4^−/−Ldlr^−/− MΦs was expressed relative to Ldlr^−/− MΦs. As shown in Fig. 5C, we found an increase in the expression of the anti-inflammatory M2 MΦ marker Mrc1 in Stat4^−/−Ldlr^−/− mice compared with Ldlr^−/− mice (p < 0.01). Thus, our results suggest that STAT4 deficiency results in a reduced population of splenic MΦs, and these MΦs display attenuated activation.

Tfh cell differentiation is regulated by IL-6, IL-21, and IL-27 via a set of transcription factors, including Bcl6, c-Maf, Batf, IRF4, and, likely, STAT4 (26). In contrast, TGF-β and STAT5 serve as negative regulators of Tfh cell differentiation. Although the major transcription factors are well characterized for Tfh cell differentiation, the transcriptional network and cytokines that drive Tfh cell differentiation in atherosclerosis have not been studied. We decided to focus on STAT4-dependent MΦ-derived cytokines and their ability to support Tfh cell differentiation. First, we tested whether MΦs isolated from DDC-fed Stat4^−/−Ldlr^−/− and Ldlr^−/− mice display a different cytokine profile upon restimulation with LPS. The analysis was focused on cytokines that are known to be responsible for Tfh cell (IL-27, IL-6, TGF-β) and CD8⁺ Treg (TGF-β) differentiation (18, 28, 29). In Fig. 5D, the ratio represents the relative expression of the analyzed genes by Stat4^−/−Ldlr^−/− MΦs versus Ldlr^−/− MΦs (set as 1). RT-PCR analysis detected no difference in the expression of Il27 or Il6 between MΦs from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice, but it revealed reduced levels of IFN-γ expression by STAT4-deficient MΦs (Fig. 5D). Importantly, STAT4-deficient peritoneal MΦs demonstrated an increased expression of Tgfβ (Fig. 5D).

To further explore the effects of STAT4-deficient MΦs on the differentiation of T cell subsets, we analyzed how conditioned media from STAT4-deficient and STAT4-sufficient MΦs isolated from Stat4^−/−Ldlr^−/− and Ldlr^−/− mice might affect T cell differentiation. CD4⁺ Ldlr^−/− T cells were cultured for 48 h and then stained with hallmark transcription factors for Th1 (T-bet) and Treg (Foxp3). The differentiation into CD4⁺Tbet⁺ or CD4⁺Foxp3⁺ cells with Ldlr^−/− MΦ media was set at 1, and the differentiation of CD4⁺Tbet⁺ or CD4⁺ Foxp3⁺ cells with Stat4^−/−Ldlr^−/− MΦ media was expressed relative to the effects of Ldlr^−/− MΦs (Fig. 5E). We found that the conditioned media from Stat4^−/−Ldlr^−/− MΦs stimulated less differentiation into CD4⁺Tbet⁺ Th1 cells; however, surprisingly, it had no significant effect on CD4⁺Foxp3⁺ Treg differentiation (Fig. 5E). Next, we analyzed the impact of conditioned media from activated Stat4^−/−Ldlr^−/− and Ldlr^−/− MΦs on the generation of CD4⁺CXCR5⁺ Tfh cells. We used the levels of Tfh cell differentiation in conditioned media from Ldlr^−/− MΦs as basal levels of Tfh cell generation (set up as 1) and calculated a relative ratio of Tfh cells generated under different conditions (Fig. 5F). TGF-β is a cytokine that has been shown to inhibit differentiation of mouse Tfh cells (15, 28) and support CD8⁺ Treg differentiation (28, 30). Therefore, anti–TGF-β blocking Abs were used to test a role for TGF-β in Tfh cell differentiation for some experiments. Our data revealed a reduction in Tfh cell differentiation using conditioned media from LPS-stimulated Stat4^−/−Ldlr^−/− MΦs (Fig. 5F). This reduction in Tfh cell differentiation was TGF-β dependent, because blocking anti-TGF-β Abs restored the rate of Tfh cell generation (Fig. 5F).

Next, we tested whether the condition media from Stat4^−/−Ldlr^−/− MΦs could modulate differentiation of CD8⁺ Tregs. The differentiation of CD8⁺ Tregs in media from Ldlr^−/− peritoneal MΦs was set as 1. We detected a significant increase in CD8⁺ Treg differentiation upon coculture of CD8⁺ cells with Stat4^−/−Ldlr^−/− MΦ-derived media, and this effect was reduced by the coculture with anti–TGF-β Abs (Fig. 5G). Thus, STAT4 deficiency in MΦs results in an altered cytokine profile that, at least via low levels of TGF-β expression, affects the generation of Tfh cells and CD8⁺ Tregs.

Our data demonstrate that STAT4 directly affects CD8⁺ Treg differentiation in a cell-intrinsic manner, as well as modulates differentiation of CD8⁺ Tregs indirectly, via MΦ-released cytokines. To further examine the direct impact of STAT4 deficiency on CD8⁺ Treg functions, we tested the ability of STAT4-deficient CD8⁺ Tregs to control Tfh cell differentiation in vivo. FACS-sorted STAT4-deficient (knockout [KO]) or STAT4-sufficient (wild-type [WT]) CD8⁺CD122⁺ Tregs (CD45.2⁺) were adoptively transferred into C57BL/6 (CD45.1⁺) mice, followed by immunization with KLH in CFA. This is a well-established protocol that stimulates generation of Tfh cells and germinal center B cells in vivo (14). Compared with mice receiving no cells (control) or STAT4-sufficient (WT) CD8⁺CD122⁺ Tregs, the recipients of STAT4-deficient CD8⁺CD122⁺ Tregs (KO) exhibited a greatly decreased percentage of CD4⁺CXCR5⁺Bcl6⁺ Tfh cells (Fig. 6A). The transfer of Stat4^−/− CD8⁺CD122⁺ Tregs also inhibited generation of B220⁺GL-7⁺Fas⁺ germinal center B cells and KLH-specific Ab production, including IgG, IgM, and IgA (Fig. 6B). These results provide, to our knowledge, a first proof-of-principle for a cell-intrinsic role for STAT4 in the negative regulation of CD8⁺ Treg functions, such as control of Tfh cell development and germinal center B cell–dependent humoral immune responses in vivo.

Our data clearly indicate that STAT4-deficient CD8⁺ Tregs suppress the formation of germinal centers upon immunization with KLH via suppression of Tfh cell differentiation in immunized C57BL/6 recipients (Fig. 6). To further investigate the impact of CD8⁺ Treg–specific STAT4 deficiency in atherogenesis, we performed adoptive transfer of sorted CD8⁺ Tregs from Stat4^−/−Ldlr^−/− or Ldlr^−/− mice into 8-wk-old Ldlr^−/− recipients. To evaluate basal atherosclerotic plaque formation, we used 8-wk-old Ldlr^−/− recipients with no injection of CD8⁺ Tregs (PBS control). After 11 wk of DDC feeding, all recipient mice were analyzed for plaque burden in aortas. On average, Ldlr^−/− mice fed a DDC for 11 wk developed 13.7 ± 0.5% plaque throughout the aorta (n = 6). In agreement with the previous report by Clement et al. (13), the adoptive transfer of WT CD8⁺CD122⁺ Tregs reduced plaque formation in Ldlr^−/− recipients versus plaque formation in PBS-injected Ldlr^−/− recipients, indicating a suppressive atheroprotective role for CD8⁺ Tregs in atherosclerosis (13.7 ± 0.5 versus 10.1 ± 1.2%, p < 0.05). Importantly, the adoptive transfer of Stat4^−/−Ldlr^−/− (KO) CD8⁺CD122⁺ Tregs resulted in a further reduction in plaque burden in Ldlr^−/− recipients in comparison with the plaque development in Ldlr^−/− mice that received Ldlr^−/− (WT) CD8⁺CD122⁺ Tregs (3.6 ± 1.2 versus 10.1 ± 1.2%, respectively, p < 0.05, Fig. 7A). Transferred Stat4^−/−Ldlr^−/− CD8⁺ Tregs also significantly diminished the percentage of CD4⁺CXCR5⁺Bcl6⁺ Tfh cells in the spleens and para-aortic LNs of Ldlr^−/− recipients in comparison with Ldlr^−/− mice that received Ldlr^−/− CD8⁺ Tregs (Fig. 7B). These results suggest that STAT4-deficient CD8⁺ Tregs impact the local LN and also effect a systemic immune response in Ldlr^−/− recipients. The transfer of Stat4^−/−Ldlr^−/− CD8⁺ Tregs also suppressed the generation of B220⁺GL-7⁺FAS⁺ germinal center B cells in para-aortic LNs of Ldlr^−/− recipients (Fig. 7B, 7C). In parallel, attenuated levels of circulating IgG1, IgG3, and IgA Abs were detected in the plasma of Ldlr^−/− mice that received Stat4^−/−Ldlr^−/− CD8⁺ Tregs versus Ldlr^−/− CD8⁺ Tregs (Fig. 7D). Overall, these results clearly demonstrate an important cell-intrinsic role for STAT4 in the regulation of CD8⁺ Treg functions and provide strong evidence for the importance of the STAT4-dependent role of CD8⁺ Tregs in the regulation of Tfh cell and germinal center B cell development in atherosclerosis.

STAT4 is a major transcription factor regulating adaptive immune responses, but it also plays a detrimental role in several autoimmune diseases (6). In this study, we investigate a role for STAT4 in a complex model of atherogenesis under conditions of IR. In this article, we demonstrate that STAT4 deficiency results in reduced plaque formation at early and advanced stages of atherosclerosis caused by a cholesterol-containing diabetogenic diet resulting in concomitant IR. A key novel result from this study is that we show unexpected effects of STAT4 deficiency on the modulation of CD8⁺ Treg and Tfh cell content in atherosclerosis-prone conditions. STAT4 deficiency supports CD8⁺ Treg differentiation in vitro and CD8⁺ Treg functions in vivo, suggesting CD8⁺ cell-intrinsic effects of STAT4. Additionally, STAT4 deficiency affects T cell fate, and it also directly diminishes the proinflammatory MΦ phenotype that, in turn, supports CD8⁺ Treg differentiation and a partial abolishment of Tfh cell development via a TGF-β–dependent mechanism. Altogether, our data uncover a proinflammatory role for STAT4 in IR-accelerated atherosclerosis and highlight a new exciting role for STAT4 in the regulation of the Tfh/CD8⁺ Treg axis.

Numerous reports indicate an association among IR, hyperlipidemia, glucose intolerance, hypertension, obesity, and cardiovascular disease (31, 32). One of the limitations in this area of investigation is a restricted number of mouse models that can be used for studying the proatherogenic effects of IR or type 2 diabetes (4). To overcome this limitation, Schreyer et al. (24) established a combined model of IR and atherosclerosis; later, Subramanian et al. (5) further characterized DDC-fed Ldlr^−/− mice with the specific focus on atherosclerosis and adipose tissue inflammation. In this study, we used this model to study the effects of STAT4 deficiency on IR-accelerated atherogenesis. Our results demonstrate that STAT4 is a powerful proatherogenic transcription factor in the conditions of IR, because Stat4^−/−Ldlr^−/− mice show significantly attenuated atherosclerotic plaque burden after 16 or 24 wk of DDC feeding.

It has been established that STAT4 and T-bet are necessary for full differentiation of Th1 cells (7, 12, 33). However, recent studies have shown that STAT4 is also involved in the regulation, differentiation, and activation of several cell types, including NK cells, monocytes, MΦs, dendritic cells, and Th cell lineages (9, 34, 35). In this study, in the model of IR-accelerated atherosclerosis, we detected a reduction in the number and activation of splenic MΦs and CD3⁺ T cells, as well as a diminished number of MΦs and Tfh cells within the aorta. In contrast, the CD8⁺ Treg population was elevated in Stat4^−/−Ldlr^−/− mice in comparison with Ldlr^−/− controls. Thus, our data indicate that STAT4 regulates the systemic proinflammatory response via the modulation of various immune cells in atherosclerosis.

CD8⁺ Tregs that are specific for the Qa-1 molecule, an unconventional MHC class I molecule in mice, control autoreactive CD4⁺ T cells and Qa-1⁺ Tfh cells (20). A large body of evidence suggests an important role for CD8⁺ Tregs in human autoimmune diseases (36), and a recent study also highlights roles for CD8⁺ Tregs and the Tfh/germinal center B cell axis in atherosclerosis (13). In this study, we underscore a critical role for STAT4 in the regulation of the CD8⁺ Treg/Tfh cell axis in atherogenesis. We demonstrate that STAT4 plays a critical role in the formation and suppressive functions of CD8⁺ Tregs during atherogenesis.

A transcriptional network that directs the differentiation of CD8⁺ Tregs has not been identified. There is evidence that Helios and STAT5 are required for CD8⁺ Treg differentiation and survival (26). Our results indicate that STAT4 deficiency leads to an increase in splenic and aortic CD8⁺ Tregs that is accompanied by a reduction in Tfh cells and plasma B cells in DDC-fed Stat4^−/−Ldlr^−/− mice. Furthermore, we present data suggesting that STAT4 can support CD8⁺ Treg differentiation from CD8⁺ cells in cultures under in vitro Treg-polarizing conditions, suggesting cell-intrinsic effects of STAT4 in the CD8⁺ cell phenotype. It is well established that STAT4 suppresses Foxp3 expression and, therefore, CD4⁺ Treg differentiation in vitro (35). The expression of Foxp3 in CD8⁺ Tregs is the subject of much debate (23). Future studies will be necessary to dissect STAT4-dependent molecular signaling pathways that drive CD8⁺ Treg differentiation. Importantly, the effects of STAT4 deficiency are not restricted to the regulation of CD8⁺ Treg differentiation, because we demonstrate that STAT4 deficiency also affects CD8⁺ Treg functions. Adoptively transferred STAT4-deficient CD8⁺ Tregs significantly suppress the development of Tfh cells, germinal center B cells, and a subsequent synthesis of Igs in KLH-immunized C57BL/6 recipients. These results provide a first proof-of-principle for a cell-intrinsic role for STAT4 in CD8⁺ Treg functions. In this article, we also demonstrate that STAT4-deficient CD8⁺ Tregs effectively suppress differentiation of Tfh cells and germinal B cell formation upon atherogenesis. Importantly, the adoptive transfer of STAT4-deficient CD8⁺ Tregs reduced atherosclerotic plaque formation to a considerable degree, indicating the functional importance of STAT4 deletion in CD8⁺ Tregs. Altogether, we demonstrate that STAT4 is a transcription factor that is crucial for the differentiation and functions of CD8⁺ Tregs in vivo under inflammatory conditions.

Tfh cell differentiation is regulated by IL-6 and IL-21 via Bcl6, which is required for Tfh cell commitment, and STAT3 is likely also involved in Tfh cell fate decision (26). Although the major transcription factors are well characterized for Tfh cell differentiation, the transcriptional network and cytokines that drive Tfh cell differentiation in atherosclerosis are under active investigation. There have been studies that indicate inverse roles for IL-27 and TGF-β in the generation and function of mouse Tfh cells (26). IL-27 is a cytokine that enhances the functionality of Tfh cells within the murine model of lupus (26, 28). In contrast, Tfh cells display reduced differentiation in the presence of TGF-β (18). Interestingly, mouse studies have indicated a strong correlation between the activation of the IL-12/STAT4 pathway and the reduction in TGF-β in chronic autoimmune diseases, such as rheumatoid arthritis (37).

MΦs are the central cells in atherosclerosis, and they regulate the development of the disease in multiple ways, such as modulation of cholesterol metabolism and production of immunomodulatory chemokines and cytokines. Because IL-27 and TGF-β expression is important for Tfh cell differentiation, we decided to further test a role for MΦ-derived cytokines and examine the levels of IL-27 and TGF-β expression in activated MΦs isolated from insulin-resistant Stat4^−/−Ldlr^−/− mice. We did not observe any difference in the expression level of Il27 upon STAT4 deficiency; however, we did detect an increase in the expression of Tgfβ in STAT4-deficient MΦs. Next, to uncover the potential impact of STAT4-dependent MΦ functions on the regulation of T cell polarization, we stimulated isolated Ldlr^−/− CD4⁺ T cells with the media collected from STAT4-sufficient and STAT4-deficient LPS-activated peritoneal MΦs. Ldlr^−/− CD4⁺ T cells plated in media collected from STAT4-deficient LPS-activated peritoneal MΦs had a reduced ability to differentiate into Tfh cells.

Interestingly, several studies have also revealed that TGF-β is necessary for the development of CD8⁺ Tregs (26). In line with this notion, we detected preferential CD8⁺ Treg generation using STAT4-deficient MΦ-released media. These results suggest that the loss of MΦ-specific STAT4 alters MΦ activation, increases TGF-β production, and fine-tunes the content of Tfh cells and CD8⁺ Tregs. Thus, STAT4 might affect the CD8⁺ Treg/Tfh cell axis in two ways: directly via the regulation of CD8⁺ Treg differentiation/functions and Tfh cell suppression and indirectly via STAT4-dependent MΦ-dependent effects on CD8⁺ Treg and Tfh cell differentiation.

In summary, STAT4 participates in atherogenesis via the support of proinflammatory MΦ activities, regulation of the CD8⁺ Treg/Tfh cell axis, and modulation of the local immune response in the aortic wall under conditions of IR and atherosclerosis. Notably, our data uncovered a key role for STAT4 in the negative regulation of CD8⁺ Treg differentiation and suppressive functions in vivo. The obtained results suggest that modulation of STAT4 expression could provide a novel therapeutic approach to reducing accelerated atherosclerosis associated with IR and diabetes.

We thank the Eastern Virginia Medical School Flow Cytometry Facility and Breanne Gjurich for technical assistance and Ciriaco Villafor for expert animal husbandry.

The authors have no financial conflicts of interest.