CD28 Individual Signaling Up-regulates Human IL-17A Expression by Promoting the Recruitment of RelA/NF-κB and STAT3 Transcription Factors on the Proximal Promoter

CD28 is an important co-stimulatory receptor for T lymphocytes that, in humans, delivers TCR-independent signal leading to the up-regulation of pro-inflammatory cytokines. We have recently reported that CD28 autonomous signaling induces the expression of IL-17A in peripheral CD4+ T lymphocytes from healthy donors, multiple sclerosis, and type 1 diabetes patients. Due to the relevance of IL-17A in the pathophysiology of several inflammatory and autoimmune diseases, we characterized the mechanisms and signaling mediators responsible for CD28-induced IL-17A expression. Here we show that CD28-mediated up-regulation of IL-17A gene expression depends on RelA/NF-κB and IL-6-associated STAT3 transcriptions factors. In particular, we found that CD28-activated RelA/NF-κB induces the expression of IL-6 that, in a positive feedback loop, mediates the activation and nuclear translocation of tyrosine phosphorylated STAT3 (pSTAT3). pSTAT3 in turn cooperates with RelA/NF-κB by binding specific sequences within the proximal promoter of human IL-17A gene, thus inducing its expression. Finally, by using specific inhibitory drugs, we also identified class 1A phosphatidylinositol 3-kinase (PI3K) as a critical upstream regulator of CD28-mediated RelA/NF-κB and STAT3 recruitments and trans-activation of IL-17A promoter. Our findings reveal a novel mechanism by which human CD28 may amplify IL-17A expression in human T lymphocytes and provide biological bases for immunotherapeutic approaches targeting CD28-associated class 1A PI3K to dampen IL-17A-mediated inflammatory response in autoimmune/inflammatory disorders.


INTRODUCTION
IL-17-producing cells constitute a pro-inflammatory Th17 cell subset that plays critical roles in adapting immune response against extracellular pathogens and, more importantly, in the pathophysiology of several inflammatory and autoimmune diseases (1)(2)(3). The hallmark of Th17 cells is the production of the pro-inflammatory cytokines, IL-17A-F (4,5). In particular IL-17A affects the functions of a wide range of cells. For instance, IL-17A enhances the secretion of pro-inflammatory cytokines and chemokines, favors neutrophil infiltration and increases the production of defensin and mucin by epithelial cells (6)(7)(8)(9). Several key transcription factors, the most reliable RORC (10), cytokines, such as IL-6, TGFβ, IL-21, IL-1β, IL-23, and associated signaling mediators have been described to promote IL-17 expression and Th17 cell differentiation in both human and mouse (11)(12)(13)(14)(15). However, while the role of cytokines and associated signaling mediators regulating Th17 differentiation and IL-17 expression has been better defined in mouse (16)(17)(18)(19)(20), in humans is still controversial (14,21,22). Moreover, key co-stimulatory molecules regulating IL-17 expression are continuingly to be identified and may be useful for the development of new antigen non-specific immunosuppressive therapies for immune based diseases.
CD28 is an important co-stimulatory receptor that cooperates with TCR for optimal T cell activation and differentiation. CD28 contribution to IL-17 expression and Th17 differentiation has been extensively analyzed in the contest of TCR stimulation, with discrepant results depending on the strength of TCR activation, the conditioning cytokines and between human and mouse (23)(24)(25)(26)(27)(28)(29)(30)(31). For instance, in the mouse system, CD28 co-stimulation or associated signaling molecules have been described to either sustain (25,26,31) or repress TCRmediated Th17 differentiation and functions (29,30). Similar contrasting results have been found in humans, where CD28 co-stimulatory signals in combination with TCR activation induce (23,28) or suppress Th17 differentiation depending on IL-1β and IL23 conditioning cytokines (27). Moreover, upon stimulation with agonistic antibodies (Ab) or with its B7.1/CD80 or B7.2/CD86 ligands expressed on the surface of professional antigen presenting cells (APC), human CD28, but not mouse CD28, is also able to act as a unique signaling receptor and to arise TCR-independent pro-inflammatory signals (32)(33)(34). Indeed, human CD28 stimulation in the absence of TCR engagement activates a NF-κB pathway in peripheral CD4 + T cells, thus leading to the expression and production of pro-inflammatory cytokine/chemokines (32,35) and triggers Th17 cells to produce IL-17A (36). This CD28 pro-inflammatory activity is particularly relevant in the context of inflammatory diseases, such as multiple sclerosis (MS) and type 1 diabetes (T1D), where CD28 individual stimulation may amplify the inflammatory response by upregulating cytokines related to the Th17 cell profile phenotype, such as IL-6, IL-21, and IL-17A (37,38). These data suggest a role of CD28 in regulating the amplification of Th17 cells in inflammatory/autoimmune diseases. For instance, several mouse models of human inflammatory/autoimmune diseases, such as autoimmune diabetes in non-obese diabetic (NOD) mice, MS in experimental autoimmune encephalomyelitis (EAE) mice or systemic autoimmune disorders have evidenced the relevance of CD28 co-stimulatory signals (39,40).
In the present study, we characterized the molecular mechanisms and signaling mediators regulating IL-17A expression in response to CD28 individual ligation. We found that in human CD4 + T cells, CD28 induced the expression of IL-6 in a NF-κB-dependent manner. IL-6 in turn acted in a positive feedback loop by inducing the activation and the nuclear translocation of tyrosine phosphorylated STAT3 (pSTAT3). pSTAT3 in synergy with RelA/NF-κB bound specific sequences within the proximal promoter of human IL-17A gene and induced its expression. Finally, by using specific inhibitory drugs, we also identified class 1A phosphatidylinositol 3-kinase (PI3K) as the upstream regulator of CD28 signals regulating IL-17A expression.

Cells Abs and Reagents
Human primary CD4 + T cells were enriched from PBMC by negative selection using an EasySep TM isolation kit (#17952, STEMCELL Technology) and cultured in RPMI-1640 supplemented with 5% human serum (Euroclone, UK), L-glutamine, penicillin and streptomycin. The purity of the sorted population was 95-99%, as evidenced by staining with anti-CD3 plus anti-CD4 Abs. Human naïve CD45RA + and effector/memory CD45RO + were enriched from purified CD4 + T cell by positive and negative selection, respectively, using a MACS anti-Phycoerythrin (PE) microbeads sorting kit (Miltenyi Biotec, Milan, Italy) after labeling with a PEconjugated anti-CD45RA primary antibody (#130-098-184, Miltenyi Biotec, Milan, Italy). PBMCs were derived from buffy coats or anonymous healthy blood (HD) donors provided by the Policlinico Umberto I (Sapienza University of Rome, Italy). Written informed consent was obtained from blood donors and both the informed consent form and procedure was approved by the Ethics Committee of Policlinico Umberto I.

Plasmids Cell Transfection and Luciferase Assays
The NF-κB luciferase gene under the control of six thymidine kinase NF-κB sites (43) was kindly provided by J. F. Peyron (Centre Méditerranéen de Médecine Moléculaire, Nice, France). The pGL3E-hIL-17prom(-1125)-Luc construct containing the luciferase construct under the control of the 1,125 bp region upstream of the transcriptional start site of IL-17A gene (44) was from Addgene (USA). pcDNA3 expressing HA-tagged RelA and RelB have been previously described (45). Constitutive active pcDNA-STAT3C-flag construct (46) has been kindly provided by V. Poli (University of Torino, Turin, Italy).

Cytokine Production
Secretion of human IL-6 and IL-17A was measured from the supernatants of CD4 + T cells cultured for 24 h or 48 h in flat bottom 48 culture wells (3 × 10 5 cells per well) either un-stimulated or stimulated with crosslinked anti-CD28.2 Ab (2 µg ml −1 ) by using human IL-6 (#HS HS600B) and IL-17A (#D1700) ELISA kits (R&D Technology). Data were analyzed on a Bio-Plex (Bio-Rad, Hercules, CA, USA). The assays were performed in duplicate. The sensitivity of the assay was 0.11 pg ml −1 for IL-6 and 15 pg ml −1 for IL-17A.

Real-time PCR
Total RNA was extracted using Trizol (Thermo Fisher Scientific CA, USA) from 2 × 10 6 cells and RNeasy MicroKit (#74004, Qiagen) from 5 × 10 5 cells according to the manufacturer's instructions and was reverse-transcribed into cDNA by using Moloney murine leukemia virus reverse transcriptase (Invitrogen). TaqMan Universal PCR Master Mix, human IL-6, IL-17A, TGFβ, RORC, and GAPDH primer/probe sets were purchased from Applied Biosystems. The relative quantification was performed using the comparative C T method. The median value of human CD4 + T cell stimulated with control isotype matched Ab was used as C T calibrator in all comparative analyses.

Cytotoxicity Assay
The cytotoxicity of inhibitory drugs on CD4 + T cells was evaluated by propidium iodide (PI) staining (10 µg ml −1 ). CD4 + T cells were plated at 2 × 10 6 cells/ml in 48-well plates and treated with the indicated doses of inhibitory drugs or DMSO, as vehicle control, for 24 h. Cytotoxicity was analyzed by a BD Biosciences FACScalibur (Mountain View, CA) by quantifying the percentage of PI positive cells. Results were calculated from at least three independent experiments.

Statistical Analysis
The sample size was chosen based on previous studies to ensure adequate power. Parametrical statistical Frontiers in Immunology | www.frontiersin.org analysis (mean and SD) was performed to evaluate differences between continuous variables through Prism 5.0 (GraphPad Software, San Diego, CA) using unpaired Student t-test. For multiple group comparisons, significant differences were calculated using the non-parametric Mann-Whitney U-test, and linear regression analyses were performed using the Pearson chi-squared test. For all tests, p < 0.05 were considered significant.

CD28 Stimulation in the Absence of TCR Engagement Up-regulates IL-17A Expression in a IL-6-dependent Manner
We have recently found that CD28 stimulation induces the expression of IL-17A in healthy donors (HD), MS and T1D patients (37,38). In order to better characterize the molecular mechanisms of CD28-mediated IL-17A expression, we performed a detailed kinetic analysis of IL-17A gene expression and secretion by stimulating human CD4 + T cells from HD with an agonistic anti-CD28 Ab (CD28.2) that has been described to bind the same epitope recognized by B7 molecules (48). CD28 stimulation by agonistic anti-CD28.2 Ab of CD4 + T cells from HD induced IL-17A gene expression within 6 h ( Figure 1A) that further increased 24-48 h (Figures 1A,B) and decreased 72 h after stimulation ( Figure 1B). CD28-induced IL-17A gene expression was also associated with a strong increase of IL-17A cytokine secretion after 48 h from stimulation ( Figure 1C). As we have previously observed for other pro-inflammatory cytokines (33), CD28-induced IL-17A expression was not related to the preferential stimulation of effector/memory T cells, since no significant differences in IL-17A gene expression were observed upon stimulation of naïve (CD45RA, Figures S1A,S1C) or effector/memory (CD45RO, Figures S1B,S1C) CD4 + T cells with anti-CD28 Abs ( Figure S1D). Furthermore, the up-regulation of IL-17A expression (Figures 1D,E) was strongly dependent on the intrinsic signaling capability of human CD28, since CD3 stimulation alone was not able to up-regulate IL-17A gene expression ( Figure S1E) and no significant differences in IL-17A mRNA levels were observed when CD3 and CD28 were coengaged compared to CD28 individual stimulation ( Figure 1E). On the contrary, a high up-regulation of IL-2 mRNA was detected only in CD3 plus CD28-stimulated human CD4 + T cells ( Figure 1F).
These data evidence that CD28-induced IL-6 cooperates in a positive feedback loop with CD28 signaling in mediating IL-17A gene expression and secretion.
To verify the physiological involvement of STAT3 and RelA in CD28-mediated trans-activation of hIL-17A promoter, we analyzed pSTAT3 and RelA recruitment on the IL-17A promoter in ex vivo CD4 + T cells stimulated with anti-CD28.2 Ab. To this aim we used ChIP assays. The kinetic analysis of both pSTAT3 and RelA recruitment on hIL-17A proximal promoter performed on CD4 + T cells from one HD evidenced an increase within 3-6 h with a maximum around 24 h and a decrease after 36 h (Figure 6A). The recruitment of both transcription factors was associated to CD28-induced transcriptional activation of IL-17A promoter, as evidenced by RNA polymerase II (pol II) promoter occupancy (Figure 6A). The results obtained from a larger sample size (n = 3) showed that pSTAT3 ( Figure 6B) and Pol II ( Figure 6D) were significantly recruited on the IL-17A promoter within 3 h from stimulation and persisted over 6-24 h, whereas a significant recruitment of RelA was observed after 24 h (Figure 6C). Consistently with the absence of RelB nuclear translocation (Figure 4A), no significant recruitment of RelB on the hIL-17A promoter was observed in CD28-stimulated T cells also after 24 h of stimulation compared to RelA and pSTAT3 ( Figure 6E).
Altogether these data demonstrate that CD28-induced IL-17A gene expression is mediated by the cooperative recruitment and transcriptional activity of pSTAT3 and RelA on the IL-17A promoter.
Altogether these data support a pivotal role of CD28associated PI3K in the transcriptional activation of IL-17A gene mediated by pSTAT3 and RelA.

DISCUSSION
The immunopathogenesis of several inflammatory and autoimmune diseases relies on both the amplification and persistence of pro-inflammatory IL-17-producing cells (13). Therefore, the characterization of the mechanisms and molecules regulating IL-17 expression could represent an important goal of the ongoing research in inflammation and autoimmunity.
The mechanisms regulating the differentiation of IL-17producing cells have been largely elucidated in mice, where IL-17 can be induced by a combination of IL-6 or IL-21 and TGFβ co-signaling (16)(17)(18)(19)(20). In contrast to mouse T cells, in humans, several discrepancies on the factors driving and/or amplifying IL-17 expression have been reported (78). Some groups found that TGFβ alone or in combination with IL-21, IL-23, or inflammatory cytokines is required for RORγt expression and human IL-17A expression (51)(52)(53). Others evidenced that TGFβ is not essential for human Th17 differentiation and can be substituted by IL-1β (49,(79)(80)(81). More recent data by Revu et al. showed that IL-23 and IL-1β promote IL-17 production and human Th17 differentiation in the presence of TCR engagement and in the absence of CD28 stimulation, thus confirming also in the human system (27) previous data on a suppressive role of murine CD28 in Th17 cell differentiation (28,29). Herewith, we evidence a novel role of human CD28 in inducing IL-17A expression and production in the absence of TCR engagement and conditioning cytokines (Figure 1). The transcriptional activation of IL-17A in response to CD28 was independent of either TGFβ or RORγτ , since no up-regulation of neither TGFβ (Figure 2D) nor RORC (Figure 2E) was The ability of human CD28 to induce the expression of IL-6 and other inflammatory cytokines has been amply described (33-35, 37, 38). The kinetic analysis of CD28induced IL-6 expression evidenced that IL-6 gene expression and ChIPs performed on CD4 + T cells from HD (n = 4) stimulated for the indicated times with isotype control or anti-CD28.2 Abs. Specific enrichment over isotype control Abs was calculated by the Cτ method. Data express the mean ± SD of four independent experiments. Significance was calculated by Student t-test. *p < 0.05, **p < 0.01, NS, not significant. (E) Real time PCR for IL-17A promoter from anti-RelA, anti-RelB, and anti-pSTAT3 ChIPs performed on CD4 + T cells from HD (n = 6) stimulated for 24 h with isotype control or anti-CD28.2 Abs. Specific enrichment over negative control Abs (anti-Lck) was calculated by the Cτ method and data expressed as fold inductions (F.I.) over isotype control Ig-stimulated cells. Data express the mean F.I. ± SD. Statistical significance was calculated by Student t-test. *p < 0.05, **p < 0.01, NS, not significant. Mean ± SD: RelA = 6.7 ± 4.8, RelB = 1 ± 0.4, pSTAT3 = 9 ± 5.9. secretion occurred earliest (Figures 2A-C) compared to IL-17A (Figures 1A-D) and was dependent on CD28-induced NF-κB activity ( Figure 4E). These data are consistent with those from Serada et al. showing that IL-6 blockade inhibited myelin-specific Th17 cells in vivo (82). Since IL-6 may regulate IL-17A gene expression by activating IL-6 receptor-associated STAT3 (19,(55)(56)(57), these data strongly support a role of IL-6-induced STAT3 activation in regulating CD28-mediated IL-17A expression. For instance, a strong and persistent increase of STAT3 phosphorylation (over 24 h) was observed in CD4 + T cells within 3 h from CD28 stimulation ( Figure 3A) and CD28-induced IL-17A expression was abrogated by the selective STAT3 inhibitor, S31-201 (Figures 3C,D). STAT3 is an important transcription factor for both murine and human Th17 cell differentiation. Deletion of STAT3 in T cells abrogates Th17 differentiation (56,83), whereas the overexpression of constitutively active STAT3 is sufficient for inducing the development of IL-17-producing cells (84). More recent data in mouse and human systems confirmed a critical role of STAT3 in regulating the transcription of several genes involved in Th17 differentiation and functions (57,67). Several STAT3 binding sites have been identified throughout the IL-17A promoter (11,56,85) and in the proximal 1-kb region upstream of the human IL-17A gene (44) (Figure 5A). For instance, we found that CD28 stimulation by either agonistic Ab or its natural ligand B7.1 strongly up-regulated the 1-kb proximal human IL-17 promoter in a STAT3-dependent manner ( Figure 5). These data are in contrast with those from Liu et al. who did not find any transactivation of the proximal IL-17A promoter following CD28 stimulation alone (44). These discrepancies could be related to the distinct binding properties and stimulating activities of the anti-CD28 Ab used by Lin et al. compared to the natural B7 ligand or the agonistic anti-CD28.2 Ab (86). Moreover, our ChIP data showing the binding of STAT3 to the proximal human IL-17A promoter ( Figure 6B) support a direct function of STAT3 in regulating IL-17A gene expression in CD28-stimulated CD4 + T cells.
The pro-inflammatory properties of human CD28 rely on its unique intrinsic capability to activate NF-κB (32)(33)(34)87). The NF-κB family consists of five members including NF-κB1 (p50 and its precursor p105), NF-κB2 (p52 and its precursor p100), RelA (p65), RelB, and c-Rel. RelA, c-Rel and RelB contain a transactivation domain and form transcriptionally active heterodimers in association with p50 and/or p52 (88). Among the NF-κB subunits, c-Rel and RelA have been described   (41,90). As previously demonstrated (35), CD28 stimulation induces a persistent nuclear translocation of RelA, but not RelB (Figures 4A, 7A) nor c-Rel (35). CD28-induced RelA/NF-κB activity was crucial for both IL-6 expression ( Figure 4G) and IL-6-associated STAT3 activation (Figures 4D,E) as well as for IL-17A expression ( Figure 4H). The ability of CD28 autonomous stimulation to induce IL-17A production was also evidenced by Santarlasci et al., who demonstrated that CD28 stimulation alone triggers Th17 T cell clones to produce IL-17A in a NF-κB-dependent manner (36). We extended these data by evidencing the dual pivotal role of RelA/NF-κB in CD28-mediated up-regulation of IL-17A expression. Early during CD28-stimulation, RelA/NF-κB mediates the transcription of IL-6 and subsequent IL-6dependent STAT3 activation (Figures 4D,F). Later, persistent RelA/NF-κB subunits bind the NF-κB consensus site at position−643 of the proximal human IL-17A promoter (44), thus cooperating with STAT3 for inducing IL-17A transcription (Figures 5, 6). For instance, an optimal recruitment of both STAT3 and RelA on the human IL-17A promoter was observed 24 h after CD28 stimulation (Figure 6). Interestingly, Whitley et al. have recently shown that, in mouse T cells, RelA/NF-κB regulates IL-17 expression in cooperation with STAT3 by binding specific regulatory elements. However, in contrast to human IL-17A gene (44), the mouse NF-κB and STAT3 binding sites are located in distal cis-regulatory elements of Il17a-Il17f loci (68), but not in the proximal promoter, thus suggesting that the regulation of IL-17A transcription may differ between human and mouse. One important signaling mediator of CD28 is the PI3K. Indeed, CD28 binds the SH2 domain of p85 subunit of class 1A PI3K (p110δ) through the phosphorylated YMNM motif (73). Class 1A PI3K phosphorylates inositol phospholipids on carbon atom 3, thus generating the phospholipids PI 3,4-biphosphate (PIP2) and PI 3,4,5-triphosphate (PIP3) that in turn recruits Akt and PDK1 to the membrane, thus favoring their activation (71-73, 91, 92). The PI3K/PDK1/Akt pathway is essential for both CD28-mediated co-stimulatory signals and pro-inflammatory functions as well as for NF-κB activation (32,93,94). Consistently, we found that the selective inhibition of class 1 PI3K activity, impaired CD28 unique signaling leading to RelA nuclear translocation and NF-κB transcriptional activity as well as IL-6 expression and IL-6-dependent STAT3 phosphorylation and nuclear translocation (Figure 7). Class 1 PI3K/PDK1/Akt pathway also governs Th17 differentiation via multiple mechanisms (75), including STAT3 phosphorylation (74) and the nuclear translocation of RORγτ (95). Recent data from Way et al. showed that PI3K p110δ inhibition impairs IL-17 production form CD28-costimulated Th17 effector cells (96). Our data further evidence a novel role of class 1A PI3K in regulating CD28-induced recruitment of both RelA and pSTAT3 on the IL-17A promoter and on its transcriptional activation (Figure 8).
The identification of a new IL-6-STAT3/NF-κB axis, often altered in several immune-based diseases (97,98), linking CD28 to IL-17A expression provide biological bases for immunotherapeutic approaches targeting CD28 and/or associated signaling mediators in order to dampen the inflammatory response in autoimmune/inflammatory disorders (39).

ETHICS STATEMENT
This study was carried out in accordance with the recommendation of Ethical Committee of the Policlinico Umberto I (Sapienza University, Rome, Italy). All subjects gave written informed consent in accordance with the Declaration of Helsinki.

AUTHOR CONTRIBUTIONS
MK and MM performed the research and data analyses. NP and SC performed parts of the research. SA contributed samples for the study. EM contributed with STAT3C construct and commented on manuscript. LT designed the study, coordinated the work, and wrote the manuscript.