FcγRIIIa-Syk Co-signal Modulates CD4+ T-cell Response and Up-regulates Toll-like Receptor (TLR) Expression*

CD4+ T-cells in systemic lupus erythematosus (SLE) patients show altered T-cell receptor signaling, which utilizes Fc-receptor γ-chain FcRγ-Syk. A role for FcγRIIIa activation from immune complex (IC) ligation and sublytic terminal complement complex (C5b-9) in CD4+ T-cell responses is not investigated. In this study, we show that the ICs present in SLE patients by ligating to FcγRIIIa on CD4+ T-cells phosphorylate Syk and provide a co-stimulatory signal to CD4+ T-cells in the absence of CD28 signal. This led to the development of pathogenic IL-17A+ and IFN-γhigh CD4+ T-cells in vitro. Cytokines IL-1β, IL-6, TGF-β1, and IL-23 were the only requirement for the development of both populations. SLE patients CD4+ T-cells that expressed CD25, CD69, and CD98 bound to ICs showed pSyk and produced IFN-γ and IL-17A. This FcγRIIIa-mediated co-signal differentially up-regulated the expression of IFN pathway genes compared with CD28 co-signal. FcγRIIIa-pSyk up-regulated several toll-like receptor genes as well as the HMGB1 and MyD88 gene transcripts. ICs co-localized with these toll-like receptor pathway proteins. These results suggest a role for the FcγRIIIa-pSyk signal in modulating adaptive immune responses.

Concurrent with the presence of aberrant T-cell responses, elevated serum levels of both immune complexes (ICs) 2 and C5b-9 (non-lytic terminal complement complex) are associated with systemic lupus erythematosus (SLE) (1,2). These immune-reactants form immune deposits at vascular sites and trigger inflammation (3). Immune deposits are also present in the ectopic germinal centers, the site for plasma B cell development (4). Formation of ICs by autoantibodies activate complement cascade and drive the formation of C5b-9 on cell membrane. We previously showed that non-lytic C5b-9 deposits trigger clustering of membrane rafts (MRs) observed in SLE T-cells. Hence, we examined the role for Fc␥RIIIa ligation by ICs in CD4 ϩ T-cell responses in the presence of sublytic C5b-9 (5,6).
T-cell receptor (TCR) engagement with peptide-MHC (pMHC) and a co-stimulation by CD28 is required for CD4 ϩ T-cell activation and differentiation into effector CD4 ϩ T-cells (T E ). This requirement of CD28 in the periphery varies based on anatomical location, stage of immune response, nature of T-cell subsets, and the activation status of the CD4 ϩ T-cells (7)(8)(9). CD28 co-signal is a quantitative signal that overcomes the signal threshold necessary for T-cell activation, otherwise unattainable by the TCR ligation alone (10). In an autoimmune background, T-cell activation occurs without the requirement of CD28 co-signal (10). The mechanisms that drive this activation are unknown. A sublytic C5b-9 deposit trigger MR clustering, a function attributed to CD28 co-signaling (11). Naïve CD4 ϩ T-cells treated with ICs and C5b-9 phosphorylate TCR signaling proteins and spleen tyrosine kinase (Syk) (11). The external and internal stimuli that trigger helper CD4 ϩ T-cell (T H ) differentiation and lineage commitment in autoimmunity still remain unclear (10,(12)(13)(14)(15).
Although the T H 1 response initiates tissue damage, T H 17 responses sustain tissue injuries during organ-specific autoimmunity, such as in synovium, heart, skin, and brain, which are also often the site of immune deposits. T H 17 responses are observed both in inflammation and in autoimmunity (20,28). IL-17 cytokines are necessary for the development of severe glomerulonephritis (29). Pathogenic T H 17 cell commitment has a unique requirement of IL-23 that is elevated in serum and tissue in SLE patients (30). Both IFN-␥ and IL-17A are thera-peutic target for intervention in many autoimmune disease conditions (31)(32)(33).
Toll-like receptor (TLR) signaling in innate cells indirectly promotes T-cell differentiation. T-cells express TLRs and promote cytokine secretion. TLR signaling augments the T H 1, T H 17, and Tregs responses (34,35). Activation of TLRs by DNA/RNA-ICs leads to autoantibody production. Fc␥RII (CD32) is a key participant for the delivery of DNA-ICs to many cell types (36). Subcellular localization of TLR9 discriminate between self and non-self DNA (37).
In this report, we demonstrate that the Fc␥RIIIa-pSyk signal successfully replaced the CD28 requirement for differentiation of CD4 ϩ T-cells. ICs ligation to Fc␥RIIIa phosphorylates Syk (pSyk), which caused the activation of CD4 ϩ T-cells. In the presence of polarizing cytokines, this activation resulted in the development of CD4 ϩ IFN-␥ high and IL-17A-producing subsets. Fc␥RIIIa-pSyk co-signal induced colony stimulating factor 2 (Csf2) and IL-2 gene expression, which are associated with pathogenic T H 17 cells. In SLE patients, CD4 ϩ T-cell showed a subset that expressed activation markers CD25, CD69, and CD98, which also bound to ICs and showed pSyk. Furthermore, these activated cells produced IFN-␥ and IL-17A. Fc␥RIIIa-pSyk-mediated signal differentially regulated the expression of IFN pathway genes. Co-signaling triggered by IC ligation of Fc␥RIIIa up-regulated expression of TLR signaling genes, suggesting a co-operation among these pathways.

Experimental Procedures
Subjects-Blood from SLE patients and normal donors was collected with informed consent in the Saint Louis University Rheumatology clinic. The peripheral blood mononuclear cells were isolated using the Histopaque gradient (Sigma). The donors 1-9 were analyzed for IFN-␥ and IL-17A (Figs. 1 and 2). IL-21 production was analyzed in donors 1-4. This analysis from these donors is presented in Figs. 1, 2, and 4. FIGURE 1. ICs؉C5b-9 triggers development of a CD4 ؉ IFN-␥ high population. A, flow cytometry analysis for IFN-␥ production in naïve CD4 ϩ T-cells on day 9 of post polarization with anti-CD3ϩICsϩC5b-9 and anti-CD3ϩanti-CD28. Treatment with anti-CD3ϩICsϩC5b-9 produced 17.2% IFN-␥ high population, shown in donor 3. B, histogram of CD4 ϩ gated T-cells showing IFN-␥ in cells treated with anti-CD3ϩICsϩC5b-9 (a) and treated with anti-CD3ϩanti-CD28 (b) in donor 3. C, percentage of IFN-␥-producing cells shown in nine donors. D, combined analysis of 9 donors for IFN-␥ high population. The anti-CD3ϩICsϩC5b-9-treated group showed a statistically significant increase at a p value of 0.0026 over cells treated with anti-CD3 alone. No significant increase was observed in other groups. Donors 8,9, and an additional donor 10 were analyzed for the IFN gene analysis (shown in Fig. 9). Results presented in Figs. 4 -6 were obtained from additional donors not represented in Figs. 1 and 2.
ICs and C5b-9 -ICs were purified from 50 ml of pooled serum or plasma from 5-10 SLE patients that showed high levels of complement opsonized ICs. The purification procedures for ICs and C5b-9 have been previously described FIGURE 2. ICs؉C5b-9 induces IL-17A expression. A, flow cytometry analysis for IL-17A production on day 9 of post polarization. Cells treated with anti-CD3ϩICsϩC5b-9 generated 7.67% IL-17A ϩ cells, and anti-CD3ϩanti-CD28 generated 3.12% IL-17A ϩ cells, shown in donor 7. B, histogram of CD4 ϩ gated cells showing IL-17A in cells treated with anti-CD3ϩICsϩC5b-9 (25.6%, IL-17A ϩ ) (a) and treated with anti-CD3ϩanti-CD28 (b) of donor 3. C, percentage of IL-17A-producing cells shown in nine individual donors. D, combined analysis of same 9 donors for IL-17A production as in Fig. 1. The anti-CD3ϩICsϩC5b-9-treated group showed a statistically significant increase for IL-17A production at a p value of 0.016 compared with anti-CD3 alone. A significant increase was not observed in other groups. E, flow analysis showing double positive IFN-␥ high IL-17A ϩ populations. A small population of IFN-␥ high IL-17A ϩ was observed from co-stimulation by ICsϩC5b-9. (11,38,39). The nature of the ICs used has been characterized for their binding to Fc␥RIII in multiple cell types, compared with AHG and anti-Fc␥RIIIa antibody (clone 3G8) (40). In addition the ICs were compared for their potential to activate CD4 ϩ T-cells with in vitro formed Ova-anti-Ova ICs (11).
T-cell Culture and Differentiation-Peripheral blood mononuclear cells were isolated within 12 h of sample collection, and monocytes were removed by overnight plating in a culture dish. The next day the CD4 ϩ CD45RA ϩ cells were purified using naïve CD4 ϩ T-cell isolation kit II (Miltenyi Biotec, Product no. 130-094-131). Purified cells were maintained in culture with 20 units of IL-2 for 2 days. Thereafter, these cells were stimulated with plate-bound ICs at 10 g/ml and using purified soluble C5b-9 at 2.5 g/ml for 1 ϫ 10 6 cells in the presence of platebound anti-CD3 (eBioscience, clone OKT3) at 0.25 g/ml. Positive control cells were stimulated with plate-bound 1 g/ml anti-CD28 (clone 28.2) and 0.25 g/ml anti-CD3. At 24 h post stimulation cells were cultured in the presence of IL-2 (20 IU), IL-1␤ (50 ng), IL-6 (50 ng), IL-23 (20 ng), and TGF-␤1 (10 ng) for each ml of medium (Peprotech, Princeton, NJ). On days 9 -11, cells were analyzed by flow cytometry for cytokine production. Cytokine levels were measured in the culture supernatants harvested on day five due to the concern for overgrowth in anti-CD3ϩanti-CD28 activation.
Thymidine Uptake-Naïve CD4 ϩ T-cells were activated for 48 h with plate-bound anti-CD3ϩanti-CD28. Cells were then cultured in the presence of 20 units IL-2 and examined for binding of labeled ICs. Cells on day 7 were activated with platebound anti-Fc␥RIIIa/b (0.5 g/ml), ICs (10 g/ml), and anti-CD3ϩanti-CD28 (0.5 and 1 g/ml). Thymidine uptake was measured using Click-iT Plus Edu Alexa-488 assay (Product no. C10632, Life Technologies) 96 h post activation. Cells alone and isotype control (0.5 g/ml) were used as negative controls.
Quantitative Real-time-PCR and PCR Array Analysis-Total RNA was prepared from cells harvested between days 4 -5 post-stimulation using kit from Agilent Technologies (Wilmington, DE). Semiquantitative analysis for gene expression was carried from cDNA generated from total RNA using a high capacity cDNA kit (Applied Biosystems) using the comparative Ct (⌬⌬Ct) method. For Rorc (Hs01076122), endogenous control GAPDH (Hs02758991) (Applied Biosystems) was used. The RQ, RQ (minimum), and RQ (maximum) were calculated by StepOne software and plotted using GraphPad Prism. For gene expression cDNA was analyzed as per the manufacturer recommendation in the TaqMan Array for human IFN pathway (product no. 4418931) and analyzed in Data-assist. For calculating RQ, the corresponding genes in CD28 co-stimulated sample were used to normalize the expression. GAPDH, GUSB, and HPRT1 were used as endogenous controls for IFN array analysis. Analysis of variance was carried out using Partek Genomic Suite (Life Technologies). cDNA prepared from anti-CD3ϩanti-CD28 treated and anti-CD3ϩICsϩC5b-9-treated cells was analyzed for human Toll-Like Receptor Signaling Pathway genes using a RT 2 Profiler PCR array plate (PAHS-018ZC, SAS Bioscience). Data were analyzed using vendor software, and ACTB, B2M, RPLP0, and GAPDH were used as endogenous controls included in the array.
Cytokine Measurement-Culture supernatants were collected from activated cells on day 5 and kept frozen at Ϫ70°C. Cytokine measurements were performed using the multiplexing assay as per the manufacturer's instruction (EMD Millipore). For statistical analysis a non-parametric t test was performed using GraphPad Prism software.
Cell Staining-P116 cells (ATCC, CRL-2676), an acute T-cell leukemia ZAP-70 mutant was grown as per the guidelines from ATCC. These cells were activated as described in the previous section. Cells were harvested and washed with PBS and fixed in 4% formaldehyde for 15 min at room temperature. Cells were permeabilized using cold methanol at Ϫ20°C for 10 min. Cells were then kept for 1 h in 1% BSA/PBS and stained using antigen-specific antibodies at a dilution of 1:100 in BSA/PBS for 1 h and developed using anti-species specific Alexa Fluor fluorochrome conjugate (Life Technologies) at appropriate dilutions. Anti-TLR antibodies were purchased from R&D Systems and eBiosciences. Anti-MyD88 and anti-HMGB1 was obtained from Cell Signaling Technologies. As a control for labeled ICs we used human IgG-conjugated with Alexa Fluor 488. Isotype controls for mouse monoclonal and purified rabbit IgG fraction were used as negative controls.

Results
ICs and C5b-9 Co-stimulation Generate CD4 ϩ IFN␥ high Population-IFN-␥ is an autocrine T H 1 differentiation factor that requires cytokine IL-12 for differentiation (19,41). To examine whether ICsϩC5b-9 contributes to CD4 ϩ T-cell mediated pathological responses, we first examined the IFN-␥ production in the presence of IL-1␤, IL-6, IL-23, and TGF-␤1 cytokines. Flow analysis showed substantial and reproducible increases in the IFN-␥ producing populations on day nine post polarization ( Fig. 1, A and B). We observed a high and moderate IFN-␥ producing population (Figs. 1A and 2E). A statistically significant increase in IFN-␥ high population upon anti-CD3ϩICsϩC5b-9 treatment was observed compared with anti-CD3 treatment in 9 of 12 subjects analyzed (Fig. 1). These donors demonstrated an IFN-␥ high population upon ICsϩC5b-9 co-stimulation ( Fig. 1C). Donors 2, 7, and 9 also showed IFN-␥ production in response to anti-CD3ϩanti-CD28 treatment. In nine donors that produced IFN-␥, combined analysis showed a statistically significant increase in IFN-␥ producing population at a p value of 0.0026 in the anti-CD3ϩICsϩC5b-9-treated group compared with anti-CD3 group (Fig. 1D). In donor 7, a higher basal level of IFN-␥ before activation was observed. This donor also showed elevated Tbx21 transcripts, suggesting an ongoing T H 1 response at the time of sample collection (not shown).
The flow data were supported by an observed increase in IFN-␥ levels in the culture supernatants post day five from the time of polarization. A statistically significant increase in IFN-␥ production from anti-CD3ϩICsϩC5b-9 treatment, 14,398 Ϯ 6,587 pg/ml (p value of Ͻ 0.0001), compared with untreated cells 1684 Ϯ 338 pg/ml was observed (Fig. 3D). When compared with the anti-CD3-treated control group, anti-CD3ϩICsϩC5b-9-treated cells showed a statistically significant increase in IFN-␥ at a p value of Ͻ0.0025. The positive control group treated with anti-CD3ϩanti-CD28 also showed an increase in IFN-␥ production compared with untreated cells, 7571 Ϯ 5887 versus 1684 Ϯ 338 pg/ml, respectively, which was significant at a p value of Ͻ0.0001 (Fig. 3D). Untreated cells maintained in IL-2 (20 units/ml) showed minimal amounts of IFN-␥. These results confirm a role for ICsϩC5b-9 for IFN-␥ production in naïve CD4 ϩ T-cells.
Flow data were reconfirmed by the observed increases in the level of the cytokines IL-17A, IL-17F, and IL-22 in the culture supernatants, measured post day five from polarization. This time point was chosen to avoid differences arising from cell division in various activations. The amount of IL-17A produced in response to treatment with anti-CD3ϩICsϩC5b-9, compared with untreated cells, showed a statistically significant increase from 290 Ϯ 169 to 2220 Ϯ 1930 pg/ml (Mean Ϯ S.E.) at a p value of Ͻ0.0169. The positive control group treated with anti-CD3ϩanti-CD28 also showed a significant increase, from 290 Ϯ 169 to 1508 Ϯ 955 pg/ml, a p value of 0.0055 (Fig. 3A). IL-17F also showed a statistically significant increase in the anti-CD3ϩICsϩC5b9-treated group compared with the untreated control, an increase from 579 Ϯ 79 to 2979 Ϯ 328 pg/ml with a p value of Ͻ 0.0001. A comparable increase was observed in the anti-CD3ϩanti-CD28-treated group, from 579 Ϯ 79 to 2653 Ϯ 2073 pg/ml, with a p value of Ͻ0.0001 (Fig.  3B). The addition of C5b-9 to the anti-CD3ϩICs-treated group showed a statistically significant increase from 1937 Ϯ 191 to 2978 Ϯ 1042 pg/ml at a p value of 0.0035, suggesting a role for complement. There was no statistically significant difference in IL-17F production from co-stimulation with either CD28 or ICsϩC5b-9. Although an increase in the IL-22 levels was observed in response to co-stimulation with CD28 as well as polarization. An increase in IL-17A was significant at p Ͻ 0.016 in anti-CD3ϩICsϩC5b-9 and Ͻ0.005 for anti-CD3ϩanti-CD28-treated groups. An increase in IL-17F was significant at p Ͻ 0.0001. An increase in IFN-␥ was significant from both co-stimulations at p Ͻ 0.0001. *, significant increase over the untreated cells. **, significant increase upon C5b-9 addition; mean Ϯ S.E. (n ϭ 5).
ICsϩC5b-9, these values were not statistically significant (Fig.  3C). IL-22 production was also observed in flow analysis (not shown).
We also examined IL-21, a cytokine produced by both T H 17 and T FH cells. IL-21 activates lymphocyte and regulates antigen specific antibody response (45,46). All four donors examined showed enhanced production of IL-21 from ICsϩC5b-9 costimulation. Donors 3 and 7 also showed IL-21 production from anti-CD3ϩanti-CD28 activation. The combined analysis of all four donors showed a statistically significant increase in the percentage of IL-21-producing cells upon ICsϩC5b-9 co-stimulation (not shown).
ICsϩC5b-9 Co-signal Triggers Expression of Genes Associated with T H 17 Terminal Differentiation-We further confirmed the identity of T H 17 cells by examining Rorc expression, a T H 17 transcriptional regulator. All five donors analyzed showed a 2-8-fold increase in Rorc gene transcripts upon ICsϩC5b-9 co-stimulation when normalized with transcript levels present in the anti-CD3-treated cells. Donors 4 and 5 were normalized using a control from another subject, as Rorc gene transcripts were not detectable in either untreated (not shown) or anti-CD3-treated cells used as negative controls (Fig. 4A). Csf2, IL-2, and Tbx21 are markers for the terminally differentiated pathogenic T H 17 population (30). ICsϩC5b-9 co-stimulation increased the expression of gene transcripts for IL-6, Csf2, IL-10, IL-12A, IL-1A, IL-1B, and IL-2 compared with the levels of transcript observed from CD28 co-stimulation (Fig. 4B, n ϭ  3). The increase in the IL-6 transcripts was 4.21-fold and for Csf2 was 3.95-fold. These data suggest that the ICsϩC5b-9 costimulation contributes to the development of the pathogenic T H 17 population.
IC Engagement Phosphorylates Syk and Triggers Thymidine Uptake-To further confirm a role for Fc␥RIIIa-Syk signal in CD4 ϩ T-cell activation, we examined the T-cell activation markers and pSyk upon in vitro activation of naïve CD4 ϩ T-cells. Co-stimulation by CD28 or ICsϩC5b-9 induced the expression of CD25 and CD69 (Fig. 5A). Co-stimulation by ICsϩC5b-9 showed an increase in pSyk population (Fig. 5B). These pSyk ϩ cells expressed IFN-␥ (Fig. 5C). IFN-␥ production was observed in the absence of PMA and ionomycin treatment.
Confocal microscopic examination of z-series sections for CD3 complex and Fc␥RIIIa staining showed co-localization of these proteins (Fig. 6F). These results are in accordance with our previous observation and published report (11,47).
CD4 ϩ pSyk ϩ T-cells in SLE Are an Activated Phenotype That Produces IFN-␥ and IL-17A-To establish a role for Syk signaling in vivo, we next examined the presence of pSyk using two antibodies in the CD4 ϩ T-cells within the peripheral blood mononuclear cells of SLE patients. Antibodies used recognized the Tyr-348 residue in Vav1 binding of human Syk and a second antibody that recognized the Tyr-525/526 residue in the kinase domain (Fig. 7A). Both of these antibodies confirmed the presence of pSyk in activated peripheral CD4 ϩ T-cells that expressed CD25 (a and b), CD69 (c and d), and CD98 (e and f), all T-cell activation markers. We then examined the IC binding to CD4 ϩ T-cells that expressed T-cell activation markers and pSyk (Fig. 7A, panels g, h, i, and j). Activated CD4 ϩ pSyk ϩ T-cells bound ICs, suggesting the presence of Fc␥RIIIa. These results suggest that in CD4 ϩ SLE T-cells, Syk signaling contributes to cell activation (11,48). Activated CD4 ϩ T-cells express Fc␥RIIIa (40). The presence of pSyk in activated CD4 ϩ T-cells in the patient population is also supported by our previous studies where we showed Syk phosphorylation during T-cell activation (38). A role for Syk in the development of T H 1 and T H 17 responses via dendritic cell activation has been also been suggested (49).  JANUARY 15, 2016 • VOLUME 291 • NUMBER 3

JOURNAL OF BIOLOGICAL CHEMISTRY 1373
We next examined whether the activated pSyk ϩ CD4 ϩ T-cells in SLE patients produced IFN-␥ and IL-17A. Flow analysis showed that the pSyk ϩ cells both at Tyr-348 or Tyr-525/ 526 produced IFN-␥ and IL-17A (Fig. 7B). Data from two patients show 7.77 and 9.31% pSyk ϩ IFN-␥ ϩ cells in donor 1 (panels a and b) and 4.48 and 4.82% in donor 2 (panels e and f), respectively. Donor 1 showed 7.10% and 7.82% of pSyk ϩ IL-17A ϩ cells (panels c and d) and 3.32% and 3.38% population in donor 2 (panels g and h). A minor population of IL-17A high was also observed in several subjects. These cells were analyzed without activation by PMA, ionomycin, and brefeldin A treatment. Analysis of 29 patients showed that pSyk ϩ cells were activated CD4 ϩ T-cells, which produced IFN-␥ and IL-17A cytokines. Individual and combined analysis of these 29 subjects as a group demonstrated that the percentage of pSyk ϩ Fc␥RIIIa ϩ (IC binding cells); pSyk ϩ CD25 ϩ , pSyk ϩ CD69 ϩ , and pSyk ϩ CD98 ϩ cells did not vary significantly (Fig.  7C). These results suggest that the activated CD4 ϩ T-cells signal via Syk and produce inflammatory cytokines. To further confirm the role for Syk signaling, we gated CD4 ϩ T-cells for pSyk ϩ and IC binding (double positive) and examined this population for IFN-␥ and IL-17A production. This analysis showed two subsets of IFN-␥-producing cells, IFN-␥ moderate and IFN-␥ high (Fig. 7D). Analysis of these IFN-␥ subsets showed various levels of IL-17A-producing cells; IFN-␥ high always showed a higher percentage of IL-17A ϩ cells, 14.2%, 29.2%, and 12.9%, compared with 8.94%, 1.31%, and 1.38% in IFN-␥ moderate cells (Fig. 7D). These results concur with results obtained from in vitro co-stimulation of naive CD4 ϩ T-cells suggesting a role for an ICsϩC5b-9-mediated signal.
ICOS ϩ but Not PD1 high Cells Show pSyk in SLE-CD4 ϩ Tcells-ICOS and PD1 are key membrane regulators of CD4 ϩ T-cell response. Thus we next examined whether pSyk ϩ Fc␥RIIIa ϩ cells express these proteins. In all 15 donors analyzed, ICOS ϩ CD4 ϩ T-cells also showed pSyk (Fig. 8A). Cells that expressed ICOS bound to ICs (Fig. 8B). However, those cells that expressed high levels of PD1 (PD1 high ) lacked pSyk. pSyk ϩ cells expressed low levels of PD1 (Fig. 8, panels C and D). A higher percentage of PD1 high cells with high mean fluorescence intensity values was observed compared with PD1 low pSyk ϩ cells (Fig. 8H). In only two patients the mean fluorescence intensity for PD1 was equal or slightly higher in cells with pSyk. A role for PD1 in down-regulation of Syk phosphorylation via SHP2 has been shown (50). PD1 high cells did not bind to ICs, although in some patients both moderate PD1 levels and an IC binding population was observed (Fig. 8 panels E and F). A paired t test showed a statistically significant correlation among pSyk and ICOS expression at a p value of Ͻ0.0001, with a strong correlation (r ϭ 0.77). However, a correlation between pSyk and PD1 expression was not observed (r ϭ 0.22) (Fig. 8, panel  G). It is likely that PD1 dephosphorylated Syk via SHP2. These data suggest a possible role for Fc␥RIIIa-Syk signaling in modulating responses of CD4 ϩ T-cell membrane regulators (Fig. 8).
ICsϩC5b-9 Up-regulate TLR Signaling Pathway Genes-TLRs play a role in adaptive immune responses (35). To examine whether TLR signaling synergizes the Fc␥RIIIa-Syk-mediated signal in modulating T-cell responses, we analyzed the expression of TLR signaling genes. We analyzed naïve CD4 ϩ T-cells from five paired samples under identical culture conditions. Cells were co-stimulated with ICsϩC5b-9, and the gene expression levels were compared with cells co-stimulated using ani-CD28 from the same subject. Combined analysis of five samples showed increased expression of TLR-interacting pro-teins and adaptors such as Bruton agammaglobulinemia tyrosine kinase (Btk) (2.92), HMGB1 (3.62), Harvey ras sarcoma virus oncogene homolog (HRAS) (4.91), and MyD88 (2.34). Myd88-dependent signaling TLRs, TLR2 (5.50), TLR4 (2.23), TLR5 (5.17), TLR7 (2.55), and TLR10 (5.16), showed significant increases, whereas TLR9 (1.03) did not show any increase. TIRAP (5.21), which is essential for TLR2 and TLR4 signaling, was up-regulated. Expression of TRAF6 (4.65), a TNF receptorassociated family factor, is an E3 ubiquitin ligase that signals via the Toll/IL-1 family, which was also increased. Proteins that influence the adaptive responses, TRAF6 (4.65), IL10 (2.99), IL12B (2.31), IL1A (3.05), and IL1B (2.68), showed increased expression (Fig. 10). TLR3 (9.89), a MyD88-independent signal, showed the highest increase in the gene expression (Fig. 10). Two donors, 8 and 12, showed the maximum up-regulation of TLRs (Fig. 10B). In these two donors we compared the up-regulation of TLR signaling genes from untreated cells with CD28 and ICsϩC5b-9 co-stimulation. IL-12A was significantly up-    Table 1). TLR3 showed the most increase and is shown to aggravate lupus nephritis (51). To examine the presence of TLR proteins and their association with Fc␥RIIIa in CD4 ϩ T-cells, we stained P116 cells after co-stimulation with ICsϩC5b-9. ICs co-localized with MyD88 (supplemental Movie 1), HMGB1 (supplemental Movie 2), TLR3 (supplemental Movie 3), TLR5 (supplemental Movie 4), and TLR9 (supplemental Movies 5 and 6), MyD88, and HMGB1 along with IC localized on the cell membrane (Fig. 11). TLR3 and TLR9 colocalized with ICs on membrane (Fig. 11A, panels e and f and panels k and l; supplemental Movies 3 and 6). Both of these proteins were also present in the endolysosome. This was confirmed using LysoTracker deep red (Molecular Probes). These proteins appear in microclusters. IC binding showed a pattern of receptor capping. Even though we did not see up-regulation of TLR9 transcripts upon ICsϩC5b-9 co-stimulation at the protein level in naïve CD4 ϩ T-cells, we observed the TLR9 protein in P116 cells and activated human CD4 ϩ T-cells. We observed two staining patterns for TLR9. The first pattern was with membrane staining and intracellular staining for ICs (Fig. 11A, panels i and j). In the second staining pattern ICs showed membrane staining and TLR9 in endolysosomes forming a ball-like structure (Fig. 11A, panels k and l, supplement Movie 5). TLR9 co-stained with ICs, suggesting their co-localization. Untreated P116 cells showed mostly membrane staining for ICs and TLR9. A similar staining pattern was also observed for TLR3. In Western blot analysis, both HMGB1 and MyD88 were observed in immunoprecipitates prepared using anti-CD16 (clone 3G8) antibody (data not shown). These data suggest a role for Fc␥RIII-Syk signaling in the up-regulation of TLR signaling pathways in CD4 ϩ T-cells.
Our results thus suggest that Fc␥RIIIa-pSyk is a distinct cosignal in CD4 ϩ T-cells that drives the differentiation of naïve cells into IFN-␥ high and IL-17A ϩ populations. Fc␥RIIIa-pSyk signal up-regulated the genes associated with terminal differentiation of pathogenic T H 17 cells. Fc␥RIIIa-pSyk is a distinct and potent signal for up-regulation of the IFN signaling pathway. The Fc␥RIIIa-pSyk population is present in SLE patients. The ligation of Fc␥RIIIa by ICs up-regulated the FIGURE 9. ICs؉C5b-9 differentially expresses IFN pathway genes. PCR array analysis shows distinct and differential expression of IFN pathway genes in three donors. IFN pathway genes up-regulated from ICsϩC5b-9 co-stimulation normalized over the level of gene transcripts expressed from CD28 co-stimulation. Donor 8 showed strong type I IFN gene signature (blue bars). Donor 9 showed dominant type II IFN gene signature (red bars). Donor 10 showed moderate expression in both IFN genes (green bars).
TLR signaling pathway genes. These data suggest a possible synergistic role of TLR and Fc␥RIIIa signaling in human CD4 ϩ T-cells.

Discussion
In this report we show that the ICsϩC5b-9 acts as a co-stimulator of naïve CD4 ϩ T-cells. ICsϩC5b-9 generates a co-stimulatory signal that is mediated via Fc␥RIIIa-Syk phosphorylation. This ICsϩC5b-9-mediated signal efficiently replaced the CD28 requirement for the development of CD4 ϩ IFN-␥ ϩhigh and a T H 17 like population. The Fc␥RIIIa-pSyk is a distinct co-signal from CD28, as it differentially expressed IFN genes and up-regulated TLR signaling pathways genes. Naïve CD4 ϩ T-cell activation, survival, subset differentiation, and effector function are regulated by the co-signaling proteins present on CD4 ϩ T-cell membrane (52). A co-stimulatory signal from CD28 (signal 2 of two signal hypothesis) is a key requirement for naïve CD4 ϩ T-cell activation without which cells become anergic. In an autoimmune background, CD4 ϩ T-cells bypass the need of CD28 co-signal to become fully activated (10). However, the mechanism underlying this activation is unknown. Our results suggest that in an autoimmune response, Fc␥RIIIa-Syk signal is important for the activation of naïve CD4 ϩ T-cells. In CD4 ϩ T-cells that express Fc␥RIIIa, ICs ligation triggers FcR␥ chain phosphorylation, which then co-localizes and signal via Syk (38,53). Although C5b-9 is essential to trigger MR clustering, ICs engage Fc␥RIIIa and trigger Syk activation (11). Both ICs and C5b-9 are required for phosphorylation of TCR signaling proteins and triggering of T-cell activation-associated changes (11). Co-localization of ICs with in situ assembled C5b-9 and CD3 complex suggests a cooperative response among these complexes (11). On CD4 ϩ T-cell membrane ICs binding occurred at the site of Fc␥RIIIa staining, confirming the presence of these receptors (40). Fc␥RIIIa colocalize with the CD3 complex on the cell membrane (Fig. 6F). Previous studies have also shown colocalization of FcR with TCR on activated T-cells (47). In SLE-CD4 ϩ T-cells, the FcR␥ chain associates with the CD3 -chain of TCR and signals via Syk (53). Rewiring of TCR-CD3 complex, where the CD3-chain is replaced by the FcR␥ chain, which signal via Syk in SLE T-cells is shown (16). Up-regulation of both the FcR␥ chain and Syk is observed in T E cells, unlike naïve cells (16). The T-cell activation via CD28 signaling upon TCR engagement in the absence of ZAP-70 signaling can utilize Syk. The engagement of Fc␥RIIIa can only generate the Syk-mediated signal. In the presence of IL-2 and IL-12, ICsϩC5b-9 co-signals with suboptimal CD3 ligation generated a T H 1-like population (40). Intravenous gamma globulin therapy, which works by blocking low affinity FcRs, reciprocally regulates human pathogenic T H 1, T H 17, and Treg cells (54). Based on our results and the existing literature, we propose that the low affinity Fc␥RIIIa-mediated phosphorylation of Syk is an important signal for the development of proinflammatory CD4 ϩ T E cells (17,48). Additionally, TLR signal up-regulation from Fc␥RIIIa ligation by ICs may result in the development of proinflammatory cells that may be refractory to suppression by Tregs. A role for IL-1, IL-6, and LPS signal that utilizes MyD88 an adaptor for Toll/IL-1 receptor is implicated in overcoming the suppression by Tregs (55)(56)(57). Cooperation among Fc␥R-TLR signaling in M1 and M2 macrophages activates Syk kinase, which then produces proinflammatory cytokines (58). An association of TLR4 with Fc␥RIII upon ICs stimulation of mouse macrophages is observed (59). A similar cooperation between Fc␥RIIIa and TLR signals in CD4 ϩ T-cells could modulate adaptive immune responses in humans.
The expression of co-stimulatory proteins on naïve CD4 ϩ T-cells is limited, and CD28 is the only primary protein known to prime these cells (60). ICsϩC5b-9 successfully primed human peripheral naïve CD4 ϩ T-cells in the absence of CD28 co-signal. The co-signal generated by ICsϩC5b-9 was efficient and potent enough to support the development of IFN-␥ high and IL17A ϩ -producing cells, which only required IL-1␤, IL-6, IL-23, and TGF-␤1, without IL-4 and IFN-␥ suppression (Figs.  1 and 2). For differentiation of mouse naïve CD4 ϩ T-cells into T H 17 cells, IFN-␥ and IL-4 suppression is required. We did not observe this requirement.
Altered CD4 ϩ T-cell responses are a common feature of autoimmune pathology, which is often accompanied by elevated IC levels and complement activation byproducts such as C5b-9 (61,62). The elevated serum and urine levels of C5b-9 are associated with disease activity. In our model the C5b-9 contributes to cell activation by lateral clustering of MRs, which brings the receptors and signaling proteins to close proximity (11). MRs are uniformly distributed on T-cells from healthy individual and are aggregated and clustered in SLE T-cells (63). MRs aggregation is observed in the mouse model of SLE (64). Atorvastatin reversed MR signaling abnormalities in SLE T-cells (65). Fc␥R engagement by ICs in many cell types drive IFN production (66). The CD4 ϩ Fc␥RIIIa ϩ IFN-␥ high cells generated via Fc␥RIIIa-pSyk signaling represent a new subset of T-cells. Co-expression of TLR proteins in these cells could render them refractory to Treg suppression (55).
ICs are often present in the immune deposits along with C5b-9 proteins. Recent studies have also shown that ICs are held without phagolysis by subcapsular sinus macrophages, B cells, and follicular dendritic cells on the cell membrane (67,68). The retention and passive exchange of intact ICs occur among several cell types within the germinal centers. Follicular dendritic cells recycle the ICs, which make them accessible to antigen-specific B cells (68). In many disease tissues the formation of ectopic germinal centers are often observed. At these sites and in systemic circulation, ICs can drive differentiation of naïve CD4 ϩ T-cells and produce IFN-␥, IL-17A, and IL-21, which can augment antigen-specific antibody responses (45).
During T H 1 response, IFN-␥ is produced in two waves, and the secondary IFN-␥ production is driven by an autocrine IFN-␥ signal (41). Two populations of IFN-␥-producing cells in SLE CD4 ϩ T-cells IFN-␥ moderate and IFN-␥ ϩhigh were observed both in vivo and upon in vitro activation (Figs. 1, 2, and 7). The IFN-␥ moderate population likely represents the bystander secondary T-cell response (Fig. 2E). In an in vitro experiment, a secondary short term challenge by ICs produced IFN-␥ moderate cells, which was accompanied by a proportionate loss of IC binding (not shown). We examined cytokines in the SLE T-cells without PMA and ionomycin activation to avoid the influence of these potent T-cell activators. Our in vitro data on cytokine production and ICs binding was supported by the in vivo presence of such cells in SLE patients. Those SLE CD4 ϩ T-cells that expressed the T-cell activation markers, CD25, CD69, and CD98, also showed pSyk, bound to labeled ICs, and produced cytokines, suggesting a role for Syk signaling (Fig. 7A). A role for anti-CD3 antibodies in rewiring of the CD3 complex with Syk has been suggested in SLE T-cells (48). However, the presence of anti-CD3 antibodies in SLE pathology has not been documented. In our experiments, anti-CD3-treated cells alone did not generate IFN-␥ or IL-17A-producing populations. Therefore, we propose that this Syk rewiring of the TCR complex occurred from IC ligation of Fc␥RIIIa. Labeled ICs co-localize with CD3 complex in activated CD4 ϩ T-cells, suggesting the presence of Fc␥RIIIa (11). ICOS-expressing cells showed pSyk, implying the co-presence of these proteins in activated T-cells. pSyk ϩ cells showed low levels of PD1 expression. Cells expressing high levels of PD1 did not show pSyk. PD1 is an inhibitory co-stimulating signal that acts by recruiting phosphatase SHP2. ZAP-70-deficient patients show abnormal peripheral CD4 ϩ T-cells and express high levels of Syk, which drives T-cell activation (17). Kinase activity of Syk is 100-fold higher than that of

Expression of TLR pathway genes from two co-stimulations
Cells co-stimulated using ICsϩC5b-9 showed higher levels of gene expression vs. CD28 compared with expression levels in untreated cells. The ICsϩC5b-9 also showed increased expression when compared to levels of genes expression in CD28 co-stimulated cells. Btk, Bruton agammagloblinemia tyrosine kinase; HRAS, Harvey ras sarcoma viral oncogene homolog. ZAP-70, and Syk demonstrates a differential intrinsic activity compared with ZAP-70 (69). Syk is essential for innate responses and is a key signaling protein in B-cells (70). Further support for the role of Syk in CD4 ϩ T-cell differentiation also comes from production of IFN-␥ by P116 cells upon co-stimulation with ICsϩC5b-9. These cells express Fc␥RIIIa upon activation (40). We speculate that the Fc␥RIIIa-pSyk-mediated IFN-␥ production observed upon IC ligation is driven by the occupancy of the Ϫ53 CpG site in the IFN-␥ promoter by ATF2 (71). A 5.3-fold increase in ATF2 (n ϭ 5) was observed upon ICϩC5b-9 co-stimulation normalized to the level of transcripts observed from CD28 co-stimulation (Fig. 9A). ATF2 also activates IL-23p19 promoter and has three binding sites in the IL-17 promoter.
A strong association of IL-17A and other T H 17 cytokines in SLE pathogenesis in mouse model has been reported (20,72). IL-23 cytokine, which is elevated in SLE patient sera, contributes to the terminal differentiation of pathogenic T H 17 cells.  (g and h), and TLR 9 co-localized with IC binding (green) (i, j, k, and i). Shown are confocal image of cells (left) and three-dimensional view of Z-stacks (right). MyD88 (red) was observed uniformly in the cell membrane, seen in DAPI blue (a and b). HMGB1 (red) stained in both membrane and cytosol (c and d) is shown. TLR3 (red) co-localized with ICs on the cell membrane and in cytosol (e and f) is shown. TLR5 (red) co-localized with ICs on the membrane (i and j) is shown. TLR9 (red) shows membrane staining with intracellular ICs, marked by arrows (k and l). TLR 9 (red) was also co-localized with ICs on the membrane with most of TLR9 forming a ball-like structure in endolysosomes, marked by arrows (k and l). Pearson correlation coefficient for IC localization was calculated in 2-3 cells with these proteins was MyD88 (0.647), HMGB1 (0.756), TLR3 (0.687), TLR5 (0.698), TLR9 (0.5240, and TLR9 (0.716). B, cells stained with mouse IgG isotypes (a and b) and purified rabbit globulin (c and d) did not show staining. Cells stained with human IgG-Alexa Fluor 488 at 33 M fluorochrome to protein ratio were similar to labeled ICs (e). Data are representative of two independent experiments, and several fields were examined. C, human activated CD4 ϩ T-cells were stained for IC binding (a, green), TLR9 (b, red), merge (c), and three-dimensional image (d). Cells stained for IC binding and HMGB1 (e, three-dimensional image) and IC binding and MyD88 (f, three-dimensional image).
IFNs are critical in RA and SLE pathogenesis (25,73,74). DNA-and RNA-containing ICs trigger IFN-␣ production from plasmacytoid dendritic cells by engaging Fc␥RIIa with TLR7 and TLR9 (34,75). We observed the expression of all 13 type I IFN genes, which were up-regulated by ICsϩC5b-9 co-stimulation in one donor (Fig. 9, blue bars). This donor also showed a 20-fold increase in IRS2. Both IRS1 and IRS2 act as an adaptor for type I IFN-mediated signaling events. IRS2 is negatively regulated by cAMP response element-binding protein 3-like 4 (CREB3L4). A genetic interaction of the CREB, a co-activator with IL-12/STAT4 protein during T H 1 differentiation, prolongs IFN-␥ synthesis (76). We are not aware of any previous report of type I IFN expression in human CD4 ϩ T-cells. IFN pathway gene array analysis showed a unique gene signature up-regulation in all three subjects analyzed (Fig. 9). The IFN gene expression profile suggests the ICsϩC5b-9 co-signal differentially up-regulated expression of IFN pathway genes, suggesting a bifurcation of signaling events.
An individual role for TLRs and FcRs in inflammatory responses has been documented. Interplay between the members of these two receptor families has started to emerge in inflammatory diseases. Simultaneous engagement of TLRs and FcRs on dendritic cells is essential for production of IL-1␤ and IL-23 (77). A cross-talk between TLRs and FcRs initiated by ICs and pathogen-associated molecular patterns (PAMPs) and danger-associated molecular patterns (DAMPs) in autoimmune diseases is now proposed (77). Human CD4 ϩ T-cells and Jurkat cells express RNA that encodes most of TLRs (78). TLRs play a role in T-cell activation and differentiation during autoimmunity (35,78). TLRs also modulate CD4 ϩ T-cell response (79,80). In human naïve CD4 ϩ T-cells, ICsϩC5b-9 co-stimulation up-regulated TLR2, -3, -5, -8, -10, HMGB1, and MyD88 gene transcripts (Fig. 10). This suggests that the activation of CD4 ϩ T-cells by ICsϩC5b-9 sensitize them for danger signals. The TLR5 ligand, flagellin in the presence of suboptimal anti-CD3 ligation and in the absence of APC triggers cell proliferation, as well produces IFN-␥ and IL-8 (81). Chromatin-IgG complexes activate B-cells by dual engagement of B-cell receptor and TLR (82). Such synergism between TCR and FcRs with TLRs in CD4 ϩ T-cells has not been shown but could occur in human Fc␥RIIIa ϩ CD4 ϩ T-cells. Fc␥RIIIa-TLR engagement will have a wide-ranging implication in autoimmunity. DNA-ICs ligate Fc␥RIIa in plasmacytoid dendritic cells and up-regulate TLR9, which then drive the IFN response in SLE. Our results suggest that in CD4 ϩ T-cells Fc␥RIIIa could co-operate with TLRs to drive proliferation and IFN production (11,40). TLR adaptor molecule-1 (TICAM1), which interacts with TLR3 and other MyD88-dependent TLRs, were up-regulated. In response to double-stranded RNA viruses, TLR3 produces IFN-␤ via IRF3. Poly I:C, a TLR3 ligand, activates human CD4 ϩ T-cells (83). In the same study the TLR5 expression was also observed on CD4 ϩ T-cells, but the activation with flagellin was not sufficient to activate these cells. Even though TLR responses in the CD4 ϩ cells has been documented, the receptor by which the nucleic acids are delivered to endolysosomes is unknown. Our result suggests a possible role of Fc␥RIIIa in delivering DNA-ICs to endolysosomes in activated CD4 ϩ T-cells where they could interact with TLR9 (Fig. 11). DNA-or RNA-containing ICs via HMGB1 can efficiently deliver self nucleic acid to TLR containing endolysosomes (84). HMGB1 is a DNA chaperon that is capable of organizing dynamic active chromatin structures. It is diffusely distributed in cytoplasm and is released from inflamed cells actively or from apoptotic and necrotic cells. Elevated serum levels of HMGB1 are observed in SLE patients during flares. HMGB1 has a proinflammatory effect, which is mediated via TLR2, -4, and -9 (78). ICsϩC5b-9 co-signal up-regulated HMGB1 gene transcripts in naïve CD4 ϩ T-cells and the HMGB1 protein co-localized with ICs in human CD4 ϩ T-cells and P116 cells. Ablation of MyD88 in CD4 ϩ T-cells impairs both T H 1 and T H 17 responses (56). We observed the overexpression of MyD88 transcripts and MyD88 protein co-localized with ICs. Both HMGB1 and MyD88 proteins were observed in co-immunoprecipitates obtained using anti-Fc␥RIIIa/b antibodies from P116 cells. 3 A role for MyD88 in proliferation and IFN-␥ production in mice infected with Ehrlichia muris has been observed (85). TLR9 agonist in CD4 ϩ T-cells enhanced proliferation, survival, and IL-2 secretion (79). Subcellular localization of TLR9 using HEK293T cells has been shown to be critical in discriminating self versus non-self DNA (37). TLR9 reside in endoplasmic reticulum, and upon stimulation from CpG DNA it is recruited to lysosomes (86). Upon ICsϩC5b-9 co-stimulation in P116 cells TLR9 protein localized in endolysosomes with ICs. This pattern was confirmed in human CD4 ϩ T-cells. We also observed membrane staining for TLR9 with cytoplasmic IC binding. This suggests a possible role for Fc␥RIIIa in recruiting TLR9 to endo lysosomes. The significance of these events in the development of autoimmune response remains to be determined. Our results suggest a critical role for Fc␥RIIIa-pSyk signal in CD4 ϩ T-cell-mediated adaptive immunity.
In summary, our results establish a co-stimulatory role for ICsϩC5b-9 in the development of the CD4 ϩ IFN-␥ high cell subset and a T H 17 like population. ICsϩC5b-9 provides a distinct co-stimulatory signal for the up-regulation of the IFN and TLR signaling pathway genes. The data provide a link for ICs in driving TLR-dependent T-cell activation in autoimmunity (35). T-cell signaling responses by TLRs result in tolerance breakdown and bystander activation of auto reactive T H 1 and T H 17 cells. An abnormal activating co-stimulatory signal from ICsϩC5b-9 during immune contraction can override the inhibition by CTLA-4 and PD1, resulting in peripheral tolerance breakdown. A further understanding of ICsϩC5b-9 signaling in CD4 ϩ T-cells will lead to a better understanding of the role of CD4 ϩ T-cells in diseases like SLE. These findings will not only be relevant to autoimmune disorders but also in cardiovascular diseases, cancers, and viral infections. Both PD1 and CTLA-4 proteins are therapeutic targets. A role for activating FcRs is also suggested in the therapies targeting CTLA-4. It is important to further explore the role of Fc␥RIIIa signaling in CD4 ϩ T-cells.
Author Contributions-A. K. C. designed research, performed experiments, analyzed and interpreted results, and wrote the manuscript. T. L. M. obtained clinical material, clinical information, and reviewed the manuscript. C. C. and Y. B. performed the experiments.