SOCS3 promotes interleukin-17 expression of human T cells

T helper type (T_H) 1 cells are crucial for protective immunity against infectious diseases caused by intracellular pathogens, such as Mycobacterium tuberculosis. In this context, exact regulation and limitation of T-cell responses are important and therefore have to be tightly regulated to avoid harmful hyper-reactivity. Important targets for regulation are cytokine receptors, mainly those containing common γ-chains. These receptors signal via the JAK and STAT pathway. Regulation of JAK/STAT signaling is mediated by members of the suppressor of cytokine signaling (SOCS) family.¹  The SOCS family is composed of 8 proteins and at least 5 (ie, CISH, SOCS1, SOCS2, SOCS3, and SOCS5) are expressed in T cells.²  SOCS molecules are induced by different stimuli (predominantly cytokines) and act as feedback inhibitors. Inhibition is exerted by different mechanisms, including (1) inhibition of STAT recruitment to cytokine receptors, (2) targeting cytokine receptors and JAK molecules for proteasomal degradation, and (3) direct inhibition of kinase function. Particularly SOCS1 and SOCS3 are crucial for immune functions because modulation of SOCS1 or SOCS3 expression leads to severe immune disorders and embryonic lethality.³ 

In T cells, the interaction of different SOCS molecules with STAT family members is crucial for proliferation and polarization toward distinct T_H cell subtypes.²  Increased SOCS3 expression has been shown to inhibit T-cell proliferation by targeting STAT5 phosphorylation in IL-2 receptor signaling^4,5  and CD28-induced IL-2 production.⁶  Furthermore, SOCS3 has been demonstrated to favor T_H2 polarization,^7-10  and increased SOCS3 expression promotes T_H2-mediated diseases, such as inflammatory arthritis,¹¹  ovalbumin-induced airway hyper-responsiveness,^12,13  and liver injuries.^14,15  More recent studies demonstrated inhibition of T_H17 by SOCS3 in animal models, and this is mediated via regulation of STAT3.^16,17  Prerequisites for differentiation of T_H17 are controversially discussed and may partly differ between mice and humans.^17,18  This situation became even more complex because instability of T_H17 cell phenotype, including coexpression of master regulators of ROR-γt (RORC in humans), and T-bet indicated that T_H17 are a transient state rather than a stable T-cell population.^19-21  Differentiation of T_H17 can be subdivided in at least 2 stages: a polarization and an expansion stage. IL-6 is an essential factor for polarization of naive CD4⁺ T cells toward (pre)T_H17, and IL-23 is a crucial for T_H17 expansion. Both cytokines, IL-6 and IL-23, exert their function in a STAT3-dependent manner^18,22 ; therefore, SOCS3-mediated regulation may affect distinct stages of T_H17 generation. CD161 is a marker of naive T_H17 precursor as well as mature T_H17 cells within the CD4⁺ T-cell population in humans.²³ 

Recently, we and others identified differential SOCS3 expression in blood and T cells of patients with active tuberculosis.^24,25  Besides its crucial role in T-cell polarization, SOCS3 is involved in the exhaustion of T cells, which is an important feature of chronic viral infections.²⁶  In this model, suppression of SOCS3 by IL-7 restored the function of antiviral T cells and led to more effective viral clearance.²⁶ 

Several studies characterizing the role of SOCS3 on T cells in animal models have been performed, but the function of SOCS3 in human T cells has not been well defined. Consequently, we aim at characterizing the role of SOCS3 in human T cells in the present study. Initially, we established a flow cytometry–based method for SOCS3 protein quantification. This method was used for determination of SOCS3 in different T-cell populations, after in vitro restimulation, and in T-cell clones. Ectopic SOCS3 expression in primary human T cells using a lentiviral-based expression system was then applied to characterize functional changes induced by SOCS3 overexpression. Finally, recombinant IL-7 was used to study the effects of SOCS3 down-regulation on T-cell function and polarization.

PBMCs were isolated from blood (20 mL) of voluntary healthy donors recruited at the Bernhard-Nocht-Institute for Tropical Medicine and from buffy coats from the Institute for blood transfusion at the University Medical Center Eppendorf in Hamburg. M tuberculosis–infected donors were identified as described previously.²⁷  All donors gave informed consent in accordance with the Declaration of Helsinki, and the local ethics committee approved this study (W-007/09).

A polyclonal rabbit-α-human SOCS3-antibody (Abcam) was labeled with DyLight 488 or DyLight 649 fluorescent dye (Pierce Chemical) following the manufacturer's instructions. For SOCS3 analysis, T cells or T-cell clones were fixed and permeabilized with Cytofix/Cytoperm solution (BD Biosciences) following the manufacturer's instructions. The cells were then stained intracellularly with different antibodies and labeled αSOCS3 antibody. Antibody specificity was determined by preincubation of 10 μg/mL αSOCS3 antibody with different concentrations of the SOCS3 peptide used to generate the antibody (C-terminal 26 aa, Abcam) or a randomly selected irrelevant peptide for 30 minutes at room termparature. Flow cytometric analyses were generally performed using LSRII flow cytometer (BD Biosciences) and FCS Express Version software (DeNovo). For detailed descriptions of different procedures concerning analysis of T-cell subpopulations, in vitro stimulated T cells, and T-cell clones, see supplemental Methods (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

CD4⁺ T cells (7.5 × 10⁵ cells/mL) were stimulated with αCD3/CD28 (Invitrogen, 4 × 10⁵ beads/mL) for 2 days. A total of 3 × 10⁶ T cells were then lysed in 150-μL sample buffer (50mM Tris-base, 2% SDS, 5% glycerol, bromophenol blue, pH 6.8, supplemented with 100mM DTT). Denaturation of proteins was performed for 15 minutes at 70°C. The soluble proteins were separated using a 15% polyacrylamide gel and subsequently blotted on a nitrocellulose membrane. After blocking for 1 hour at room temperature (PBS, 3% BSA, 0.5% Tween 20) the following primary antibodies were incubated overnight at 4°C: polyclonal antibody rabbit-αSOCS3 (0.5 μg/mL, Abcam) alone or preincubated (30 minutes at room temperature) with the SOCS3 peptide (0.5 μg/mL), the antibody was raised against (C-terminal 26 aa, Abcam), or monoclonal antibody mouse-α-β-actin (0.5 μg/mL, Santa Cruz Biotechnology). Afterward, the following secondary antibodies were incubated for 45 minutes at room temperature: goat-αrabbit-IgG-HRP (Dako North America), rabbit-αmouse-IgG-HRP (Dako North America). Blots were developed for 5 minutes using 1 mL ECL (Pierce Chemical).

Purified PBMCs (2 × 10⁷) were stained with mAbs (αCD4 APC-Cy7, αCD45RO PE-Cy-7, both BD and αCD161 FITC, Miltenyi Biotec), and CD4⁺ CD45RO^highCD161^high or CD4⁺ CD45RO^highCD161^low T cells were sorted into 96-well, round-bottom plates (1, 2, 3, 5, or 10 cells/well) using a FACSAria II (BD Biosciences). The plates were precoated with the T-cell stimulating αCD3 antibody (OKT3) before sorting and irradiated heterologous antigen-presenting cells (1 × 10⁵ cells/well) were added. On days 0, 7, 14, and 21, IL-2 (20 U/mL) was added. After 14 days, expanded T-cell clones were counted for each plate. Only T-cell clones from plates with < 33% positive wells were selected to ensure derivation from a single precursor cell.

Freshly isolated PBMCs (2 × 10⁶ cells/well) from latently M tuberculosis infected donors were stimulated with purified protein derivative (PPD) of M tuberculosis (2 μg/mL) for 6 days. Concomitantly autologous monocytes were isolated using CD14-labeled magnetic beads (BD Biosciences) and the IMag system (BD Biosciences) following the manufacturer's instructions. The average purity determined by flow cytometry was > 95%. Enriched monocytes (1 × 10⁵/well) were cultured in RPMI 1640 medium (PAA; supplemented with 7.5% FCS, 1% L-glutamine) cells for 6 days to generate monocyte-derived macrophages (MDMs). Afterward, MDMs were infected with GFP-expressing Mycobacterium bovis BCG (MOI 5:1) for 3 hours at 37°C. Then 5 × 10⁵ cells/well of the PPD-stimulated PBMCs were cocultured with BCG-GFP–infected or noninfected MDMs for 3, 8, and 20 hours. Afterward, SOCS3 expression was analyzed by flow cytometry, and relative expression values have been compared between T cells cocultured with BCG-GFP infected and noninfected MDMs.

For ectopic expression of SOCS3, the open reading frame of human SOCS3 was cloned into the lentiviral LeGO-iG2 vector²⁸  using the restriction sites BamHI and EcoRI (oligos 5′-GCGCGGATCCATGGTCACCCACAGC-AAG-3′ and 5′-GCGCGAATTCTTAAAGC-GGGGCATCGTACT-3′). In this vector, SOCS3 was expressed under the control of a retroviral spleen focus-forming virus promoter concomitantly with eGFP under the control of an internal ribosomal entry side. As a control, vector LeGO-G2 was used, which encodes for eGFP under a spleen focus-forming virus promoter. Lentiviral particles were produced by transient transfection of HEK293T cells as described previously²⁹  using a third-generation lentiviral transduction system.³⁰  Briefly, 5 × 10⁶ HEK293T cells were cotransfected with the SOCS3 expression vector (20 μg), pMDLg/pRRE (10 μg; Addgene; which encodes for lentiviral gag/pol), pRSV-Rev (5 μg; Addgene; which encodes for lentiviral rev), and phCMV-GALV-C4070A (5 μg; which encodes for the gibbon ape leukemia virus (GALV)–envelope protein, kindly provided by C. Stocking, Heinrich-Pette-Institut, Hamburg, Germany). The cotransfection was performed using the calciumphosphate transfection method with 2× HBS (Sigma-Aldrich) and CaCl₂ (2.5M). Virus-containing supernatants were harvested twice at 24 and 48 hours after transfection, respectively. The relative virus titer was determined by transduction of HEK293T cells in the presence of Polybrene (8 μg/mL) and flow cytometry analysis of eGFP expression 3 days after infection. Titers of 5 × 10⁶-2 × 10⁷ infectious units/mL were obtained. CD4⁺ T cells were transduced with lentiviral particles at an MOI of 10 using retronectin (TaKaRa) according to the manufacturer's instructions. In brief, CD4⁺ T cells were isolated with the IMag system (BD Biosciences) using anti-CD4 antibodies attached to magnetically labeled beads following the manufacturer's instructions. Then 7.5 × 10⁵ cells/mL were prestimulated for 2 days with αCD3/CD28 (see “SOCS3 analysis by Western blot”). The 24-well plates were precoated with 9 μg/cm² Retronectin (TaKaRa). Plate preloading with viral particles was performed by centrifugation (60 minutes, 1000g, 4°C). The supernatant was discarded, and 2.5 × 10⁵ stimulated T cells in X-Vivo 15 medium (500 μL) were added. On day 1, fresh X-Vivo medium was added.

For analysis of IL-7–dependent changes in SOCS3 expression and cytokine production, recombinant IL-7 (20 ng/mL; Sigma-Aldrich) was added once on the day of transduction. For cytokine expression analyses in IL-7–treated and non–IL-7–treated samples, ratios (“IL-7–treated”/“non-IL-7–treated” T cells) of IFN-γ, TNF-α, and IL-17-expressing T-cell proportions were determined for samples of individual donors. IL-2 was not added in these experiments.

For analyses of T-cell proliferation in transduced T cells, the cell proliferation dye eFluor670 (eBioscience) was applied following the manufacturer's instructions. Proliferation of T cells was analyzed by flow cytometry calculating the median fluorescence intensities (MFIs) of cell proliferation dye eFluor670 on days 2-7 after transduction. Decreased MFIs correlate with increased proliferation. For cytokine analyses in “fast” and “slow” proliferating T cells, proliferation dye labeled T cells were restimulated on day 5 after transduction with phorbol myristate acetate (PMA; 10 ng/mL) and ionomycin (1 μg/mL) and IFN-γ, TNF-α, and IL-17–expressing T cells were determined by intracellular cytokine staining (as described in “Analyses of cytokine expression by intracellular cytokine staining”). Proliferation dye^high (“slow” proliferating) and dye^low (“fast” proliferating) T cells were gated and analyzed for the proportions of cytokine expression. Afterward, the ratio (proliferation dye^high/dye^low) was calculated for each cytokine. Quadruplicates of 4 donors have been included in these experiments.

Cytokine expression of transduced T cells was determined after 6 hours in vitro restimulation with PMA (10 ng/mL) and ionomycin (1 μg/mL) in presence of brefeldin A 2 days after transduction. After fixation and permeabilization (as described in “SOCS3 expression by flow cytometry”), the following antibodies were used for staining: αTNF-α AlexaFluor-700 (BD Biosciences), αIFN-γ PE-Cy7 (BD Biosciences), αIL-4 PE (BD Biosciences), and αIL-17 Horizon V450 (BD Biosciences).

For analysis of cytokine expression in the supernatant, eGFP^positive (SOCS3 and control vector) and eGFP^negative T cells were sorted 2 days after transduction by FACS. Purified T cells (3 × 10⁴/well) were stimulated with PMA/ionomycin (see paragraph: “Analyses of cytokine expression by intracellular cytokine staining”) without the Golgi inhibitor brefeldin A. Supernatants were then harvested 3 days after stimulation, and cytokine analysis was performed with the cytometric beads array (human T_H1/T_H2/T_H17 Kit; BD Biosciences) following the manufacturer's instructions. For data analyses, the FCAP Array Software Version 1.0.2 was used. A factor of 3.5, used for normalization of differential cell numbers between SOCS3-transduced T cells and controls, was calculated of 4 independent experiments. Because of the high numbers of T cells needed for sorting in the cytometric beads array experiments, only 3 experiments were performed.

Statistical analyses of comparisons between 2 independent groups have been performed using the Student t test (for Gaussian distributions) or the Mann-Whitney U test (for non-Gaussian distributions) according to Kolmogorov-Smirnow normality testing (SigmaStat Software Version 3.5). The paired t test was applied to assess differences of SOCS3 expression induced by in vitro stimulation with PPD and αCD3/CD28, in T_H17 (CD161^high), and T_H2 (CRTH2^high) precursor cells, for cytokine analysis after SOCS3 lentiviral transduction, and stimulations in presence or absence of IL-7. The Mann-Whitney U test was applied to assess SOCS3 expression differences in T-cell clones. Statistical analyses were accomplished using the GraphPad Prism Software Version 5.0 or the SigmaStat Software Version 3.5. Nominal 2-sided P values given were considered significant if P < .05.

Previous studies determined SOCS3 expression using real-time PCR for RNA quantification or Western blots for semiquantitative protein analysis. To precisely quantify SOCS3 protein expression in a cell-specific manner, we established antibody-based flow cytometric analysis for quantification of SOCS3. A polyclonal rabbit-α-human SOCS3 antibody was fluorescently labeled, and MFIs of different antibody concentrations were determined intracellularly. Results of a representative experiment indicating SOCS3 MFIs of PBMCs are shown as Figure 1A. Depending on the labeling efficiency, an antibody concentration of ∼ 10 μg/mL was sufficient to detect SOCS3, and adjusted antibody concentrations were therefore used throughout the experiments. The capability of this flow cytometric assay to specifically quantify SOCS3 protein expression based on MFI was determined by applying antibodies preincubated with increasing concentrations of the SOCS3 peptide epitope. The SOCS3 peptide epitope markedly reduced the MFI, and 0.5 μg/mL of the peptide was sufficient to block SOCS3 binding at most (Figure 1B). Higher concentrations had no additional effect, and a randomly chosen control peptide did not affect binding of the antibody at any concentration (Figure 1B). A 50% inhibition level was detected for a peptide concentration of 0.1 μg/mL. Because SDs of MFI between replicates in the same experiment are moderate (Figure 1B), we conclude that MFI measure is a highly accurate method for quantification of SOCS3 protein expression.

Protein analysis using Western blot detected a band at ∼ 25 kDa in accordance to the predicted molecular weight of SOCS3 and a second, slightly smaller, band matched to a described SOCS3 splice variant (Figure 1C left panel second lane).³¹  Western blot analyses hardly detected SOCS3 in primary human T cells ex vivo (Figure 1C first lane), whereas markedly increased SOCS3 expression was detected after in vitro T-cell activation (Figure 1C second lane). Again, antibody preincubation with the SOCS3 peptide epitope largely diminished intensities of assumed SOCS3 bands but not β-actin used as a control (Figure 1C fourth lane). Notably, a very moderate decrease was detected in the nonactivated T cells (Figure 1C third lane) compared with the nonpeptide control (Figure 1C first lane). Concomitant flow cytometric analysis proved the marked increase of SOCS3 MFI after T-cell activation and a decrease of SOCS3 MFI by the blocking peptide (Figure 1D). Notably, MFI of nonstimulated T cells, although ∼ 2-fold lower (compared with the stimulated sample), decreased markedly (∼ 14-fold) when SOCS3 peptide was added. These results prove SOCS3 antibody specificity and usability in flow cytometric assays for determination of protein concentrations by MFI. In addition, these results demonstrated the higher sensitivity and accuracy of the flow cytometry–based assay compared with semiquantitative Western blot.

Previous studies indicated a role of SOCS3 in human tuberculosis.^24,25  Therefore, we characterized SOCS3 expression in activated T cells (CD40L^high) after in vitro stimulation with M tuberculosis antigen (PPD). PPD induced moderately increased SOCS3 expression in activated T cells with a maximum of 2.5-fold after 48 hours compared with CD40L^low (nonactivated) T cells (Figure 1E). The PPD-specific SOCS3 induction was comparable with the polyclonal T-cell activator αCD3/CD28, which induced up to 2-fold increased SOCS3 expression in CD40L^high T cells (Figure 1E). Increased SOCS3 expression levels of in vitro activated T cells were stable for at least a week (data not shown).

To characterize the direct effect of mycobacterial infection on SOCS3, we established an in vitro assay based on T-cell interaction with M bovis BCG-GFP (BCG) infected monocyte-derived macrophages (MDMs). T cells specifically stimulated with mycobacterial proteins were coincubated with autologous MDMs with or without BCG infection. Coincubation with infected MDMs induced a tendency of transient down-regulation of SOCS3 in T cells (P = .1) during the first 3 hours (Figure 1F). We conclude that T-cell activation led to increased expression of SOCS3 and that transient down-regulation of SOCS3 in T cells occurred early during interaction with BCG infected MDMs.

SOCS3 has been described to polarize T-cell immunity, leading to a different cytokine expression pattern. Therefore, we performed concomitant analyses of SOCS3 and cytokine expression after in vitro T-cell stimulation. αCD3/CD28 stimulation induced similar SOCS3 expression levels in IFN-γ, TNF-α, and IL-2–producing T cells (Figure 2A). Hardly any IL-17 and IL-4 was induced after short-term in vitro stimulation. Consequently, we generated T-cell clones based on a single-cell sorting approach using irradiated feeder cells and CD3 cross-linking (for details see “Methods”). This method induced T_H1, T_H2, and T_H17 cells without application of polarizing culture conditions. T-cell clones were then characterized for cytokine expression pattern (ie, TNF-α, IFN-γ, IL-17, IL-4). Representative examples of T-cell clones are shown as Figure 2B. T-cell clones were then classified according to high or low expression of TNF-α, IFN-γ, IL-4, and IL-17, respectively. Similar SOCS3 expression was detected in T-cell clones expressing high or low levels of IFN-γ or TNF-α (Figure 2C). In contrast, significantly higher SOCS3 expression was detected in IL-4^high and IL-17^high T-cell clones (Figure 2C). Because inhibition of T_H17 polarization by SOCS3 has been described before,^16,32  we additionally analyzed the ex vivo expression of SOCS3 in CD161^high T helper cells (the T_H17 originating population in humans³³ ). We detected higher SOCS3 expression levels in CD161^high T cells compared with CD161^low T cells (P = .02, Figure 2D left graph) and a tendency of higher SOCS3 expression in described CRTH2^high T_H2 cells³⁴  (P = .067, Figure 2D middle graph), whereas no difference was detected between control T-cell populations discriminated by CD45RA expression (Figure 2D right graph). We conclude that SOCS3 expression is increased in IL-17^high T-cell clones as well as in primary CD161^high polarized T_H17 cells.

To investigate functional effects of differential SOCS3 expression, we used a lentiviral transduction system for ectopic SOCS3 expression in activated primary CD4⁺ T cells. Because T-cell receptor specific activation was essential for transduction, we were not able to modulate SOCS3 expression in nonpolarized naive T cells (data not shown). Ectopic SOCS3 was indicated by concomitant eGFP expression. SOCS3 expression analysis by flow cytometry revealed significantly increased expression after lentiviral transduction (P = .008; Figure 3A). The transduction efficiency varied between 50% and 90% for the control vector and 30% and 60% for the SOCS3 vector in different experiments (data not shown). Proliferation analyses revealed a marked reduction of cell divisions in CD4⁺ T cells with ectopic SOCS3 expression (Figure 3B top histogram) compared with vector control–transduced T cells (Figure 3B bottom histogram). This was accompanied by decreased proportions of SOCS3-transduced T cells during in vitro culture (Figure 3C). We did not detect increased expression of apoptosis- or necrosis-associated markers on SOCS3-transduced T cells (data not shown).

Next, we analyzed cytokine expression of SOCS3-transduced T cells after in vitro restimulation. Polyclonal T-cell activation revealed similar proportions of TNF-α and IFN-γ in T cells with ectopic SOCS3 expression and vector control–transduced T cells (Figure 4A left and middle graphs). Notably, we detected increased proportions of IL-17-producing CD4⁺ T cells with ectopic SOCS3 expression (P = .015; Figure 4A right graph). IL-4 was not detectable after restimulation. Because differences in cytokine expression induced by SOCS3 can be the result of differential proliferation of distinct T_H-cell subpopulations, we next compared cytokine expression of “slow” and “fast” proliferating T cells. Concomitant analyses for cell proliferation and cytokine-expression detected cytokine-positive T cells predominantly in the “fast” proliferating population (Figure 4B). The median ratios (“fast”/“slow” proliferating cells) were between 2/1 and 3/1 for TNF-α and IFN-γ. Notably, the ratios were significantly higher (> 3/1) for IL-17–expressing T cells. Comparable results were gained for nontransduced, vector-transduced, and SOCS3-transduced T cells (Figure 4B). This demonstrated that T_H17 cells are even faster proliferating than T_H1 cells and does not support the notion that increased IL-17 expression induced by SOCS3 is the result of reduced proliferation.

Next, we determined cytokines in the supernatants of transduced T cells. For these experiments, eGFP⁺ (transduced) T cells were enriched by FACS before restimulation. We detected comparable IFN-γ and TNF-α levels between T cells with ectopic SOCS3 expression and controls (values were normalized for differential cell counts; for details see “Methods”; Figure 4C). Notably, as for intracellular cytokines, a marked increase of IL-17 was detected in the supernatant of T cells with ectopic SOCS3 expression (Figure 4C). We conclude that ectopic SOCS3 expression reduced proliferation of activated T cells and induced increased proportions of IL-17–producing T cells.

The effects of SOCS3 down-regulation were addressed using described small hairpin (sh)RNAs for SOCS3 in a lentiviral vector (for details see supplemental Methods). We analyzed 5 commercially available SOCS3-specific shRNAs for reduction of SOCS3 expression in HEK293T cells using flow cytometry (supplemental Figure 1A) and Western blot (supplemental Figure 1B). None of these shRNAs down-regulated SOCS3 protein expression consistently, whereas eGFP-specific control shRNA reduced eGFP expression (supplemental Figure 1C). Similar results were gained for CD4⁺ T cells (data not shown). Therefore, we were not able to analyze the influence of SOCS3 down-regulation using shRNAs.

Recently, it has been shown that IL-7 inhibits SOCS3 expression in exhausted T cells from mice with chronic viral infections.²⁶  Therefore, we determined the effect of IL-7 on SOCS3 expression of human CD4⁺ T cells and accompanied functional differences. Treatment with recombinant IL-7 before or after lentiviral transduction of T cells did not affect SOCS3 transduction efficiency (data not shown) but led to reduced endogenous expression of SOCS3 (∼ 20%; P = .045; Figure 4D). Analyses of cytokine expressing T-cell proportions (ratios of “IL-7–treated”/“non-IL-7–treated” samples) revealed slightly increased proportions of TNF-α-expressing T cells on IL-7 treatment (Figure 4E), whereas IFN-γ-expressing T cells were similar between IL-7–treated and nontreated T cells (Figure 4E). Notably, IL-17-expressing T-cell proportions were significantly reduced compared with IFN-γ and TNF-α-producing T cells when IL-7 was added to culture (Figure 4E). These results were concordant between SOCS3-transduced T cells and controls. Therefore, IL-7–treated T cells with reduced SOCS3 expression comprised decreased T_H17 proportions. These results strengthen the hypothesis that SOCS3 promotes T_H17 cells.

In the present study, we show that SOCS3 influences human T-cell functions and polarization. Our data suggest that increased SOCS3 expression promotes (1) IL-4–producing T-cell clones, (2) IL-17–producing primary T cells, T-cell clones, and T_H17 precursor cells, and (3) decreased proliferation of T cells in vitro. IL-7–induced inhibition of SOCS3 expression leads to reduced proportions of IL-17–producing T_H17 cells.

SOCS3 has been shown to influence polarization of CD4⁺ T cells. In this context, an enhancing effect of SOCS3 on the generation of IL-4–producing T_H2 cells has been demonstrated.^7-10  This SOCS3-mediated T_H2 promoting effect is the result of inhibition of T_H1 by blocking STAT4 phosphorylation.⁷  In accordance, we detected higher SOCS3 expression in IL-4–expressing T-cell clones as well as a tendency of higher SOCS3 expression in CRTH2^high-expressing T_H2 cells³⁴  in the present study. Because short-term in vitro stimulation hardly induced any IL-4–producing T_H2 cells, we were not able to determine T_H2 proportions in response to ectopic SOCS3 expression by lentiviral transduction of primary T cells.

More recent studies did not observe an influence of SOCS3 on T_H2 polarization, but rather an inhibitory effect of SOCS3 on T_H17 cells.^16,35  In contrast, the results from the present study did not suggest an inhibitory effect of SOCS3 on T_H17. We identified increased SOCS3 expression in T cell clones with predominant T_H2 and T_H17 cytokine expression pattern. In addition, ex vivo analyzed T_H17 (CD161^high) cells exhibited increased SOCS3 expression and T_H2 (CRTH2^high) cells showed a tendency of higher SOCS3 expression. In these experiments SOCS3 expression differences may be underestimated because T_H17 cells are only a fraction of CD161^high T cells.²³  Increased proportions of T_H17 cell were also found induced by ectopic SOCS3 expression. Altogether, these results suggested a promoting role for SOCS3 in human T_H17 development.

The molecular mechanisms how increased SOCS3 expression favors human T_H17 cells remain elusive. The immune polarizing effect of SOCS3 is mainly assigned to its STAT inhibitory functions. In this context, SOCS3 blocks phosphorylation of STAT4, the main inducer of T_H1,^7,36  and phosphorylation of STAT3, the critical inducer of T_H17.^16,37  A possible explanation for our contrary findings concerning T_H17 is that SOCS3 inhibits polarization of noncommitted naive T cells (T_H0) toward T_H17, whereas differentiated T_H17 or preT_H17 cells are not (or less) affected by SOCS3-mediated inhibition of STAT3 phosphorylation (eg, by differential effects of SOCS3 on distinct cytokine receptor–mediated STAT3 phosphorylation³⁸ ). This thesis could not be examined in the present study because of assay restrictions and limitations of the human model. Among others, analysis of CD161 cannot distinguish between T-cell differentiation stages because CD161 is expressed on T_H17 as well as on naive T_H17 precursors. Furthermore, activation of T cells before lentiviral transduction is a prerequisite for the method; therefore, we were not able to analyze the influence of SOCS3 on nonpolarized T cells in this assay. Nevertheless, because short-term in vitro stimulation with low concentrations of T-cell stimulating beads activates predominantly effector and memory cells (M.J., unpublished data, March 2008), it is justified to conclude that SOCS3 modulation by lentiviral transduction affected precommitted T_H17 cells and not naive T_H17 precursors in our assay. Experiments that address the point whether distinct cytokine receptors (eg, IL-23 or IL-6 receptors) are differentially affected by SOCS3 are difficult to perform in our experimental setting because SOCS3 simultaneously inhibits proliferation (Figure 3B) and SOCS3-transduced T cells disappear in culture within a few days (Figure 3C).

Besides STAT-dependent regulation of T-cell differentiation by SOCS3, other mechanisms are also possible. Quintana et al described very recently that down-regulation of IL-2 expression by the transcription factor Aiolos promoted T_H17 differentiation.³⁹  SOCS3 has also been shown to inhibit IL-2 by directly blocking activation of CD28.⁶  We conclude that SOCS3 may affect T_H17 polarization in different ways. Ongoing studies aim at deciphering this complex picture to clarify the role of SOCS3 in T-cell polarization.

Recombinant IL-7 has recently been described to directly inhibit SOCS3 expression of CD8⁺ T cells in animal models of chronic viral infections.²⁶  In the present study, we demonstrated that IL-7 inhibited SOCS3 also in human CD4⁺ T cells. Because we were not able to down-regulate SOCS3 using shRNAs, we applied recombinant IL-7 to suppress SOCS3. Here we observed a promoting effect of IL-7 predominantly on TNF-α-producing T cells, which were more frequent in the presence of IL-7, whereas IFN-γ–expressing T cells were found at similar frequencies. In contrast, IL-17–producing T cells decreased in the presence of IL-7, and this has been shown by others, too.⁴⁰  We cannot exclude that IL-7 has additional effects on T-cell functions besides SOCS3 suppression; therefore, exclusive inhibition of SOCS3 would be important to verify these results. Despite the limitations of this assay, we conclude that the IL-17–suppressive effect of IL-7 supported our hypothesis for a T_H17 promoting role of SOCS3.

In the present study, we showed that ectopic SOCS3 expression in CD4⁺ T cells strongly inhibited proliferation. This is in accordance with previous studies that showed reduced proliferation of T cells caused by SOCS3-mediated blocking of CD28 coreceptor activation.^4,6,41,42  Defective proliferation is a feature of T-cell exhaustion that has been attributed to T cells in chronic viral diseases, and SOCS3 plays a central role in these processes.²⁶  It is tempting to speculate that T-cell exhaustion also plays a role in human tuberculosis where SOCS3 expression is increased during active disease.^24,25  Previously, we identified exhausted CD8⁺ T cells in children with severe forms of tuberculosis.⁴³ 

Besides infectious diseases, SOCS3 plays an important role in autoimmunity⁴⁴  and tumors.⁴⁵  Consequently, SOCS3 has become a target for interventive therapy.^2,46  Against this background and controversial results in animal models, we would strongly suggest to place the role of SOCS3 in T-cell polarization and exhaustion on today's agenda. In this context, we consider the present study an important step to elucidate the function of SOCS3 in human T-cell polarization.

The authors thank B. Kretschmer for comments on the manuscript and M. Milsom for technical help, particularly the protocol for cloning of shRNAs from pLKO.1 into LeGO-G.

This work is part of the PhD thesis of K.K.

This work was supported by the Deutsche Forschungsgemeinschaft project, Immune polarization in childhood TB (JA 1479/3-1; S.S. and M.J.). U.M. was supported by the BMBF within the iGene consortium.

Contribution: K.K. performed experiments, analyzed the data, and wrote the manuscript; K.H. contributed to the establishment and generation of T-cell clones; S.S. performed the generation and infection of human MDMs; C.S.-J. performed fluorescence-activated single cell sorting; U.M., K.R., and B. Fehse supported establishment of T-cell transduction with lentiviral LeGO vectors; B. Fleischer contributed to the concept and supervision; M.J. performed experiments, analyzed the data, supervised the study, and wrote the manuscript; and all authors read and approved all versions of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Marc Jacobsen, Pediatric Infectious Diseases Group, Department of General Pediatrics, Neonatology, and Pediatric Cardiology, University Children's Hospital, Moorenstrasse 5, 40225 Duesseldorf, Germany; e-mail: marc.jacobsen@med.uni-duesseldorf.de.