Th17 cells are defined by their capacity to produce IL-17, and are important mediators of inflammation and autoimmunity. Human Th17 cells express high levels of the retinoic acid-related orphan receptor variant 2 (RORC2), but it is currently unclear whether expression of this transcription factor alone is sufficient to recapitulate all the known properties of Th17 cells. We used lentivirus-mediated transduction to investigate the role of RORC2 in defining aspects of the human Th17 cell lineage. Expression of RORC2 induced production of IL-17A, IL-22, IL-6 and TNF-α, a Th17-cell-associated chemokine receptor profile and upregulation of CD161. RORC2-transduced T cells were hypo-responsive to TCR-mediated stimulation, a property shared with ex vivo Th17 cells and overcome by addition of exogenous IL-2 or IL-15. Co-culture experiments revealed that RORC2-expressing cells were partially resistant to Treg cells since production of IL-17 and proliferation were not suppressed. Evidence that IL-17 stimulates CD4+ T cells to produce IL-2 and proliferate suggested that the resistance of Th17 cells to Treg-mediated suppression may be partly attributed to IL-17 itself. These findings demonstrate that expression of RORC2 in T cells has functional consequences beyond altering cytokine production and provides insight into the factors regulating the development of human Th17 cells.
CD4+ Th cells are classically divided into functional subsets based on cytokine production: Th1 cells regulate cellular immunity and produce IL-2 and IFN-γ, while Th2 cells regulate humoral immunity and produce IL-4, IL-5 and IL-13 1, 2. More recently, a Th-cell lineage defined by its capacity to secrete IL-17 (Th17 cells) has been identified 3–5. Mouse models have demonstrated that Th17 cells are critical for host defense against extracellular pathogens 6, 7, whereas their aberrant expansion is linked to the pathogenesis of inflammatory autoimmune disorders 8. Although Th17 cells produce several inflammatory cytokines, many of their effector functions are attributed to IL-17 production. IL-17 is known to recruit neutrophils and stimulate the production of pro-inflammatory cytokines, chemokines and antimicrobial peptides from a variety of immune and non-immune cells 8–11. Evidence from studies documenting increased levels of IL-17 in the peripheral blood and tissues of patients with rheumatoid arthritis, multiple sclerosis and inflammatory bowel disease 12–14, and experimental models demonstrating a clear role for Th17 cells in the pathogenesis of these diseases, suggests that strategies to inhibit the differentiation and/or function of Th17 cells in vivo may result in new therapies for inflammatory and autoimmune diseases.
Most of the current knowledge regarding the phenotype and function of Th17 cells is based on studies carried out in mice. Mouse Th17 cells are defined by their capacity to produce IL-17 (IL-17A), IL-17F, TNF-α, IL-6 and IL-22, but not IFN-γ 10, 15–17. In vitro, TGF-β1 and IL-6 polarize naive mouse CD4+ T cells into Th17 cells 18–20, whereas IL-23 is thought to be important for their expansion, survival, effector function and pathogenicity in vivo20–22. At the molecular level, at least five transcription factors are involved in normal Th17-cell development: Stat3 17, 23, 24, a T-cell-specific splice isoform of retinoic acid receptor-related orphan receptor (Ror) known as Ror-γt or Rorc2 25, 26, a splice isoform of Ror-α (Ror-αd) 27, Runx1 28, and IFN-regulatory factor 4 29. Deficiency in ror-γt results in a major reduction in the number of Th17 cells and protection from experimental autoimmune encephalomyelitis; by corollary, over-expression of Ror-γt promotes Th17 development. Recently, it was shown that Ror-αd or Runx 1 can synergize with Ror-γt to promote IL-17 expression 27, 28.
Human Th17 cells express IL-17A, IL-17F, TNF-α, IL-6 and IL-22 30–35, as well as IL-26 26. Many groups have also examined the expression of chemokine and cytokine receptors on human Th17 cells 26, 30, 31, 34–37, with the overall conclusion that co-expression of CCR4, CCR6 and the IL-23R, but not CXCR3, is characteristic and best distinguishes Th17 cells from other human Th-cell subsets. More recently, Th17 cells were reported to express CD161, the human homolog of NK1.1 38.
The cytokine combination that optimally induces the development of human Th17 cells in vitro is a subject of much debate. Initial studies found that T-cell activation in the presence of IL-1β, IL-6 and/or IL-23 was sufficient to induce Th17 cells, and that TGF-β1 inhibited this process 30, 32, 35, 39–41. In contrast, other groups found that TGF-β1 was indeed important for the development of human IL-17-producing cells 33, 42 and expression or retinoic acid-related orphan receptor variant 2 (RORC2), the human ortholog of Ror-γt 26. This discrepancy could be explained by a recent report showing that the requirement for TGF-β in differentiation of human Th17 cells in vitro is indirect, and related to suppression of Th1 development 43.
It is well known that like their mouse counterparts, human Th17 cells express RORC2 31, 32, 39, and over-expression of RORC2 in cord blood CD4+ T cells induces expression of IL-17A, IL-17F, IL-26 and CCR6, but not IL-22, CCR4 or CCR2 26, lending support to the notion that RORC2 is a Th17-cell-specific lineage transcription factor. However, cells ectopically expressing RORC2 have neither been directly compared with cytokine-induced or ex vivo Th17 cells, nor have they been tested for alterations in biological function. Additionally, little is known about the direct effects of IL-17 on human CD4+ T cells, which are known to express the IL-17R 44. To address these questions we performed a comprehensive investigation of the role of RORC2 in human Th17 lineage commitment and examined the immunoregulatory activity of IL-17 on CD4+ T-cell subsets.
Enforced expression of RORC2 induces IL-17 production in human CD4+ T cells
In order to test if expression of RORC2 was sufficient to induce the development of human Th17 cells, we developed a bi-directional lentiviral vector in which expression of RORC2 was placed under control of the EF1-α promoter. A non-signaling, truncated nerve growth factor receptor (ΔNGFR) was co-expressed as a cell-surface marker for tracking and sorting of transduced cells (Fig. 1A). Naive CD4+CD25−CD45RO− T cells were activated and transduced with either the pCCL control vector or pCCL.RORC2 and expanded in the presence of IL-2 45. After 6 days, 70–86% (mean 78±8%, n=5) of control pCCL-transduced cells were ΔNGFR+ compared with 48–82% (mean 65±17%, n=5) ΔNGFR+ cells in cultures transduced with pCCL.RORC2 (Fig. 1B), a difference consistent with less efficient packaging of the larger viral genome 46 resulting from the additional RORC2 cDNA sequence. Eight days after transduction, ΔNGFR+ T cells were purified and expanded for analyses. Following purification, transduced cells were >98% positive for NGFR expression. All experiments to assess phenotypic and functional changes induced by RORC2 over-expression were performed following purification of transduced cells. Untransduced controls were included for all subsequent experiments and no differences with pCCL control transduced were observed.
RORC2 expression was measured by quantitative RT-PCR in resting ΔNGFR+ T cells to confirm increased expression of this transcription factor in pCCL.RORC2-transduced cells. On average, resting RORC2-transduced T cells had ∼387-fold higher RORC2 expression than pCCL-transduced cells (p=0.0027, n=5) (Fig. 1C). Expression of RORC2 in pCCL.RORC2-transduced T cells remained stable in subsequent rounds of expansion (data not shown).
To determine whether expression of RORC2 induced IL-17 production by T cells, ΔNGFR+ T cells were stimulated with α-CD3/α-CD28-coated beads, and supernatants were analyzed by ELISA. pCCL.RORC2-transduced T cells produced significantly higher amounts of IL-17 than controls: after 72 h, pCCL-transduced T cells produced, on average, 0.34±0.28 ng/mL IL-17, whereas pCCL.RORC2-transduced cells produced 3.2±1.1 ng/mL (n=5, p=0.0057) (Fig. 1D). Optimal production of IL-17 by pCCL.RORC2-transduced T cells required co-stimulation with α-CD28; at 10 μg/mL of immobilized α-CD3, pCCL.RORC2-transduced T cells produced on average 2.0±0.9 ng/mL of IL-17 compared with 4.6±2.2 ng/mL in the presence of α-CD28 (n=5, p=0.008) (Fig. 1D). In order to confirm these results, the capacity of transduced T cells to produce IL-17 was also determined by intracellular cytokine staining (Fig. 1E). The percentage of IL-17-producing T cells was significantly higher in pCCL.RORC2-transduced T cells compared with pCCL-transduced T cells (mean IL-17+ cells was 23.2±10.9% for pCCL.RORC2-transduced T cells versus 2.0%±1.9% for control cells; n=6, p=0.0002), whether or not the cells were activated with PMA/Ca2+ ionophore or with α-CD3/α-CD28-coated beads (Fig. 1E and data not shown). The percent of IL-17+ cells in pCCL-transduced T cells was equivalent to untransduced controls and consistent with previous reports from T cells kept in culture for 12 days 39.
We also examined whether the amount of IL-17 produced by RORC2-transduced cells could be further enhanced by cytokines associated with in vitro differentiation of Th17 cells. As shown in Fig. 1F, IL-1β significantly increased IL-17 production by RORC2-transduced cells by 2.0±0.4 fold (n=3, p=0.02). Addition of IL-23, IL-6 or TGF-β1 did not induce significant changes in IL-17 production.
Over-expression of RORC2 induces a Th17 cytokine profile
In addition to IL-17, human Th17 cells are well known to produce IL-6, IL-22 and TNF-α, but not IL-4, and little or no IFN-γ 17, 26, 31–33, 39. To define whether over-expression of RORC2 fully recapitulated the Th17 cytokine production phenotype, we directly compared cytokine production by pCCL.RORC2-transduced T cells with that of Th17 cells generated in vitro by cytokine-induced polarization. Naive CD4+CD25−CD45RO− T cells were activated with IL-2 (Th0), IL-12 and IL-2 (Th1) or a combination of IL-1β, IL-6 and IL-23 (Th17) 32, 35, 39. Although, as expected, cytokine-induced Th17 cells upregulated RORC2, the level of RORC2 mRNA in these cells was consistently lower than in RORC2-transduced cells, which on average, expressed 72-fold more RORC2 than donor-matched Th17-polarized cells (p=0.01, n=4, data not shown). Due to recent evidence regarding a possible requirement for TGF-β 26, 33, 42, we also performed parallel differentiation assays with IL-1β, IL-6, IL-23 and TGF-β1 at concentrations ranging from 0.1 to 10 ng/mL, but found that TGF-β1 inhibited IL-17 expression (data not shown), consistent with previous reports 32, 35, 39.
Following stimulation with α-CD3/α-CD28-coated beads, cytokine-induced Th17 cells and pCCL.RORC2-transduced T cells produced similar amounts of IL-17 (pCCL.RORC2: 3.2±1.5 ng/mL versus Th17: 2.3±1.4 ng/mL IL-17, n=3) (Fig. 2A). In contrast to the similar levels of secreted IL-17, the percentage of IL-17+ cells in cultures of cytokine-induced Th17 cells (5.6±2.8%, data not shown) was significantly lower than in RORC2-transduced cells (Fig. 1E). As expected, cells differentiated in Th0- or Th1-polarizing conditions, or pCCL-transduced T cells expressed low amounts of IL-17. Similar to cytokine-induced Th17 cells, over-expression of RORC2 also resulted in a significant increase in the proportion of cells producing IL-22 (pCCL mean 8.6.±4.5% versus pCCL.RORC2 mean 14.7±0.7%, n=4, p=0.048) and a decrease in IFN-γ (pCCL mean 38.7±3.3% versus pCCL.RORC2 mean 11.5±4.3%, n=4, p<0.0001) (Fig. 2B). Notably, although RORC2-transduced T cells showed significantly increased IL-22 expression, the frequency of IL-22-producing cells was consistently lower than amongst cytokine-induced Th17 cells.
To more fully evaluate the effects of RORC2 on cytokine expression, supernatants from transduced T cells activated with α-CD3/α-CD28-coated beads were analyzed using a cytometric bead array (CBA). Over-expression of RORC2 resulted in a significant increase in TNF-α (6.8-fold higher, p=0.028, n=5), IL-6 (13.8-fold higher, p=0.01, n=5) and IL-2 (2.1-fold higher, p=0.02, n=5) production, with a parallel decrease in IFN-γ (3.2-fold decrease, p=0.028, n=5), IL-4 (10.7-fold decrease, p=0.04, n=5) and IL-10 (3.8-fold decrease, p=0.05, n=5) production compared with pCCL-transduced cells (Fig. 2C). In comparison to Th0 controls, cytokine-induced Th17 cells also expressed more IL-6 (18.4-fold increase, p=0.002, n=3) and markedly less IFN-γ (68.2-fold decrease, p<0.0001, n=3) and IL-4 (2.7-fold decrease, p=0.002, n=3) (Fig. 2D). Surprisingly, and in contrast to pCCL.RORC2-transduced T cells, cytokine-induced Th17 cells did not express TNF-α and showed increased expression of IL-10 (5.5-fold increase, p=0.01, n=3) compared with cells in Th0-polarizing conditions. Thus, over-expression of RORC2 is sufficient to induce a cytokine expression profile consistent with the expected phenotype of Th17 cells: IL-17+IL-22+TNF-α+IL-6+IFN-γlow.
Expression of RORC2 alters chemokine receptor, CD161 and granzyme expression
Since homing receptors are important for T-cell effector function, and recent evidence suggests that human Th17 cells have a unique profile of chemokine receptor expression 30–32, 34, 36, 47, we next examined whether RORC2 influenced expression of CCR2, CCR4, CCR6 and CXCR3. In comparison to pCCL-transduced cells, over-expression of RORC2 resulted in a significant increase in the proportion of CCR6+ T cells, with an average of 67.7±15.1% CCR6+ T cells in pCCL.RORC2-transduced populations compared with 13.6±3.1% CCR6+ T cells in pCCL-transduced populations (n=4, p=0.0035) (Fig. 3A). In addition, the proportion of CCR4+ T cells was also increased in pCCL.RORC2-transduced populations (mean % CCR4+ for pCCL.RORC2-transduced T cells was 64.5±10.9% versus 40.6±4.8% for pCCL-transduced T cells; n=4, p=0.0064). In parallel, over-expression of RORC2 strongly inhibited expression of CXCR3 (mean % CXCR3+ for pCCL.RORC2-transduced T cells 6.6±2.7% versus 49.6±9.7% for pCCL-transduced T cells; n=4, p=0.0017). No statistically significant differences in CCR2 expression were observed following pCCL.RORC2 transduction. These findings correlate with reports that Th17 cells express CCR6 30–32, 34, 36, 47, and that ex vivo IL-17+IFN-γ− cells are CXCR3−CCR4+CCR6+30.
We also analyzed expression of the same chemokine receptors in cytokine-induced Th17 cells. In contrast to pCCL.RORC2-transduced cells, Th17-polarized cells did not display statistically significant changes in the expression of CCR4 or CCR6 (data not shown) under our culture conditions. Similar to pCCL.RORC2-transduced T cells, however, CXCR3 expression was consistently inhibited in Th17-polarized cells (mean % CXCR3+ for Th17 cells was 41.5±18.8% versus 56.3±14.7% for Th1 cells; p=0.0179, n=3, data not shown). These data suggest that inhibition of the Th1-associated trafficking receptor CXCR3 is characteristic of IL-17-producing cells.
Th17 cells were recently reported to express CD161, a molecule typically associated with NK and NKT cells 38. To ask if RORC2 could induce the expression of CD161, we examined the expression of CD161 on transduced cells, and found that expression of RORC2 resulted in strong upregulation of CD161, but only in a small subset of cells (pCCL mean 0.4±0.4%, RORC2 mean 5.8±3.9%; n=4, p=0.03) (Fig. 3B). Thus, it is unlikely that expression of RORC2 alone is sufficient to drive CD161 expression.
Several groups have also reported that Th17 cells have altered expression of granzymes 16, 31, 48 and that this can lead to functional changes 48. We therefore asked whether RORC2 could directly influence expression of granzymes. pCCL.RORC2-transduced T cells showed decreased expression of granzyme A and B: there was a 2.6±0.9-fold decrease in the number of cells expressing granzyme A (p=0.02, n=4) and a 9.4±6.4-fold decrease in granzyme B expressing cells (p=0.04, n=4) (Fig. 3C). A similar trend was observed in cytokine-induced Th17 cells. Compared with Th0 cells, Th17 cells had a 1.6-fold decrease in granzyme A (n=4, p=0.003) and 3.3-fold decrease in granzyme B, (p=0.049, n=4; data not shown). These data suggest that a downstream consequence of RORC2 expression could be lower levels of cytotoxic activity.
Expression of RORC2 inhibits T-cell proliferation
We next examined whether over-expression of RORC2 resulted in functional consequences beyond alterations in cytokine production and chemokine receptor expression. We first assessed the proliferative capacity of pCCL.RORC2-transduced cells under different activation conditions and found that, regardless of whether co-stimulation was provided, pCCL.RORC2-transduced cells proliferated less than pCCL-transduced cells. When stimulated with immobilized α-CD3 and soluble α-CD28, pCCL.RORC2-transduced cells proliferated 44.7% less, on average, than pCCL-transduced cells (p=0.0004, n=5) (Fig. 4A). Similar results were obtained upon stimulation with APC and α-CD3 (data not shown and Fig. 5A). We then examined whether exogenous cytokines could reverse the hypo-responsiveness of RORC2-transduced cells and found that addition of IL-2 or IL-15 restored their proliferative capacity to a level equivalent to that of pCCL-transduced T cells (Fig. 4B). Since IL-23 is involved in the survival and expansion of mouse Th17 cells 16, 20, 49, we also examined whether IL-23 or other cytokines associated with Th17 cell differentiation could reverse the anergic state of RORC2-transduced cells. Although IL-23 did significantly increase proliferation of RORC2-transduced cells by 2.2±0.35-fold (n=3, p=0.01), in comparison to cultures with IL-2 or IL-15, the effects were minimal. TGF-β tended to inhibit proliferation of both control and RORC2-transduced cells, whereas neither IL-1β nor IL-6 had a significant effect.
In order to ask whether this hypo-responsiveness was characteristic of Th17 cells, we investigated whether cytokine-induced Th17 cells had a similar phenotype and found that they did not differ significantly from Th1 or Th0 cells in this respect (data not shown). We therefore established conditions for testing the proliferation of ex vivo Th17 cells. Since Th17 cells selectively express CCR4 and CCR6 30, populations of CD4+CCR4+CCR6+, CD4+CCR4−CCR6+, CD4+CCR4+CCR6-−, and CD4+CCR4−CCR6− T cells were sorted and tested following expansion for 2 wk. Intracellular staining confirmed enrichment for Th17 cells since CCR4+CCR6+ T cells contained a significant population of cells expressing IL-17 and IL-22, but low IFN-γ (Fig. 4C). CD4+CCR4+CCR6+ T cells expressed RORC2 mRNA and high expression of IL-17 was confirmed by ELISA (data not shown). In contrast, the CCR4−CCR6− T cells primarily expressed IFN-γ and IL-22, but little lL-17. CCR4+CCR6− T cells analyzed in parallel produced very low amounts of IL-17, IFN-γ and IL-22. CCR4−CCR6+ T cells expressed high levels of IL-17 and IL-22, but also produced high amounts of IFN-γ, thus displaying a Th1/Th17 intermediate phenotype. When we assessed the proliferative capacity of CCR4+CCR6+ T cells following stimulation with immobilized α-CD3 and soluble α-CD28, CCR4+CCR6+ T cells were hypo-responsive in both the absence and presence of co-stimulation (Fig. 4D and data not shown), a phenotype not displayed by other sorted subsets. These data confirm that the reduced proliferative capacity of RORC2-transduced cells is likely characteristic of in vivo differentiated Th17 cells.
Expression of RORC2 results in altered susceptibility to suppression by Treg cells
It has been reported that human Th17 cells are resistant to Treg-cell-mediated suppression of proliferation 31 and IL-17 production 40. In mice, in vivo susceptibility of Th17 cells to Treg-mediated suppression appears to depend on antigen specificity and stage of disease 50, 51. In order to determine whether the proliferation of RORC2-expressing T cells was subject to suppression, pCCL- and pCCL.RORC2-transduced T cells were stimulated with APC and α-CD3 in the presence or absence of Treg cells. Surprisingly, co-culture of RORC2-transduced cells with Treg cells resulted in a net increase in proliferation in these cultures (Fig. 5A). As expected, addition of Treg cells to cultures of pCCL-transduced cells resulted in suppression of proliferation (at a 1:1 ratio pCCL-transduced proliferation was suppressed on average by 53.9±7.3%, n=4).
To determine whether the increased thymidine incorporation resulted from enhanced proliferation of the RORC2-transduced cells, and/or reversal of Treg cell anergy, we repeated the experiments using CFSE-labeled transduced cells or Treg cells. Shown in Fig. 5B, Treg cells did not significantly alter the proliferation of RORC2-transduced cells: on average, 12.4±8.1% of pCCL.RORC2-transduced cells divided compared with 12.1±7.1% when co-cultured with Treg cells (n=4, p=0.37). In contrast, co-culture with T effector (Teff) cells stimulated a significant increase in proliferation of the RORC-transduced cells, likely due to IL-2 production from the Teff cells: 30.0±14.5% of pCCL.RORC2-transduced cells co-cultured with Teff cells divided (n=4, p=0.01), compared with pCCL RORC2 alone.
We next investigated whether the net increase in thymidine incorporation from the Treg/pCCL-RORC2-transduced cell co-cultures was due to reversal of Treg anergy. Indeed, as shown in Fig. 5C, when proliferation of CFSE-labeled Treg cells was tracked, we found that the RORC2-transduced cells caused a significant increase in Treg proliferation. On average, 7.2±1.5% of Treg cells divided compared with 14.9±4.5% when co-cultured with pCCL.RORC2-transduced cells (n=3, p<0.05).
We also examined the capacity of Treg cells to regulate cytokine production by pCCL.RORC2-transduced T cells. When added at a 1:1 ratio, Treg cells significantly suppressed the capacity of pCCL.RORC2-transduced T cells to produce TNF-α 48.4±16.9% suppression, p=0.01, n=3) IL-6 (61.4±14.3% suppression, p=0.009, n=3), and the low amounts of IFN-γ (30.3±6.4% suppression, p=0.009, n=3), but not IL-17 (Fig. 5D). Interestingly, Treg cells also failed to suppress the low amounts of IL-17 produced by pCCL-transduced cells, suggesting that the failure of Treg cells to suppress IL-17 production is not specific for RORC-expressing cells. As expected, Treg cells suppressed the production of IFN-γ (38.4±2.4 % suppression, p=0.0007, n=3), IL-6 (50.4±29.4% suppression, p=0.03, n=3), and TNF-α (36.0±25.5 % suppression, p=0.04, n=3) by pCCL-transduced cells.
Recent evidence has shown that mouse and human Treg cells have the capacity to convert into Th17 cells under inflammatory conditions 52–56. To exclude the possibility that the lack of IL-17 suppression was due to conversion of Treg cells into IL-17-producing cells, we examined suppression of IL-17 by intracellular staining in co-cultures of CFSE-labeled transduced T cells and Treg cells. As shown in Fig. 5E, Treg cells efficiently suppressed IFN-γ production by pCCL-transduced cells, but did not affect IL-17 production by RORC2-transduced cells. Moreover, there was no significant increase in IL-17 production from the Treg cells themselves, indicating that the lack of IL-17 suppression was not due to Treg-cell conversion.
Direct effects of IL-17 on ex vivo CD4+ T cells
Based on the finding that Treg cells did not suppress IL-17 production and had increased proliferation when cultured with pCCL.RORC2-transduced T cells, we investigated whether IL-17 directly affected cytokine production and/or proliferation of ex vivo T-cell subsets. CD4+CD25− Teff cells were activated with α-CD3/α-CD28-coated beads in the absence or presence of IL-2 or IL-17 for 3 days, and analyzed by intracellular cytokine staining for expression of IL-2 and IFN-γ. Addition of exogenous IL-17 resulted in a significant increase in the percentage of IL-2-producing Teff cells; 48.6±15.3% in the presence of IL-17 versus 35.3%±18.0% in its absence (n=3, p=0.04) (Fig. 6A). A corresponding IL-17-induced increase in secreted IL-2 was also observed in T-cell supernatants (data not shown). Exogenous IL-17 did not significantly alter expression of IFN-γ IL-4, IL-5, TNF-α or IL-10 by Treg or Teff cells (Fig. 6A and data not shown). Addition of IL-17 to Treg cells did not induce cytokine production by Treg cells (Fig. 6A and data not shown).
We next tested the effect of IL-17 on the proliferation of ex vivo Teff and Treg cells by activation in the absence or presence of increasing amounts of IL-17. Teff and Treg cells displayed a dose-dependent increase in proliferation in response to IL-17 (Fig. 6B). At 50 ng/mL of IL-17, Teff cells on average proliferated 3.5±0.5-fold more than cells in the absence of exogenous cytokine (n=4, p=0.002) and displayed an increase in proliferation equivalent to addition of IL-2. At 50 ng/mL of IL-17, Treg cells on average proliferated 3.8±2.0-fold more than cells cultured without exogenous cytokine (n=6, p=0.02) Similar results were also observed with cells stimulated with immobilized α-CD3 and soluble α-CD28 indicating the effect was APC-independent (data not shown).
We hypothesized that the increase in proliferation induced by exogenous IL-17 may be indirectly due enhanced production of autocrine and/or paracrine IL-2. To investigate this possibility, we examined proliferation of Teff and Treg cells stimulated in the presence of IL-17, in the absence or presence of neutralizing α-IL-2 mAb. As shown in Fig. 6C, neutralizing IL-2 in these cultures significantly, but not completely, reversed the IL-17-induced proliferation in both Teff and Treg cells. On average there was 47.3% decrease in proliferation of Teff cells (n=3, p=0.05) and a 32.9% decrease in proliferation of Treg cells (n=3, p<0.05) at 100 ng/mL of IL-17 when IL-2 was neutralized.
Since IL-17 can increase IL-2 production and stimulate proliferation of Teff cells, we hypothesized that it may also alter the capacity of Treg cells to suppress proliferation. To address this possibility, ex vivo Teff cells were activated with increasing numbers of Treg cells in the absence or presence of IL-2 or IL-17. As shown in Figs. 6D and E, exogenous IL-17 resulted in a significant reduction in the capacity of Treg cells to suppress Teff-cell proliferation. Together, these data suggest that the resistance of Th17 cells to Treg-mediated suppression may be related to the effects of IL-17 on IL-2 production and T-cell proliferation.
We investigated the role of RORC2 in the development of human Th17 cells and for the first time determined whether over-expression of RORC2 was sufficient to induce functional changes beyond alterations in cytokine production. Consistent with the expected phenotype of Th17 cells, RORC2-expressing T cells produced high levels of IL-17, IL-22, IL-6, TNF-α but not IFN-γ. In addition, over-expression of RORC2 resulted in a Th17-like pattern of chemokine receptor expression (CCR4+CCR6+CXCR3−), induction of CD161 expression on a subset of cells, reduced expression of granzyme A and B, decreased proliferative capacity and altered susceptibility to suppression by Treg cells. Taken together with previous reports 25–27, these findings provide compelling evidence that expression of RORC2 regulates the phenotype and function of Th17 cells in humans.
Human Th17 cells are currently defined on the basis of cytokine production. In order to evaluate the effects of RORC2, we directly compared the cytokine production profile of RORC2-transduced T cells with donor-matched cytokine-induced Th17 cells that were generated in vitro by activation in the presence of IL-1β, IL-6 and IL-23. TGF-β1 was not included since we found that under these conditions TGF-β1 inhibited IL-17 production, as previously reported 32, 39, 40, 57. Both cytokine-induced Th17 cells and RORC2-transduced T cells produced IL-17, IL-22 and IL-6 and downregulated expression of IFN-γ. The frequency of IL-22-expressing cells in RORC2-transduced cultures, however, was consistently lower than in cytokine-induced Th17 cells, suggesting that additional factors, such as ROR-αd 26, 27, could be required for the full induction of this cytokine. In contrast, Manel et al. reported that ectopic expression of RORC2 did not stimulate IL-22 production 26, a discrepancy likely related to differences in the source of cells (cord blood versus adult peripheral cells) and the timing of assays. Moreover, since both Th1 and pCCL-transduced control cells also expressed significant levels of IL-22, it remains unclear whether IL-22 should be considered as a Th17-defining cytokine.
RORC2-transduced T cells express high levels of TNF-α and low amounts of IL-10, consistent with recent reports in human 33, 34 and mouse Th17 cells 5, 8, whereas cytokine-induced Th17 cells did not produce TNF-α, and expressed elevated levels of IL-10. Since the expression of these two cytokines in human Th17 cells has not been consistently reported, it is currently not clear whether the production of these two cytokines may vary depending on the precise combination of polarizing cytokines and culture conditions, as reported for IL-10 in mouse Th17 cells 21. Evidence that there is no significant positive or negative correlation between production of IL-10 and IL-17 in Th17 cell clones 34 suggests that levels of IL-10 are not a key feature of human Th17 cells.
Human Th17 cells have also been characterized and isolated on the basis of chemokine and cytokine receptor expression 30, 31, 36, 37, 47. We found that over-expression of RORC2 resulted in potent induction of CCR4 and CCR6 expression and repression of CXCR3 but had no effect on CCR2. Currently, there are conflicting reports regarding the combination of CCR that specifically mark human Th17 cells. Ex vivo Th17 cells have been reported to be enriched in CCR6+ cells 47, CCR4+CCR6+ T cells 30, CCR5+CCR6+34 and CCR2+CCR5− T cells 37. By contrast, Lim et al. did not find a unique Th17-specific pattern of chemokine receptor expression 36. Considering the effects of RORC2 expression, and the finding that CCR4+CCR6+ T cells, but not CCR4−CCR6+ T cells are significantly enriched in IL-17+IFN-γ− cells 30, co-expression of CCR6 and CCR4 in the absence of CXCR3 seems to be most characteristic of human Th17 cells. In terms of cytokine receptors, specific expression of IL-23R has been associated with Th17 cells 32, but we could not detect a significant change in this protein upon RORC2 expression (data not shown). This lack of effect of RORC2 is in accordance with the finding that knockdown of RORC2 in polarized Th17 cells did not alter expression of IL-23R 33.
Over-expression of RORC2 also resulted in significant upregulation of CD161, but only in a small (∼6%) subset of cells. It remains unclear if CD161 specifically defines Th17 cells, as T cells isolated on the basis of this marker also had a high percent of cells that co-expressed both IL-17 and IFN-γ 38. Our data suggest that factors in addition to RORC2 are required for expression of CD161.
The reduced granzyme expression observed with over-expression of RORC2 is consistent with those of Annunziato et al. who reported reduced expression of granzyme A in Th17 cell clones compared with Th1 and Th2 cell clones 31. In contrast, Kebir et al. reported that IL-17-producing cells generated in vitro by polarization with IL-23 expressed high levels of granzyme B 48. These authors did not, however, directly compare granzyme B expression in polarized Th17 cells with that in Th1 or unpolarized cells, and since IL-23 can induce IFN-γ 58 it is possible that these cells possessed an intermediate Th1/Th17 phenotype. Further studies will be required to determine whether the RORC2-induced downregulation of granzymes A and B is consistent with the phenotype of ex vivo human Th17 cells and if so, to define the functional relevance of their reduced expression.
Since Th17 cells produce many pro-inflammatory cytokines, we initially hypothesized that they would be highly proliferative. Surprisingly, RORC2-transduced cells exhibited a greatly reduced proliferative capacity, which was restored by addition of IL-2 or IL-15. Evidence that this hypo-responsiveness of RORC2-transduced cells was also characteristic of in vivo differentiated Th17 cells came from the finding that ex vivo CD4+CCR4+CCR6+ T cells, which are significantly enriched in Th17 cells, displayed a similar reduction in proliferation compared with CD4+CCR4−CCR6−, CD4+CCR4+CCR6− or CD4+CCR4−CCR6+ T cells. That the proliferative defect of RORC2-transduced cells can be overcome by IL-2 or IL-15, is consistent with reports that ex vivo Th17 cells are effectively expanded by these cytokines 31, 59, and contrasts with data in mice showing that IL-2 inhibits IL-17 expression 60.
In view of the role of Th17 cells in autoimmunity, we investigated whether co-culture with Treg cells altered the function of RORC2-expressing T cells. While proliferation in co-cultures of Treg and RORC2-transduced T cells was difficult to assess as both cell types are hypo-responsive, the minimal proliferation of the RORC2-transduced T cells was not suppressed by Treg cells. By contrast, CFSE-labeling experiments revealed that co-culture with RORC2-transduced cells increased Treg cell proliferation. As the anergic state of Treg cells can be reversed by IL-2 61 and IL-6 62 (both of which are expressed by RORC2-transduced cells), this finding is not entirely surprising.
Like ex vivo and cytokine-induced Th17 cells 31, 40, 63, the capacity of pCCL.RORC2-transduced T cells to produce IL-17 was not suppressed by Treg cells. Treg cells could, however, effectively suppress production of TNF-α and IL-6 by RORC2-transduced cells. Intracellular staining experiments revealed that the lack of IL-17 suppression was not due to parallel conversion of Treg cells into IL-17-producing cells. Further investigation will be required to define the molecular basis for the resistance of IL-17 to suppression by Treg cells.
To define whether the effects of RORC2-transduced cells on other cells could be attributed to IL-17, we examined the effects of IL-17 on Teff and Treg cells. Although human T cells express IL-17R 64, previously there was limited evidence that IL-17 had biological effects on these cells 9. We found that IL-17 stimulated IL-2 production and proliferation by Teff cells, and also inhibited suppression of Teff-cell proliferation by Treg cells. Evidence that neutralizing anti-IL-2 mAb partially reversed the pro-proliferative effects of IL-17 on both Teff and Treg cells indicates that part of this effect is due to induction of autocrine and/or paracrine IL-2. These data also suggest that autocrine and/or paracrine IL-17 may be at least partly responsible for the resistance of Th17 cells to suppression by Treg cells. Further research will be required to define the biological significance and the mechanistic basis for the specific inability of Treg cells to suppress the production of IL-17 by Th17 cells, and to define ways to effectively inhibit Th17-cell dysregulation in immune disorders.
In conclusion, we have shown that over-expression of RORC2 is sufficient not only for appropriate cytokine production but also for phenocopying in T cells, many of the known aspects of the Th17 lineage, including chemokine receptor expression, suppression of granzymes, hypo-responsiveness and susceptibility to suppression by Treg cells. In mice it has recently been reported that ectopic expression of ROR-αd can also induce a Th17-like profile of cytokine production 26, 27, and that Runx1 can both induce and work in synergy with ROR-γt to induce Th17 cells 28. Further investigation will be required to determine the relative role of these transcription factors in the normal development and function of human Th17 cells.
Materials and methods
Peripheral blood was obtained from healthy volunteers following approval by the University of British Columbia Clinical Research Ethics Board and after obtaining written informed consent. PBMC were isolated by Ficoll separation, and CD4+ T cells were purified by negative selection (StemCell Technologies, Vancouver, Canada). Naive CD4+CD25−CD45RO− T cells were isolated by incubation with CD25 and CD45RO beads (Miltenyi Biotec, Auburn, CA)and passed over an LD depletion column to achieve purities ≥95%. CD4+CD25+ Treg cells were purified from CD4+ T cells by positive selection for CD25 (Miltenyi Biotec) over two columns to ensure 81–90% purity based on expression of CD25 (BD Pharmingen); >95% pure CD4+CD25− Teff cells were obtained by passing the CD25− fraction over an LD depletion column (Miltenyi Biotec).
Lentivirus production and transduction
The hRORC2 cDNA was cloned into the third generation pCCL bidirectional lentiviral vector 45 under control of the EF1-α promoter. Co-expression of ΔNGFR driven by the mCMV promoter served as a transduction marker. Lentivirus was produced by transient 4-plasmid overnight transfection of HEK 293T cells and the titers of concentrated virus were determined by limiting dilution on 293T cells 45. Naive T cells were activated with α-CD3 (1 μg/mL OKT3, Orthoclone) and autologous irradiated APC in complete medium (X-VIVO 15 (Cambrex) with 5% pooled AB human serum (Cambrex), and penicillin/streptomycin (Invitrogen)), containing rhIL-2 (100 U/mL, Chiron). The amount of TGF-β in this medium was measured to be 187±49 pg/mL (n=3 independent determinations). After 16 h, lentivirus was added at a multiplicity of infection of 5. Transduced T cells were purified and expanded as previously described 45, with purities greater than 98% consistently achieved. Prior to testing in functional and phenotypic assays, T cells in the resting phase (12–14 days following activation) were washed and rested in IL-2-free medium overnight. Untransduced cells tested in parallel were equivalent to control-transduced cells in all aspects of phenotype and function (data not shown).
Cytokine polarization of Teff cells
CD4+CD25−CD45RO− T cells were activated with soluble α-CD3 mAb (1 μg/mL OKT3) and autologous irradiated APC at a 1:5 ratio of T cells to APC in complete medium. Th1 polarized cell lines were generated by addition of rhIL-12 (5 ng/mL, BD Biosciences) and IL-2 (100 U/mL, Chiron). To generate Th17 polarized cell lines, rhIL-1βC10 (ng/mL, Sigma-Aldrich), rhIL-6 (10 ng/mL, R&D Systems), rhIL-23 (10 ng/mL, R&D Systems), and neutralizing α-IFN-γ and α-IL-4 mAb (10 μg/mL each, eBioscience) were added. Polarizing cytokines were replenished on days 2 and 4, then on day 5 IL-2 (100 U/mL) was added and cells were expanded for an additional 9 days.
Isolation of CD4+CCR4+CCR6+ T cells by flow cytometry
To isolate CD4+CCR4+CCR6+ T cells, CD4+ T cells were labeled with antibodies to CD4 (eBiosciences), CCR4 and CCR6 (both BD Pharmingen) and sorted using a FACS Aria (BD Biosciences) to a purity >96%. Sorted cells were expanded with α-CD3 (1 μg/mL OKT3, Orthoclone) and autologous irradiated (50 Gy) APC in complete medium.
Total RNA was prepared from T cells with TriZol (Invitrogen). cDNA was synthesized with Superscript reverse transcriptase II (Invitrogen) and oligo(dT) primers (Invitrogen) in the presence of DNAse1 (Promega). Gene expression was examined with a Bio-Rad iCycler Optical System with iQ SYBR green real-time PCR kit (Bio-Rad Laboratories); RORC2 forward primers: 5′-TGGAAGTGGTGCTGGTTAGGA-3′, and reverse primer: 5′-AAGGCTCGGAACAGCTCCAT-3′. All samples were run in triplicate, and relative expression of RORC2 was determined by normalizing to GAPDH to calculate a fold change in value.
Flow cytometric analyses
Staining for cell-surface markers CD4 (eBiosciences), ΔNGFR (BD Pharmingen), CCR4 (BD Pharmingen), CCR6 (BD Pharmingen), CXCR3 (R&D Systems), CCR2 (R&D Systems), CD161 (BD Pharmingen) and IL-23R (R&D Systems) was carried out prior to intracellular staining for granzymes A and B (both BD Pharmingen). For analysis of intracellular cytokine production, T cells were activated with 10 ng/mL PMA and 500 ng/mL Ca2+ ionophore (both Sigma-Aldrich) or with a 1:1 ratio of α-CD3/α-CD28-coated beads (Invitrogen) for 6 h, and brefeldin A (10 μg/mL, Sigma-Aldrich) was added half-way through activation. Following surface staining, cells were fixed in 2% formaldehyde and permeabilized with 0.5% saponin. Intracellular cytokine staining was performed with antibodies against IL-17 (eBiosciences or R&D Systems), IL-4 (BD Pharmingen), IL-22 (R&D Systems), IL-2 (BD Pharmingen) and/or IFN-γ (BD Pharmingen). Samples were acquired on a BD FACSCanto and analyzed with FCS Express Pro Software Version 3 (De Novo Software, Thornhill, Canada).
Determination of cytokine concentration
To quantify amounts of TNF-α, IL-6, IL-2, IFN-γ, IL-4 and IL-10, Th1/Th2 II cytometric bead arrays (BD Biosciences) were performed on supernatants after activation with immobilized α-CD3 (10 μg/mL) and soluble α-CD28 (1 μg/mL) (50 000 cell/well), α-CD3/α-CD28 T cells expander beads (100 000 cells/well) or α-CD3 (10 μg/mL) plus allogenic APC (50 000 cells/well) for 24 h (for IL-2) or 48 h (for IFN-γ, IL-4, IL-6, TNF-α and IL-10). ELISA was used to detect IL-17 (eBiosciences).
Proliferation and suppression assays
To test proliferative capacity, 50 000 T cells/well were stimulated with immobilized α-CD3 (1 μg/mL OKT3, Ortho Biotech) in the presence or absence of soluble α-CD28 (1 μg/mL; BD Biosciences), α-CD3/α-CD28 expander beads (Invitrogen) or α-CD3 (1 μg/mL OKT3, Ortho Biotech) plus irradiated APC. Proliferation was assessed after 72 h for α-CD3/α-CD28 stimulations, and at 96 h after α-CD3/APC stimulation, by [3H]thymidine incorporation (1 μCi/well, Amersham Biosciences), added for the final 16 h of culture. Where indicated, rhIL-2 (100 U/mL, Chiron) or rhIL-15 (20 ng/mL, StemCell Technologies), rhIL-1β10 ng/mL, Sigma-Aldrich), rhIL-6 (10 ng/mL, R&D Systems), rhIL-23 (10 ng/mL, R&D Systems), TGF-β1 (1 ng/mL, R&D Systems), rhIL-17 (R&D Systems), or α-IL-2 (Clone 5334, 5 μg/mL, R&D Systems) were included in proliferation assays.
To test for suppressive capacity, T cells were stimulated at 50 000 cells/well with α-CD3 (1 μg/mL OKT3) and 50 000 irradiated APC in the presence or absence of various amounts of CD4+CD25+ Treg cells corresponding to the indicated ratios. Suppression was assessed by measuring the amount of [3H]thymidine incorporation in the final 16 h of a 96 h culture period. Cytokine suppression was determined by examination of culture supernatants after 48 h using a Th1/Th2 II cytometric bead array or ELISA. In some experiments, pCCL-, pCCL-RORC2-transduced or Treg cells were labeled with 2.5 mM 5- (and 6-) CFSE (Molecular Probes) and activated after 96 h of stimulation with α-CD3 and APC. CFSE dilution was assessed by flow cytometry with parallel staining for intracellular cytokine production as detailed in the Flow Cytometric Analyses section.
All analyses for statistically significant differences were performed with 1-tailed paired Student's t test. p-Values of less than 0.05 were considered significant. All error bars represent standard deviations.
This work was supported by grants from the Canadian Institutes for Health Research (MOP 191221) and StemCell Technologies Inc. Core support for lentivirus production by Rupi Dhesi and flow cytometry by Lixin Xu was funded by the Immunity and Infection Research Centre MSFHR Research Unit. M.K.L. holds a Canada Research Chair in Transplantation and is an MSFHR scholar. S.Q.C. holds an MSFHR Junior Graduate Studentship award, a CIHR/MSFHR Transplantation Training Program award and a CIHR/UBC Skin Research Trainee award. A.Y.W. holds a CIHR/MSFHR Transplantation Training Program award. C.Y.K. holds a CIHR/UBC Translation Research in Infectious Disease award. We thank Sarah Allan and Paul Orban for critical reading of the manuscript.
Conflict of interest: The authors declare no financial or commercial conflict of interest.