Interleukin-23 Promotes a Distinct CD4 T Cell Activation State Characterized by the Production of Interleukin-17*

Interleukin (IL)-17 is a pro-inflammatory cytokine that is produced by activated T cells. Despite increasing evidence that high levels of IL-17 are associated with several chronic inflammatory diseases including rheumatoid arthritis, psoriasis, and multiple sclerosis, the regulation of its expression is not well characterized. We observe that IL-17 production is increased in response to the recently described cytokine IL-23. We present evidence that murine IL-23, which is produced by activated dendritic cells, acts on memory T cells, resulting in elevated IL-17 secretion. IL-23 also induced expression of the related cytokine IL-17F. IL-23 is a heterodimeric cytokine and shares a subunit, p40, with IL-12. In contrast to IL-23, IL-12 had only marginal effects on IL-17 production. These data suggest that during a secondary immune response, IL-23 can promote an activation state with features distinct from the well characterized Th1 and Th2 profiles.

We have examined the possibility that IL-17 is expressed in response to signals distinct from those associated with the Th1 or Th2 response. We observe a previously unrecognized activity of the recently identified cytokine IL-23 (20). IL-23 is a heterodimeric cytokine that shares one subunit, p40, with IL-12. The initial characterization of this cytokine has suggested it can promote proliferation within the memory T cell population. Subsequent work demonstrated that transgenic over-expression of the second component of IL-23, p19, was sufficient to induce systemic inflammation and premature death (21). In addition, the mice had markedly elevated levels of circulating neutrophils. Interestingly they did not exhibit consistent elevation of IFN-␥, a hallmark effect of IL-12. These data suggest that IL-23 may have a biological role substantially distinct from that of IL-12. In this report we present evidence that IL-23 acts to induce a distinct T cell activation state that produces IL-17 as a principle effector cytokine.

EXPERIMENTAL PROCEDURES
Cell Culture-Single cell suspensions of spleen were prepared from C57/BL-6 mice, and mononuclear cells were isolated from suspended splenocytes by density gradient centrifugation. 2 ϫ 10 6 cells/ml were cultured with IL-2 (100 units/ml) in the presence or absence of various stimuli (for times indicated in figure legends), following which the cells were collected and analyzed for IL-17 using ELISA (R&D Systems, Minneapolis, MN). Dendritic cells were derived from macrophages (obtained as adherent population from splenocyte suspension) by treating macrophages with rGM-CSF (2 ng/ml) and rIL-4 (1000 units/ml) for 4 days, washing and re-activating using LPS (0.5 g/ml). Memory and naïve T cells were isolated by staining mononuclear cells isolated from single cell suspension of murine splenocytes with CyC-CD4 ϩ PE-CD44 or CyC-CD4 ϩ PE-CD62L and sorting for CD4 ϩ cells that were either CD44 high /CD62L low for memory phenotype or CD44 low /CD62 high for naïve phenotype.
Purification of IL-23-Murine IL-23 component was produced by co-expression of carboxyl-terminal His-tagged p19 and FLAG-tagged p40 in human embryonic kidney cells (293 cells), and secreted protein was purified by nickel affinity resin. Endotoxin levels were undetectable at less than 0.2 endotoxin units per g.

RESULTS
We first examined the ability of various microbial products to stimulate the production of IL-17. Increased IL-17 has recently been observed by Infante-Duarte et al. (22) in response to microbial lipopeptides from a Lyme disease causing spirochete, * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Borrelia burgdorferi. We observed that spleen cell cultures in the presence of various microbial products including LPS (Gram-negative bacteria), lipoteichoic acid (Gram-positive bac-teria) or lipopeptide (bacterial lipopeptide) resulted in the production of IL-17 ( Fig. 1). Neither purified T cells alone nor purified macrophages themselves produced IL-17. Purified T FIG. 1. IL-17 production in different cell types. A, mononuclear splenocytes were cultured in the presence or absence of microbial lipopeptide (100 ng/ml), LPS (100 ng/ml), or lipoteichoic acid (100 ng/ ml) for 3 days, following which the cells were collected and analyzed for IL-17 using ELISA. B, purified T cells were obtained from murine splenocytes following positive selection of fluorescence-activated cell-sorted, CD4-labeled cells. These cells were cultured (1 ϫ 10 6 cells/ ml) in presence or absence of plate-bound anti-CD3 (5 g/ml) or supernatant from activated dendritic cells (LPS-treated) for 3 days and culture supernatants collected and analyzed for IL-17 levels using an ELISA kit. Representative results from three independent experiments are shown. cells, upon receptor cross-linking using plate-bound anti-CD3 and treatment with supernatants from activated macrophages/ dendritic cells, produced increased IL-17, indicating the pres-ence of an unidentified factor(s) released by these cells that acts on T cells to promote IL-17 production.
In profiling the expression of candidate molecules that might be responsible for this IL-17 promoting activity, we observed 100 -1000-fold increased mRNA expression of the IL-23 (20) components, p19 and p40, in activated dendritic cells using real-time RT-PCR (not shown), hence, the effect of IL-23 was examined. Murine spleen cell cultures, in the presence of IL-23, resulted in high levels of IL-17 production in a dose-dependent manner ( Fig. 2A). However, when these cells were cultured under IL-12-stimulated Th1-inducing conditions, they resulted in marginal IL-17 production, whereas under Th2-inducing conditions there was no increased production of IL-17 over controls (Table I). IL-23 also resulted in higher levels of GM-CSF than observed under Th1-inducing conditions. In contrast, IFN-␥ levels were significantly lower than those obtained under Th1-inducing conditions. Tumor necrosis factor-␣ levels were similar to Th1 conditions. IL-12p40 alone did not result in any IL-17 production (data not shown). IL-23 promoted ele-

FIG. 3. IL-23 acts on memory T cells to induce IL-17 production.
CD4-positive memory (CD44 high /CD62L low ) or naïve (CD44 low /CD62L high ) T cells were cultured with IL-2 in the presence or absence of IL-23 (or its boiled prep as a control) and plate-bound anti-CD3 (5 g/ ml) and anti-CD28 (1 g/ml) for 3 days (6 days for naïve cell population), washed, and re-stimulated with anti-CD3 antibody for another 24 h. Supernatants were collected, and IL-17 levels were measured using ELISA.

TABLE I Cytokine levels in Th cells in response to IL-23
Purified CD4ϩ T cells were incubated in presence of IL-2 (100 units/ ml) and activated as described under "Experimental Procedures" under Th1-inducing conditions (IL-12ϩ anti-IL-4), Th2-inducing conditions (IL-4ϩ anti-IFN-␥), or purified IL-23 (10 nM) for 3 days, following which the cultures were washed and re-stimulated with anti-CD3 for another 24 hours. Levels of various cytokines were measured using ELISA. The levels less than the lowest dilution of the standard curve range of ELISA kit were recorded as not detectable (N.D.) vated levels of IL-17 mRNA (Fig. 2B). IL-17 mRNA levels were increased several hundred-fold within 6 h of IL-23 exposure and remained elevated in the continued presence of IL-23. This effect was not inhibited by the presence of an antibody against IL-17, suggesting that the IL-17 itself was not contributing to this process (not shown). In addition, mRNA for IL-17F, a recently identified IL-17 family member (23,24), was also found to be up-regulated in response to IL-23 (Fig. 2C). IL-23 has been reported to promote the proliferation of memory but not naïve T cells (20,25). We therefore examined the effect of IL-23 on IL-17 production from naïve versus memory T cell populations. Purified CD4 ϩ T cells were isolated from splenocytes by fluorescence-activated cell sorting. The memory cell population was selected as CD4 ϩ CD44 high (26) or CD4 ϩ CD62L low (27), and naïve cell population was selected as CD4 ϩ CD44 low or CD4 ϩ CD62L high . As seen in (Fig. 3), IL-23 stimulated robust IL-17 production within the memory cell population (CD44 high and CD62L low ) and little or no IL-17 production within the naïve (CD44 low or CD62L high ) cell population.
IL-23-mediated IL-17 production was completely blocked in the presence of a neutralizing IL-12 antibody (R&D Systems) that interacts with the p40 subunit shared with IL-23 (used due to unavailability of neutralizing antibodies specific to IL-23p19 or IL-23) (Fig. 4A and 4B, left panel). This effect was not due to ligation of Fc receptors on antigen-presenting cells as there was no change in IL-17 production in the presence of unrelated antibody. This antibody also inhibited Ͼ50% the induction of IL-17 production observed in response to conditioned media from LPS-stimulated dendritic cells (Fig. 4B,  right panel). A marked reduction, but not abrogation, of IL-17 production was seen in response to ConA stimulation from spleen cell cultures of mice lacking IL-12p40 component (strain B6.129S1-IL12b tm1Jm ) as compared with wild type mice or mice lacking IL-12p35 component (strain B6.129S1-IL12a tm1Jm ) (Fig. 4B). Thus, IL-23 plays a substantial role in promoting IL-17 production, although it is clearly not absolutely required.
Finally, to examine the role of IL-12 in IL-17 production, we added increasing amounts (0.001-1 nM) of murine IL-12 to IL-23 (1 nM) containing cultures. As seen in (Fig. 5A), IL-12 decreased IL-17 levels in a dose-dependent manner. Additionally, we treated splenocytes from mice lacking IL-12 receptor beta chain 2 (IL-12R␤2) (28), the specific receptor component of  with purified IL-23. Splenocytes from IL-12R␤2 Ϫ/Ϫ mice responded to IL-23 stimulus by increasing IL-17 production over the un-stimulated control (Fig. 5B) without affecting IFN-␥ levels. Surprisingly, the background levels of IL-17 in these mice were more than 10-fold elevated as compared with wild type mice, suggesting a possible negative regulation by IL-12 of IL-23-induced IL-17 production. However, in contrast to IL-12R␤2 knockout mice, we did not observe increased IL-17 in spleen cultures from IL-12p35 knockout mice. The reasons for this difference are not known, but could relate to alteration in IL-12p40 function in the absence of p35, or differences in genetic background or pathogen exposure.

DISCUSSION
Taken together these data suggest a role for IL-23 in the promotion of a distinct T cell activation state that expresses IL-17 as an effector cytokine. The Th1 and Th2 paradigms have been described as promoting cell-mediated versus humoral immune responses. These responses provide important defense for intracellular and extracellular pathogens, respectively, and defects in either of these responses are associated with increased susceptibility to specific pathogens. In contrast, IL-23 may serve to promote an adaptive immune response to pathogens that is characterized by a heavy reliance on cells thought to function primarily as mediators of the innate immune response. IL-17, as a principle effector cytokine of this response, is able to promote the more rapid recruitment of monocytes and neutrophils through induced chemokine production. In addition, the GM-CSF production observed in response to IL-23 supports the production of additional myeloid cells. This is further augmented by high level G-CSF production from local IL-17-stimulated stromal cells. The character of this adaptive response is, however, not an exclusive reliance on phagocytic cells of the myeloid lineage as IL-17 is known to promote the induction of ICAM thereby providing important co-stimulation of further T cell responses. The actions of IL-23 appear to be restricted to memory T cells. However, it remains to be determined whether there exist co-stimulatory signals that enable action of IL-23 on naïve T cell populations. Further analysis with TCR transgenic animals will help to clarify these issues.
Recently several studies (30 -33) have pointed out significant differences between mice deficient in p35 and mice deficient in p40. These studies share the observation that loss of p40 is generally more deleterious than loss of p35 in the immunemediated clearance of a variety of model organisms. For example, Elkins et al. (32) found that although p35 KO mice were readily able to clear the intracellular bacterium Francisella tularensis LVS, p40 KO mice were unable to completely clear the bacteria and exhibited chronic infection. In light of the results presented here, these observations may reflect the capacity of IL-12 and IL-23 to promote unique and complementary immune responses that facilitate the complete clearance of bacteria and perhaps other pathogens. Interestingly, recently reported p19 transgenic mice display a profound systemic inflammation and neutrophilia (21). This phenotype is consistent with the results reported here and further indicates the potential of this signal to mediate profound changes in the balance of immune function. IL-17 production is not completely abrogated in p40 KO mice, indicating that there exist other IL-23-independent pathways to its expression. It will be very important to investigate whether IL-23 has similar functions in human and mice and to establish its relevant role in human immune function. A great deal of attention has been given to the Th1 paradigm and its relationship to major unmet human diseases. Given the clear association of IL-17 expression with most of these same inflammatory diseases it may be that the relationship between IL-12 and these diseases should be re-evaluated and additional work undertaken to understand the role of this new cytokine axis in human disease.