IL-17 drives psoriatic inflammation via distinct, target cell-specific mechanisms

Significance

Psoriasis is an inflammatory disease affecting the skin, a barrier site. The disease is characterized by abnormal growth of keratinocytes and infiltration of inflammatory cells. Clinical trials targeting the IL-17 cytokine have shown remarkable efficacy, and IL-17 also has been strongly implicated in the imiquimod-induced mouse model of psoriasis. However why IL-17 cytokines should be so central is not known, because target cells and their functions have not been clearly delineated. Here we demonstrate that IL-17 signaling into nonkeratinocytes, specifically dermal fibroblasts, induces mediators that further increase IL-17 production by innate γδT cells and promote cellular infiltration, whereas IL-17 signaling into keratinocytes aids proliferation and blocks their differentiation. These findings reveal the circuitry underpinning critical disease-driving effects of IL-17.

Abstract

Psoriasis is a chronic inflammatory skin disease characterized by abnormal keratinocyte proliferation and differentiation and by an influx of inflammatory cells. The mechanisms underlying psoriasis in humans and in mouse models are poorly understood, although evidence strongly points to crucial contributions of IL-17 cytokines, which signal via the obligatory adaptor CIKS/Act1. Here we identify critical roles of CIKS/Act1-mediated signaling in imiquimod-induced psoriatic inflammation, a mouse model that shares features with the human disease. We found that IL-17 cytokines/CIKS-mediated signaling into keratinocytes is essential for neutrophilic microabscess formation and contributes to hyperproliferation and markedly attenuated differentiation of keratinocytes, at least in part via direct effects. In contrast, IL-17 cytokines/CIKS-mediated signaling into nonkeratinocytes, particularly into dermal fibroblasts, promotes cellular infiltration and, importantly, leads to enhanced the accumulation of IL-17–producing γδT cells in skin, comprising a positive feed-forward mechanism. Thus, CIKS-mediated signaling is central in the development of both dermal and epidermal hallmarks of psoriasis, inducing distinct pathologies via target cell-specific effects. CIKS-mediated signaling represents a potential therapeutic target in psoriasis.

Psoriasis is a common chronic inflammatory skin disease affecting up to 2–3% of the population worldwide. This disease has an incompletely defined etiology and is characterized by dysregulated proliferation and differentiation of keratinocytes, dermal angiogenesis, and immune cell infiltration. Currently there is no cure, but efficacious treatments to control symptoms exist. The development of psoriasis involves an intricate interplay of genetic and environmental factors. Multiple susceptibility gene loci have been identified, including genes encoding proteins involved in keratinocyte differentiation and function as well as proteins involved in inflammatory responses. Psoriasis is associated with serious comorbidities, such as metabolic syndrome and cardiovascular disease, likely reflecting a systemic inflammatory component of the disease (comprehensively reviewed in ref. 1).

The choreography underpinning the development of pathology in psoriasis is incompletely understood and may differ among individuals and among the various mouse models of this disease. Nevertheless, a number of inflammatory cytokines, including IFN-γ, IL-1 (α,β), IL-36 (α,β,γ), IL-22, and, especially, TNF-α and the IL-23/IL-17 axis, have been implicated in psoriatic inflammation in mouse models and/or humans (2, 3). IL-17–producing cells are found in lesions from psoriatic patients and may be key factors driving pathology (4–6). Importantly, recent clinical trials have shown therapies targeting IL-12/IL-23 (p40), IL-23 (p19), and IL-17 or its receptor (IL-17RA) to be highly efficacious (7–9). Furthermore, comprehensive genomewide association studies have identified Il12b and Il23a, which encode the IL-23 subunits, and Traf3ip2, which encodes CIKS [connection to IκB kinase and stress-activated protein kinase, also known as “Act1” (NF-κB activator 1)], the obligate adaptor for IL-17 receptor signaling, as strong susceptibility loci for psoriasis (1, 10, 11).

IL-17 (IL-17A) is the signature cytokine of Th17 cells and usually is coproduced together with the closely related IL-17F, with which it can form heterodimers. Increased expression of these cytokines has been linked not only to psoriasis but also to other inflammatory diseases (reviewed in ref. 12). In addition to Th17 cells and CD8⁺ cells secreting IL-17 (Tc17 cells), innate lymphoid γδT cells, invariant natural killer T (iNKT) cells, and innate lymphoid cells-3 (ILC3) also produce IL-17 cytokines (13). Indeed dermal γδT cells appear to be the main producers of IL-17 in several mouse models of psoriasis (5). IL-17–producing dermal γδT cells also are present in humans, and their numbers are increased in psoriatic lesions (5, 14).

IL-17C is an additional member of the IL-17 cytokine family that has been implicated in psoriasis (15–17). IL-17C seems to overlap functionally with IL-17A but is primarily produced by epithelial cells, including keratinocytes (18). Among the three IL-17 cytokines associated with psoriasis—A, C, and F—IL-17A appears to be the most potent. All three are members of an extended family (IL-17A–F) that signals via cognate heteromeric receptors composed of members of the IL-17 receptor family (RA–RE), with the IL-17RA chain likely common to all receptors (19). The IL-17 receptor family members encode a SEFIR-like (similar expression to fibroblast growth factor genes and IL-17 receptors) domain in their cytoplasmic tails, which also is present within the adaptor protein CIKS (20, 21). Upon ligand engagement, CIKS is recruited to IL-17 cytokine receptors via heterotypic SEFIR domain association (22, 23). Consistent with a pathogenic role for IL-17 cytokines in various diseases, CIKS is critical for the development of collagen-induced arthritis, lupus, and asthma in mouse models (24–26).

The importance of IL-17 cytokines in psoriatic inflammation has been addressed in several mouse models, although questions remain. Psoriasis-like inflammation induced upon intradermal injections of IL-23 in mice was reported to be largely dependent on the presence of IL-17RA; however other studies have highlighted the critical contributions of IL-22 (27–29). Topical application of the TLR7/8 ligand imiquimod (IMQ) induces skin inflammation in mice that mimics various aspects of human psoriasis (5, 30). This potent immune activator also is used for the treatment of warts and superficial basal cell carcinomas but can induce psoriasis-like skin flares as a side effect in susceptible patients (31). IMQ-induced psoriatic inflammation was reported to be critically dependent on IL-23 as well as on the presence of IL-17RA (30), although whether IL-17RA signaling is absolutely required in this model has been questioned recently (32).

Here we describe previously undescribed mechanisms by which CIKS-mediated signaling contributes to specific psoriatic pathologies in the IMQ model. We demonstrate that CIKS-mediated signaling downstream of IL-17 cytokines critically disrupts the regulation of keratinocyte proliferation and differentiation and is essential for the formation of neutrophilic microabscesses. Notably, these effects are completely dependent on CIKS signaling within keratinocytes. We also demonstrate that CIKS contributes to the recruitment of inflammatory cells, especially to the IMQ-induced increase in IL-17–expressing cells, but, in contrast to the epidermal pathologies, these latter effects depend on CIKS signaling in nonkeratinocytes, specifically in dermal fibroblasts. These findings reveal cellular targets and mechanisms by which CIKS-mediated signaling effects specific psoriatic pathologies, and they identify CIKS as a potential therapeutic target in psoriasis.

Results

Absence of CIKS Protects Mice from IMQ-Induced Psoriasis.

To determine roles of CIKS-mediated signaling in IMQ-induced psoriasis, we applied IMQ-containing cream daily for 5 d to the shaved backs and ears of 8- to 10-wk-old WT and CIKS-deficient (Traf3ip2^−/−; hereafter referred to as “CIKS-KO”) mice. Ear thickness was measured every other day, and tissue was obtained for analysis on day 6. The psoriasis-like skin pathology induced by IMQ in WT mice was greatly attenuated, but not abolished, in CIKS-KO mice. Erythema, easily discernable on the backs of treated WT mice, was reduced in treated CIKS-KO mice (Fig. 1A). Similarly, treatment-induced thickening of the epidermis (with acanthosis and hyperkeratosis) and cellular infiltration were reduced in CIKS-KO mice relative to WT mice, and epidermal neutrophilic abscesses no longer were present in mutant mice (Fig. 1 B and C). The reduction in thickening of the dorsal epidermis in IMQ-treated CIKS-KO mice was significant compared with treated WT mice, as determined by quantitation of the epidermal areas in H&E-stained slices from multiple mice (Fig. 1D). Treated CIKS-KO mice also exhibited significantly reduced thickening of ears (Fig. 1E). However, treated CIKS-KO mice still developed some epidermal thickening compared with untreated controls.

Fig. 1.

CIKS-dependent IMQ-induced psoriatic phenotypes. (A) Dorsal areas of WT and CIKS-KO (Traf3ip2^−/−) mice after 5-d IMQ treatment or no treatment (Control) of the shaved dorsal skin. (B) H&E-stained dorsal sections showing more pronounced IMQ-induced epidermal thickening in WT mice than in CIKS-KO mice. The arrow points to a microabscess that is shown in higher magnification at right. Data are representative of 10–15 mice per group. (C) Neutrophilic microabscesses were quantitated by counting in 1,000 × 5 µm areas. Mean ± SEM; **P < 0.01; n = 12. (D) Epidermal thickening quantitated via epidermal area measurements from sections as shown in B. Data are shown as mean ± SEM; **P < 0.01; n = 10. (E) Incremental increase in ear thickness after IMQ or control treatment. Mean ± SEM; **P < 0.01; n = 6–9. (F) Anti-CD31–stained sections of dorsal skin from WT and CIKS-KO mice after IMQ or control treatment. (G) Anti-CD45–stained sections of dorsal skin from WT and CIKS-KO mice after IMQ or control treatment. (H) Quantitative flow cytometric analysis of CD45⁺ cells from whole (2 × 3 cm) dorsal skin samples. Mean ± SEM; **P < 0.01; n = 6–9. (Scale bars: 100 µm.)

Skin sections were stained with antibodies to CD31 and CD45 to assess angiogenesis/dilatation of microvessels and hematopoietic cell infiltration, respectively, which are two hallmarks of psoriasis. IMQ induced a clear increase in the size and density of cutaneous blood vessels in the dermis of the dorsal skin in WT mice, a pathology largely absent in treated CIKS-KO mice (Fig. 1F). IMQ also led to the infiltration of substantially more CD45⁺ cells into the skin of WT than CIKS-KO mice (Fig. 1G); this infiltration was quantitated with flow cytometric analysis of single-cell suspensions of skin (Fig. 1H). In contrast to these profound differences in skin, the spleens of WT and CIKS-KO mice were equally enlarged (Fig. S1). In sum, loss of CIKS markedly attenuated IMQ-induced psoriatic phenotypes, suggesting that the signaling of IL-17 cytokines via CIKS is critical for the development of skin pathology in this model.

IMQ-Induced Production of IL-17 Is Amplified by Positive Feedback.

To elucidate the mechanisms by which CIKS-mediated signaling contributes to psoriasis in this model, we examined the mRNA expression of several cytokines, chemokines, and receptors relevant to skin inflammation. The 5-d IMQ treatment of WT mice led to significant increases in expression of IL-17A, IL-17F, CCR6, IL-36 (α, β, and γ), as well as IL-1β and, to a lesser extent, IL-17C and IL-22 in the dorsal skin of WT mice (absolute levels of IL-22 and IL-17C were low). In contrast, IMQ treatment of CIKS-KO mice resulted in significantly reduced induction of IL-17A, IL-17F, CCR6, IL-1β, and IL-22 with little change in the induction of IL-36 cytokines and IL-17C (Fig. 2A). It was surprising that IMQ failed to induce significant levels of IL-17A and F in the absence of CIKS, because CIKS mediates signaling downstream of these cytokines. These findings suggest that the expression of IL-17A and F feeds forward via CIKS to drive greater expression of these cytokines.

Fig. 2.

IMQ-induced inflammatory cytokines in skin: CIKS-dependent IL-17 production. (A) Relative mRNA expression for indicated genes in the dorsal skin of WT and CIKS-KO mice after IMQ treatment or no treatment (Control). Mean ± SEM; *P < 0.05, **P < 0.01; n = 8–12. (B, D, and F) Representative flow cytometric analyses of IMQ-treated or untreated dorsal skin cells from WT and CIKS-KO mice analyzed for expression of markers as shown. (C, E, and G) Percentages (among CD45⁺ cells) and numbers of IL-17A⁺ cells (C), IL-17A⁺ γδT cells (E), and neutrophils (CD45⁺, CD11b⁺, Ly-6g⁺) (G). Data were generated as shown in B, D, and F, respectively. Mean ± SEM; *P < 0.05, **P < 0.01; n = 6–9.

CCR6 is a chemokine receptor found on cells homing to mucosal tissues and skin and is expressed particularly on IL-17–producing cells such as Th17, γδT, iNKT, and ILC3 (13). To determine whether IMQ led to the accumulation of IL-17A/F–producing CCR6⁺ cells in skin of WT, but not CIKS-KO, mice, we assayed IL-17A protein expression with intracellular staining assays of cell suspensions from skin. IMQ induced a profound increase in the total and relative number of IL-17–producing cells in the skin of WT mice, but this increase was prevented in the absence of CIKS (i.e., in CIKS-KO mice) (Fig. 2 B and C). Next we determined that IL-17⁺ cells, so prominent after IMQ treatment of WT mice, were primarily dermal γδT cells. This finding is consistent with a recent report that dermal γδT cells express relatively lower T-cell receptor (TCR) levels than dendritic epidermal γδT cells (5). Importantly, the number of dermal γδT cells was markedly increased in IMQ-treated WT mice, and this increase was almost fully prevented in treated CIKS-KO mice (Fig. 2 D and E). Thus, CIKS-mediated signaling by IL-17 cytokines results in the accumulation of more IL-17–expressing dermal γδT cells in this model, generating a positive feedback.

We detected increased mRNA expression for IL-17A and F and IL-23 receptor (generally expressed by IL-17–producing cells) in the spleens and skin-draining lymph nodes of IMQ-treated WT but not CIKS-KO mice (Fig. S2 A and B). Most of the IMQ-induced IL-17A was derived from γδT cells as well, and, as in skin, γδT cell numbers increased after treatment in the lymph nodes of WT but not CIKS-KO mice (Fig. S2 C–E).

Consistent with the appearance of microabscesses in the epidermis of IMQ-treated WT but not CIKS-KO mice, the total and relative numbers of neutrophils (CD11b⁺Ly6g⁺) (Fig. 2 F and G) and the mRNA expression of the neutrophil chemoattractant KC/CXCL1 (Fig. 2A) were elevated significantly in the skin of IMQ-treated WT but not CIKS-KO mice.

IMQ-Induced Epidermal Changes Are Altered Markedly in the Absence of CIKS.

The antimicrobial peptides S100A8, S100A9, and Lcn2 are reported targets of IL-17 cytokines in keratinocytes and are highly expressed in psoriatic lesions (33, 34). S100A8, S100A9, and Lcn2 were induced significantly in IMQ-treated dorsal skin of WT but not CIKS-KO mice (Fig. 3A). To investigate the proliferation and differentiation of keratinocytes in the IMQ model, we first determined the mRNA expression levels of K10, an early differentiation marker expressed in suprabasal layers, and the levels of the terminal differentiation markers filaggrin and loricrin. Surprisingly, we found that expression of all three differentiation markers was increased significantly in IMQ-treated CIKS-KO mice compared with treated WT mice (Fig. 3B; of note, K10 and loricrin expression already trended higher in untreated mutant mice compared with untreated WT mice).

Fig. 3.

CIKS-dependent hyperplasia and disrupted differentiation of keratinocytes. (A and B) Relative mRNA expression of indicated antimicrobial genes (A) and differentiation markers (B) in IMQ-treated or untreated (Control) dorsal skin of WT and CIKS-KO mice. Mean ± SEM; *P < 0.05, **P < 0.01; n = 8–12. (C and E–G) Immunofluorescence images of dorsal skin sections from IMQ-treated or untreated (Control) WT and CIKS-KO mice stained with DAPI (blue) and with anti-Ki67 (green) (C); with DAPI (blue), anti-K10 (green), and anti-K5 (red) (E); with DAPI (blue), anti-filaggrin (green), and anti-K5 (red) (F); and with DAPI (blue), anti-loricrin (green), and anti-K5 (red) (G). (Scale bars: 100 µm.) Data are representative of 6–10 mice per group. (D) Ki67⁺ cells were quantitated by counting fluorescent dots in 200 × 200 μm areas from three different sections per mouse as shown in C. Mean ± SEM; **P < 0.01; n = 6–10 mice.

Next we stained skin sections for the proliferation-associated protein Ki67 (Fig. 3 C and D) and the basal keratinocyte marker K5 in combination with K10, filaggrin, or loricrin (Fig. 3 E–G). Untreated WT and CIKS-KO mice contained only rare Ki67⁺ cells, but IMQ treatment of WT mice resulted in uniform expression of Ki67 in the basal keratinocyte layer, with some suprabasal cells also staining positive. In contrast, IMQ-treated CIKS-KO mice contained many fewer Ki67⁺ basal cells than treated WT mice, although their numbers still were increased relative to untreated controls (Fig. 3 C and D). Consistent with these results and the IMQ-induced epidermal thickening shown in Fig. 1D, we observed a large expansion of K5-staining cells in the skin of IMQ-treated WT mice compared with untreated controls and a smaller expansion in IMQ-treated CIKS-KO mice compared with untreated controls (Fig. 3 E–G). Despite the large increase in epidermal thickening, IMQ-treated WT skin contained few K10⁺ cells (Fig. 3E) and hardly any cells staining positive for filaggrin or loricrin (Fig. 3 F and G), whereas such cells were present in untreated controls. Unexpectedly, the opposite effect was seen with IMQ treatment of CIKS-KO skin, with many keratinocytes (even more than in untreated controls) staining positive for K10 as well as filaggrin and loricrin (Fig. 3 E–G). These findings indicate that IMQ treatment of WT skin led to a large increase in proliferating keratinocytes and a steep decrease in differentiated keratinocytes. IMQ treatment of CIKS-KO skin also led to an increase in proliferating keratinocytes, but the increase was significantly smaller than in the treated WT skin. Most remarkably and in contrast to WT mice, in CIKS-KO mice IMQ treatment resulted in a readily apparent increase in cells expressing early and late differentiation markers. Even though epidermal thickening was reduced (but not abolished) in the absence of CIKS, the epidermis in mutant mice was qualitatively very different from that observed in WT skin. Therefore, CIKS-mediated signaling contributes to the IMQ-induced proliferation of keratinocytes and is critical for the decrease in differentiation.

IMQ-Induced Skin Phenotypes in Mice with Epithelial-Specific Ablation of CIKS.

To delineate cell type-specific contributions of IL-17/CIKS-mediated signaling in IMQ-induced psoriasis, we generated mice in which CIKS was conditionally deleted in keratinocytes via a K5-driven Cre recombinase (Traf3ip2^-/flx; K5-cre), hereafter referred to as “CIKS K5-KO” mice. Although IMQ treatment of these mice did result in some visible erythema, the back skin was less affected than that of treated WT mice (Fig. 4A). Consistent with this finding, H&E staining of skin sections revealed reduced epidermal thickening (acanthosis and hyperkeratosis) in treated CIKS K5-KO mice as compared with treated WT mice (Fig. 4 B and D). Furthermore, neutrophilic microabscesses were present only in treated WT mice, not in treated CIKS K5-KO mice (Fig. 4 B and C). However, in contrast with global CIKS deficiency, there was no significant difference in the IMQ-induced increase in ear thickness in CIKS K5-KO and WT mice (Fig. 4E); although the epidermal thickening was reduced in the ears of CIKS K5-KO mice, the dermal thickening was increased (see below). Although epidermal pathology was notably reduced in the absence of CIKS in keratinocytes, the IMQ-induced increase in angiogenesis/blood vessel dilatation as assessed with CD31 antibody staining was not attenuated in these mice compared with treated WT mice (Fig. 4F). This finding is consistent with the visible erythema noted above. Also, in stark contrast to mice globally lacking CIKS (i.e., CIKS-KO mice), we observed not a reduction but instead a modest increase in infiltrating CD45⁺ cells in the skin of IMQ-treated CIKS K5-KO mice as compared with treated WT mice (Fig. 4G). This finding is consistent with the increased dermal ear thickness noted above. The mild increase in infiltrating cells in CIKS K5-KO mice over that seen in WT mice was confirmed with cytometric analysis of single-cell suspensions (of note, basal levels already trended slightly higher in CIKS K5-KO mice) (Fig. 4H). In aggregate, these data indicate that CIKS-mediated signaling within keratinocytes plays an important role in IMQ-induced epidermal pathology but, surprisingly, is not required for cellular infiltration into skin. In fact, CIKS-mediated signaling within keratinocytes appears to dampen IMQ-induced infiltration into skin slightly but has no obvious effect on IMQ-induced changes in blood vessels.

Fig. 4.

Select IMQ-induced psoriatic phenotypes dependent on CIKS in keratinocytes. (A) Dorsal areas of WT and CIKS K5-KO (K5-cre;Traf3ip2^−/flx) mice after IMQ treatment or no treatment (Control) of shaved dorsal skin. (B) H&E-stained dorsal sections showing more pronounced IMQ-induced epidermal thickening in WT mice than in CIKS K5-KO mice and microabscesses in WT mice. The arrow points to a microabscess that is shown in higher magnification at right. Data are representative of 10–15 mice per group. (C) Neutrophilic microabscesses were quantitated by counting in 1,000 × 5 µm areas. Mean ± SEM; **P < 0.01; n = 6. (D) Epidermal thickening was quantitated via epidermal area measurements from sections as shown in B. Mean ± SEM; **P < 0.01; n = 10. (E) Incremental increase in ear thickness after IMQ treatment or no treatment (Control). Mean ± SEM; n = 4–6. (F) Anti-CD31–stained sections of dorsal skin from WT and CIKS K5-KO mice after IMQ treatment or no treatment (Control). (G) Anti-CD45–stained sections of dorsal skin from WT and CIKS K5-KO mice after IMQ treatment or no treatment (Control). (H) Quantitative flow cytometric analysis of CD45⁺ cells from whole (2 × 3 cm) dorsal skin samples. Mean ± SEM; *P < 0.05, **P < 0.01; n = 6–9. (Scale bars: 100 µm.)

CIKS-Mediated Signaling in Keratinocytes Promotes Epidermal Pathology but Not Cellular Infiltration into Skin.

To define further the specific epidermal and dermal changes induced by IMQ in CIKS K5-KO mice, we measured mRNA expression levels of genes relevant to psoriasis. We detected highly significant (or nearly significant) increases in IMQ-induced expression of IL-17A, IL-17F, IL-22, CCR6, IL-17C, IL-36 (α, β, and γ), and IL-1β in WT as well as in CIKS K5-KO mice. In particular, expression levels of IL-17A and CCR6 and, to a lesser extent, IL-22 and IL-17F were higher in CIKS K5-KO mice than in WT mice (Fig. 5A). The increase in IL-17A and CCR6 was consistent with the increased skin infiltration in CIKS K5-KO above that in WT mice and contrasts starkly with the significantly reduced expression of these markers in globally deficient CIKS-KO mice (Fig. 2). Although the previously mentioned hematopoietic cell-expressed genes were highly increased upon IMQ treatment in both WT and conditional mutant mice, the keratinocyte-expressed genes IL-17C and IL-36α/β/γ, as well as IL-1β, were induced to a lesser degree in CIKS K5-KO mice than in WT mice (although the levels still were above those seen in untreated skin) (Fig. 5A).

Fig. 5.

IMQ-induced inflammatory cytokines in skin: IL-17 production independent of CIKS in keratinocytes. (A) Relative mRNA expression for indicated genes in the dorsal skin of WT and CIKS K5-KO mice after IMQ treatment or no treatment (Control). Mean ± SEM; *P < 0.05, **P < 0.01; n = 8–12. (B and E) Representative flow cytometric analyses of dorsal skin cells from WT and CIKS K5-KO mice treated with IMQ or left untreated (Control), analyzed for expression of markers as shown. (C, D, and F) Percentages (among CD45⁺ cells) and numbers of IL-17A⁺ cells (C), IL-17A⁺ γδT cells (E), and neutrophils (CD45⁺, CD11b⁺, Ly-6g⁺) (F) generated from data as shown in B (for C and D) and E (for F). Mean ± SEM; *P < 0.05, **P < 0.01; n = 6–9.

Analysis of the expression of IL-17 protein upon IMQ treatment also showed a significant increase in the total and relative numbers of IL-17–expressing cells in CIKS K5-KO mice compared with untreated mice, and these numbers were higher than those observed in treated WT mice (Fig. 5 B and C). As in WT mice, the great majority of IL-17–producing cells in CIKS K5-KO mice were dermal γδT cells (Fig. 5 B and D). Thus, loss of CIKS in keratinocytes did not mirror global loss of CIKS. The absence of CIKS in keratinocytes not only failed to reduce IMQ-induced increases in IL-17–producing cells in skin but instead resulted in slightly higher levels. This finding is the opposite of the effect observed in mice globally lacking CIKS.

As in treated WT mice, the mRNA expression levels of IL-17A and F were increased in the lymph nodes of IMQ-treated CIKS K5-KO mice, as were the relative numbers of IL-17–producing and γδT cells (Fig. S3).

In contrast to the increased overall skin infiltration in IMQ-treated CIKS K5-KO mice, there was a highly significant reduction in the total and relative numbers of infiltrating neutrophils (Fig. 5 E and F), akin to the reduction observed in CIKS-KO mice. Consistent with this reduction, the expression of the neutrophil chemoattractant KC/CXCL1 also was significantly lower in the skin of IMQ-treated CIKS K5-KO mice than in treated WT mice (Fig. 5A). These data suggest CIKS-mediated signaling in keratinocytes is important for neutrophil accumulation in skin.

mRNA expression levels of S100A8, S100A9, and, to lesser extent, Lcn2 also were reduced in CIKS K5-KO mice as compared with WT mice (Fig. 6A), similar to the reductions seen in CIKS-KO mice. In addition, mRNA expression levels of the keratinocyte differentiation markers K10, filaggrin, and loricrin were higher in the IMQ-treated skin of CIKS K5-KO mice than in the treated skin of WT mice (Fig. 6B) (of note, loricrin levels already were higher in untreated CIKS K5-KO mice than in untreated WT mice). Furthermore, although IMQ treatment of the skin of CIKS K5-KO mice resulted in a clear increase of Ki67⁺ proliferating keratinocytes, this increase was notably lower than in IMQ-treated WT mice (Fig. 6 C and D). This observation is consistent with the epidermal thickening in IMQ-treated CIKS K5-KO mice noted above: Although thickening clearly occurred in IMQ-treated CIKS K5-KO mice relative to untreated mice, it was significantly less than in treated WT mice. As already noted above, IMQ treatment of WT skin led to a remarkable decrease in keratinocyte differentiation, evidenced by the loss of keratinocytes staining positive for the early differentiation marker K10 and the late differentiation markers filaggrin and loricrin. In contrast, IMQ treatment of CIKS K5-KO mice resulted in an expansion of keratinocytes staining positive for these differentiation markers (Fig. 6 C–G). The epidermal changes in CIKS K5-KO thus phenocopy the effects observed in CIKS-KO mice (Fig. 3 C–G).

Fig. 6.

Hyperplasia and disrupted differentiation in keratinocytes dependent on CIKS. (A and B) Relative mRNA expression of indicated antimicrobial genes (A) and differentiation markers (B) in IMQ-treated or untreated (Control) dorsal skin of WT and CIKS K5-KO mice. Mean ± SEM; *P < 0.05, **P < 0.01; n = 8–12. (C and E–G) Immunofluorescence images of dorsal skin sections from IMQ-treated or untreated (Control) WT and CIKS K5-KO mice stained with the indicated makers as in Fig. 3. (Scale bars: 100 µm.) Data are representative of 6–10 mice per group. (D) Ki67⁺ cells were quantitated from sections as shown in C. Mean ± SEM; **P < 0.01; n = 6–10.

IMQ-induced pathology in the epidermis thus is largely dependent on CIKS-mediated signaling in keratinocytes, because it led to enhanced proliferation and greatly reduced differentiation of keratinocytes, neutrophil accumulation, and microabscess formation. In this respect, mice lacking CIKS in keratinocytes were similar to mice globally deficient in CIKS. However, CIKS-KO and CIKS K5-KO mice differed with respect to overall cellular infiltration into skin. Compared with WT mice, IMQ treatment of mice lacking CIKS in keratinocytes resulted in higher levels of infiltration and accumulation of IL-17–expressing γδT cells, whereas mice globally lacking CIKS failed to accumulate these cells and showed overall reduced cellular infiltration. Therefore, signaling by IL-17 cytokines via CIKS in cells other than keratinocytes is essential for the feed-forward mechanism that results in the accumulation of IL-17–producing cells, whereas signaling via CIKS in keratinocytes tends to dampen this process in the context of repeated IMQ applications.

IL-17–Stimulated Dermal Fibroblasts Promote Production of IL-17 by γδT Cells.

To investigate how IL-17 signaling in nonkeratinocytes might have enhanced inflammation and increased the production of IL-17, we examined IL-17–induced responses of primary WT and CIKS-deficient dermal fibroblasts; the latter cells are located next to IL-17–producing dermal γδT cells in tissue. IL-17 induced the expression of inflammatory mediators such as IL-1α, IL-6, CCL20, and CXCL1 in WT fibroblasts but failed to do so in CIKS-deficient fibroblasts (Fig. 7A). TNF-α is an inflammatory cytokine also implicated in psoriatic inflammation; it synergized with IL-17 in a CIKS-dependent manner to increase expression of IL-1α, IL-1β, and CXCL1, and it added to the IL-17–induced expression of IL-6. IL-17–induced production of such mediators likely contributed to cellular infiltration into the dermis. γδT cells readily produce IL-17 in response to IL-23 plus IL-1 (35); therefore we asked whether IL-17–stimulated dermal fibroblasts might feed forward to help induce the production of IL-17 by γδT cells. Primary cell cultures of WT and CIKS-deficient dermal fibroblasts were stimulated with IL-17 plus TNF-α and cocultured with primary γδT cells in the presence of IL-23; then IL-17 production by γδT cells was assessed via intracellular staining (Fig. 7 B and C). Under these conditions stimulation of WT dermal fibroblasts led to a significant increase in IL-17 production by γδT cells, whereas stimulation of CIKS-deficient dermal fibroblasts failed to do so. In the absence of dermal fibroblasts, γδT cells did not increase the production of IL-17, demonstrating that IL-17 was unable to promote its own production in γδT cells directly, even in the presence of IL-23 and TNF-α (Fig. 7B). As expected, γδT cells proficiently produced IL-17 when stimulated directly with IL-1β plus IL-23, regardless of whether fibroblasts were present (Fig. 7 B and C). We next established that IL-1 cytokines produced by stimulated dermal fibroblasts helped mediate the production of IL-17 by γδT cells, because both anti–IL-1 receptor antibodies and anakinra largely abolished this effect when added to the cocultures (Fig. 7 D and E). These findings identify a mechanism by which IL-17 signaling into dermal fibroblasts (nonkeratinocytes) leads to a further increase in production of IL-17.

Fig. 7.

IL-17–stimulated dermal fibroblasts promote production of IL-17 by γδT cells. (A) Relative mRNA expression for indicated genes in primary dermal fibroblast (FB) cultures from WT and CIKS-KO mice, stimulated with recombinant IL-17A, TNF-α, or both for 6 h. Mean ± SEM; *P < 0.05, **P < 0.01; n = 3. (B) Representative flow cytometric analyses of WT γδT cells for intracellular IL-17 expression after coculture (or no coculture) with primary dermal fibroblasts from WT or CIKS-KO mice in the presence or absence of IL-23 or IL-23+IL-1β, as indicated. The primary dermal fibroblasts were pretreated with IL-17A+TNF-α before coculture or were left untreated, as shown (Control = absence of any cytokines). (C) Percentage of γδT cells producing IL-17, based on analyses of cultured or cocultured γδT cells shown in B. Mean ± SEM; *P < 0.05, **P < 0.01; n = 3. (D) Representative flow cytometric analyses of intracellular IL-17 expression in WT γδT cells after coculture with pretreated (IL-17A+TNF-α) primary dermal fibroblasts from WT or CIKS-KO mice, in the presence of IL-23 with or without anti–IL-1R or anakinra. (E) Percentage ofγδT cells producing IL-17, based on analyses of γδT cells shown in D. Mean ± SEM; *P < 0.05, **P < 0.01; n = 4.

IL-17 Modulates Proliferation and Differentiation of Keratinocytes.

The data presented above demonstrate that IL-17 cytokine-dependent signaling to keratinocytes via CIKS promotes proliferation and reduces differentiation of keratinocytes in vivo. IL-17 cytokines might modulate these processes directly, or they might do so indirectly via mechanisms involving other cells in vivo. For example, IL-17 may induce keratinocytes to produce factors that target or recruit other cells, which in turn produce cytokines or growth factors that modulate proliferation and differentiation of keratinocytes. To address this issue, we investigated the response of primary keratinocytes to IL-17 in culture. As expected, the responses to IL-17 in the keratinocyte cultures were fully dependent on the presence of CIKS (Fig. S4). Stimulation of proliferating mouse keratinocytes (cultured in low Ca²⁺ conditions) with IL-17 for 48 and 72 h resulted in significantly decreased mRNA expression of K10, filaggrin, and loricrin, especially at the latter time point (Fig. 8 A and B). Cultured primary keratinocytes rapidly undergo terminal differentiation in elevated Ca²⁺ culture conditions. Stimulation with IL-17 concurrent with the addition of Ca²⁺ also significantly reduced the expression of the differentiation marker K10 after 48 h, and there was a strong trend toward reduced expression of the differentiation markers loricrin and filaggrin after 72 h (Fig. 8 A and B).

Fig. 8.

IL-17A regulates the proliferation and differentiation of primary keratinocytes. (A and B) Relative mRNA expression of indicated differentiation markers in primary WT keratinocytes cultured in low Ca²⁺ or 0.12 mM Ca²⁺ and left without stimulation or stimulated with IL-17A for 48 h (A) or 72 h (B). (C and D) Changes in the percentage of primary keratinocytes in S-phase cultured in low or 0.12 mM Ca²⁺ and left without stimulation or stimulated with IL-17A for 24 h (C) or 48 h (D), relative to untreated keratinocytes cultured in low Ca²⁺. Mean ± SEM; *P < 0.05, **P < 0.01; n = 4.

We performed cell-cycle analysis of keratinocytes cultured in high and low Ca²⁺ conditions in the presence or absence of IL-17. As expected, more cells were in proliferation-associated S-phase when cultured under low Ca²⁺ conditions. The addition of IL-17 increased the proportion of proliferating keratinocytes under both Ca²⁺ conditions, reaching significance after 24 and 48 h of stimulation, respectively (Fig. 8 C and D). Taken together, these findings suggest that IL-17 cytokine signaling via CIKS in cultured primary keratinocytes can directly promote their proliferation and inhibit their differentiation. These mechanisms likely contribute to the IMQ-induced epidermal changes mediated by CIKS-dependent signaling in keratinocytes in vivo.

Discussion

Many findings implicate IL-17 cytokines in psoriasis in humans and mouse models, although relatively little is known about precisely how these cytokines contribute to this inflammatory disorder. In addition to IL-17A, the less potent family members IL-17F and IL-17C also may play a role in psoriasis. CIKS is the obligate adaptor for signaling by these cytokines and thus represents a potential therapeutic target to block their functions. The specific role of CIKS in psoriasis has not been explored previously. Here we investigated whether CIKS-mediated signaling is critical in the IMQ-induced psoriasis model and what its specific contributions and underlying mechanisms of action are. We demonstrate that CIKS-mediated signaling is a major contributor to many distinct components of the pathologic phenotype in this model. It promotes proliferation of keratinocytes and, unexpectedly, leads to a block in terminal differentiation. Furthermore, CIKS-mediated signaling is required for the formation of neutrophilic microabscesses, and it strongly promotes cellular infiltration into skin, in particular the accumulation of IL-17–producing dermal γδT cells. Thus, a positive feedback loop amplifies IL-17 production, an unexpected finding. Also surprising was the finding that IL-17 cytokines target nonkeratinocytes to enable this feedback. Specifically we were able to demonstrate that IL-17 signaling into primary dermal fibroblasts enables these cells to promote the production of IL-17 by cocultured primary γδT cells, mediated largely via IL-1 cytokines. On the other hand, IL-17/CIKS-mediated signaling into keratinocytes is required for the formation of neutrophilic microabscesses, contributes to increased proliferation, and reduces differentiation of keratinocytes. The latter phenotypes may be controlled, at least in part, by IL-17 cytokines via cell-intrinsic mechanisms, thus possibly expanding the spectrum of functions these cytokines may execute.

Global loss of CIKS leads to reduced expression of IL-17 in response to IMQ; this finding was unexpected, because CIKS functions downstream of IL-17. We demonstrate that this effect is caused by the markedly reduced accumulation of IL-17–producing dermal γδT cells, the main source of IL-17 in this model of psoriasis. IL-17 cytokines thus feed forward to induce the production of more IL-17. Because IL-17 is known to induce the expression of chemokines and cytokines in epithelial cells, it was reasonable to postulate that IL-17 might have recruited/activated γδT cells (and other cells) in this fashion. However, conditional ablation of CIKS in keratinocytes failed to reduce IL-17 or the accumulation of IL-17 producers. Therefore, IL-17 cytokines target cells other than keratinocytes to increase the numbers of IL-17–producing dermal γδT cells, and we identified a cell type that is able to mediate this positive feedback mechanism. IL-17 signaling into primary dermal fibroblasts, especially in combination with TNF-α, induces the expression of various inflammatory mediators, in particular IL-1, enabling these fibroblasts to promote the production of IL-17 in cocultured primary γδT cells in the presence of IL-23. The observed feed-forward effect is entirely dependent on the presence of CIKS in dermal fibroblasts. This finding does not rule out possible additional mechanisms by which IL-17 may promote inflammation and its own production, such as the ability of IL-17–induced mediators to recruit or amplify γδT cells, or the existence of still other cell types that may function in a manner analogous to that of dermal fibroblasts.

The noted accumulation of IL-17–producing γδT cells in IMQ-treated WT mice was increased further, albeit modestly, in mice with conditional, epithelial-specific ablation of CIKS. This finding implies the existence of a negative feedback mechanism in WT mice that impacts IL-17 production and is dependent on IL-17/CIKS-signaling in keratinocytes. Although accumulation was dampened only modestly, and the dampening may have occurred only after repeated IMQ applications, it will be instructive to learn the molecular mechanisms by which IL-17–activated keratinocytes help down-modulate IL-17 production. In summary, IL-17 production is greatly amplified via CIKS-mediated signaling in nonkeratinocytes, particularly in dermal fibroblasts, and can be partially limited via CIKS-mediated signaling in keratinocytes.

The phenotypic difference in the epidermal layers of IMQ-treated WT mice and those of treated mice lacking CIKS either globally or selectively in keratinocytes was striking. Both CIKS-deficient models developed notably less epidermal thickening than WT mice, as is consistent with a reduced rate of proliferation in basal keratinocytes. Nevertheless, both CIKS-deficient models still exhibited some epidermal thickening upon treatment as compared with untreated mice, possibly indicating a merely quantitative difference between treated WT and treated CIKS-deficient mice. However, we also observed a prominent qualitative difference in the epidermis in the two mutant strains compared with WT mice. Although IMQ treatment of WT mice inhibited the differentiation of keratinocytes—only scant cells expressed early and late keratinocyte differentiation markers—treatment of both CIKS-deficient mouse models significantly expanded the numbers of cells expressing these differentiation markers. Taken together, the findings indicate that CIKS-mediated signaling in keratinocytes caused increased proliferation and markedly reduced differentiation in response to IMQ. This result was unexpected and expands our knowledge of functions of IL-17 cytokines.

There are several modes of action by which IL-17/CIKS signaling in keratinocytes could control their proliferation/differentiation. It could act indirectly via extracellular factors targeting other cell types and/or it could act more directly, either cell-autonomously or via production of autocrine/paracrine-acting factors. Because IL-17 is able to increase the proliferation and reduce the differentiation of primary keratinocytes in culture, this cytokine is capable of modulating these processes directly. Although this finding does not rule out indirect mechanisms via other cells in vivo, it suggests that contributions via direct mechanisms are possible. This result is unexpected, because IL-17 has been thought to control primarily the expression of inflammatory and antimicrobial products. However, it was reported recently that the keratinocyte-derived antimicrobial agent RegIIIγ might act as an autocrine/paracrine-acting inducer of keratinocyte proliferation and inhibitor of differentiation (36). Thus it is conceivable that IL-17–induced antimicrobial products also may be involved in regulating the proliferation and differentiation of keratinocytes.

IL-17/CIKS-mediated signaling in keratinocytes is essential for the formation of IMQ-induced neutrophilic microabscesses and neutrophil accumulation in skin. Several studies have highlighted the importance of neutrophils for development of psoriasis. Drug-induced agranulocytosis was reported to lead to rapid and dramatic improvement of psoriatic plaques, and psoriatic lesions quickly returned upon full recovery of neutrophil numbers in blood (37). Such findings support the notion that IL-17/CIKS–mediated signaling in keratinocytes also may regulate proliferation/differentiation of these cells in an indirect manner, possibly via neutrophils.

It must be acknowledged that CIKS may, at least theoretically, contribute to the psoriatic phenotype in ways other than exclusively via its proven role in conveying signals downstream of IL-17 cytokines. It has been reported that increased expression levels of CIKS might be sufficient to generate downstream signals in some settings (38). However, at present there is no evidence in the IMQ-induced model of psoriasis that might implicate CIKS-mediated signaling in an IL-17 cytokine-independent fashion.

Our findings provide proof of principle that CIKS is a potential therapeutic target in psoriasis-like inflammation. Anti–IL-17 and anti–IL-17RA antibodies are now in phase III clinical trials and have shown exceptional promise in earlier trials, although small-molecule inhibitors targeting intracellular signaling components ultimately could prove superior. Initial studies of anti–IL-17 or anti–IL-17RA antibody therapies in psoriatic patients indicated that numerous pathologic measures were improved dramatically, mirrored by normalization in the expression of a very wide range of genes known to be dysregulated in psoriasis. Although mouse models only approximate human psoriasis, the present study may inform the interpretation of the remarkable results achieved in recent human trials. As demonstrated here, IL-17/CIKS–mediated signaling in more than one cell type makes critical contributions to the IMQ-induced psoriasis model. Signaling in keratinocytes is critical for their proliferation and loss of differentiation, and signaling in nonkeratinocytes, particularly in dermal fibroblasts, is critical for cellular infiltration and the positive feedback leading to increased numbers of IL-17–producing cells. Similar circuitry may exist in humans that would help explain the great efficacy of anti–IL-17 signaling therapies. Indeed, initial limited studies on skin biopsies have shown that anti–IL-17/IL-17RA treatment rapidly reverses disease-associated keratinocyte proliferation and cellular infiltration and, remarkably, also rapidly diminishes the production of IL-17 itself (39). Thus, the 5-d IMQ model analyzed here may phenocopy the human disease in fundamentally important ways.

Materials and Methods

Mice.

C57BL/6NTac Traf3ip2^−/− (CIKS-KO) mice have been described (26); they were cohoused and experimentally matched with C57BL/6NTac Traf3ip2^+/+ (WT) mice. C57BL/6J Traf3ip2^flx/flx (WT) and C57BL/6J Traf3ip2^−/− mice also have been described (25); they were cohoused and experimentally matched with C57BL/6J mice in which CIKS was conditionally ablated in epithelial cells via K5-Cre (K5-cre; Traf3ip2^−/flx) (i.e., CIKS K5-KO mice). CIKS K5-KO mice were generated by intercrosses with C57BL/6J K5-Cre transgenic mice (40), with K5-Cre transmission via male mice only; final crosses also yielded the following littermate controls which behaved similarly: Traf3ip2^flx/+, Traf3ip2^flx/flx, Traf3ip2^−/+, and Traf3ip2^−/fl. Experiments were conducted with 8- to 10-wk-old mice. Mice were bred and housed in a facility at the National Institute of Allergy and Infectious Diseases (NIAID), and all experiments were performed with the approval of the NIAID Animal Care and Use Committee and in accordance with all relevant institutional guidelines.

Experimentally Induced Psoriasis.

Aldara cream (Medicis) containing 5% (wt/wt) IMQ was applied to the shaved dorsal skin and to ears daily for 5 d, as described previously (30). Control mice also were shaved but otherwise were left untreated. Ear thickness was measured every other day, and mice were killed on day 6.

Cellular Analysis.

Lymph nodes and spleen were mechanically dissociated to obtain single-cell suspensions, as described (24). For single-cell suspensions from dorsal skin, 2 × 3 cm sections were incubated for 1.75 h at 37 °C with 5 mL RPMI medium containing 500 μg/mL of Liberase (Roche) and then were minced with sharp scissors, incubated for an additional 15 min with 0.05% DNase I (Sigma-Aldrich), and filtered sequentially through 100-, 70-, and 40-μm nylon mesh (41). All cell suspensions were treated with red blood cell lysis buffer (Lonza), and cells were stained with Aqua (Invitrogen) and with antibodies against one or more of the following proteins for cytometric analyses: B220(RA3-6B2), CD3ε(145-2C11), CD4(RM4-5), CD8(53-6.7), IL-17A, and IL-17F(O79-289) (BD Biosciences); TCRγδ(UC7-13D5), IL-17A(eBio17B7), IL-17F(eBio18F10), and IFNγ(XMG.2) (eBioscience); Ly6g(IA8), CD45.2(104), TCRγδ(GL3), TCRvγ4(UC3-10A6), TCRvγ5(536), CD11b(M1/70), TCRβ(H57-597), and F4/80(CI:A3-1) (BioLegend). For intracellular staining, cells first were stimulated with phorbol12-myristate13-acetate (5 ng/mL) and ionomycin (500 ng/mL) and were treated with protein transport inhibitor mixture (eBioscience) for 4 h. Data were collected with a FACSCanto instrument (BD Biosciences) and were analyzed using FlowJo software (Tree Star).

Histology.

Mouse dorsal skin and ear tissue were fixed in 4% formaldehyde. Tissue sections were stained with H&E or anti-CD45 (1:50; BD Bioscience), which was visualized after diaminobenzidine reaction (Vector Laboratories) with an Olympus BX50. Epidermal areas were quantitated on H&E-stained slides with ImageJ software (National Institutes of Health). For immunofluorescence, 4% formaldehyde-fixed skin sections were deparaffinized, blocked with 1% BSA in PBS, and stained with primary antibodies against the following proteins: K5 (1:100; Lifespan Biosciences), K10 (1:100; Covance), filaggrin (1:100; Covance), loricrin (1:100; Covance), and Ki67 (1:100; BD Pharmingen). Secondary antibodies were labeled with either Alexa Fluor 488 or Alexa Fluor 546 and directed against goat IgG, guinea pig IgG, and mouse IgG (1:250; Molecular Probes). Slides were mounted with Vectashield with or without DAPI (Vector Labs) and visualized with a Leica AF6000LX fluorescence microscope. Frozen sections were fixed in acetone and blocked with 1% BSA in PBS. Endogenous biotin was blocked using the Streptavidin-Biotin blocking kit (Vector Labs). Sections were incubated for 2 h with biotin-conjugated anti-CD31 antibodies (1:50; BD Biosciences) and visualized with streptavidin-conjugated Alexa Fluor 568 (1:1,000; Molecular Probes).

Quantitative Real-Time PCR.

RNA was purified using TRIzol (Invitrogen) and the RNeasy kit (Qiagen); cDNA was generated with the cDNA synthesis kit (Qiagen), and quantitative real-time PCR was performed using the Taqman protocol. The mouse primers for Gapdh, IL-17, IL-17F, IL-17C, CCR6, Lcn2, S100a8, S100a9, K10, filaggrin, loricrin, and CCL20 were obtained from Applied Biosystems. All values were normalized to Gapdh.

In Vitro Culture of Keratinocytes, Dermal Fibroblasts, and γδT Cells.

Isolation of keratinocytes and fibroblasts from neonatal mice and their culture conditions, including the culture of keratinocytes with or without external calcium, were as described previously (42, 43). Recombinant IL-17A (100 ng/mL; R&D Systems) was used to stimulate primary keratinocytes, whereupon cells were harvested for RT-PCR analysis or cell-cycle analysis. Cells in S-phase were considered to be proliferating. For cell-cycle analysis, cells were suspended in a hypotonic fluorochrome solution (50 μg/mL propidium iodide, 0.1% sodium citrate, and 0.1% Triton X-100) at 4 °C for 15–20 min and then were analyzed with the FACSCanto instrument (BD Biosciences). For cocultures of dermal fibroblasts and γδT cells, WT and CIKS-KO primary dermal fibroblasts (1 × 10⁴) were seeded in a flat-bottom plate. After 24 h, fibroblasts were treated for 24 h with IL-17A (100 ng/mL; R&D Systems) plus TNF-α (20 ng/mL; PeproTech) or were left untreated; then WT γδT cells (2 × 10⁴) were added. The cocultures were grown for an additional 20 h in the presence of IL-23 (10 ng/mL; eBioscience) with the protein transport inhibitor mixture (eBioscience) present during the final 4 h. For some coculture experiments α-IL-1 receptor antibodies (JAMA-147) or isotype control monoclonal antibodies (both 10 μg/mL; Bio-XCell) or anakinra (Kineret; 25 μg/mL; Sobi) were added 30 min before the addition of γδT cells. γδT cells were isolated from skin-draining lymph nodes with a TCRγ/δ⁺ T-cell isolation kit (Miltenyi). For each experiment γδT cells from 10 to 15 mice were pooled.

Statistical Analyses.

All data are presented as the mean ± SEM. A two-tailed Student t test was used to evaluate significance; P values <0.05 were considered statistically significant, and values <0.01 were considered highly significant.