Osteopontin regulates development and function of invariant natural killer T cells

Abstract

Invariant natural killer T (iNKT) cells belong to a subset of lymphocytes bridging innate and acquired immunity. We demonstrated that osteopontin (OPN) is involved in the activation of iNKT cells. In the present work, we examined whether OPN affects development and function of iNKT cells. We found that the number of peripheral iNKT cells was significantly reduced in OPN-deficient mice compared with wild-type mice. Although the number of thymic iNKT cells was not different between WT and OPN-deficient mice, intrathymic iNKT cell maturation was impaired in OPN-deficient mice. iNKT cell function was also significantly altered in OPN-deficient mice, as evidenced by (i) deficient down-regulation of iNKT cell receptor, (ii) reduction of IL-4 production while preserving production of IFN-γ, and (iii) reduction of Fas ligand (FasL) expression, leading to reduced Fas/FasL-dependent cytotoxicity against hepatocytes. Importantly, activation of the transcription factors NFAT2 (nuclear factor of activated T cells 2) and GATA-3 was impaired, whereas activation of T-bet was preserved in iNKT cells of OPN-deficient mice. These data collectively indicate that OPN plays a pivotal role not only in the development, but also in the function of iNKT cells.

Invariant natural killer T (iNKT) cells constitute a subset of T cells and specifically recognize glycolipid antigens presented by CD1d (1), regulating T helper (Th)1 and Th2 immune responses by producing various cytokines (1, 2). iNKT cells are involved in the development of several human diseases such as type I diabetes, multiple sclerosis, and systemic lupus erythematosus (3–5).

It has been shown that specific transcription factors are critical for the development of polarized Th1 and Th2 responses (6) in conventional T cells. The T-bet is a master coordinator of gene expression in T cells and is critical for the development of Th1 T cells (7). T-bet-deficient CD4⁺ T cells exhibit a marked decrease in IFN-γ production. Interestingly, mature iNKT cells cannot be formed in the absence of T-bet (8). GATA-3 is a key transcription factor that controls Th2 cell development and Th2 cytokine production in conventional T cells (9). CD4⁺ T cell numbers are markedly reduced in GATA-3-deficient mice, and in vitro differentiated CD4⁺ T cells from these animals have defects in Th2 cytokine production. Similarly, mature iNKT cells are drastically reduced in GATA-3-deficient mice. In addition, GATA-3 mRNA expression correlates with IL-4 expression in iNKT cells (10). STAT6 is critical for the development of Th2 responses in conventional T cells, yet iNKT cells do not require STAT6 for IL-4 production (11). The nuclear factor of activated T cells (NFAT) family is also involved in regulating the Th1/Th2 balance (6). NFAT activation leads to the production of a variety of cytokines in conventional T cells. In contrast, NFAT2 activation leads to the up-regulation of IL-4 expression, but not TNF-α or IFN-γ in NKT cells (12). Thus, T-bet, STAT6, GATA-3, and NFAT2 behave differently between iNKT and conventional T cells.

Osteopontin (OPN) skews T cell differentiation toward Th1 (13). OPN gene expression is regulated by T-bet, and, importantly, T-bet-dependent expression of OPN in T cells is essential for skewing of T cells toward Th1 cell differentiation (14). T-bet is also critical for OPN expression in plasmacytoid dendritic cells (pDC) (15). Intracellular OPN acts as an integral part of the signal transduction machinery in pDC. OPN has been implicated in the pathogenesis of various immunological disorders (16, 17). Therefore, it is important to examine whether OPN affects the development and function of immune cells beside T cells and pDC. In this regard, iNKT cells are involved in the pathogenesis of hepatic injury (18). We demonstrated that OPN-deficient mice were protected from Con A-induced hepatic injury and OPN is critically involved in iNKT cell activation (17). Therefore, we attempted to clarify how OPN affects the development and function of iNKT cells. Here, we show that iNKT cell number in the liver and spleen of OPN-deficient mice, compared with WT mice, is significantly reduced. Although the thymic iNKT cell number was not different between wild-type (WT) and OPN-deficient mice, intrathymic development of iNKT cells was blocked in OPN-deficient mice. iNKT cell function was also significantly altered in OPN-deficient mice, as evidenced by (i) deficient down-regulation of iNKT cell receptor upon activation; (ii) reduction of IL-4 secretion and production, while preserving production of IFN-γ; and (iii) reduction of Fas ligand (FasL) expression, leading to reduced Fas/FasL-dependent cytotoxicity against hepatocytes. Importantly, activation of the transcription factors NFAT2 and GATA-3 was impaired, whereas activation of T-bet was preserved in iNKT cells of OPN-deficient mice. These data collectively indicate that OPN plays a pivotal role not only in the development but also in the function of iNKT cells.

Results and Discussion

Involvement of OPN in iNKT Cell Development.

We isolated iNKT cells from normal WT mice and studied OPN expression at the mRNA level. Resident intrahepatic iNKT cells clearly expressed OPN, whereas iNKT cells from OPN-deficient mice did not (Fig. 1A). OPN is involved in the survival of cells (16) and differentiation and development of dendritic cells (DCs) (19). Thus, we examined whether OPN deficiency impacted the development of iNKT cells. In our previous preliminary examination, we detected only a slight decrease in the numbers of hepatic iNKT cells in OPN-deficient mice (17). However, in those experiments OPN-deficient mice were studied at early backcrosses to C57BL/6 mice. In the present work, we used OPN-deficient mice, backcrossed to C57BL/6 at least 11 generations. Then, we compared the number of CD1d-restricted iNKT cells in liver, spleen, and thymus between adult OPN-deficient and WT mice. The percentages of CD1d-restricted iNKT cells in the liver and spleen of OPN-deficient mice were significantly lower than those in WT mice. In the liver and spleen, there was an ≈2-fold reduction compared with WT iNKT cells, but there was no significant difference in thymic iNKT cell prevalence between OPN-deficient and WT mice [supporting information (SI) Fig. S1]. To characterize the involvement of OPN during iNKT cell development further, we examined the total numbers of iNKT cells (Fig. S2). There was no significant difference in terms of iNKT cell numbers in thymus between OPN-deficient and WT mice from 1 to 4 weeks (Fig. S2 and Fig. 1B). iNKT cells in the spleen of OPN-deficient mice were lower than those in WT mice at 4 weeks, but not from 1 to 3 weeks. The cell number of iNKT cells in the liver was significantly reduced in OPN-deficient mice, compared with those in WT mice at 3 and 4 weeks, but not at 1 and 2 weeks. Thus, the number of peripheral iNKT cells is significantly reduced, whereas that of intrathymic iNKT cells is not reduced in OPN-deficient mice. This reduction of peripheral iNKT cell could be caused by a blockade in iNKT cell maturation. This possibility prompted us to examine whether intrathymic maturation of iNKT cells in OPN-deficient mice was impaired. Immature iNKT cells undergo several well defined developmental stages in the thymus. The most immature iNKT cells are at CD44⁻NK1.1⁻ (stage 1), which differentiate through the CD44⁺NK1.1⁻ (stage 2) to become CD44⁺NK1.1⁺ (stage 3) cells (20). At 1 week, ≈33, 60, and 6% of iNKT cells are at stage 1, 2, and 3, respectively, in WT mice (Fig. 1C). The maturation of thymic iNKT cells was altered greatly in the absence of OPN. The percentage of immature stage 1 iNKT cells gradually decreased as mice aged, which contrasts with the gradual increase of stage 3 iNKT cells in WT mice. At 1 week, ≈6% and 33% of iNKT are at stage 3 and stage 1, respectively, and at 4 weeks, 32% and 7% of iNKT are at stage 3 and stage 1, respectively, in WT mice. In contrast, 20% of thymic iNKT cells remained in stage 1 at 4 weeks in OPN-deficient mice. The number of stage 3 iNKT cells was significantly reduced at 3 and 4 weeks in OPN-deficient mice. It is known that stage 1 thymic iNKT cells cannot exit the thymus (20). Thus, our data indicate that the reduction of peripheral iNKT cells in OPN-deficient mice can be partly explained by deficient intrathymic iNKT cell maturation. After exiting the thymus, iNKT cells undergo further maturation by up-regulating CD69 expression (20). Therefore, we compared the expression of CD69 on peripheral iNKT cells between WT and OPN-deficient mice. We found no significant difference in terms of CD69 expression (Fig. S3), indicating that the iNKT cell maturational step involving CD69 up-regulation does not depend on the presence of OPN. One may argue that reduction of peripheral iNKT cells can be explained by an increase in apoptosis of peripheral iNKT cells in OPN-deficient mice. Thus, we compared apoptosis of iNKT cells in peripheral organs and found that the number of apoptotic iNKT cells was not significantly different between WT and OPN-deficient mice (Fig. S4).

Fig. 1.

Deficient intrathymic maturation of iNKT cells in OPN-deficient mice. (A) RT-PCR analyses of mRNA expression of OPN in iNKT cells, isolated from liver of WT and OPN^−/− adult mice. Data are representative of three independent experiments. (B) Percentage of iNKT cells in various organs of WT and OPN^−/− mice at the indicated time points after birth. Data are presented as mean ± SEM, n = 5 per group. (C) iNKT cells from B were further fractionated by CD44-FITC and NK1.1-PerCP-Cy5.5. (Upper) Data are representative of three independent experiments. (Lower) Data represent mean ± SEM. n = 4 per group. *, P < 0.05; **, P < 0.005.

Impairment of Activation-Induced T Cell Receptor (TCR) Down-Regulation on Peripheral iNKT Cells in OPN-Deficient Mice.

We next investigated whether OPN is involved in iNKT cell function. Stimulation of iNKT cells with anti-CD3 or specific ligand, α-galactosylceramide (α-GalCer), results in the rapid disappearance of these cells in vivo (21). In addition, iNKT cell activation by Con A leads to a rapid reduction in iNKT cell numbers, caused by profound down-regulation of iNKT cell receptors (18, 22). Therefore, we used α-GalCer-loaded CD1d-dimer or NK1.1 and TCRβ to trace the response of iNKT cells to Con A and examined whether down-regulation of the iNKT cell receptor is impaired in OPN-deficient mice. As shown in Fig. 2A, injection of Con A led to a significant reduction of percentage of iNKT cells in the liver and spleen of WT mice at 2 h. In contrast, such a reduction of percentage of splenic iNKT cells was not observed in OPN-deficient mice. A significant reduction of percentage of hepatic iNKT cells was observed in OPN-deficient mice, but this was much milder compared with WT mice. To characterize the down-regulation of iNKT cells in the presence or absence of OPN further, we monitored the actual number of iNKT cells at various time points after Con A injection. After an initial reduction of iNKT cell number at 24 h, there was a significant increase in iNKT cell number at 48 and 72 h in spleen and liver of WT mice (Fig. 2B). In contrast, although iNKT cell number increased at 48 and 72 h in spleen and liver of OPN-deficient mice, this increase was substantially lower than in WT mice. It should be noted that at 24 h, iNKT cell numbers in OPN-deficient mice were much higher than in WT mice. This finding is explained by the lack of significant receptor down-regulation in OPN-deficient mice. Thus, our data demonstrated that iNKT cell function, as judged by TCR down-regulation, is impaired in OPN-deficient mice. These data led us to hypothesize that other iNKT cell functions might be impaired in the absence of OPN as well.

Fig. 2.

Deficient TCR-mediated signaling in OPN-deficient mice. (A) OPN^−/− and WT mice were injected with Con A (10 mg/kg) or PBS, and percentages of NKT cells were analyzed at 2 h after injection. (Left) Cells from liver. (Right) Cells from spleen. Data represent mean ± SEM, n = 5 per group. (B) OPN^−/− and WT mice were injected with Con A (10 mg/kg) or PBS, and iNKT cell numbers in liver and spleen were analyzed at the indicated time points. Data represent mean ± SEM, n = 5 per group. (C) Serum cytokine levels were measured after α-GalCer injection at the indicated time points. Data represent mean ± SEM, n = 8 per group. (D) Production of cytokines by iNKT cells from WT or OPN^−/− mice. iNKT cells were stimulated for 3 days in vitro with α-GalCer. Cytokine production in the culture supernatant was measured by ELISA. Data represent mean ± SEM of three independent experiments. *, P < 0.05; **, P < 0.005; NS, not significant.

Impaired Secretion of IL-4 by OPN-Deficient iNKT Cells.

One unique feature of iNKT cells is the rapid secretion of cytokines, including IFN-γ and IL-4, upon activation. Cytokine secretion by iNKT cells plays a critical role in the pathogenesis of various diseases (23, 24). iNKT cells produce high levels of cytokines, especially IFN-γ and IL-4, in response to α-GalCer both in vivo and in vitro (1). To determine whether deficiency of OPN also affects cytokine secretion by iNKT cells in vivo, we injected i.v. α-GalCer into OPN-deficient and WT mice. IL-4 levels reached a peak in the serum of WT mice at 2 h after injection (Fig. 2C). In OPN-deficient mice, IL-4 levels were significantly lower than those in WT mice. In contrast, there was no significant difference in IFN-γ levels between WT and OPN-deficient mice. Nevertheless, neither IL-4 nor IFN-γ was detected in serum of α-GalCer-treated CD1d-deficient (thus, iNKT cell-deficient) mice, indicating that those cytokines are derived from iNKT cells. However, it is possible that other cells were activated by iNKT cell-derived cytokines, leading to the production of IL-4 or IFN-γ. Thus, purified iNKT cells were stimulated in vitro with α-GalCer. Consistent with our in vivo work, significantly lower amounts of IL-4 were secreted by iNKT cells from OPN-deficient mice than WT mice, but similar levels of IFN-γ were secreted (Fig. 2D). Thus, it appears that the lack of OPN results in reduced IL-4 secretion by iNKT cells. To confirm this finding further, purified iNKT cells were stimulated in vitro with α-GalCer in the presence or absence of exogenous recombinant OPN. We found that treatment of iNKT cells of OPN-deficient mice with exogenous OPN resulted in significant augmentation of IL-4, but not IFN-γ production (Fig. S5). Nevertheless, iNKT cells derived from WT mice secreted similar amounts of IL-4 in the absence or presence of exogenous OPN. We then tested whether the lack of OPN simply affected secretion or also production of IL-4. Therefore, purified iNKT cells were stimulated in vitro with α-GalCer, and cytokine production was examined by real-time PCR. Results showed impaired production of IL-4 but not IFN-γ in OPN-deficient mice (Fig. S6).

Reduced Cytotoxicity of OPN-Deficient iNKT Cells Against Hepatocytes.

iNKT cells exhibit cytotoxicity against hepatocytes in a Fas/FasL-dependent manner in a Con A-induced hepatic injury model (18, 25, 26). In addition, we demonstrated that hepatic injury in this model is significantly reduced in OPN-deficient mice (17). We reasoned that reduced hepatic injury after Con A injection in OPN-deficient mice is caused by reduced OPN-mediated neutrophil infiltration into the liver and thus reduced neutrophil-mediated hepatic injury (17). However, it is also possible that Fas/FasL-dependent iNKT cell cytotoxicity is impaired in OPN-deficient mice, and thus, we assessed the cytotoxic activity of iNKT cells against a murine hepatocyte cell line (TLR2). iNKT cells derived from WT mice exhibited significant cytotoxicity (Fig. 3A). There was a statistically significant reduction in the cytotoxicity of iNKT cells from OPN-deficient mice. We then tested whether the iNKT cell cytotoxicity against TLR2 cells was Fas/FasL-dependent. TLR2 cells expressed high levels of Fas (Fig. 3B). The knockdown of Fas expression on TLR2 cells by small interference RNA (Fas siRNA) significantly reduced cytotoxicity of WT iNKT cells against TLR2 cells (Fig. 3C). Importantly, iNKT cells derived from OPN-deficient mice expressed significantly lower amounts of FasL mRNA compared with those from WT mice (Fig. 3D). Collectively, these data suggested that reduced cytotoxicity of iNKT cells from OPN-deficient mice against hepatocytes is caused by reduced expression of FasL, although we could not eliminate the possible involvement of other factors. To address directly whether OPN is critical for FasL expression in iNKT cells, we used iNKT cell hybridoma 1B6, which has been described in ref. 27. The expression of OPN by 1B6 cells was silenced by OPN-specific siRNA. FasL protein expression on 1B6 cells was significantly reduced by OPN-specific siRNA (Fig. S7). It is known that IL-4 regulates FasL expression and augments the cytotoxicity of iNKT cells (28). In the present study, we found that activated iNKT cells derived from OPN-deficient mice secrete reduced amounts of IL-4. Therefore, we reasoned that reduced expression of FasL on iNKT cells is caused by reduced secretion of IL-4 by iNKT cells in OPN-deficient mice. To test this hypothesis, we treated iNKT cells from WT or OPN-deficient mice with exogenous IL-4 and studied FasL mRNA expression. Expression of FasL mRNA by iNKT cells of OPN-deficient mice, but not WT mice, was significantly increased upon stimulation with IL-4 (Fig. 3E).

Fig. 3.

Reduced cytotoxic ability of iNKT cells in OPN-deficient mice. (A) iNKT cells of WT or OPN^−/− mice were stimulated for 3 days in vitro as described in Materials and Methods. The cells were subjected to a standard cytotoxic assay with target TLR2 cells. Data represent mean ± SEM of three independent experiments. (B) Fas expression of TLR2 was analyzed by flow cytometry. Data are representative of three independent experiments. (C) iNKT cells of WT mice were cultured as described above. The target cells were treated with Fas-siRNA or control siRNA. A standard cytotoxic assay against TLR2 cells was performed. Cont siRNA, control siRNA. Data represent mean ± SEM of three independent experiments. (D) FasL expression of iNKT cells from WT or OPN^−/− mice. FasL mRNA expression was determined by quantitative real-time PCR. The relative value of gene expression was normalized against the gene expression levels of G3PDH. Data represent mean ± SEM of three independent experiments. (E) Up-regulation of FasL mRNA expression on iNKT cells by exogenous IL-4. iNKT cells of OPN^−/− mice or WT mice were cultured with or without exogenous recombinant IL-4 (200 ng/ml) for 5 h. FasL expression by iNKT cells was determined by quantitative real-time PCR. The relative value of gene expression was normalized against the gene expression levels of G3PDH. Data represent mean ± SEM of three independent experiments. *, P < 0.05; **, P < 0.005; NS, not significant.

Defective Activation of NFAT2 in iNKT Cells of OPN-Deficient Mice.

Our work demonstrated that signaling through the iNKT cell receptor was impaired in OPN-deficient mice, as judged by impaired IL-4 secretion and defective receptor down-regulation. It is known that the transcription factor NFAT2 plays a critical role in IL-4 production by iNKT cells (12). Therefore, we examined whether activation of NFAT2 is defective in α-GalCer-activated iNKT cells from OPN-deficient mice. NFAT2 protein was clearly up-regulated in nuclear extracts of spleen obtained from WT mice at 30 min after α-GalCer injection (Fig. 4A), indicating that iNKT cell stimulation in vivo leads to activation of NFAT2. However, in OPN-deficient mice, NFAT2 protein was not increased, demonstrating that NFAT2 activation is defective in the absence of OPN. Consistent with our Western blot analyses, NFAT2 gene expression in α-GalCer-treated WT mice was 2-fold higher than in α-GalCer-treated OPN-deficient mice (Fig. 4B). As expected, augmentation of NFAT2 gene expression was not detected in OPN-deficient mice. In conventional T cells, GATA-3 is also involved in Th2 cytokine expression (9), whereas T-bet is involved in Th1 cytokine expression (7). We examined GATA-3 and T-bet gene expression before and after α-GalCer injection and found that GATA-3 expression was also impaired in OPN-deficient mice, whereas T-bet expression was preserved in these animals (Fig. 4C). It is interesting to note that T-bet expression is essential for OPN expression (14), whereas OPN expression is not required for T-bet expression. To confirm further that OPN is involved in NFAT activation in iNKT cells, iNKT cell hybridoma 1B6 cells were stimulated in vitro with α-GalCer. The expression of OPN was silenced by OPN-specific siRNA. The intranuclear NFAT2 protein expression was clearly increased in the 1B6 cells after stimulation, whereas silencing of OPN gene expression by OPN-specific siRNA led to suppression of NFAT2 transcription, thus reduced expression of intranuclear NFAT2 by 1B6 cells (Fig. 4D). We also found that NFAT2 mRNA expression in the 1B6 cells was augmented upon stimulation by α-GalCer and that the augmentation of NFAT2 mRNA expression was abrogated by OPN gene silencing (Fig. S8). Importantly, production of IL-4 by activated 1B6 cells was significantly reduced by OPN gene silencing, whereas production of IFN-γ remained unaltered (Fig. 4D).

Fig. 4.

Deficient activation of NFAT2 by iNKT cells in OPN-deficient mice. (A) Western blot analysis of splenic nuclear extracts. Nuclear extracts were prepared from WT and OPN-deficient mice 30 min after α-GalCer injection. Data are representative of three independent experiments. (B and C) Quantitative real-time PCR analyses of gene expression of NFAT2, GATA-3, and T-bet. The relative value of gene expression was normalized against the gene expression levels of G3PDH. Data represent mean ± SEM of three independent experiments. (D) NFAT2 expression by 1B6 hybridoma cells. 1B6 hybridoma cells were treated with OPN-siRNA and stimulated with α-GalCer and irradiated CD1d-transfectants. IL-4 and IFN-γ production in the culture supernatant was measured by ELISA. (Upper) Data are representative of three independent experiments. (Lower) Data represent mean ± SEM of three independent experiments. *, P < 0.05; **, P < 0.005; NS, not significant.

In conclusion, our proposed model for the role of OPN in iNKT cells is illustrated in Fig. 5. The absence of OPN results in impaired iNKT cell development and defective peripheral iNKT cell activation, as evidenced by deficient receptor down-regulation and IL-4 production. iNKT cell antigen receptor-mediated signaling pathways, such as activation of NFAT2, was also impaired. Reduced IL-4 production caused by defective activation of NFAT2 led to reduced FasL expression on iNKT cells, which resulted in attenuated cytotoxic activity of iNKT cells against Fas-positive hepatocytes. Collectively, our findings have revealed a critical contribution of OPN in the development and function of iNKT cells, which might be exploited for treatment of diseases that are controlled by OPN or iNKT cells.

Fig. 5.

Schematic representation of the link between OPN, NFAT activation, IL-4 production, and FasL expression in iNKT cells.

Materials and Methods

Animals.

Specific pathogen-free female C57BL/6 (B6) mice were purchased from Japan SLC. OPN-deficient (OPN^−/−) mice (29), backcrossed 11 times to B6 mice at the Institute for Genetic Medicine, Hokkaido University, were used. All mice were maintained under specific pathogen-free conditions and used according to institutional guidelines.

Flow Cytometric Analysis.

Hepatic lymphocytes were isolated as described in ref. 17. Liver, spleen, and thymus cells were first incubated with anti-FcγR (2.4G2; BD Biosciences). Cells were stained with a primary antibody (α-GalCer-loaded mouse CD1d-IgG1 fusion protein DimerX) for 60 min at 4°C, followed by a secondary anti-mouse IgG1-phycoerythrin (PE) and anti-TCRβ-antigen-presenting cell (H57–597) antibody for 30 min at 4°C. iNKT cells were defined as α-GalCer-loaded CD1d dimer⁺ TCRβ⁺ cells. Apoptosis was assayed by annexinV-FITC apoptosis detection kit I. Other antibodies used for staining were anti-NK1.1-PerCP-Cy5.5 (PK136), anti-FasL-PE (MFL3), anti-Fas-PE (Jo2), anti-CD69-FITC (H1.2F3), and anti-CD44-FITC (IM7) (all from BD Biosciences). All analyses were performed on a FACSCalibur (BD Biosciences) with CellQuest software.

In Vitro and in Vivo iNKT Cell Stimulation.

iNKT cells (gated as α-GalCer-loaded CD1d-dimer⁺ TCRβ⁺) were isolated by a FACS Vantage instrument (Becton Dickinson). The sorted iNKT cells were cultured overnight with recombinant IL-2. NKT cells were further stimulated with 100 ng/ml α-GalCer (Pharmaceutical Research Laboratories, Kirin Brewery) in the presence of 100-Gy-irradiated CD11c⁺ DCs. In some experiments, mice were injected i.v. with 100 μg/kg α-GalCer or vehicle. CD1d-transfected cells were provided by Albert Bendelac (University of Chicago, Chicago, IL).

Analysis of mRNA Expression.

Total RNA was isolated by using TRIzol (Invitrogen). The specific primers used were as follows: glyceraldehyde-3-phosphate dehydrogenase (G3PDH), 5′-ACCACAGTCCATGCCATCAC-3′ (sense), 5′-TCCACCACCCCTGTTGCTGTA-3′ (antisense); Fas, 5′-GAGAATTGCTGAAGACATGACAATCC-3′ (sense), 5′-GTAGTTTTCACTCCAGACATTGTCC-3′ (antisense); FasL, 5′-ATCCCTCTGGAATGGGAAGA-3′ (sense) and 5′-CCATATCTGTCCAGTAGTGC-3′ (antisense); GATA-3, 5′-AGAACCGGCCCCTTATCAA-3′ (sense) and 5′-AGTTCGCGCAGGATGTCC-3′ (antisense); T-bet, 5′-CAACAACCCCTTTGCCAAAG-3′ (sense) and 5′-TCCCCCAAGCAGTTGACAGT-3′ (antisense). Quantitative real-time PCR analysis of mRNA expression was carried out by using LightCycler Fast Start DNA Master SYBR Green I systems (Roche Diagnostics). The expression of mRNA was calculated by LightCycler software version 3. Data were standardized by G3PDH.

Protein Extraction and Western Blot Analyses.

Transcription factors were analyzed by extracting nuclear protein with reagent NE-PER (nuclear and cytoplasmic extraction reagents; Pierce) according to the manufacturer's instructions. After calibration of protein content, 15 μg of each of protein extract was electrophoresed through 10–20% polyacrylamide Tris·HCl Ready Gels (Bio-Rad) and probed with polyclonal anti-NFAT2 (7A6) as described in ref. 17.

siRNA Design.

siRNA against Fas was procured from B-Bridge, Inc.. The sequences of the selected region to be targeted by siRNA for Fas were: 5′-GUGCAAGUGCAAACCAGACdTdT-3′ (sense) and 5′-GUCUGGUUUGCACUUGCACdTdT-3′ (antisense); mOPN siRNA, 5′-GCCAUGACCACAUGGACGAdTdT-3′ (sense) and 5′-dTdTCGGUACUGGUGUACCUGCU-3′ (antisense). Irrelevant control siRNA (nonspecific control VIII) was purchased from Fisher Scientific.

⁵¹Cr-Release Cytotoxic Assay.

The cytotoxic activity of iNKT cells was assayed by a 4-h ⁵¹Cr-release cytotoxic assay with the TLR2 hepatocyte cell line (RIKEN Cell Bank) as described in ref. 30. The cells were harvested and seeded at the indicated E/T ratios.

Statistics.

Data are presented as means ± SEM and are representative of at least two independent experiments. The significance of differences between two groups was determined by using Student's t test.