IL-18 is a potent coinducer of IL-13 in NK and T cells: a new potential role for IL-18 in modulating the immune response

IL-13 and IL-4 have similar biological activities and are characteristic of cytokines expressed by Th2 cells. In contrast, IL-12 and IL-18 have been shown to be strong cofactors for Th1 cell development. In this study, we found strong induction of IL-13 mRNA and protein by IL-2 + IL-18 in NK and T cells. In contrast, IL-12 did not enhance the IL-13 production induced by IL-2 alone. Moreover, IL-13 mRNA and protein expression induced by IL-2 + IL-18 in purified NK and T cells obtained from IFN-gamma knockout (-/-) mice were greater than seen in purified cells from normal controls. In contrast, IL-10 production induced by IL-2 and/or IL-12 was not significantly different in IFN-gamma (-/-) mice and normal controls. These results suggest IL-13 expression induced by IL-2 + IL-18 may be regulated by IFN-gamma in vivo, while IL-10 expression may be IFN-gamma-independent. Thus, depending upon the cell type, IL-18 may act as a strong coinducer of Th1 or Th2 cytokines. Our findings suggest that IL-12 and IL-18 have different roles in the regulation of gene expression in NK and T cells.     

Role of IFN-gamma in Th1 differentiation: IFN-gamma regulates IL-18R alpha expression by preventing the negative effects of IL-4 and by inducing/maintaining IL-12 receptor beta 2 expression

Two key events occur during the differentiation of IFN-gamma-secreting Th1 cells: up-regulation of IL-12Rbeta2 and IL-12-driven up-regulation of IL-18Ralpha. We previously demonstrated that IL-12-driven up-regulation of IL-18Ralpha expression is severely impaired in IFN-gamma(-/-) mice. However, it was unclear from these studies how IFN-gamma influenced IL-18Ralpha since IFN-gamma alone had no direct effect on IL-18Ralpha expression. In the absence of IL-4, IL-12-dependent up-regulation of IL-18Ralpha/IL-12Rbeta2 was independent of IFN-gamma. However, in the presence of IL-4, IFN-gamma functions to limit the negative effects of IL-4 on both IL-18Ralpha and IL-12Rbeta2. Neutralization of IL-4 restored IL-12-driven up-regulation of IL-18Ralpha/IL-12Rbeta2 in an IFN-gamma-independent fashion. In the absence of both IL-12 and IL-4, IFN-gamma up-regulates IL-12beta2 expression and primes IFN-gamma-producing Th1 cells. When T cells were primed in the presence of IL-4, no correlation was found between the levels of expression of the IL-18Ralpha or the IL-12Rbeta2 and the capacity of these cells to produce IFN-gamma, suggesting that IL-4 may also negatively affect IL-12-mediated signal transduction and thus Th1 differentiation. These data clarify the role of IFN-gamma in regulation of IL-18Ralpha/IL-12Rbeta2 during both IL-12-dependent and IL-12-independent Th1 differentiation.     

CD8 T cells inhibit IgE via dendritic cell IL-12 induction that promotes Th1 T cell counter-regulation.

Th1 and Th2 cells are counterinhibitory; their balance determines allergic sensitization. We show here that CD8 T cell subsets break these rules as both T cytotoxic (Tc)1 and Tc2 cells promote Th1 over Th2 immunity. Using IL-12(-/-), IFN-gamma(-/-), and OVA(257-264)-specific Valpha2Vbeta5 TCR-transgenic mice, we have identified the key steps involved. OVA-specific IFN-gamma(-/-) CD8 T cells inhibited IgE responses equivalent to wild-type CD8 T cells (up to 98% suppression), indicating that CD8 T cell-derived IFN-gamma was not required. However, OVA-specific CD8 T cells could not inhibit IgE in IFN-gamma(-/-) recipients unless reconstituted with naive, wild-type CD4 T cells, suggesting that CD4 T cell-derived IFN-gamma did play a role. Transfer of either Tc1 or Tc2 Valpha2Vbeta5 TCR-transgenic CD8 T cells inhibited IgE and OVA-specific Th2 cells while promoting OVA-specific Th1 cell responses, suggesting a potential role for a type 1 inducing cytokine such as IL-12. CD8 T cells were shown to induce IL-12 in OVA(257-264)-pulsed dendritic cells (DC) in vitro. Furthermore, CD8 T cells were unable to inhibit IgE responses in IL-12(-/-) recipients without the addition of naive, wild-type DC, thus demonstrating a pivotal role for IL-12 in this mechanism. These data reveal a mechanism of IgE regulation in which CD8 T cells induce DC IL-12 by an IFN-gamma-independent process that subsequently induces Th1 and inhibits Th2 cells. Th1 cell IFN-gamma is the final step that inhibits B cell IgE class switching. This demonstrates a novel regulatory network through which CD8 T cells inhibit allergic sensitization.     

IL-21 induces tumor rejection by specific CTL and IFN-gamma-dependent CXC chemokines in syngeneic mice

IL-21 is an immune-stimulatory four alpha helix cytokine produced by activated T cells. To study the in vivo antitumor activities of IL-21, TS/A murine mammary adenocarcinoma cells were genetically modified to secrete IL-21 (TS/A-IL-21). These cells developed small tumors that were subsequently rejected by 90% of s.c. injected syngeneic mice. Five days after injection, TS/A-IL-21 tumors showed numerous infiltrating granulocytes, NK cells, and to a lesser extent CD8(+) T cells, along with the expression of TNF-alpha, IFN-gamma, and endothelial adhesion molecules ICAM-1 and VCAM-1. At day 7, CD8(+) and CD4(+) T cells increased together with IFN-gamma, and the CXC chemokines IFN-gamma-inducible protein 10, monokine induced by IFN-gamma, and IFN-inducible T cell alpha-chemoattractant. The TS/A-IL-21 tumor displayed a disrupted vascular network with abortive sprouting and signs of endothelial cell damage. In vivo depletion experiments by specific Abs showed that rejection of TS/A-IL-21 cells required CD8(+) T lymphocytes and granulocytes. When injected in IFN-gamma-deficient mice, TS/A-IL-21 cells formed tumors that regressed in only 29% of animals, indicating a role for IFN-gamma in IL-21-mediated antitumor response, but also the existence of IFN-gamma-independent effects. Most immunocompetent mice rejecting TS/A-IL-21 cells developed protective immunity against TS/A-pc (75%) and against the antigenically related C26 colon carcinoma cells (61%), as indicated by rechallenge experiments. A specific CTL response against the gp70-env protein of an endogenous murine retrovirus coexpressed by TS/A and C26 cells was detected in mice rejecting TS/A-IL-21 cells. These data suggest that IL-21 represents a suitable adjuvant in inducing specific CTL responses.     

Systemic administration of IL-23 induces potent antitumor immunity primarily mediated through Th1-type response in association with the endogenously expressed IL-12

IL-23, a cytokine, which is composed of the p40 subunit shared with IL-12 and the IL-23-specific p19 subunit, has been shown to preferentially act on Th1 effector/memory CD4+ T cells and to induce their proliferation and IFN-gamma production. The IL-23 is also reported to act on Th17-CD4+ T cells, which are involved in inducing tissue injury. In this study, we examined the antitumor effects associated with systemic administration of IL-23 and their mechanisms in mouse tumor system. Systemic administration of high-dose IL-23 was achieved using in vivo electroporation of IL-23 plasmid DNA into the pretibial muscles of C57BL/6 mice. The IL-23 treatment was associated with significant suppression of the growth of pre-existing MCA205 fibrosarcoma and prolongation of the survival of treated mice without significant toxicity when compared with those of the mice treated with EGFP. Although the therapeutic outcomes were similar to those with the IL-12 treatment, the IL-23 treatment induced characteristic immune responses distinctive to those of IL-12 treatment. The IL-23 administration even at the therapeutic levels did not induce detectable IFN-gamma concentration in the serum. In vivo depletion of CD4+ T cells, CD8+ T cells, or NK cells significantly inhibited the antitumor effects of IL-23. Furthermore, the CD4+ T cells in the lymph nodes in the IL-23-treated mice showed significant IFN-gamma and IL-17 response upon anti-CD3 mAb stimulation in vitro. These results and the ones in the IFN-gamma or IL-12 gene knockout mice suggest that potent antitumor effects of IL-23 treatment could be achieved when the Th1-type response is fully promoted in the presence of endogenously expressed IL-12.     

IL-12 administration accelerates autoimmune diabetes in both wild-type and IFN-gamma-deficient nonobese diabetic mice,revealing pathogenic and protective effects of IL-12-induced IFN-gamma

IL-12 administration to nonobese diabetic (NOD) mice induces IFN-gamma-secreting type 1 T cells and high circulating IFN-gamma levels and accelerates insulin-dependent diabetes mellitus (IDDM). Here we show that IL-12-induced IFN-gamma production is dispensable for diabetes acceleration, because exogenous IL-12 could enhance IDDM development in IFN-gamma-deficient as well as in IFN-gamma-sufficient NOD mice. Both in IFN-gamma(+/-) and IFN-gamma(-/-) NOD mice, IL-12 administration generates a massive and destructive insulitis characterized by T cells, macrophages, and CD11c(+) dendritic cells, and increases the number of pancreatic CD4(+) cells secreting IL-2 and TNF-alpha. Surprisingly, IL-12-induced IFN-gamma hinders pancreatic B cell infiltration and inhibits the capacity of APCs to activate T cells. Although pancreatic CD4(+) T cells from IL-12-treated IFN-gamma(-/-) mice fail to up-regulate the P-selectin ligand, suggesting that their entry into the pancreas may be impaired, T cell expansion is favored in these mice compared with IL-12-treated IFN-gamma(+/-) mice because IL-12 administration in the absence of IFN-gamma leads to enhanced cell proliferation and reduced T cell apoptosis. NO, an effector molecule in beta cell destruction, is produced ex vivo in high quantity by pancreas-infiltrating cells through a mechanism involving IL-12-induced IFN-gamma. Conversely, in IL-12-treated IFN-gamma-deficient mice, other pathways of beta cell death appear to be increased, as indicated by the up-regulated expression of Fas ligand on Th1 cells in the absence of IFN-gamma. These data demonstrate that IFN-gamma has a dual role, pathogenic and protective, in IDDM development, and its deletion allows IL-12 to establish alternative pathways leading to diabetes acceleration.     

Impaired accumulation and function of memory CD4 T cells in human IL-12 receptor beta 1 deficiency

Defects in IL-12 production or IL-12 responsiveness result in a vulnerability to infection with non-viral intracellular organisms, but the immunological mechanisms responsible for this susceptibility remain poorly understood. We present an immunological analysis of a patient with disseminated Salmonella enteritidis and a homozygous splice acceptor mutation in the IL-12Rbeta1-chain gene. This mutation resulted in the absence of IL-12Rbeta1 protein on PBMC and an inability of T cells to specifically bind IL-12 or produce IFN-gamma in response to either IL-12 or IL-23. The accumulation of memory (CD45R0(high)) CD4 T cells that were CCR7(high) (putative central memory cells) was normal or increased for age. Central memory CD4 T cells of the patient and age-matched controls were similar in having a low to undetectable capacity to produce IFN-gamma after polyclonal stimulation. In contrast, the patient had a substantial decrease in the number of CCR7(neg/dull) CD45R0(high) memory CD4 T cells (putative effector memory cells), and these differed from control cells in having a minimal ability to produce IFN-gamma after polyclonal stimulation. Importantly, tetanus toxoid-specific IFN-gamma production by PBMC from the patient was also significantly reduced compared with that in age-matched controls, indicating that signaling via the IL-12Rbeta1-chain is generally necessary for the in vivo accumulation of human memory CD4 T cells with Th1 function. These results are also consistent with a model in which the IL-12Rbeta1 subunit is necessary for the conversion of central memory CD4 T cells into effector memory cells.     

Human CD4+ T cells present within the microenvironment of human lung tumors are mobilized by the local and sustained release of IL-12 to kill tumors in situ by indirect effects of IFN-gamma

By implanting nondisrupted pieces of human lung tumor biopsy tissues into SCID mice, it has been possible to establish viable grafts of the tumor, as well as the tumor-associated microenvironment, including inflammatory cells, fibroblasts, tumor vasculature, and the extracellular matrix. Using this xenograft model, we have evaluated and characterized the effects of a local and sustained release of human rIL-12 (rhIL-12) from biodegradable microspheres. In response to rhIL-12, the human CD45+ inflammatory cells present within the xenograft mediate the suppression or the complete arrest of tumor growth in SCID mice. Analysis of the cellular events reveals that human CD4+ and CD8+ T cells are induced by rhIL-12 to produce and secrete IFN-gamma. Serum levels of human IFN-gamma in mice bearing rhIL-12-treated tumor xenografts correlate directly with the degree of tumor suppression, while neutralizing Abs to human IFN-gamma abrogate the IL-12-mediated tumor suppression. Gene expression profiling of tumors responding to intratumoral rhIL-12 demonstrates an up-regulation of IFN-gamma and IFN-gamma-dependent genes not observed in control-treated tumors. Genes encoding a number of proinflammatory cytokines, chemokines (and their receptors), adhesion molecules, activation markers, and the inducible NO synthase are up-regulated following the introduction of rhIL-12, while genes associated with tumor growth, angiogenesis, and metastasis are decreased in expression. NO contributes to the tumor killing because an inhibitor of inducible NO synthase prevents IL-12-induced tumor suppression. Cell depletion studies reveal that the IL-12-induced tumor suppression, IFN-gamma production, and the associated changes in gene expression are all dependent upon CD4+ T cells.     

Therapeutic protective effects of IL-12 combined with soluble IL-4 receptor against established infections of herpes simplex virus type 1 in thermally injured mice.

The effect of combination therapy between IL-12 and soluble IL-4R (sIL-4R) on the established infection of HSV-1 in thermally injured mice (TI mice) was investigated. All of the TI mice infected with lethal amounts of HSV-1 died when IL-12 was given therapeutically at a dose of 500 U/mouse. However, 80% of these mice treated prophylactically with IL-12 survived compared with 0% survival of the same mice treated with saline. The therapeutic administration of IL-12 to TI mice currently infected with HSV-1 caused an 80% survival of these mice when the treatment was combined with sIL-4R. Although IL-12 did not stimulate IFN-gamma production in cultures of splenic T cells from TI mice, IFN-gamma was produced by stimulation with IL-12 when the producer cells were prepared from TI mice that had been treated previously with sIL-4R. After stimulation with anti-CD3 mAb, splenic T cells from TI mice with the established infection of HSV-1 produced IL-4 into their culture fluids. However, IL-4 was not produced by splenic T cells that were prepared from the same infected mice treated with IL-12 and sIL-4R in combination. The results obtained herein indicate that the efficacies of the combination therapy against the established infection of HSV-1 may result from the IFN-gamma production stimulated by IL-12 in TI mice that are treated with sIL-4R for reducing burn-associated type 2 T cell responses.     

IL-28 elicits antitumor responses against murine fibrosarcoma

IL-28 is a recently described antiviral cytokine. In this study, we investigated the biological effects of IL-28 on tumor growth to evaluate its antitumor activity. IL-28 or retroviral transduction of the IL-28 gene into MCA205 cells did not affect in vitro growth, whereas in vivo growth of MCA205IL-28 was markedly suppressed along with survival advantages when compared with that of controls. When the metastatic ability of IL-28-secreting MCA205 cells was compared with that of controls, the expression of IL-28 resulted in a potent inhibition of metastases formation in the lungs. IL-28-mediated suppression of tumor growth was mostly abolished in irradiated mice, indicating that irradiation-sensitive cells, presumably immune cells, are primarily involved in the IL-28-induced suppression of tumor growth. In vivo cell depletion experiments displayed that polymorphonuclear neutrophils, NK cells, and CD8 T cells, but not CD4 T cells, play an equal role in the IL-28-mediated inhibition of in vivo tumor growth. Consistent with these findings, inoculation of MCA205IL-28 into mice evoked enhanced IFN-gamma production and cytotoxic T cell activity in spleen cells. Antitumor action of IL-28 is partially dependent on IFN-gamma and is independent of IL-12, IL-17, and IL-23. IL-28 increased the total number of splenic NK cells in SCID mice and enhanced IL-12-induced IFN-gamma production in vivo and expanded spleen cells in C57BL/6 mice. Moreover, IL-12 augmented IL-28-mediated antitumor activity in the presence or absence of IFN-gamma. These findings indicate that IL-28 has bioactivities that induce innate and adaptive immune responses against tumors.     

IL-2 receptor blockade inhibits late,but not early,IFN-gamma and CD40 ligand expression in human T cells: disruption of both IL-12-dependent and -independent pathways of IFN-gamma production

mAbs directed against the alpha-chain (Tac/CD25) of the IL-2R are an emerging therapy in both transplantation and autoimmune disease. However, the mechanisms underlying their therapeutic efficacy have not been fully elucidated. Therefore, we examined the effect of IL-2R blockade on Th1 and Th2 cytokine production from human PBMC. Addition of a humanized anti-Tac Ab (HAT) to activated PBMC cultures inhibited IFN-gamma production from CD4 and CD8 T cells by 80-90%. HAT partially inhibited production of TNF-alpha and completely inhibited production of IL-4, IL-5, and IL-10. Furthermore, IL-12, a central regulatory cytokine that induces IFN-gamma, was undetectable in treated cultures. As T cell-dependent induction of IL-12 is regulated via CD40/CD40 ligand (CD40L) interactions, we examined the effect of HAT on CD40L expression. We found CD40L expression to be biphasic with an early (6 h) peak that is CD28/IL-2-independent, but a later peak (48 h) being CD28/IL-2-dependent and inhibited by HAT. Similarly, IFN-gamma production at 6 h was CD28/IL-2-independent but CD28/IL-2-dependent and inhibited by HAT at 48 h. Nonetheless, addition of rCD40L or exogenous IL-12 to HAT-treated cultures could not restore IFN-gamma production. The IFN-gamma deficit in such cultures appears to be due to a direct inhibition by HAT of IL-12-independent IFN-gamma production from T cells rather than altered expression of either the IL-12Rbeta1 or IL-12Rbeta2 chains. These data demonstrate that IL-2 plays a critical role in the regulation of Th1 and Th2 responses and impacts both IL-12-dependent and -independent IFN-gamma production.     

Stat4 regulates multiple components of IFN-gamma-inducing signaling pathways

Stat4 is activated in response to IL-12. Most functions of IL-12, including the induction of IFN-gamma, are compromised in the absence of Stat4. Since the precise role of Stat4 in IFN-gamma induction has not been established, experiments were conducted to examine Stat4 activation of IFN-gamma and other genes required for cytokine-induced expression of IFN-gamma. We first examined IL-12 signaling components. Basal expression of IL-12Rss1 and IL-12Rss2 is decreased in Stat4-deficient cells compared with that in control cells. However, IL-12 was still capable of inducing equivalent phosphorylation of Jak2 and Tyk2 in wild-type and Stat4-deficient activated T cells. We have further determined that other cytokine signaling pathways that induce IFN-gamma production are defective in the absence of Stat4. IL-18 induces minimal IFN-gamma production from Stat4-deficient activated T cells compared with control cells. This is due to defective IL-18 signaling, which results from the lack of IL-12-induced, and Stat4-dependent, expression of the IL-18R. Following IL-12 pretreatment to induce IL-18R, wild-type, but not Stat4-deficient, activated T cells demonstrated IL-18-induced NF-kappaB DNA-binding activity. In addition, IL-12-pretreated Stat4-deficient activated T cells have minimal IFN-gamma production followed by stimulation with IL-18 alone or in combination with IL-12 compared with control cells. Thus, Stat4 activation by IL-12 is required for the function of multiple cytokine pathways that result in induction of IFN-gamma.     

Modulation of CLA, IL-12R, CD40L, and IL-2Ralpha expression and inhibition of IL-12- and IL-23-induced cytokine secretion by CNTO 1275

Cytokines interleukin (IL)-12 and IL-23 are implicated in the pathogenesis of psoriasis. IL-12 causes differentiation of CD4+ T cells to interferon-gamma (IFN-gamma)-producing T helper 1 (Th1) cells, while IL-23 induces differentiation to IL-17-producing pathogenic Th17 cells. The effects of the monoclonal antibody to IL-12/23 p40 subunit (CNTO 1275) on IL-12 receptor (IL-12R) expression, markers associated with skin homing, activation, and cytokine secretion were investigated in vitro using human peripheral blood mononuclear cells (PBMCs) from healthy donors. PBMCs were activated in the presence or absence of recombinant human (rh) IL-12 or rhIL-23, with or without CNTO 1275. CNTO 1275 inhibited upregulation of CLA, IL-12R, IL-2Ralpha and CD40L expression and also inhibited IL-12- and IL-23-induced IFN-gamma, IL-17A, tumor necrosis factor (TNF)-alpha, IL-2, and IL-10 secretion. Thus, the therapeutic effect of CNTO 1275 may be attributed to the IL-12/23 neutralization, resulting in decreased expression of skin homing and activation markers, and IL-12- and IL-23-induced cytokine secretion.     

Synergistic induction of antigen-specific CTL by fusions of TLR-stimulated dendritic cells and heat-stressed tumor cells

Dendritic cell (DC)/tumor cell fusion cells (FCs) can induce potent CTL responses. The therapeutic efficacy of a vaccine requires the improved immunogenicity of both DCs and tumor cells. The DCs stimulated with the TLR agonist penicillin-killed Streptococcus pyogenes (OK-432; OK-DCs) showed higher expression levels of MHC class I and II, CD80, CD86, CD83, IL-12, and heat shock proteins (HSPs) than did immature DCs. Moreover, heat-treated autologous tumor cells displayed a characteristic phenotype with increased expression of HSPs, carcinoembryonic Ag (CEA), MUC1, and MHC class I (HLA-A2 and/or A24). In this study, we have created four types of FC preparation by alternating fusion cell partners: 1) immature DCs fused with unheated tumor cells; 2) immature DCs fused with heat-treated tumor cells; 3) OK-DCs fused with unheated tumor cells; and 4) OK-DCs fused with heat-treated tumor cells. Although OK-DCs fused with unheated tumor cells efficiently enhanced CTL induction, OK-DCs fused with heat-treated tumor cells were most active, as demonstrated by: 1) up-regulation of multiple HSPs, MHC class I and II, CEA, CD80, CD86, CD83, and IL-12; 2) activation of CD4+ and CD8+ T cells able to produce IFN- gamma at higher levels; 3) efficient induction of CTL activity specific for CEA or MUC1 or both against autologous tumor; and 4) superior abilities to induce CD107+ IFN-gamma+ CD8+ T cells and CD154+ IFN-gamma+ CD4+ T cells. These results strongly suggest that synergism between OK-DCs and heat-treated tumor cells enhances the immunogenicity of FCs and provides a promising means of inducing therapeutic antitumor immunity.     

CD4+ T cells mediate IFN-gamma-independent control of Mycobacterium tuberculosis infection both in vitro and in vivo

Although IFN-gamma is necessary for survival of Mycobacterium tuberculosis infection in people and animal models, it may not be sufficient to clear the infection, and IFN-gamma is not a reliable correlate of protection. To determine whether IFN-gamma-independent mechanisms of immunity exist, we developed a murine ex vivo culture system that directly evaluates the ability of splenic or lung lymphocytes to control the growth of M. tuberculosis within infected macrophages, and that models in vivo immunity to tuberculosis. Surprisingly, CD4(+) T cells controlled >90% of intracellular M. tuberculosis growth in the complete absence of IFN-gamma stimulation of macrophages, via a NO-dependent mechanism. Furthermore, bacillus Calmette-Guerin-vaccinated IFN-gamma-deficient mice exhibited significant protection against M. tuberculosis challenge that was lost upon depletion of CD4(+) T cells. These findings demonstrate that CD4(+) T cells possess IFN-gamma-independent mechanisms that can limit the growth of an intracellular pathogen and are dominant in secondary responses to M. tuberculosis.     

Synergistic effects of IL-4 and IL-18 on IL-12-dependent IFN-gamma production by dendritic cells

Mouse splenic dendritic cells (DCs) produce IFN-gamma in response to IL-12. In the present study, we analyzed effects of Th1 and Th2 cytokines on IFN-gamma production by DCs. IL-18 produced by DCs and macrophages acts in an autocrine manner and augments IL-12-induced IFN-gamma production by DCs as also observed in T and NK cells. Surprisingly, IL-4, a Th2 cytokine, also acts synergistically with IL-12 on IFN-gamma production by DCs. In addition, IL-4 markedly enhances IFN-gamma production when DCs are stimulated through CD40 or MHC class II. These results indicate that both Th1 and Th2 cytokines act on DCs during T cell-DC interaction upon Ag presentation. p38 mitogen-activated protein kinase is constitutively activated in mature DCs and is required for IFN-gamma production by DCs. IL-18 but not IL-4 or IL-12 further activates the p38 mitogen-activated protein kinase activity, suggesting that IL-4 and IL-18 enhance IFN-gamma production through distinct intracellular signal transduction pathways in DCs.