IL-4-Transfected Tumor Cell Vaccines Activate Tumor-Infiltrating Dendritic Cells and Promote Type-1 Immunity1
http://www.100md.com
免疫学杂志 2005年第11期
Abstract
We previously demonstrated that IL-4 gene-transfected glioma cell vaccines induce effective therapeutic immunity in preclinical glioma models, and have initiated phase I trials of these vaccines in patients with malignant gliomas. To gain additional mechanistic insight into the efficacy of this approach, we have treated mice bearing the MCA205 (H-2b) or CMS-4 (H-2d) sarcomas. IL-12/23 p40–/– and IFN-–/– mice, which were able to reject the initial inoculation of IL-4 expressing tumors, failed to mount a sustained systemic response against parental (nontransfected) tumor cells. Paracrine production of IL-4 in vaccine sites promoted the accumulation and maturation of IL-12p70-secreting tumor-infiltrating dendritic cells (TIDCs). Adoptive transfer of TIDCs isolated from vaccinated wild-type, but not IL-12/23 p40–/–, mice were capable of promoting tumor-specific CTL responses in syngeneic recipient animals. Interestingly, both STAT4–/– and STAT6–/– mice failed to reject IL-4-transfected tumors in concert with the reduced capacity of TIDCs to produce IL-12p70 and to promote specific antitumor CTL reactivity. These results suggest that vaccines consisting of tumor cells engineered to produce the type 2 cytokine, IL-4, critically depend on type 1 immunity for their observed therapeutic efficacy.
Introduction
Based on their profile of secreted cytokines, T cell responses have been frequently subdivided into two types, type 1 and 2, with the type 1 response being represented by predominant expression of IL-2 and IFN- and the type 2 response preferentially producing IL-4, IL-5, and IL-10 (1). Although IL-4 is generally regarded as an inducer of type 2 responses, recent studies clearly demonstrated that IL-4 has pleiotropic effects on immune cells of multiple lineages (reviewed in Refs. 2 and 3) and that it plays an important role as an inducer of type 1 T cell immunity (4). In particular, IL-4 supports dendritic cell (DC)4 maturation (5) and promotes enhanced IL-12p70 secretion from DCs (6, 7). Local delivery of IL-4 at an immunization site in a Leishmania major model results in augmented IL-12 production and Th1-type responses in vivo (8), which contrasts with the type 2 immunity induced by systemic or prolonged rIL-4 delivery in the same disease model.
IL-4 gene therapy (GT) of cancer using genetically engineered tumor cells induces potent protective as well as therapeutic antitumor immunity in animal models (reviewed in Refs. 2 and 3). Interestingly, IL-4 transduced cancer cells display increased lesional infiltration by DCs, relative to other cytokines (9), which may result in enhanced cross-presentation of tumor-associated Ags.
In comparing various cytokine gene therapies in the rat 9L gliosarcoma model, we observed that peripheral vaccination with IL-4-transfected 9L cells yielded the most effective therapeutic effects among the cytokines examined (i.e., IL-4, IL-12, GM-CSF, or IFN-) (10). These results prompted our initiation of phase I clinical trials of vaccinations with autologous glioma cells engineered to produce IL-4 (10, 11, 12). Our preliminary analyses of underlying mechanisms critical to antitumor immunity induced by these vaccines suggested that remarkably high levels of IFN- were produced by splenocytes (SPCs), suggesting that this modality induced type 1 T cell responses (13).
In the present study, we tested our hypothesis that IL-4 GT promotes IL-12p70 expression from tumor-infiltrating APCs, which may play a critical role in the development of durable antitumor immunity. Because STAT4 and 6 are required for intracellular signaling following IL-12 and IL-4 stimulation, respectively (reviewed in Refs. 14 and 15), we examined and confirmed roles for both STAT4 and 6 in the effectiveness of IL-4 GT in our tumor model.
Materials and Methods
Statistical analyses
Comparative growth of s.c. tumors and T cell responses were assessed using a Student’s t test for two samples with unequal variances. Statistical significance was determined at the <0.05 level.
Results
Syngeneic fibrosarcoma cells stably transfected with IL-4 or IL-12 are rejected in immunocompetent mice
As an initial evaluation of the effects of paracrine cytokine production within the tumor microenvironment, fibrosarcoma cells were stably transfected with cDNAs for IL-4 or IL-12 by recombinant retroviral infection and were inoculated s.c. into syngeneic mice (Fig. 1). Because the overall goal of the study was to determine the role of type 1 and type 2 cytokines in IL-4 GT, we also included IL-12 GT as a prototypical type 1 cytokine model to compare with IL-4 GT in the initial sets of experiments. Although cytokine-gene transfection did not alter the growth rate of parental MCA205 or CMS4 cell lines in vitro (data not shown), in both C57BL/6 and BALB/c mice, IL-4 or IL-12 cDNA-transfected tumors were rejected, whereas mock-transfected tumors grew aggressively in all mice tested. These findings indicate that local production of IL-4 or IL-12 within the tumor microenvironment dramatically reduces the tumorigenicity of the parental sarcoma cells in vivo.
Rejection of IL-4 cDNA-transfected tumors is independent of endogenous IL-12/23p40 and IFN-
To assess the roles of host type 1 cytokines in the rejection of cytokine gene engineered tumors, C57BL/6-background (H-2b) IL-12/23 p40–/– or IFN- –/– mice received s.c. injections with 2 x 105 MCA205 stably transfected to secrete IL-4 or IL-12. As shown in Fig. 2A, p40–/– mice rejected both MCA205-IL-4 and MCA205-IL-12, suggesting that rejection was independent of endogenous IL-12 or IL-23 production. In contrast, as shown in Fig. 2B, IFN- –/– mice allowed the growth of MCA205-IL-12 tumors, suggesting that the antitumor effects associated with locally secreted IL-12p70 was at least partially dependent on the capacity of the host to produce IFN-. Interestingly, the growth of MCA205-IL-4, but not control MCA205-Neo tumors was suppressed in IFN-–/– mice (p = 0.0007 vs Neo), as only one of five mice with MCA205-IL-4 challenge had palpable tumors, suggesting that effective host immunity against MCA205-IL-4 was largely independent of endogenous IL-12 and IFN- production.
T cells, but not NKT cells, are required for rejection of IL-4 expressing tumors
To examine the immune cells responsible for the rejection of MCA205-IL-4 tumors, wild-type C57BL/6, NKT–/–, and C57BL/6-background athymic mice were injected with 2 x 105 MCA205-IL-4 s.c. As shown in Fig. 2C, long-term observation of up to 60 days following tumor-challenge indicated that MCA205-IL-4 tumors failed to progress in NKT–/– mice, but eventually grew out by day 35 in athymic mice. This indicates that NKT cells were not responsible for the acute rejection of IL-4 expressing tumors. In contrast, in the nu/nu recipients, T cells appeared to be responsible for sustained, long-term antitumor immunity, although the delayed tumor growth in these mice suggested the potentially important role of alternate innate immune effectors, such as eosinophils, for the early phase of antitumor response (19, 20).
Endogenous IL-12/23p40 and IFN- production are required for long-term systemic immunity
To address the roles of endogenous IL-12/23p40 and IFN- in the IL-4 GT-induced generation and maintenance of long-term memory T cell-mediated immunity against parental MCA205, we rechallenged each of the animals that had rejected an initial inoculation of MCA205-IL-4 cells with a s.c. injection of 2 x 105 parental MCA205 cells. Fig. 3A depicts the growth of parental MCA205 tumors in the flank of C57BL/6 wild-type or p40–/– mice that had previously rejected MCA205-IL-4 tumors. All wild-type mice rejected the rechallenged tumors, whereas parental MCA205 cells grew progressively in all of the primed IL-12/23 p40–/– (or IFN- –/–; data not shown) mice. To evaluate the profile of cytokine-responses in SPCs isolated from mice vaccinated with MCA205-IL-4 cells, SPCs were stimulated with irradiated MCA205 cells in vitro and assessed for their ability to produce IFN- and IL-4 as measured by specific ELISA (Fig. 3B). SPCs from IL-12/23p40–/– mice produced a severely reduced level of IFN- in comparison to wild-type SPCs, consistent with the inability of these animals to produce biologically active IL-12p70, a potent inducer of IFN- (21, 22). In contrast, IL-4 production levels by SPCs were comparable between wild-type, p40–/–, and IFN- –/– mice. Hence, although IL-12/23p40 and IFN- did not appear to have major roles in the initial rejection of IL-4-secreting tumor cells, these two cytokines are clearly important in supporting memory responses against the parental tumor.
IL-4 GT enhances the accumulation of TIDCs and their expression of IL-12p70
To investigate the role of IL-4 GT on APCs in the vaccine site, we analyzed TIDCs. In comparison to nonengineered tumors, the secretion of transgene-derived IL-4 within the vaccine site attracted enhanced numbers of CD11c+ TIDCs (Fig. 4A), that were competent to produce IL-12p70 (Fig. 4B). We also evaluated surface markers on freshly isolated CD11c+ TIDCs to assess their maturation status (Fig. 4C). Although the CD11c+ TIDCs demonstrated heterogeneous subpopulations with regard to the I-Ab and CD86 markers in both groups, IL-4-transfected tumors contained a higher frequency of I-Ab high and CD86high cells (i.e., matured DCs) than the control tumors. These data suggest that IL-4 GT enhances the recruitment and maturation of IL-12p70 producing TIDCs in association with enhanced therapeutic efficacy.
Adoptive transfer of TIDCs promotes tumor-specific CTL response in recipient mice in an IL-12/23p40-dependent manner
To further determine the functional roles of TIDCs within IL-4 GT treatment sites, TIDCs were isolated from IL-4-transfected tumors in wild-type mice or IL-12/23p40–/– mice and adoptively transferred into recipient C57BL/6 mice as a vaccine. The size of tumors and the density of TIDC obtained from wild-type or IL-12/23p40–/– mice were comparable, and the lack of IL-12p70 production in IL-12/23p40–/– mice-derived DC was confirmed (data not shown). Control mice received a mock i.p. injection with PBS. SPCs from the recipient mice were in vitro restimulated with MCA205 cells for 6 days and then subjected to standard 4-h 51Cr-release assays to evaluate the levels of specific CTL responses (Fig. 5C). TIDCs isolated from wild-type mice induced tumor-specific CTL responses against parental MCA205 (p < 0.05 at all E:T ratios vs p40–/– mice and mock-injected mice), whereas TIDCs isolated from MCA205-IL-4 tumor grown in IL-12/23p40–/– mice (p40–/– TIDC) failed to expand any tumor-specific T cell responses. The lytic activities against MCA205 were considered to be tumor-specific as control target YAC-1 cells demonstrated only background levels of susceptibility to the killing effects of the SPCs. These data suggest that TIDCs and their ability to produce IL-12/23 are important for developing adaptive immunity against the parental tumor cells.
Host STAT4 and STAT6 are required for the antitumor effectiveness of IL-4 GT
Subsequently, we examined the role of STAT pathway using BALB/c-background STAT4–/– and STAT6–/– mice as recipients. Syngeneic CMS4 tumors grew aggressively and comparably in BALB/c, STAT4–/–, and STAT6–/– mice (data not shown). To evaluate the effect of local IL-4 expression on TIDC status in these mice, IL-4-transfected and nontransfected CMS4 fibrosarcoma cells were admixed and injected s.c., thereby allowing for the harvest of sufficient TIDC even in wild-type mice. As shown in Fig. 6A, tumors grew progressively in both STAT4–/– and STAT6–/– mice in comparison to wild-type mice (p = 0.022 for BALB/c vs STAT4–/–, p = 0.046 for BALB/c vs STAT6–/–), suggesting that both STAT4 and STAT6 are required for the induction of antitumor immunity by IL-4 GT. As depicted in Fig. 6B, TIDC secretion of IL-12p70 was moderately reduced in STAT4–/– mice, and severely diminished in STAT6–/– mice when compared with TIDC isolated from wild-type mice. We also evaluated the maturation status of TIDCs by flow cytometry (Fig. 6C). As was also the case in Fig. 4, the CD11c+ TIDC populations were heterogeneous in their composition. Nevertheless, there appeared to be a tendency for TIDCs isolated from STAT6–/– mice to display lower MHC class II (I-Ad) and CD86 levels than TIDCs harvested from wild-type mice or STAT4–/– mice.
Both type 1 and type 2 immune responses are required for the optimal efficacy of IL-4 GT
To further delineate the roles of type 1 and 2 cytokine responses in the effectiveness of IL-4 GT, we examined the CTL responses and cytokine production profiles from SPCs obtained from STAT4–/–, STAT6–/–, or wild-type mice, all of which had been previously inoculated s.c. with CMS4-IL-4 fibrosarcoma cells. As shown in Fig. 7A, and as expected, SPCs from STAT4–/– mice produced a reduced level of IFN- in comparison to the wild-type mice, while SPCs from STAT6–/– mice failed to produce detectable levels of IL-4. CTL assays using these SPCs demonstrated that both STAT4–/– and STAT6–/– mice mounted CTL responses of a reduced magnitude against CMS4 fibrosarcoma cells in comparison to wild-type mice (Fig. 7B). All these SPCs demonstrated only a background level of killing activity against YAC-1 cells, suggesting a CMS4-specific killing activity of the SPCs. These results suggest that the abrogation of antitumor efficacy in STAT4–/– and STAT6–/– mice (Fig. 6) may be due, at least partially, to the impaired development of tumor-specific CTL in these animals; and that host type 1 cytokine responses as well as type 2 responses play critical roles in the efficacy of IL-4 GT.
Discussion
The most significant findings in the present study are that local IL-4 production within the site of a tumor vaccine leads to the activation of TIDCs through a STAT6-dependent pathway, and that these activated TIDCs induce antitumor T cell responses in an IL-12p40/IFN--dependent manner. The critical role of host-derived type 1 cytokines was further supported by the abrogation of IL-4 GT-induced antitumor effects in STAT4–/– mice.
Although long-term antitumor protection of animals required host type 1 and type 2 cytokines, our data in establishment models indicated that the initial rejection of IL-4 gene-transfected tumors was independent of host type 1 cytokines and NKT cells. We hypothesize that this early event was most likely mediated by eosinophils and macrophages as demonstrated by experiments using Abs that specifically blocked the accumulation of eosinophils at the site of IL-4-expressing tumors, as we (20) and others (19) have previously demonstrated. The role of the innate immunity against IL-4-expressing tumors is also suggested in our current study by the delayed growth of MCA205-IL-4 tumors in athymic mice.
In contrast to IL-4 GT, IL-12 GT was dependent on the ability of the host to produce IFN-. This result is expected given the known role for IL-12 as a major inducer of IFN- and IFN-inducible protein 10, both of which play major roles in IL-12 mediated antitumor responses (21, 22). Although NKT cells did not appear to be critical mediators or effectors in our IL-4 GT establishment model, antitumor immunity stimulated by vaccination with GM-CSF-secreting tumor cells has been reported to be abrogated in CD1d–/– and NKT cell–/– mice (23). In this referenced study, the lack of antitumor immunity was accompanied by impaired tumor-induced type 2 cytokine production, although IFN- secretion and cytotoxicity appeared to be preserved. We are currently evaluating whether IL-4 GT requires NKT cells for systemic, long-term antitumor responses.
Our results demonstrate the critical role of TIDCs that express IL-12p70. Local IL-4 production within the tumor-vaccine site increased the number of TIDCs, and up-regulated IL-12p70 secretion from TIDC. These TIDCs must have engulfed and presented tumor Ags, as demonstrated by our results obtained in adoptive transfer experiments. It has been demonstrated that GM-CSF and CD40L-cotransfected C-26 tumors attract DCs that are able to stimulate specific CTL clones and to prime naive mice for a CTL response against the parental tumor (24). Our results also support the notion that tumor cell-based vaccines engineered to express DC-stimulating cytokines (i.e., GM-CSF or IL-4) are able to enhance the accumulation and activation of TIDCs that are critical for cross-priming tumor-specific T cell responses.
Our data also demonstrate the critical role of host-derived IL-12/23, particularly that expressed by TIDCs, in supporting the ability of IL-4GT to induce tumor-specific T cell-mediated immunity in vivo. Using bone marrow-derived DCs, it has been previously reported that the absence of IL-12/23 p40 does not affect the ability of vaccination with Ag-loaded DCs to elicit OVA-specific CTL cytotoxicity (25), which contrasts with our current results. Although this referenced study demonstrated a requirement for CD4+ Th cells and CD40L in vaccine effectiveness, immunizations with IL-12/23 p40–/– DC were performed using an H-2Kb-restricted peptide epitope in the absence of Th epitopes. Therefore, it is possible that this study may have underestimated the central importance of CD40-CD40L interactions in triggering DC of IL-12p70 (26). In contrast, in our present study, we presume that TIDCs have both engulfed tumor Ags and cross-presented both MHC class I and II epitopes to responder CD8+ and CD4+ T cells, respectively.
These data suggest that IL-4 GT relies upon host type 1 cytokine responses in mediating its antitumor effects. Indeed, there appears to be accumulating evidence that type 1 responses are promoted by IL-4 GT. IL-4–/– mice are severely impaired in their development of antitumor immunity to tumor lines that are typically rejected by control, wild-type animals (27). The lack of antitumor immunity in IL-4–/– mice was associated with reduced IFN- production and undetectable CTL activity, indicating that defective type 1 responses occur in the absence of endogenous IL-4 production. When IL-4 is provided in the vicinity of tumor cells, tumor immunity was induced in IL-4–/– mice, supporting a role for IL-4 in the priming phase of effector T cells (27). In a colon 26 tumor model, IL-4 GT has been reported to enhance IFN-, IL-12 p35, and p40 mRNA expression in the draining lymph nodes (28). Furthermore, mice deficient in the IFN- gene did not reject colon 26/IL-4 cells, suggesting that IL-4-induced generation of memory CTL requires IFN- production in the draining lymph nodes to generate protective immunity.
Based on our current study using both STAT4–/– and STAT6–/– mice, we propose the following mechanisms related to how IL-4 GT induces type 1 dependent antitumor immunity. First, local production of IL-4 in the tumor vaccine site activates TIDCs through STAT6-dependent pathways. Activated TIDCs become polarized to induce type 1 effector responses as demonstrated by their enhanced IL-12p70 production. In the induction of adaptive antitumor response, IL-12 and STAT4-mediated signals play pivotal roles as observed in our TIDC adoptive transfer experiments and the failure of IL-4 GT to induce antitumor immune responses in STAT4–/– mice.
Although STATs 1, 3, 4, and 5 have all been reported to be activated by IL-12, STAT4 appears to be the major player in IL-12-mediated responses (29) and STAT4–/– mice exhibit defective type 1 T cell proliferation, IFN- production, and NK cell activity in response to IL-12 stimulation (reviewed in Refs. 14 and 15).
STAT6 is required for IL-4R signaling and for the activation and differentiation of Ag-specific type 2 T cells (30). In our current study, TIDCs isolated from STAT6–/– mice demonstrated an immature phenotype and a diminished capacity for IL-12p70 production, which is in accordance with a previous study by Lutz et al. (31) demonstrating that STAT6–/– DC did not become matured or produce enhanced levels of IL-12p70 in response to IL-4 when they were cultured in low concentrations of GM-CSF.
Based on our results, we believe that STAT6–/– mice fail to reject the IL-4-transfected tumors for at least two reasons. First, as discussed above, at the level of Ag-presentation, STAT6–/– TIDCs fail to respond to local IL-4 stimulation, appear to remain functionally immature, and produce suboptimal levels of IL-12p70 required for the support of specific CTLs. Secondarily, STAT6–/– mice-derived SPCs do not secrete detectable levels of IL-4. Although Kacha et al. (32) demonstrated accelerated tumor rejection and augmented IFN- producing tumor-specific CTL activity in STAT6–/– mice, in our IL-4 GT model, we hypothesize that IL-4 production from T effector cells may be important for optimal antitumor effectiveness. Indeed, as has been demonstrated by others, tumor eradication resulting from IL-4 GT is not restricted to a type 1 response, but can also be mediated by a type 2-biased T cell response. Indeed, IL-4-transfected C26/FR tumor cell vaccines induced antitumor type 2 effector T cells that released IL-4 upon specific stimulation, and these T cells were capable of mediating the rejection of lung metastasis (33). Interestingly, CD8+ T cells isolated from IL-4-vaccinated IFN-–/–, but not from IL-4–/–, mice were able to cure established lung metastases, supporting the conclusion that IL-4 produced by Tc2 cells was critical for tumor rejection in this model system (33). Anti-TS/A tumor memory immune responses induced by IL-4-transfected TS/A tumor cells were more potent than those elicited by IL-2 gene-transduced TS/A cells, with the former responses characterized as type 2 (34). Our previous studies also suggested that both type 1 and type 2 responses play coordinated roles in host protection against the rat 9L glioma (13, 35). Lastly, it has been demonstrated that STAT6-mediated type 2 responses attract eosinophils that possess potent cytolytic activity against tumor cells (36) and that the effective generation of antitumor CTLs requires IL-4 production from CD8+ T cells (37).
Overall, our study strongly suggests that the efficacy of IL-4 GT relies upon intact expression of both type 1 and type 2 cytokines by host immune cells. We believe our observations are clinically relevant because it has been well-documented that cancer patients, including glioma (38), display suppressed type 1 immune-reactivity (39, 40), but prevalent type 2 and/or regulatory type T cell responses against tumor Ags (41, 42). In addition, expression of IL-12 from PBMC is decreased in cancer patients (43), and patients with advanced metastatic cancers often present with significantly lower frequencies of circulating DCs and serum concentrations of IL-12 (44, 45). In this light, analyses of both type 1 and type 2 cytokine response status in patients undergoing IL-4 GT trials (10, 11, 12) are clearly warranted to determine those factors that are critically linked to immunological and clinical responses to this treatment modality.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by P01 NS40923 (to H.O.) and a grant from the Copeland Fund of the Pittsburgh Foundation (to H.O.).
2 J.E. and N.K. contributed equally to this article.
3 Address correspondence and reprint requests to Dr. Hideho Okada, Department of Neurological Surgery, University of Pittsburgh School of Medicine, G12a Hillman Cancer Center, 5117 Center Avenue, Pittsburgh, PA 15213-1863. E-mail address: okadah{at}msx.upmc.edu
4 Abbreviations used in this paper: DC, dendritic cell; SPC, splenocyte; TIDC, tumor-infiltrating DC; GT, gene therapy.
Received for publication January 25, 2005. Accepted for publication March 25, 2005.
References
Morel, P. A., T. B. Oriss. 1998. Crossregulation between Th1 and Th2 cells. Crit. Rev. Immunol. 18: 275-303.
Okada, H., J. Banchereau, M. T. Lotze. 2003. Interleukin-4. A. W. Thomson, and M. T. Lotze, eds. The Cytokine Handbook 227-262 Elsevier Science, London. .
Okada, H., N. Kuwashima. 2002. Gene therapy and biologic therapy with interleukin-4. Curr. Gene Ther. 2: 437-450.
Lingnau, K., P. Hoehn, S. Kerdine, S. Koelsch, C. Neudoerfl, N. Palm, E. Ruede, E. Schmitt. 1998. IL-4 in combination with TGF- favors an alternative pathway of Th1 development independent of IL-12. J. Immunol. 161: 4709-4718.
Nestle, F. O., L. Filgueira, B. J. Nickoloff, G. Burg. 1998. Human dermal dendritic cells process and present soluble protein antigens. J. Invest. Dermatol. 110: 762-766.
Kalinski, P., H. H. Smits, J. H. Schuitemaker, P. L. Vieira, M. van Eijk, E. C. de Jong, E. A. Wierenga, M. L. Kapsenberg. 2000. IL-4 is a mediator of IL-12p70 induction by human Th2 cells: reversal of polarized Th2 phenotype by dendritic cells. J. Immunol. 165: 1877-1881.
Hockrein, H., M. O’Keeffe, T. Luft, S. Vandenabeele, R. J. Grumont, E. Maraskovsky, K. Shortman. 2000. Interleukin (IL)-4 is a major regulatory cytokine governing bioactive IL-12 production by mouse and human dendritic cells. J. Exp. Med. 192: 823-833(Junichi Eguchi2,*, Naruo )
We previously demonstrated that IL-4 gene-transfected glioma cell vaccines induce effective therapeutic immunity in preclinical glioma models, and have initiated phase I trials of these vaccines in patients with malignant gliomas. To gain additional mechanistic insight into the efficacy of this approach, we have treated mice bearing the MCA205 (H-2b) or CMS-4 (H-2d) sarcomas. IL-12/23 p40–/– and IFN-–/– mice, which were able to reject the initial inoculation of IL-4 expressing tumors, failed to mount a sustained systemic response against parental (nontransfected) tumor cells. Paracrine production of IL-4 in vaccine sites promoted the accumulation and maturation of IL-12p70-secreting tumor-infiltrating dendritic cells (TIDCs). Adoptive transfer of TIDCs isolated from vaccinated wild-type, but not IL-12/23 p40–/–, mice were capable of promoting tumor-specific CTL responses in syngeneic recipient animals. Interestingly, both STAT4–/– and STAT6–/– mice failed to reject IL-4-transfected tumors in concert with the reduced capacity of TIDCs to produce IL-12p70 and to promote specific antitumor CTL reactivity. These results suggest that vaccines consisting of tumor cells engineered to produce the type 2 cytokine, IL-4, critically depend on type 1 immunity for their observed therapeutic efficacy.
Introduction
Based on their profile of secreted cytokines, T cell responses have been frequently subdivided into two types, type 1 and 2, with the type 1 response being represented by predominant expression of IL-2 and IFN- and the type 2 response preferentially producing IL-4, IL-5, and IL-10 (1). Although IL-4 is generally regarded as an inducer of type 2 responses, recent studies clearly demonstrated that IL-4 has pleiotropic effects on immune cells of multiple lineages (reviewed in Refs. 2 and 3) and that it plays an important role as an inducer of type 1 T cell immunity (4). In particular, IL-4 supports dendritic cell (DC)4 maturation (5) and promotes enhanced IL-12p70 secretion from DCs (6, 7). Local delivery of IL-4 at an immunization site in a Leishmania major model results in augmented IL-12 production and Th1-type responses in vivo (8), which contrasts with the type 2 immunity induced by systemic or prolonged rIL-4 delivery in the same disease model.
IL-4 gene therapy (GT) of cancer using genetically engineered tumor cells induces potent protective as well as therapeutic antitumor immunity in animal models (reviewed in Refs. 2 and 3). Interestingly, IL-4 transduced cancer cells display increased lesional infiltration by DCs, relative to other cytokines (9), which may result in enhanced cross-presentation of tumor-associated Ags.
In comparing various cytokine gene therapies in the rat 9L gliosarcoma model, we observed that peripheral vaccination with IL-4-transfected 9L cells yielded the most effective therapeutic effects among the cytokines examined (i.e., IL-4, IL-12, GM-CSF, or IFN-) (10). These results prompted our initiation of phase I clinical trials of vaccinations with autologous glioma cells engineered to produce IL-4 (10, 11, 12). Our preliminary analyses of underlying mechanisms critical to antitumor immunity induced by these vaccines suggested that remarkably high levels of IFN- were produced by splenocytes (SPCs), suggesting that this modality induced type 1 T cell responses (13).
In the present study, we tested our hypothesis that IL-4 GT promotes IL-12p70 expression from tumor-infiltrating APCs, which may play a critical role in the development of durable antitumor immunity. Because STAT4 and 6 are required for intracellular signaling following IL-12 and IL-4 stimulation, respectively (reviewed in Refs. 14 and 15), we examined and confirmed roles for both STAT4 and 6 in the effectiveness of IL-4 GT in our tumor model.
Materials and Methods
Statistical analyses
Comparative growth of s.c. tumors and T cell responses were assessed using a Student’s t test for two samples with unequal variances. Statistical significance was determined at the <0.05 level.
Results
Syngeneic fibrosarcoma cells stably transfected with IL-4 or IL-12 are rejected in immunocompetent mice
As an initial evaluation of the effects of paracrine cytokine production within the tumor microenvironment, fibrosarcoma cells were stably transfected with cDNAs for IL-4 or IL-12 by recombinant retroviral infection and were inoculated s.c. into syngeneic mice (Fig. 1). Because the overall goal of the study was to determine the role of type 1 and type 2 cytokines in IL-4 GT, we also included IL-12 GT as a prototypical type 1 cytokine model to compare with IL-4 GT in the initial sets of experiments. Although cytokine-gene transfection did not alter the growth rate of parental MCA205 or CMS4 cell lines in vitro (data not shown), in both C57BL/6 and BALB/c mice, IL-4 or IL-12 cDNA-transfected tumors were rejected, whereas mock-transfected tumors grew aggressively in all mice tested. These findings indicate that local production of IL-4 or IL-12 within the tumor microenvironment dramatically reduces the tumorigenicity of the parental sarcoma cells in vivo.
Rejection of IL-4 cDNA-transfected tumors is independent of endogenous IL-12/23p40 and IFN-
To assess the roles of host type 1 cytokines in the rejection of cytokine gene engineered tumors, C57BL/6-background (H-2b) IL-12/23 p40–/– or IFN- –/– mice received s.c. injections with 2 x 105 MCA205 stably transfected to secrete IL-4 or IL-12. As shown in Fig. 2A, p40–/– mice rejected both MCA205-IL-4 and MCA205-IL-12, suggesting that rejection was independent of endogenous IL-12 or IL-23 production. In contrast, as shown in Fig. 2B, IFN- –/– mice allowed the growth of MCA205-IL-12 tumors, suggesting that the antitumor effects associated with locally secreted IL-12p70 was at least partially dependent on the capacity of the host to produce IFN-. Interestingly, the growth of MCA205-IL-4, but not control MCA205-Neo tumors was suppressed in IFN-–/– mice (p = 0.0007 vs Neo), as only one of five mice with MCA205-IL-4 challenge had palpable tumors, suggesting that effective host immunity against MCA205-IL-4 was largely independent of endogenous IL-12 and IFN- production.
T cells, but not NKT cells, are required for rejection of IL-4 expressing tumors
To examine the immune cells responsible for the rejection of MCA205-IL-4 tumors, wild-type C57BL/6, NKT–/–, and C57BL/6-background athymic mice were injected with 2 x 105 MCA205-IL-4 s.c. As shown in Fig. 2C, long-term observation of up to 60 days following tumor-challenge indicated that MCA205-IL-4 tumors failed to progress in NKT–/– mice, but eventually grew out by day 35 in athymic mice. This indicates that NKT cells were not responsible for the acute rejection of IL-4 expressing tumors. In contrast, in the nu/nu recipients, T cells appeared to be responsible for sustained, long-term antitumor immunity, although the delayed tumor growth in these mice suggested the potentially important role of alternate innate immune effectors, such as eosinophils, for the early phase of antitumor response (19, 20).
Endogenous IL-12/23p40 and IFN- production are required for long-term systemic immunity
To address the roles of endogenous IL-12/23p40 and IFN- in the IL-4 GT-induced generation and maintenance of long-term memory T cell-mediated immunity against parental MCA205, we rechallenged each of the animals that had rejected an initial inoculation of MCA205-IL-4 cells with a s.c. injection of 2 x 105 parental MCA205 cells. Fig. 3A depicts the growth of parental MCA205 tumors in the flank of C57BL/6 wild-type or p40–/– mice that had previously rejected MCA205-IL-4 tumors. All wild-type mice rejected the rechallenged tumors, whereas parental MCA205 cells grew progressively in all of the primed IL-12/23 p40–/– (or IFN- –/–; data not shown) mice. To evaluate the profile of cytokine-responses in SPCs isolated from mice vaccinated with MCA205-IL-4 cells, SPCs were stimulated with irradiated MCA205 cells in vitro and assessed for their ability to produce IFN- and IL-4 as measured by specific ELISA (Fig. 3B). SPCs from IL-12/23p40–/– mice produced a severely reduced level of IFN- in comparison to wild-type SPCs, consistent with the inability of these animals to produce biologically active IL-12p70, a potent inducer of IFN- (21, 22). In contrast, IL-4 production levels by SPCs were comparable between wild-type, p40–/–, and IFN- –/– mice. Hence, although IL-12/23p40 and IFN- did not appear to have major roles in the initial rejection of IL-4-secreting tumor cells, these two cytokines are clearly important in supporting memory responses against the parental tumor.
IL-4 GT enhances the accumulation of TIDCs and their expression of IL-12p70
To investigate the role of IL-4 GT on APCs in the vaccine site, we analyzed TIDCs. In comparison to nonengineered tumors, the secretion of transgene-derived IL-4 within the vaccine site attracted enhanced numbers of CD11c+ TIDCs (Fig. 4A), that were competent to produce IL-12p70 (Fig. 4B). We also evaluated surface markers on freshly isolated CD11c+ TIDCs to assess their maturation status (Fig. 4C). Although the CD11c+ TIDCs demonstrated heterogeneous subpopulations with regard to the I-Ab and CD86 markers in both groups, IL-4-transfected tumors contained a higher frequency of I-Ab high and CD86high cells (i.e., matured DCs) than the control tumors. These data suggest that IL-4 GT enhances the recruitment and maturation of IL-12p70 producing TIDCs in association with enhanced therapeutic efficacy.
Adoptive transfer of TIDCs promotes tumor-specific CTL response in recipient mice in an IL-12/23p40-dependent manner
To further determine the functional roles of TIDCs within IL-4 GT treatment sites, TIDCs were isolated from IL-4-transfected tumors in wild-type mice or IL-12/23p40–/– mice and adoptively transferred into recipient C57BL/6 mice as a vaccine. The size of tumors and the density of TIDC obtained from wild-type or IL-12/23p40–/– mice were comparable, and the lack of IL-12p70 production in IL-12/23p40–/– mice-derived DC was confirmed (data not shown). Control mice received a mock i.p. injection with PBS. SPCs from the recipient mice were in vitro restimulated with MCA205 cells for 6 days and then subjected to standard 4-h 51Cr-release assays to evaluate the levels of specific CTL responses (Fig. 5C). TIDCs isolated from wild-type mice induced tumor-specific CTL responses against parental MCA205 (p < 0.05 at all E:T ratios vs p40–/– mice and mock-injected mice), whereas TIDCs isolated from MCA205-IL-4 tumor grown in IL-12/23p40–/– mice (p40–/– TIDC) failed to expand any tumor-specific T cell responses. The lytic activities against MCA205 were considered to be tumor-specific as control target YAC-1 cells demonstrated only background levels of susceptibility to the killing effects of the SPCs. These data suggest that TIDCs and their ability to produce IL-12/23 are important for developing adaptive immunity against the parental tumor cells.
Host STAT4 and STAT6 are required for the antitumor effectiveness of IL-4 GT
Subsequently, we examined the role of STAT pathway using BALB/c-background STAT4–/– and STAT6–/– mice as recipients. Syngeneic CMS4 tumors grew aggressively and comparably in BALB/c, STAT4–/–, and STAT6–/– mice (data not shown). To evaluate the effect of local IL-4 expression on TIDC status in these mice, IL-4-transfected and nontransfected CMS4 fibrosarcoma cells were admixed and injected s.c., thereby allowing for the harvest of sufficient TIDC even in wild-type mice. As shown in Fig. 6A, tumors grew progressively in both STAT4–/– and STAT6–/– mice in comparison to wild-type mice (p = 0.022 for BALB/c vs STAT4–/–, p = 0.046 for BALB/c vs STAT6–/–), suggesting that both STAT4 and STAT6 are required for the induction of antitumor immunity by IL-4 GT. As depicted in Fig. 6B, TIDC secretion of IL-12p70 was moderately reduced in STAT4–/– mice, and severely diminished in STAT6–/– mice when compared with TIDC isolated from wild-type mice. We also evaluated the maturation status of TIDCs by flow cytometry (Fig. 6C). As was also the case in Fig. 4, the CD11c+ TIDC populations were heterogeneous in their composition. Nevertheless, there appeared to be a tendency for TIDCs isolated from STAT6–/– mice to display lower MHC class II (I-Ad) and CD86 levels than TIDCs harvested from wild-type mice or STAT4–/– mice.
Both type 1 and type 2 immune responses are required for the optimal efficacy of IL-4 GT
To further delineate the roles of type 1 and 2 cytokine responses in the effectiveness of IL-4 GT, we examined the CTL responses and cytokine production profiles from SPCs obtained from STAT4–/–, STAT6–/–, or wild-type mice, all of which had been previously inoculated s.c. with CMS4-IL-4 fibrosarcoma cells. As shown in Fig. 7A, and as expected, SPCs from STAT4–/– mice produced a reduced level of IFN- in comparison to the wild-type mice, while SPCs from STAT6–/– mice failed to produce detectable levels of IL-4. CTL assays using these SPCs demonstrated that both STAT4–/– and STAT6–/– mice mounted CTL responses of a reduced magnitude against CMS4 fibrosarcoma cells in comparison to wild-type mice (Fig. 7B). All these SPCs demonstrated only a background level of killing activity against YAC-1 cells, suggesting a CMS4-specific killing activity of the SPCs. These results suggest that the abrogation of antitumor efficacy in STAT4–/– and STAT6–/– mice (Fig. 6) may be due, at least partially, to the impaired development of tumor-specific CTL in these animals; and that host type 1 cytokine responses as well as type 2 responses play critical roles in the efficacy of IL-4 GT.
Discussion
The most significant findings in the present study are that local IL-4 production within the site of a tumor vaccine leads to the activation of TIDCs through a STAT6-dependent pathway, and that these activated TIDCs induce antitumor T cell responses in an IL-12p40/IFN--dependent manner. The critical role of host-derived type 1 cytokines was further supported by the abrogation of IL-4 GT-induced antitumor effects in STAT4–/– mice.
Although long-term antitumor protection of animals required host type 1 and type 2 cytokines, our data in establishment models indicated that the initial rejection of IL-4 gene-transfected tumors was independent of host type 1 cytokines and NKT cells. We hypothesize that this early event was most likely mediated by eosinophils and macrophages as demonstrated by experiments using Abs that specifically blocked the accumulation of eosinophils at the site of IL-4-expressing tumors, as we (20) and others (19) have previously demonstrated. The role of the innate immunity against IL-4-expressing tumors is also suggested in our current study by the delayed growth of MCA205-IL-4 tumors in athymic mice.
In contrast to IL-4 GT, IL-12 GT was dependent on the ability of the host to produce IFN-. This result is expected given the known role for IL-12 as a major inducer of IFN- and IFN-inducible protein 10, both of which play major roles in IL-12 mediated antitumor responses (21, 22). Although NKT cells did not appear to be critical mediators or effectors in our IL-4 GT establishment model, antitumor immunity stimulated by vaccination with GM-CSF-secreting tumor cells has been reported to be abrogated in CD1d–/– and NKT cell–/– mice (23). In this referenced study, the lack of antitumor immunity was accompanied by impaired tumor-induced type 2 cytokine production, although IFN- secretion and cytotoxicity appeared to be preserved. We are currently evaluating whether IL-4 GT requires NKT cells for systemic, long-term antitumor responses.
Our results demonstrate the critical role of TIDCs that express IL-12p70. Local IL-4 production within the tumor-vaccine site increased the number of TIDCs, and up-regulated IL-12p70 secretion from TIDC. These TIDCs must have engulfed and presented tumor Ags, as demonstrated by our results obtained in adoptive transfer experiments. It has been demonstrated that GM-CSF and CD40L-cotransfected C-26 tumors attract DCs that are able to stimulate specific CTL clones and to prime naive mice for a CTL response against the parental tumor (24). Our results also support the notion that tumor cell-based vaccines engineered to express DC-stimulating cytokines (i.e., GM-CSF or IL-4) are able to enhance the accumulation and activation of TIDCs that are critical for cross-priming tumor-specific T cell responses.
Our data also demonstrate the critical role of host-derived IL-12/23, particularly that expressed by TIDCs, in supporting the ability of IL-4GT to induce tumor-specific T cell-mediated immunity in vivo. Using bone marrow-derived DCs, it has been previously reported that the absence of IL-12/23 p40 does not affect the ability of vaccination with Ag-loaded DCs to elicit OVA-specific CTL cytotoxicity (25), which contrasts with our current results. Although this referenced study demonstrated a requirement for CD4+ Th cells and CD40L in vaccine effectiveness, immunizations with IL-12/23 p40–/– DC were performed using an H-2Kb-restricted peptide epitope in the absence of Th epitopes. Therefore, it is possible that this study may have underestimated the central importance of CD40-CD40L interactions in triggering DC of IL-12p70 (26). In contrast, in our present study, we presume that TIDCs have both engulfed tumor Ags and cross-presented both MHC class I and II epitopes to responder CD8+ and CD4+ T cells, respectively.
These data suggest that IL-4 GT relies upon host type 1 cytokine responses in mediating its antitumor effects. Indeed, there appears to be accumulating evidence that type 1 responses are promoted by IL-4 GT. IL-4–/– mice are severely impaired in their development of antitumor immunity to tumor lines that are typically rejected by control, wild-type animals (27). The lack of antitumor immunity in IL-4–/– mice was associated with reduced IFN- production and undetectable CTL activity, indicating that defective type 1 responses occur in the absence of endogenous IL-4 production. When IL-4 is provided in the vicinity of tumor cells, tumor immunity was induced in IL-4–/– mice, supporting a role for IL-4 in the priming phase of effector T cells (27). In a colon 26 tumor model, IL-4 GT has been reported to enhance IFN-, IL-12 p35, and p40 mRNA expression in the draining lymph nodes (28). Furthermore, mice deficient in the IFN- gene did not reject colon 26/IL-4 cells, suggesting that IL-4-induced generation of memory CTL requires IFN- production in the draining lymph nodes to generate protective immunity.
Based on our current study using both STAT4–/– and STAT6–/– mice, we propose the following mechanisms related to how IL-4 GT induces type 1 dependent antitumor immunity. First, local production of IL-4 in the tumor vaccine site activates TIDCs through STAT6-dependent pathways. Activated TIDCs become polarized to induce type 1 effector responses as demonstrated by their enhanced IL-12p70 production. In the induction of adaptive antitumor response, IL-12 and STAT4-mediated signals play pivotal roles as observed in our TIDC adoptive transfer experiments and the failure of IL-4 GT to induce antitumor immune responses in STAT4–/– mice.
Although STATs 1, 3, 4, and 5 have all been reported to be activated by IL-12, STAT4 appears to be the major player in IL-12-mediated responses (29) and STAT4–/– mice exhibit defective type 1 T cell proliferation, IFN- production, and NK cell activity in response to IL-12 stimulation (reviewed in Refs. 14 and 15).
STAT6 is required for IL-4R signaling and for the activation and differentiation of Ag-specific type 2 T cells (30). In our current study, TIDCs isolated from STAT6–/– mice demonstrated an immature phenotype and a diminished capacity for IL-12p70 production, which is in accordance with a previous study by Lutz et al. (31) demonstrating that STAT6–/– DC did not become matured or produce enhanced levels of IL-12p70 in response to IL-4 when they were cultured in low concentrations of GM-CSF.
Based on our results, we believe that STAT6–/– mice fail to reject the IL-4-transfected tumors for at least two reasons. First, as discussed above, at the level of Ag-presentation, STAT6–/– TIDCs fail to respond to local IL-4 stimulation, appear to remain functionally immature, and produce suboptimal levels of IL-12p70 required for the support of specific CTLs. Secondarily, STAT6–/– mice-derived SPCs do not secrete detectable levels of IL-4. Although Kacha et al. (32) demonstrated accelerated tumor rejection and augmented IFN- producing tumor-specific CTL activity in STAT6–/– mice, in our IL-4 GT model, we hypothesize that IL-4 production from T effector cells may be important for optimal antitumor effectiveness. Indeed, as has been demonstrated by others, tumor eradication resulting from IL-4 GT is not restricted to a type 1 response, but can also be mediated by a type 2-biased T cell response. Indeed, IL-4-transfected C26/FR tumor cell vaccines induced antitumor type 2 effector T cells that released IL-4 upon specific stimulation, and these T cells were capable of mediating the rejection of lung metastasis (33). Interestingly, CD8+ T cells isolated from IL-4-vaccinated IFN-–/–, but not from IL-4–/–, mice were able to cure established lung metastases, supporting the conclusion that IL-4 produced by Tc2 cells was critical for tumor rejection in this model system (33). Anti-TS/A tumor memory immune responses induced by IL-4-transfected TS/A tumor cells were more potent than those elicited by IL-2 gene-transduced TS/A cells, with the former responses characterized as type 2 (34). Our previous studies also suggested that both type 1 and type 2 responses play coordinated roles in host protection against the rat 9L glioma (13, 35). Lastly, it has been demonstrated that STAT6-mediated type 2 responses attract eosinophils that possess potent cytolytic activity against tumor cells (36) and that the effective generation of antitumor CTLs requires IL-4 production from CD8+ T cells (37).
Overall, our study strongly suggests that the efficacy of IL-4 GT relies upon intact expression of both type 1 and type 2 cytokines by host immune cells. We believe our observations are clinically relevant because it has been well-documented that cancer patients, including glioma (38), display suppressed type 1 immune-reactivity (39, 40), but prevalent type 2 and/or regulatory type T cell responses against tumor Ags (41, 42). In addition, expression of IL-12 from PBMC is decreased in cancer patients (43), and patients with advanced metastatic cancers often present with significantly lower frequencies of circulating DCs and serum concentrations of IL-12 (44, 45). In this light, analyses of both type 1 and type 2 cytokine response status in patients undergoing IL-4 GT trials (10, 11, 12) are clearly warranted to determine those factors that are critically linked to immunological and clinical responses to this treatment modality.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by P01 NS40923 (to H.O.) and a grant from the Copeland Fund of the Pittsburgh Foundation (to H.O.).
2 J.E. and N.K. contributed equally to this article.
3 Address correspondence and reprint requests to Dr. Hideho Okada, Department of Neurological Surgery, University of Pittsburgh School of Medicine, G12a Hillman Cancer Center, 5117 Center Avenue, Pittsburgh, PA 15213-1863. E-mail address: okadah{at}msx.upmc.edu
4 Abbreviations used in this paper: DC, dendritic cell; SPC, splenocyte; TIDC, tumor-infiltrating DC; GT, gene therapy.
Received for publication January 25, 2005. Accepted for publication March 25, 2005.
References
Morel, P. A., T. B. Oriss. 1998. Crossregulation between Th1 and Th2 cells. Crit. Rev. Immunol. 18: 275-303.
Okada, H., J. Banchereau, M. T. Lotze. 2003. Interleukin-4. A. W. Thomson, and M. T. Lotze, eds. The Cytokine Handbook 227-262 Elsevier Science, London. .
Okada, H., N. Kuwashima. 2002. Gene therapy and biologic therapy with interleukin-4. Curr. Gene Ther. 2: 437-450.
Lingnau, K., P. Hoehn, S. Kerdine, S. Koelsch, C. Neudoerfl, N. Palm, E. Ruede, E. Schmitt. 1998. IL-4 in combination with TGF- favors an alternative pathway of Th1 development independent of IL-12. J. Immunol. 161: 4709-4718.
Nestle, F. O., L. Filgueira, B. J. Nickoloff, G. Burg. 1998. Human dermal dendritic cells process and present soluble protein antigens. J. Invest. Dermatol. 110: 762-766.
Kalinski, P., H. H. Smits, J. H. Schuitemaker, P. L. Vieira, M. van Eijk, E. C. de Jong, E. A. Wierenga, M. L. Kapsenberg. 2000. IL-4 is a mediator of IL-12p70 induction by human Th2 cells: reversal of polarized Th2 phenotype by dendritic cells. J. Immunol. 165: 1877-1881.
Hockrein, H., M. O’Keeffe, T. Luft, S. Vandenabeele, R. J. Grumont, E. Maraskovsky, K. Shortman. 2000. Interleukin (IL)-4 is a major regulatory cytokine governing bioactive IL-12 production by mouse and human dendritic cells. J. Exp. Med. 192: 823-833(Junichi Eguchi2,*, Naruo )