Toll-like Receptor 3-mediated Suppression of TRAMP Prostate Cancer Demonstrates the Critical Role of Type I Interferons in Tumor Immune Surveillance

Abstract

Inflammation has increasingly been recognized as a critical component influencing tumor growth. Recent reports have revealed conflicting evidence for the role of Toll-like receptors (TLRs) in modulating tumorigenesis. In our study, we implicate TLR3 in mediating immune surveillance with increased growth of implanted transgenic adenocarcinoma of the mouse prostate (TRAMP) tumors in TLR3^−/− compared to TLR3^+/+ mice. Activation of TLR3 by polyinosinic-polycytidylic acid (polyI:C) leads to induction of multiple inflammatory pathways including nuclear factor kappa-B (NF-κB), mitogen-activated protein (MAP) kinases, and interferon (IFN) regulatory factors (IRFs). We explored the potential of TLR3 stimulation in prostate cancer immunotherapy, and showed that treatment with polyI:C can strongly suppress both subcutaneously implanted TRAMP tumors in syngenic mice, as well as orthotopic prostate cancers in TRAMP C57Bl6 × FvB F1 Tg^+/− transgenic mice. Treated tumors remained well to moderately differentiated with increased infiltration of T lymphocytes and natural killer (NK) cells compared to poorly differentiated adenocarcinoma observed in untreated tumors. Like TLR3^−/− mice, interferon-alpha receptor 1 (IFNAR1)^−/− mice exhibited reduced tumor surveillance and impaired tumor suppression following polyI:C treatment. We observed that type I interferon-dependent induction of cytokines was responsible for NK activation, with depletion of NK cells leading to increased tumor growth as well as expansion of CD4⁺CD25⁺Foxp3⁺ T regulatory (Treg) lymphocytes. Our study therefore delineates the importance of IFNAR-dependent functions in TLR3-mediated tumor suppression and supports the use of TLR3 agonists for prostate cancer immune-based therapies.

Introduction

In advanced prostate cancer, the failure of hormone therapy or chemotherapy to provide durable responses has forced investigations into other modalities such as tumor immunotherapy. Although the mechanisms of tumor surveillance remain unclear, innate immune recognition of tumors and subsequent activation of adaptive immunity within the microenvironment may modulate tumor growth. Areas of malignant glands in prostate cancer are frequently juxtaposed to inflamed areas of prostate tissue, implying that either inflammation propagates the cancer or is responding to it (1). Potential candidate receptors in regulating these inflammatory processes include the family of TLRs. The TLR family of pattern-recognition receptors (PRRs) consists of 13 mammalian receptors that recognize conserved microbial motifs, or pathogen-associated molecular patterns (PAMPs) specific for microbial components (2). Predominantly expressed on innate immune cells, TLRs recognize these motifs and trigger innate immune activation with a subsequent direction of adaptive immunity. Several endogenous ligands such as heat-shock proteins, the chromatin component HMGB1, and other components of injured cells collectively termed danger-associated molecular patterns (DAMPs), have also been characterized suggesting a role for this receptor family in inflammatory responses resulting from tissue damage even in the absence of infection, such as the transformed cell (3).

Most TLRs interact with and recruit the adaptor proteins myeloid differentiation primary response gene 88 (MyD88) and then serine kinase IL-1 receptor-associated kinase (IRAK), leading to activation of MAP kinases, NF-κB, and expression of inflammatory genes (2). Negative regulatory mechanisms exist to balance the robust immune activation to prevent tissue damage and autoimmunity. This includes negative regulation of intracellular signaling pathways at the level of IRAK by IRAK-M or through activation of Treg cells (4–6). TLR3 is unique in that it signals exclusively through MyD88-independent pathways and activates IRF-3, thereby inducing type I interferons and IFN-inducible genes. This occurs through the adaptor protein TIR-domain-containing adapter-inducing interferon-β (TRIF) also known as toll-like receptor adaptor molecule 1 (TICAM-1). Best known for the regulation of type I interferons in anti-viral responses, TLR3 has also been implicated in NK cell activation and dendritic cell maturation (7–9). Whereas MyD88-dependent pathways largely regulate cytotoxic T lymphocyte (CTL) induction, natural killer (NK) activation relies on TRIF-dependent pathways (8).

The role of TLRs in mediating immune surveillance is conflicting, as prior reports have suggested tumor promoting as well as suppressing effects. Deficiency in MyD88 confers decreased development of tumors in a mouse model of spontaneous intestinal tumorigenesis and diethylnitrosamine-induced hepatocellular tumors, whereas deficiency in IRAK-M also impairs growth of implanted tumor cells, although these adaptor molecules have opposing effects on TLR signaling (10–12). Nonetheless, TLR agonists have been used as adjuvants to enhance host immunity; with the TLR7 agonist imiquimod approved for use in treatment of basal cell cancer and TLR9 agonists in phase I–III trials against multiple malignancies including breast, melanoma, and lymphomas (13). Although growth of tumors in the absence of TLR3 has not been studied, the absence of type I interferons in IFNAR1^−/− mice resulted in increased melanomas in a syngenic mouse model (14).

In this study, we examined the role of MyD88-independent pathways in tumor surveillance and showed increased growth of tumors in the absence of TLR3. Exploring the ability of polyI:C to suppress prostate cancer growth, we demonstrated growth retardation of autochthonous TRAMP tumors through both IFN-dependent and -independent pathways. TLR3-dependent activation of NK cells may in part mediate this effect, while depletion of NK cells resulted in an induction of T regulatory lymphocytes. This work further implicates and highlights TLR3 as a mediator of tumor immune surveillance and provides mechanistic insight in the balance between inflammation and immune tolerance.

Materials and Methods

Mice

TRAMP Tg^+/− mice on a C57Bl6 background (Jackson Labs) were generated and genotyped as previously described (15). For the assessment of autochthonous TRAMP tumors by polyI:C, 10 week old TRAMP C57Bl6 × FvB F1 Tg^+/− males were administered intraperitoneal saline or 250 µg of polyI:C (Sigma) reconstituted in 100 µL sterile normal saline weekly for six consecutive weeks. TLR3^−/− mice backcrossed to a C57Bl6 background over 10 generations and IFNAR1^−/− mice backcrossed to a C57Bl6 background for six generations were generated and genotyped as previously described (gift from Shizuo Akira, (16)). Male six to eight week old mice were used for experiments. Age-matched IFNAR1^+/+ male littermates and C57Bl6 wild-type males for TLR3^−/− mice were used as controls. All mice were kept in SPF conditions in the UCLA-DLAM facilities according to ARC protocols.

Cell Culture

TRAMP C2 cell lines (ATCC) were grown in DMEM supplemented with 0.005 mg/ml bovine insulin, 10 mM DHEA, 5% fetal bovine serum, and 5% NuSerum IV (Gibco BRL). For cell viability determination, 4000 cells were plated in each well of a 96-well plate overnight and treated with incremental concentrations of polyI:C for 48 hours. Incorporation of MTT (Sigma) was assessed by spectroscopy at 600 λ and compared to a standard curve.

Subcutaneous Tumor Implantation Model

TRAMP C2 cells grown to log phase were harvested with trypsinized media and washed in DMEM twice. Following subcutaneous implantation of 5 × 10⁶ cells in 100 µL normal saline in the shaved flank of male mice, tumor size was determined by measuring the length, width, and depth of tumor and multiplying the product by pi/6 to estimate tumor volume in an ellipsoid shape. To assess tumor regression by polyI:C, mice were administered intraperitoneal saline or 250 µg of polyI:C weekly for four consecutive weeks starting on day five following tumor cell implantation.

Quantitative PCR

To assess induction of gene expression, mice implanted with TRAMP cells were subject to four weekly treatments of saline or 250 µg of polyI:C, and sacrificed four hours prior to the final treatment. RNA from splenocytes prepared by Trizol (Invitrogen) per manufacturer’s instructions was used to generate cDNA for quantitative real time PCR (qPCR) using iScript RT-PCR (Bio-Rad). A Bio-Rad iCycler was used to analyze the samples under the following conditions: 95°C (5 min), 55 cycles of 95°C (20 sec), 55°C (30 sec), and 72°C (10 sec). Expression of GAPDH was used for normalization. The qPCR reaction total volume was 19 µl and consisted of 2 µl cDNA, 8 µl H20, 8 µl of 2xSYBR Green Master Mix (Applied Biosystems), and 1 µl of a 10 µM oligo set for the following genes: IL-6 5’ CACAGAGGATACCACTCCCAACA, 3’ TCCACGATTTCCCAGAGAACA; Mx-1 5’ AAACCTGATCCGACTTCACTTCC, 3’ TGATCGTCTTCAAGGTTTCCTTGT; IP-10 5’ CCAGTGAGAATGAGGGCCATA, 3’ TCGTGGCAATGATCTCAACAC; IL-15 5’ CACTTTTTAACTGAGGCTGGCATT, 3’ TCCAGTTGGCCTCTGTTTTAGG; IL-1β 5’ GAGCTGAAAGCTCTCCACCTCA, 3’ TCGTTGCTTGGTTCTCCTTGTAC; TNFα 5’ GGTGCCTATGTCTCAGCCTCTT, 3’ CGATCACCCCGAAGTTCAGTA; TGFβ 5’ GCTGCTGACCCCCACTGATA, 3’ CGTCTCCTTGGTTCAGCCAC; IL-12 5’ CCCATTCCTACTTCTCCCTCAA, 3’ CCTTTCTGGTTACACCCCTCCT; GAPDH 5’ ACTCCACTCACGGCAAATTCA, 3’ CGCTCCTGGAAGATGGTGAT.

Intracellular Staining and Flow Cytometry

Single cell suspensions prepared from spleens were used for flow cytometry. Intracellular staining was performed according to manufacturer’s instructions (BD Bioscience). Splenocytes were stained with CD4 (RM4-5), CD25 (PC61.5), and control IgG2a and IgG1 antibodies at 1:500 dilution for 20 min, and stained with FoxP3 (FJK-16s) or IgG2a antibodies at 1:200 dilution for 30 min. To assess induction of effector cells following polyI:C stimulation in vivo, age-matched 6 week old C57Bl6 male mice were administered an intraperitoneal dose of 250 µg of polyI:C at 0, 24, and 48 hours prior to harvest. Labeled splenocytes were analyzed using BD FACSCalibur flow cytometry and CellQuest Pro software. Antibodies include anti-mouse CD8 (53–6.7), CD4 (RM4-5), CD25 (PC61.5), Foxp3 (FJK-16s), NK1.1 (PK136), and respective IgG2a and IgG1 controls (eBioscience).

Antibody Depletion

NK cell depletion monoclonal antibody clone PK136 was expanded in hybridoma media (Gibco BRL) and purified by precipitation in 50% saturating ammonium sulfate followed by dialysis against PBS. Following implantation of TRAMP C2 tumor cells in six week old male C57Bl6 mice, groups of mice were administered 100 µg IgG control antibody or 100 µg PK136 by intraperitoneal injection starting on day two, three, and four after tumor challenge and weekly starting at day 11. Subsequently, intraperitoneal saline or 250 µg of polyI:C was administered weekly for four consecutive weeks starting on day five, with mice sacrificed four hours following the final treatment. Splenocytes were analyzed by flow cytometry, while cytokine levels were measured from serum. NK cell depletion was confirmed by flow cytometry to be greater than 95% (data not shown).

Enzyme-linked immunosorbent assay

Serum harvested from cardiac puncture of euthanized animals was used for detecting circulating levels of TGFβ and IL-6 per manufacturers instructions (R&D Systems).

Immunohistochemistry

Prostate and seminal vesicles were extracted and fixed in 10% buffered formalin, paraffin imbedded, and sectioned at 5 µm. Hematoxylin and eosin staining was performed and slides visualized using an Olympus BX41 microscope and captured using an Olympus DP71 digital camera and software. For immunohistochemistry, paraffin embedded sections were dissolved in xylene and rehydrated in graded ethanol. For anti-CD3 (A0452, Dako) and anti-NK (Asialo GM1, Cedarlane) staining, antigen retrieval by steaming in 0.01 M sodium citrate ph 6.0 at 95° C for 25 minutes was followed by primary antibody staining at 1:100 and 1:1000 dilution respectfully for 60 minutes. Secondary antibody staining was performed using an anti-rabbit HRP labeled polymer (Dako) for 30 minutes. Anti-CD45R (RA3-6B2, eBioscience) staining was completed at a 1:50 dilution for two hours. For anti-Foxp3 (FJK-16s, eBioscience) staining, antigen retrieval consisting of steaming in 0.001 M EDTA pH 8.0 at 115°C for three minutes was completed followed by primary antibody staining at a 1:1000 dilution for one hour. Secondary antibody staining for both anti-CD45R and anti-Foxp3 included two-steps consisting of a 30 minute incubation using a 1:200 dilution of a rabbit anti-rat antibody (E0468, Dako), followed by incubation with an anti-rabbit HRP labeled polymer (Dako) for 30 minutes. Staining was completed by incubation with 3, 3’-diaminobenzidine for 10 minutes and hematoxylin counterstaining.

Results

TLR3 Regulates Prostate Tumor Growth

To examine the role of TLR3 in tumor surveillance, we first implanted syngenic tumor cells in TLR3^−/− mice and TLR3^+/+ counterparts to observe tumor growth in the absence of exogenous stimulation of TLR3. Tumors from subcutaneously implanted TRAMP cells in TLR3^−/− mice grew markedly larger compared to TLR3^+/+ mice (Fig 1a). Since tumors exhibited enhanced basal growth in the absence of TLR3, we then assessed the therapeutic potential of TLR3 agonists. Ligands to TLR3 include dsRNA or the synthetic polymer polyI:C. Serial intraperitoneal administration of polyI:C in mice with subcutaneously implanted TRAMP cells completely suppressed tumor growth, using a dose previously shown to induce NK activity and be well tolerated (Fig 1b) (8).

Figure 1.

Modulation of TRAMP tumor growth by TLR3. A, subcutaneously implanted TRAMP C2 cells in TLR3^+/+ or TLR3^−/− mice were assessed for tumor growth. Each treatment group represents a mean of three mice with standard deviations shown. Data shown is representative of two independent experiments. B, wild-type C57Bl6 mice or C, TLR3^−/− mice subcutaneously implanted with TRAMP C2 cells were challenged with intraperitoneal saline or polyI:C on days five, 12, 19, and 26 following implantation and assessed for tumor growth. Each treatment group represents a mean of three mice with standard deviations shown. Data shown is representative of two independent experiments. D, TRAMP C2 cells stimulated with polyI:C for 48 hours as indicated were assessed for viability by MTT incorporation. Samples were performed in quadruplicate with mean and standard deviations as shown. Data shown is representative of two independent experiments. *, p < 0.06 for polyI:C at 0 versus 10 µg/ml. P values were determined by two-tailed Student’s t-test.

PolyI:C signals through TLR3 to activate NF-κB, MAPK, and IRF-3 to induce inflammatory mediators, but may also activate alternative receptors such as cytoplasmic PRRs retinoic acid-inducible protein I (RIG-I) and melanoma differentiation-associated protein 5 (MDA5) (17). Furthermore, the tumor suppression by polyI:C may be induced not only by acting on immune cells, but also through direct action of polyI:C on TRAMP cells. To investigate whether polyI:C therapy works through host immune cells, we examined the effect of polyI:C treatment on the tumor growth of implanted TRAMP cells, containing a TLR3^+/+ background, in TLR3^−/− hosts. Only a limited suppression of tumor growth was seen with sequential treatments by polyI:C, supporting a major contribution of host TLR3 in modulating tumor growth (Fig 1c). To further investigate the direct role of polyI:C on prostate cancer epithelium, TRAMP C2 cells were stimulated with incremental doses of polyI:C in vitro and proliferation was assessed. Statistically insignificant reductions in proliferation occurred even at the highest doses of polyI:C, supporting minimal direct effects of polyI:C on prostate epithelial tumor cells in our model (Fig 1d).

Suppression of Autochthonous Prostate Cancer Growth in polyI:C-treated Mice

To further assess polyI:C-induced tumor suppression activity in vivo, we used the established murine autochthonous prostate cancer TRAMP model (15). When crossed with FVB mice, the TRAMP C57Bl6 × FVB F1 Tg^+/− transgenic mice develop a more rapidly progressing phenotype with consistent development of localized adenocarcinoma by 12 weeks. In our studies, 10 week old mice were subject to weekly intraperitoneal administration of polyI:C and sacrificed six weeks after initiation of treatment. Examination revealed no gross metastases in the liver, spleen, or lungs in all animals (data not shown). When compared to saline treated animals, the prostate, bladder, and seminal vesicles in the polyI:C treated group were approximately half the weight at 16 weeks (Fig 2a). Furthermore, prostate and seminal vesicles from polyI:C treated mice at 16 weeks appeared similar in size to glands from animals of pre-treated age at 10 weeks (Fig 2b).

Figure 2.

Suppression of autothchonous prostate cancer growth by polyI:C. Prostate and seminal vesicles from 10 week old C57Bl6 × FvB F1 Tg^+/− transgenic mice, and 16 week old mice treated with saline or polyI:C weekly for six weeks were examined for A, weight, B, gross examination, and C, histology with hematoxylin and eosin staining at 20 times magnification. Representative examples are shown. Each treatment group represents a mean of three mice with standard deviations shown. Data shown is representative of two independent experiments.

Under hematoxylin and eosin staining, 10 week old TRAMP C57Bl6 × FVB prostates contained normal glands juxtaposed with areas of well differentiated adenocarcinoma. At 16 weeks, prostates from TRAMP C57Bl6 × FVB mice showed progression to moderate and poorly differentiated adenocarcinoma with areas involving sheets of anaplastic cells. Strikingly, following six weekly treatments with polyI:C, minimal progression compared to 10 week old TRAMP C57Bl6 × FVB prostates was observed (Fig 2c).

Type I Interferon-dependent and –independent Requirements in polyI:C-mediated Tumor Inhibition

TLR3 is unique to TLRs in that it transduces its signal exclusively through TRIF to activate both NF-κB and IRF-3 (18). To differentiate between type I interferon-dependent and -independent functions, we examined the growth of tumors from subcutaneously implanted TRAMP cells in IFNAR1^−/− mice and IFNAR1^+/+ counterparts and found increased growth in the absence of IFNAR1 (Fig 3a). In addition, IFNAR1^−/− mice exhibited reduced response to polyI:C- mediated suppression in the growth of subcutaneous TRAMP tumors (Fig 3b).

Figure 3.

Role of type I interferons in polyI:C-mediated tumor suppression. A, IFNAR1^+/+ and IFNAR1^−/− mice subcutaneously implanted with TRAMP C2 cells were assessed for tumor growth. Data are means with standard deviations from three to five mice in each group. This experiment was repeated four times with consistent results. B, IFNAR1^−/− mice subcutaneously implanted with TRAMP C2 cells were challenged with intraperitoneal saline or polyI:C on days five, 12, 19, and 26 following implantation and assessed for tumor growth. Data are means with standard deviations from three to five mice in each group, and is representative of four independent experiments. C, quantitative RT-PCR performed for the indicated genes on splenocytes from TRAMP-tumor bearing IFNAR1^+/+ and IFNAR1^−/− mice treated with saline or polyI:C weekly for four weeks with the final stimulation four hours prior to sacrifice. Data are means with standard deviations from three mice in each group, and representative of two independent experiments. All p values were determined by two-tailed Student’s t-test.

To further investigate type I interferon-dependent mechanisms, we next examined the gene expression of a panel of inflammatory genes in vivo from IFNAR1^+/+ and IFNAR1^−/− mice with subcutaneously implanted TRAMP tumors in response to intraperitoneal polyI:C stimulation after four hours (19). Categorizing genes based on dependency on type I interferons may give insight into TLR3-mediated effector functions. We found induction of IL-6, myxovirus resistance-1 (Mx-1), interferon-inducible protein-10 (IP-10), and IL-15 by polyI:C to be type I interferon-dependent, while expression of IL-1β, TNFα, and IL-12 was independent of type I interferons. We also examined the gene induction of TGFβ, a cytokine important in the development of T regulatory cells. Minimal gene induction was observed at four hours following challenge by polyI:C with and without IFNAR1 (Fig 3c).

Regulation of CD4⁺ CD25⁺ T Regulatory Cells by NK cells

Previous studies have implicated NK cell activation in response to TLR3 stimulation to be largely TRIF-mediated (8). Consistent with this finding, we found strong type I-interferon dependent induction of IL-15, a potent activator of NK and T cells (20). This appears specific to TLR3, as no splenocytic induction of IL-6 and IL-15 in response to polyI:C was observed in TLR3^−/− mice (data not shown). To confirm the ability of polyI:C to induce NK cells in vivo, we stimulated C57Bl6 mice with polyI:C and examined levels of splenocytic CD4⁺, CD8⁺, NK, and CD4⁺CD25⁺Foxp3⁺ Tregs at 0, 24, and 48 hours after stimulation. Increased levels of splenocytic NK cells were observed at 48 hours after polyI:C stimulation (Fig 4a). To assess the relevance of NK cells in our model of polyI:C-mediated tumor suppression, we examined the response to polyI:C upon NK cell depletion using monoclonal antibodies. Significantly enhanced tumor growth compared to polyI:C alone was observed with depletion of NK cells (Fig 4b). This confirmed previous data supporting a major role for NK cells in TLR3-mediated anti-tumor effects (8).

Figure 4.

Contributions of NK cells in polyI:C-mediated tumor suppression. A, Groups of four C57Bl6 male mice were treated with polyI:C for 0, 24, and 48 hours and assessed for splenocytic induction of CD4⁺, CD8⁺, NK, and CD4⁺CD25⁺Foxp3⁺ (Treg) cells by flow cytometry. *, p < 0.01 for NK levels at 0 compared to 48 hours. B–D, Groups of three to four C57Bl6 male mice with subcutaneously implanted TRAMP cells were treated with saline or polyI:C on days five, 12, 19, and 26 following implantation. Additionally, groups of polyI:C treated mice were administered control IgG or NK depleting antibodies. B, mice were assessed for tumor growth. P value represents data between polyI:C and control IgG versus polyI:C and anti-NK antibody. C, serum was used to determine levels of cytokines by ELISA as indicated. D, percentage of CD4⁺CD25⁺Foxp3⁺ cells in total splenocytes was assessed by flow cytometry. *, p < 0.02 for NK levels treated with saline versus polyI:C and anti-NK antibodies. Each treatment group represents a mean of three mice with standard deviations shown. Data shown is representative of four independent experiments. P values were determined by two-tailed Student’s t-test.

Prior reports have implicated cross-regulation between NK cells and Tregs. Tregs have been shown to inhibit dendritic cell-mediated activation of NK cells, while others have linked a direct inhibitory effect on NK cells to a TGFβ dependent mechanism (21, 22). Although IFNγ has been implicated in inhibiting Treg development, little is know how activation or depletion of NK cells influences Tregs (23). When we examined the serum in mice depleted of NK cells in our tumor suppression model, we found elevated levels of TGFβ and lower levels of IL-6 from mice depleted of NK cells and stimulated with polyI:C compared to mice stimulated with polyI:C alone (Fig 4c). Consistent with prior gene induction results, no significant elevation of TGFβ levels were observed with polyI:C treatment alone. This suggested that following depletion of NK cells, the systemic environment favors the development of Tregs. To investigate further, we examined the levels of CD4⁺CD25⁺Foxp3⁺ cells in the splenocytes of mice with and without NK depletion in our polyI:C-induced tumor suppression model. As expected, no statistically significant increase in Tregs was observed with polyI:C stimulation alone, but elevated Treg levels were detected with the combination of polyI:C stimulation and NK cell depletion compared with no treatment (Fig 4d).

Autochthonous Prostate Cancer Inflammatory Infiltrate in polyI:C-treated Mice

In order to correlate the systemic effects with the tumor microenvironment, we examined the autochthonous TRAMP tumors for infiltration of inflammatory cells by immunohistochemistry. We observed that at 10 weeks, prostates from TRAMP C57Bl6 × FvB F1 Tg^+/− mice exhibited small numbers of infiltrating CD3⁺ and NK cells, with no detectable CD45R and Foxp3 staining cells. As the tumors progressed at 16 weeks, similar numbers of inflammatory infiltrates were present. Notably, with weekly polyI:C stimulation, a marked influx of CD3⁺ cells within both the prostate glands as well as stroma, and an increase in NK cells in the prostatic stroma were observed, with no detectable CD45R and rare Foxp3 staining cells (Fig 5).

Figure 5.

Inflammatory infiltrate in polyI:C-treated autothchonous prostate cancer. Infiltration of CD3, CD45R, NK, and Foxp3 expression cells as indicated were performed by immunohistochemistry in prostates from 10 week old, and 16 week old C57Bl6 × FvB F1 Tg^+/− transgenic mice treated with saline or polyI:C weekly for six weeks. Spleens were used for positive controls. Arrows point out rare positive staining cells for respective antibodies. Representative examples from each group of three mice are shown at 20 times magnification.

Discussion

Originally described as receptors on immune cells responsible for sensing pathogens, TLRs have now been implicated in broader biologic functions. Using the TRAMP orthotopic and subcutaneous tumor models, we provide evidence for TLR3 as a sensor in tumor immune surveillance, with the absence of TLR3 leading to accelerated tumor growth without exogenous stimulation of the receptor. Consistently, activation of TLR3 with exogenous ligand suppressed growth of TRAMP tumors. In orthotopic prostate tumors, serial polyI:C treatments leads to recruitment of T lymphocytes and NK cells within the tumor microenvironment resulting in a striking suppression of tumor growth both in gross size as well as histologic progression of adenocarcinoma. We demonstrated that polyI:C contributes to tumor suppression through a type I interferon-dependent activation of NK cells, as well as implicate a cross regulation of Treg lymphocytes by NK cells.

In our studies, increased growth of wild-type TRAMP tumors in TLR3^−/− compared to TLR3^+/+ mice implicates TLR3 expressed by immune cells in mediating tumor surveillance. The question of how TLRs sense tumor cells is the subject of ongoing investigations, potentially involving DAMPs released from tumor cells, such as components of tumor mRNA (24, 25). Although we did not observe significant differences in TRAMP cell growth upon direct stimulation with polyI:C, the expression of multiple TLRs has been shown on prostate epithelial cells and recent in vitro studies suggested the potential for direct apoptotic effects of TLR3 stimulation (26). PolyI:C may induce autophagy in tumor cells by expression of inflammatory cytokines or apoptosis through a RIP/FADD/Caspase8-dependent pathway (27). Furthermore, reports have shown pro-tumor effects of TLR activation on tumor cells by enhancing immune evasion or promoting tumor growth through tissue repair mechanisms (3, 10, 28). In humans, a sequence variant in a 3’-untranslated region of TLR4 as well as polymorphisms in the TLR gene cluster encoding TLR1, 6, and 10, and the downstream signaling mediators IRAK1 and IRAK4 confer increased prostate cancer risk (29–31). How these polymorphisms affect TLR signaling within the tumor microenvironment have yet to be determined. Whether these mutations affect tumor initiation or progression of cancers will also be the subject of future studies.

Although classically involved in anti-viral responses, type I interferons have been implicated in immune surveillance of tumors with increased tumor formation in MCA-induced sarcomas and growth of syngenic tumor cells in IFNAR1^−/− compared to IFNAR1^+/+ mice (14, 32). Endogenous IFN responsiveness in hematopoietic cells has been demonstrated to be important for anti-cancer immune responses (33). The lack of type I interferon responsiveness in IFNAR1^−/− mice encompasses both MyD88-independent type I IFN production by TLR3 and 4 through TRIF as well as MyD88-dependent type I IFN production by TLR7 and 9 through IRF-7. However, the partial tumor suppression by polyI:C in IFNAR1^−/− mice compared to IFNAR1^+/+ mice still implicates the importance of IFN-independent pathways. We also show a major contribution of NK cells in polyI:C-mediated tumor suppression, which may also potentiate the effects of antigen-specific CD8⁺ cells (34).

To further discriminate the potential mechanisms of IFN-dependent and -independent effector functions, we surveyed the expression of various genes in our tumor model in IFNAR1^+/+ and IFNAR1^−/− mice. These genes can be categorized into families that are dependent and independent on type I interferons. We showed IFNAR1-dependent gene induction of IL-6, IL-15, IP-10, and Mx-1, supporting NK cell proliferation, and recruitment of activated T and NK cells to areas of inflammation, by IL-15 and IP-10 respectively. In contrast, gene expression of IL-1β, TNFα, and IL-12 was induced by polyI:C independent of type I interferons, suggesting a role for these cytokines in mediating tumor toxicity in the type I interferon-independent pathways.

We showed that depletion of NK cells led to an increase in splenic Tregs, demonstrating a balance between immune activation and immune suppression following the acute activation of NK cells. This suggests that either inhibition of Tregs or direct activation of NK cells may enhance immune-based tumor therapies. Suppression of Tregs may be sufficient to break tolerance or induce anergy to self-antigens and points to the importance of type I interferons in the development of immune-based cancer therapy (35). In patients with systemic lupus erythematosus, IFNα secreted by antigen presenting cells (APCs) may potentially inhibit Treg activity (36). This is in accordance with clinical observations of elevated levels of Treg cells in patients with prostate cancer, and suppression of human prostate tumor infiltrating CD8⁺ Treg cells by TLR8 in vitro (37, 38). Alternatively, our findings also suggest that persistent activation of NK cells may suppress the development of Tregs. However, following pathogen challenge, induction of Treg cells may play a critical role in protecting against tissue injury and inhibiting autoimmunity (6). Thus, activation of TLRs in cancer therapy may potentially interfere with normal tissue homeostasis and will need to be carefully monitored.

Our future studies will also focus on the tumor microenvironment, which may differ from systemic activation of cytokines and splenocyte effector cells. Although in response to polyI:C we did not observe a decrease in circulating TGFβ levels or Treg levels, Tregs may still be influenced at the tumor microenvironment. We have shown recruitment of T lymphocytes and NK cells within the glands and stroma of the prostate upon serial polyI:C therapy. Studies will need to define the kinetics of such infiltration to the tumor microenvironment and use more sensitive approaches to follow the activation of Tregs and local cytokine levels.

Emerging studies indicate that TLR activation of multiple inflammatory pathways through the recognition of both pathogen-associated molecular patterns and endogenous ligands may modulate the development and progression of cancer. We have provided evidence supporting a role of TLR3 in promoting tumor immune surveillance, and the use of TLR3 agonists as immunotherapy in prostate cancer. While other TLRs may contribute to tumor progression by overly enhancing inflammation through MyD88-dependent signaling pathways, TLR3 may provide an optimal balance of type I interferon induction and inflammatory cytokine activity. Thus, comparison of type I interferon and pro-inflammatory effects of different TLRs in tumor settings deserves further study. With the caveats of direct stimulation and activation of cancer cells by TLRs and of disrupting systemic regulation of immune functions, we believe that defining the TLR-dependent pathways responsible for immune surveillance will contribute to developing novel therapeutics. At the same time, use of TLR3 agonists may be promising in suppressing progression of prostate cancer.