The NLRP3 inflammasome acts as a danger signal sensor that triggers and coordinates the inflammatory response upon infectious insults or tissue injury and damage. However, the role of the NLRP3 inflammasome in natural killer (NK) cell–mediated control of tumor immunity is poorly understood. Here, we show in a model of chemical-induced carcinogenesis and a series of experimental and spontaneous metastases models that mice lacking NLRP3 display significantly reduced tumor burden than control wild-type (WT) mice. The suppression of spontaneous and experimental tumor metastases and methylcholanthrene (MCA)-induced sarcomas in mice deficient for NLRP3 was NK cell and IFN-γ–dependent. Focusing on the amenable B16F10 experimental lung metastases model, we determined that expression of NLRP3 in bone marrow–derived cells was necessary for optimal tumor metastasis. Tumor-driven expansion of CD11b+Gr-1intermediate (Gr-1int) myeloid cells within the lung tumor microenvironment of NLRP3−/− mice was coincident with increased lung infiltrating activated NK cells and an enhanced antimetastatic response. The CD11b+Gr-1int myeloid cells displayed a unique cell surface phenotype and were characterized by their elevated production of CCL5 and CXCL9 chemokines. Adoptive transfer of this population into WT mice enhanced NK cell numbers in, and suppression of, B16F10 lung metastases. Together, these data suggested that NLRP3 is an important suppressor of NK cell–mediated control of carcinogenesis and metastases and identify CD11b+Gr-1int myeloid cells that promote NK cell antimetastatic function. Cancer Res; 72(22); 5721–32. ©2012 AACR.
NOD-like receptors (NLR) are a class of sentinel receptors that are pivotal in the detection of pathogenic microbes and danger signals (1). NLRs have been identified to play a role in various diseases, such as autoinflammatory diseases, autoimmune diseases, GVHD, and bacterial or viral infections (2). In response to danger signals, some of the NLRs interact with the adaptor molecule apoptosis-associated speck-like protein (ASC) and procaspase-1 to form the inflammasome that stimulates caspase-1 activation, triggering the maturation and secretion of proinflammatory cytokines interleukin-1β (IL-1β) and IL-18 (3).
The NLRP3 inflammasome is currently the best-characterized inflammasome. NLRP3 is expressed by nonimmune cells, such as epithelial cells (4), and antigen-presenting cells (APC) defined by the myeloid cell marker CD11b, including neutrophils, macrophages, monocytes, and conventional dendritic cells (5). The pathogenic microbes capable of activating the NLRP3 inflammasome include the fungi Candida albicans (6), pore-forming toxins producing bacteria such as Staphylococcus aureus (7), and viruses including influenza (8–10) and adenovirus (11). In addition to pathogens, the NLRP3 inflammasome can be activated by endogenous danger signals, such as ATP and monosodium uric acid that are released as a result of tissue damage or injury (12, 13). Indigestible particulate substances, such as silica and asbestos, which can cause pulmonary fibrosis and mesothelioma, have been shown to induce the activation of the NLRP3 inflammasome, triggering chronic inflammation (14). Recent studies showed that the NLRP3 inflammasome is emerging as a key player in a broad range of diseases, such as type II diabetes (15), gout (13), atherosclerosis (16), and ulcerative colitis (4, 17–19), but its role in tumorigenesis (17, 20) and stimulating antitumor immunity (21, 22) has only recently begun to be explored and seems complex.
Natural killer (NK) cells control the initiation of methylcholanthrene (MCA)-induced sarcomas (23) and B16F10 experimental metastases (24), yet the role of the host NLRP3 inflammasome in NK cell–mediated control of tumor initiation, growth, and metastasis is unknown. Here, we showed that NLRP3 inflammasome signaling is required for NK cell–mediated control of MCA-induced carcinogenesis and experimental primary tumor growth and metastases. In particular, we show that loss of NLPR3 facilitates the intratumoral accumulation of CD11b+Gr-1int myeloid cells that promote NK cell antimetastatic function.
Materials and Methods
C57BL/6 wild-type (WT) mice were purchased from the Walter and Eliza Hall Institute of Medical Research (Melbourne, Australia) and maintained at the Peter MacCallum Cancer Centre (Peter Mac, Melbourne, Australia). C57BL/6 NLRP3-deficient (NLRP3−/−) mice (13) were bred and maintained at the Peter Mac. Mice 6 to 14 weeks of age were used in all experiments according to Peter Mac Animal Experimental Ethics Committee guidelines.
B16F10 melanoma, RM-1 prostate adenocarcinoma, and E0771 mammary adenocarcinoma cell lines were cultured in Dulbecco's Modified Eagle Medium supplemented with 10% heat-inactivated fetal calf serum (FCS), 2 mmol/L glutamax, 100 U/mL penicillin, 100 μg/mL streptomycin, and were sourced from American Type Culture Collection (tested by Standard Tandem Repeat profiling) or Dr. Robin Anderson (Peter MacCallum Cancer Centre, East Melbourne, Australia) as previously described (25, 26).
MCA-induced fibrosarcoma model
WT or NLRP3 mice were injected subcutaneously with 100 μL of corn oil containing MCA (25, 100, or 400 μg) on the right hind flank. Some WT and NLRP3−/− mice were treated with control Ig (cIg) or depleted of NK cells by treatment with anti-asialoGM1 (Wako Chemicals; 100 μg injected intraperitoneally (i.p.) at day −1, 0, and then weekly until week 8) or neutralized for IFN-γ (H22, 250 μg injected i.p. at day −1, 0, and then weekly until week 8). Development of fibrosarcomas was monitored weekly over the course of 300 days. Measurements were made with a caliper square as the product of 2 perpendicular diameters (mm2) and individual mice are represented.
Experimental tumor models
In experimental metastasis experiments, WT or NLRP3−/− mice were injected intravenously with B16F10 melanoma or RM-1 prostate carcinoma at different doses, from 1 × 104 to 5 × 105 cells. Mice were sacrificed and lungs were harvested on day 14 postinoculation. For spontaneous metastases, the indicated doses of E0771 mammary carcinoma cells were injected into the fourth mammary fat pad of WT or NLRP3−/− mice. Primary tumor growth was measured every 4 and 20 days after tumor inoculation, the primary tumors (mean = 0.68–0.80 cm2 in both groups of mice) were resected. Mice were sacrificed at day 35 and lungs harvested and the colonies counted and recorded for individual mice as shown. Some mice in these experiments were treated with cIg, anti-CD4 (GK1.5), anti-CD8α (53.6.7) or anti-CD8β (53.5.8), anti-NK1.1 (PK136), or anti-asialoGM1, as indicated in the legends, to deplete cell subsets as previously described (25). Tumor nodules were counted with the aid of a dissection microscope.
After euthanizing mice, lungs were perfused with PBS via the right ventricle to remove peripheral blood. The perfused lungs were removed, minced finely and incubated in an enzyme solution containing 1 mg/mL collagenase type IV (Sigma-Aldrich), 0.1 mg/mL hyaluronidase (Sigma-Aldrich), and 0.02 mg/mL DNase I (Roche) for 45 minutes at 37°C. Digested lung tissue was passed through a 70 μmol/L cell strainer and washed with PBS. Single-cell suspensions were generated and used for fluorescence-activated cell sorting (FACS) analysis or sorting.
Tumor-induced interstitial fluid isolation
Lungs were harvested from the tumor bearing mice, minced finely and incubated in PBS for 2 hours at 37°C. The supernatants were harvested for cytokines/chemokines measurement using Proteome Profiler Array kits (R&D Systems) or Cytometric Bead Array kits (BD Biosciences), according to the manufacturers' instructions.
Single-cell suspensions obtained from lungs of tumor-bearing mice were analyzed using FACS LSR II (BD Biosciences). Cells were first stained with anti-CD16/32 (2.4G2) for 10 minutes at 4°C, then with specifically conjugated antibodies for 30 minutes at 4°C in the dark. The following anti-mouse antibodies were used in the analysis: PE-conjugated, CD69 (H1.2F3), Ly6G (1A8), CD11c (N418), I-A/I-E (2G9), CD124 (mIL4-Rm1), F4/80 (BM8), CD80 (16-10A1), c-kit (2B8), Siglec-F (E50-2440), B220 (RA3-6B2), CCR2 (475301), and CD86 (GL1); PE-Cy7–conjugated, Ly6C (AL-21); APC-conjugated, NK1.1 (PK136), CD11b (M1/70); Pacific Blue–conjugated, Gr-1 (RB6-8C5) and CD3 (145-2C11); APC-Cy7–conjugated, CD45.2 (104), biotin-conjugated, CD40 (3/23), and CD103 (2E7). Nonviable cells were excluded on the basis of staining with 7-aminoactinomycin D (BD Pharmingen) or FluoroGold (Sigma-Aldrich). Datasets were analyzed using Flowjo software (Tree Star).
Bone marrow chimera
For the reconstitution of WT or NLRP3−/− recipients, mice were lethally irradiated using a 2-step irradiation regimen of 8 Gy then 3 hours later, with 4 Gy. Bone marrow cells harvested from donor tibia and femur were i.v. transferred through the tail vein. Four chimera groups were generated as indicated (CD45.1 vs. CD45.2). Reconstituted mice were placed on antibiotic water. Six weeks after transplantation, blood samples were obtained to assess reconstitution efficiency by staining for CD45.1 and CD45.2 with PE-conjugated anti-mouse CD45.1 and PE-conjugated APC-conjugated anti-mouse CD45.2.
Adoptive transfer of NLRP3−/− CD11b+Gr-1+ cells
NLRP3−/− mice were injected intravenously with 2 × 105 B16F10 melanoma cells. On day 3, lungs from the mice were harvested and processed as described earlier. Cells were stained with APC-Cy7–conjugated CD45.2, APC-conjugated CD11b, and Pacific Blue–conjugated Gr-1 monoclonal antibodies (mAb). Nonviable cells were excluded on the basis of staining with 7-aminoactinomycin D or FluoroGold. Various populations of CD11b and Gr-1 double positive cells were isolated by using FACS Aria (BD Biosciences). WT mice were injected intravenously with 2 × 105 B16F10 melanoma cells on day 0 and CD11b+Gr-1int or Gr-1dim cells on day 0, 1, and 2. On day 14 after the tumor challenge, mice were sacrificed, lungs were harvested and the number of metastasis per lung was counted using a dissecting microscope.
Statistical analysis was conducted using GraphPad Prism software. Mann–Whitney U test or Students t test was used for statistical evaluations as indicated in each experiment. Mantel–Cox log-rank test was used to evaluate statistical differences in Kaplan–Meier analysis. Data are generally shown as mean ± SEM unless otherwise stated. Value of P < 0.05 was considered significant.
NLRP3 inflammasome–deficient mice resist MCA-induced carcinogenesis
To date, no studies have illustrated the effect of the NLRP3 inflammasome signaling in carcinogen-induced tumor initiation. We used the MCA-induced fibrosarcoma model, as inflammation and immunity, and in particular host NK cells, have been shown as critical regulators of tumor initiation (23, 27, 28). Compared with WT mice, tumor incidence was reduced in NLRP3−/− mice, most evident at the highest carcinogen dose (Fig. 1A–C). A similar result was observed in the BALB/c strain background (Supplementary Fig. S1). Interestingly, NK cell depletion over the first 8 weeks significantly abrogated the relative resistance observed in NLRP3-deficient mice (Fig. 1D). In a similar manner, neutralization of IFN-γ also significantly abrogated the resistance observed in NLRP3−/− mice, with anti–IFN-γ treated WT and NLRP3−/− mice displaying similar tumor onset and incidence (Fig. 1D). Further analysis of the growth rate of tumors that emerged in the experiment depicted in Fig. 1C and D revealed that, once MCA-induced sarcomas were initiated, they had a slightly slower growth rate in NLRP3−/− and WT hosts (Supplementary Fig. S2). The growth of tumors once established also reflected a similar impact of NK cell depletion or IFN-γ neutralization on WT and NLRP3−/− mice (Supplementary Fig. S2). This data suggested that the effect of NK cells, IFN-γ, and NLRP3 were not independent.
Mice targeted for other elements of the NLRP3 inflammasome pathway, including caspase-1 and IL-1R1 were more resistant to MCA-induced carcinogenesis than WT mice at all doses of carcinogen (Supplementary Fig. S3A and S3B). Greater resistance of caspase-1−/− and IL-1R−/− mice suggested that the NLRP3 inflammasome might not be the only inflammasome that mediates the activation of caspase-1 in MCA-induced carcinogenesis. Neutralizing IL-1β or IL-1R at early times during the development of tumors (from week 0 to week 4, initiation stage) resulted in a lower percentage of mice with sarcoma compared with the control group (Supplementary Fig. S3C). However, when neutralizing IL-1β or IL-1R at a later time point (from week 4 to week 8, promotion stage), the proportion of mice with sarcoma was similar between all groups (Supplementary Fig. S3D). Importantly, depletion of IL-1α early or late did not affect the survival (Supplementary Fig. S3C and S3D), thus suggesting that IL-1β was mandatory for the initiation of MCA-induced carcinogenesis.
Antimetastatic effect of NLRP3 deficiency is NK cell–dependent
We next examined whether NLRP3 plays a role in NK cell–mediated control of experimental tumor metastasis. It is known that the suppression of B16F10 and RM-1 lung metastases in WT mice is mediated by NK cells (24, 29–31). WT mice had significantly higher numbers of lung metastases than NLRP3−/− mice following intravenous inoculation of 3 different doses of B16F10 melanoma cells (Fig. 2A), suggesting that host NLRP3 promoted experimental metastases development. Strikingly, in contrast, B16F10 grew equivalently when inoculated subcutaneously in WT and NLRP3−/− mice (Supplementary Fig. S4). Similar results were obtained using the RM-1 prostate carcinoma experimental lung metastasis model, in which at all tumor doses inoculated NLRP3−/− mice had a significantly lower number of RM-1 lung metastases than WT mice (Fig. 2B). B16F10 tumor cells were also injected i.v. into IL-18−/−, IL-1R1−/−, and caspase-1−/− mice to determine whether these molecules were important in B16F10 tumor metastasis. IL-18−/− mice had a significantly higher number of lung metastases than WT mice, whereas surprisingly IL-1R1−/− and caspase-1−/− mice had a similar number of lung metastases to WT mice (Supplementary Fig. S5A). The higher number of lung metastases in IL-18−/− mice might be due to the reported defective NK cell activity in these mice (32). Inoculation of RM-1 i.v. into WT, NLRP3−/−, and caspase-1−/− mice also supported the likelihood that NLRP3 might independently promote metastasis development (Supplementary Fig. S5B). Experiments using anti-IL-1α, anti-IL-1β, or anti-IL-1R1 neutralizing antibodies in WT mice inoculated with B16F10 or B16 tumor cells supported these findings (Supplementary Fig. S5C and S5D). The number of lung metastases in NLRP3−/− mice depleted of NK cells increased to the same level as NK cell–depleted WT mice (Fig. 2C). In both WT mice and NLRP3−/− mice, depletion of CD4+ and CD8+ T cells did not affect the development of B16F10 lung metastases (Fig. 2C). Thus, the protective effect of loss of NLRP3 against B16F10 lung metastases was NK cell–dependent, but independent of CD4+ and CD8+ T cells.
To appreciate whether NLRP3 plays a role in the control of spontaneous tumor metastasis, the E0771 mammary adenocarcinoma was orthotopically injected into the mammary fat pad of WT and NLRP3−/− mice. Tumors grew at similar rates in both groups over the first 20 days before surgical resection, suggesting that NLRP3 was not critical for primary tumor growth (Fig. 3A and B). In contrast, when spontaneous metastasis was evaluated on day 35 (15 days after surgical resection), WT mice had significantly higher number of lung metastases than NLRP3−/− mice (Fig. 3C and D). Similar to observations in the experimental B16F10 metastases model, the reduction in metastases in NLRP3−/− mice compared with WT mice was abrogated in NK cell–depleted, but not CD8+ T-cell–depleted, mice (Fig. 3C and D), whereas primary mammary tumor growth was unaffected by these same depletions (Fig. 3A and B).
Activated NK cells increase in lungs of tumor-burdened NLRP3-deficient mice
An analysis of the number and activity of NK cells in the lungs from B1F10 tumor-bearing mice at day 3 showed that lungs from tumor-bearing NLRP3−/− mice contained more NK cells than the WT mice and there were more CD69+ NK cells in the tumor bearing NLRP3−/− mice (Fig. 4A and B). This suggested that NK cell activity was enhanced in the NLRP3−/− mice and this could play a part in the suppressed lung metastases in these mice. NK cells are recognized for their ability to secrete IFN-γ and kill tumor cells and both of these effector mechanisms contribute to their control of tumor metastases (24, 33). Importantly, we were able to show that IFN-γ neutralization could significantly increase B16F10 lung metastases in both WT and NLRP3−/− mice (Fig. 4C). Furthermore, NK cells isolated from the lungs of B16F10 inoculated NLRP3−/− mice were more cytotoxic than those from WT mice (Fig. 4D). Thus, it is most likely that NLRP3 loss contributes to reduced metastases via the greater numbers of more activated NK cells that kill and secrete IFN-γ.
NLRP3 expression in hematopoietic cells is required for metastasis development
NLRP3 is not expressed in NK cells (5) and the homeostasis of NK cells is normal in the NLRP3−/− mice (Supplementary Fig. S6A and S6B). In addition, the intrinsic IFN-γ producing and cytotoxic-capacity of NK cells from WT and NLRP3−/− mice is equivalent (Supplementary Fig. S6C and S6D). Furthermore, NLRP3 is expressed not only in hematopoietic cells, but also in nonimmune cells, such as epithelial cells (4). To determine if NLRP3 expression in host hematopoietic cells or nonhematopoietic cells was important in regulating tumor metastasis, we generated 4 groups of bone marrow (BM) chimeras. All mice were reconstituted effectively with no obvious perturbations in the hematopoietic compartments (data not shown). In concert with previous results, NLRP3−/− mice receiving NLRP3−/− BM had a reduced number of B16F10 lung metatsases compared with WT mice transplanted with WT BM cells (Fig. 5A). The number of lung metastases in WT mice receiving NLRP3−/− BM cells was significantly decreased and was comparable with NLRP3−/− mice transplanted with NLRP3−/− BM cells (Fig. 5A). This suggested that NLRP3 expression in hematopoietic cells was critical in the development of experimental tumor metastasis.
CD11b+Gr-1int myeloid cells are expanded in tumor-bearing NLRP3−/− mice
The dependence of metastatic growth suppression on the hematopoietic cells in NLRP3-deficient mice indicated the lungs of tumor-bearing NLRP3−/− mice might be characterized by the presence of cells that enhance NK cell–mediated antitumor immunity. As shown in Fig. 5B, we discovered that the suppression of metastases growth in NLRP3-deficient mice was closely accompanied by the expansion of a CD11b+Gr-1int cell population, whereas the lungs from tumor-bearing WT mice contained very few of these cells. These CD11b+Gr-1int cells expressed NLRP3 (as detected by GFP in NLRP3−/− mice; Supplementary Fig. S7A) and displayed the morphology of eosinophils as indicated by Giemsa staining (Supplementary Fig. S7B). They were a distinct population from CD11b+Gr-1bright neutrophils and CD11b+Gr-1dim monocytes (Supplementary Fig. S7B). This CD11b+Gr-1int population was not found in the lymphoid organs of the same B16F10 inoculated NLRP3−/− mice or lungs of naïve NLRP3−/− mice (data not shown). We next investigated the nature of the CD11b+Gr-1int cells by analyzing cell surface markers associated with various myeloid cell types. Flow cytometry analysis showed that CD11b+Gr-1int cells were Ly6C+ F4/80+ CCR2+ MHCII+ CD11clow Ly-6G− B220− CD206− CD124− (Fig. 5C). These cells also weakly expressed CD40, CD80, CD86, CD103, and Siglec-F. These findings suggested that CD11b+Gr-1int cells shared markers with M1-like macrophages and eosinophils (34, 35). A sufficiently clear and numerous population of similar CD11b+Gr-1int cells could not be isolated, characterized, and functionally compared in WT mice.
CD11b+Gr-1int cells control the growth of B16F10 lung metastases in NLRP3-deficient mice
The association between the expansion of CD11b+Gr-1int cells and the suppression of B16F10 tumor metastasis led us to directly investigate if this myeloid cell subset was involved in the antitumor response in NLRP3−/− mice. The adoptive transfer of NLRP3−/− CD11b+Gr-1int cells into WT mice strongly suppressed the growth of B16F10 metastases (Fig. 6). However, metastases suppression did not occur in the tumor-bearing WT mice that received Gr-1dim population. No tumor cells were transferred from the tumor bearing NLRP3−/− mice into the recipient mice, as the same Gr-1dim or Gr-1int cells were transferred into naïve WT mice and B16F10 lung metastases did not develop. NK cells from naïve WT mice were cultured in the absence or presence of sorted populations of CD11b+Gr-1int cells or CD11b+Gr-1dim cells from the lungs of NLRP3−/− mice with B16F10 tumor. Neither population suppressed or enhanced NK cell–mediated cytotoxicity of YAC-1 target cells (Supplementary Fig. S8A), suggesting that CD11b+Gr-1int cells did not directly regulate NK cell effector function.
CD11b+Gr-1int cells secrete CCL5 and CXCL9
To further attempt to characterize a link between the CD11b+Gr-1int cells and NK cells in controlling lung metastases in NLRP3−/− mice, we next used an array to next compare the cytokine and chemokine production in the tumor-induced lung interstitial fluid of WT and NLRP3−/− mice (Supplementary Fig. 8B). Two chemokines, CCL5 (RANTES) and CXCL9 (MIG), emerged as being more greatly secreted in the NLRP3−/− mice and this result was validated and quantitated by measuring CCL5 and CXCL9 levels using CBA analysis (Fig. 7A). Importantly, we next showed that it was the CD11b+Gr-1dim population that produced the most CCL5 and CXCL9 among the CD11b+Gr-1+ populations in the lungs of B16F10-inoculated NLRP3−/− mice (Fig. 7B). This analysis also revealed that CD11b+Gr-1int cells produced low levels of IL-1β, IL-2, and IL-12p70 and comparatively more IL-17A than the other subsets (Supplementary Fig. S9). The CD11b+Gr-1dim cells made far more IL-6 than the other subsets. Collectively, these findings were supportive of the CD11b+Gr-1int cells promoting NK cell activation and recruitment. CCL5 and CXCL9 are chemokines known to recruit and activate NK cells (36–39) and when adoptively transferred into B16F10-inoculated WT mice the CD11b+Gr-1int cells, but not the CD11b+Gr-1dim cells, from the lungs of B16F10-inoculated NLRP3−/− mice, enhanced the frequency of NK cells within the lung (Fig. 7C).
Although it was recently shown that NLRP3 inflammasome plays a protective role in colitis-associated cancer (17, 20), its role in many other types of carcinogenesis has not been reported. In particular, despite the important cross-talk between NK cells and various APCs of the myeloid lineage, the role of NLRP3 in regulating immune response to cancer initiation and metastasis has been poorly explored. Our results here show that NLRP3 inflammasome signaling is required for NK cell–mediated control of MCA-induced carcinogenesis and experimental primary tumor growth and metastases. In particular, we show that loss of NLRP3 facilitates the intratumoral accumulation of CCL5 and CXCL9 producing CD11b+Gr-1int myeloid cells that promote NK cell antimetastatic function.
We show an important role for NLRP3 in promoting metastases in both experimental (B16F10, B16, and RM-1) and spontaneous models (EO771). Specifically, NLRP3 in hematopoietic cells was identified as a critical in the development of experimental B16F10 metastasis. In addition, using the GFP reporter function of the NLRP3−/− mice we discovered that CD11b+Gr-1int cells from B16F10 tumor-bearing NLRP3−/− mice may normally express NLRP3, accumulated in the lungs of B16F10 inoculated NLRP3−/− mice, and suppress the number of B16F10 lung metastases upon adoptive transfer into WT mice. CD11b+Gr-1int cells possess a unique myeloid cell surface, morphologic and cytokine/chemokine phenotype that has M1 macrophage and eosinophil-like features. M1-polarized macrophages have previously been shown to be involved in antitumor responses (40). We were unable to confirm any direct regulatory function of these CD11b+Gr-1int myeloid cells on naïve NK cell cytotoxic function, but their ability to produce and secrete high levels of CCL5 and CXCL9 and low levels of IL-1β, IL-2, and IL-12p70 was all consistent with a capacity to recruit and activate NK cells. Certainly, the cytotoxicity of NK cells derived from the lungs of B16F10 tumor-bearing NLRP3−/− mice was enhanced above that observed for the same NK cells from WT mice. Furthermore, NK cell frequency in the lungs of B16F10-inoculated WT mice was enhanced postadoptive transfer of CD11b+Gr-1int myeloid cells, but not CD11b+Gr-1dim cells. Exploration of the impact of CCL5 and CXCL9 on NK cell-mediated control of metastases deserves further attention. Our work now provides data on a new myeloid population that might enhance NK cell antitumor function as opposed to the heterogeneous myeloid-derived suppressor cell CD11b+Gr-1+ populations whose ability to suppress NK cell antitumor function remains disputed (41–43).
One critical outstanding issue is how the expansion of CD11b+Gr-1int cells is normally suppressed in WT mice with lung metastases development. Previous studies showed that Klebsiella pneumonia–induced macrophage cell death was dependent on NLRP3, but independent of caspase-1 (44). Thus, it is possible that danger signals released from tumor cells modulate myeloid cell death via the NLRP3 signaling pathway. Another possible mechanism includes the polarization of CD11b+Gr-1int cells to an immunosuppressive phenotype by tumor cells via NLRP3. In this light, the potential of NLRP3 to act as a switch of the M1/M2 phenotype is intriguing, as the IL-1R/MyD88 signaling pathway has been shown to be critical in the polarization of macrophages toward an immunosuppressive M2 phenotype (45). Unravelling how NLRP3 loss in the context of tumor metastases in the lung drives the expansion of tumor suppressive CD11b+Gr-1int cells will be challenging, but potentially exciting given the strong antitumor potential of small numbers of these cells when adoptively transferred.
Although NLRP3 affected the ability of host NK cells to control the tumors as experimental or spontaneous metastases, the NLRP3 inflammasome did not play a detectable role in regulating primary subcutaneous or orthotopic growth of the same established tumor cell lines. This may reflect the greater role of NK cells in suppressing metastases compared with primary tumors. Whether the CD11b+Gr-1int myeloid cells we have described here can be found in any primary tumors or can regulate T-cell–mediated effector function will be of future interest. For technical reasons, the presence and role of CD11b+Gr-1int myeloid cells was not explored in the MCA-induced fibrosarcoma model. The earliest subcutaneous lesions in which the MCA is inoculated and NK cells are active are not amenable to flow-cytometric analysis or immunohistology. The mechanism of action of NLRP3 in suppressing NK cell immune surveillance (elimination phase; ref. 46) may be distinct from that observed in a very different biologic process such as metastasis. Importantly, however, NLRP3 is regulating NK cells and the role of IFN-γ in both of these quite distinct stages of tumor development.
Interestingly, NLRP3-deficient mice were less resistant to MCA-induced fibrosarcoma than caspase-1- or IL-1R1–deficient mice, whereas B16F10 lung metastases burden was independent of caspase-1 and IL-1 cytokines. Our data are consistent with the fact that IL-1β has been shown to be involved in the pathogenesis of MCA-induced fibrosarcoma (47). We showed that IL-1β was mandatory for the initiation of carcinogenesis, but had no critical impact on the promotion stage. Given the abundance of different types of danger signals released during carcinogenesis, it is possible that multiple types of inflammasomes that require caspase-1 may work together to promote tumor growth. It was a more surprising observation that caspase-1–deficient mice displayed a comparable number of B16F10 and RM-1 metastases to WT mice. Thus, here NLRP3 might promote the tumor metastases independently of caspase-1, indicating an inflammasome-independent role of NLRP3. The role of IL-1β in metastasis development is controversial. IL-1 has been shown to have a metastasis-promoting role (48, 49) and a recent study showed that IL-1β−/− mice had increased number of metastatic nodules after B16F10 or 3LL tumor inoculation (50). However, we have been unable to reproduce these results as we found no role for IL-1 in B16F10 tumor metastases. While these observations dissecting NLRP3 from caspase-1 and IL-1β are interesting, they were not the focus of this study and will require far more follow up using other gene-targeted mice for the various NLR family members and adaptors (e.g., ASC).
The distinguishing feature of our study is that we have shown the role of NLRP3 suppression of NK cell–mediated tumor immunity. At face value this seems counterintuitive given the loss of host NLRP3 compromises the release of proinflammatory cytokines, IL-1, and IL-18, both known to contribute to the activation of NK cells. But tumor microenvironments are unique and clearly NLRP3 is normally a central regulator of cross-talk between myeloid cells and NK cells in these settings. When it is missing, new populations of myeloid cells may emerge that strikingly promote NK cell antitumor function. Intervention upon this NLRP3-dependent immunosuppressive pathway may be a potential therapeutic strategy to enhance host antimetastatic immunity.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: M.T. Chow, A. Möller, M.J. Smyth
Development of methodology: M.T. Chow
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.T. Chow, C.S.F. Wong, J. Tschopp, A. Möller, M.J. Smyth
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.T. Chow, J. Sceneay, C. Paget, A. Möller, M.J. Smyth
Writing, review, and/or revision of the manuscript: M.T. Chow, J. Sceneay, C. Paget, A. Möller, M.J. Smyth
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.T. Chow, J. Sceneay, C. Paget, C.S.F. Wong, H. Duret
Study supervision: J. Sceneay, C. Paget, A. Möller, M.J. Smyth
The authors thank Janelle Sharkey and Qerime Mundrea for technical assistance. The authors also thank Robert Schreiber for providing the anti–IL-1 mAbs.
This work was supported by the National Health and Medical Research Council of Australia (NH&MRC) Program Grant (no. 454569) and the Victorian Cancer Agency. M.J. Smyth received support from an NH&MRC Australia Fellowship. A. Möller was supported by an NBCF Research Fellowship (no. ECF-11-09) and an AICR project grant (no. 09-0676). M.T. Chow was supported by a Cancer Research Institute PhD scholarship. C. Paget was supported by a postdoctoral fellowship from the US Department of Defense (GRANT10752705). J. Sceneay was supported by a State Trustees Australia Foundation scholarship.
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.