Nod1 Limits Colitis-Associated Tumorigenesis by Regulating IFN-γ Production

Patients with inflammatory bowel disease have a substantially increased risk for developing colorectal cancer. The mechanisms by which chronic intestinal inflammation leads to cancer remain to be fully understood, but it is widely accepted that the persistent upregulation of proinflammatory mediators during the acute inflammatory response promotes the production of DNA-damaging oxygen radical species, growth factors, and antiapoptotic factors that facilitate tumor initiation and progression (1). However, self-limited inflammation that remains tightly regulated is critical for tissue repair as well as tumor immunity. Recent data have suggested the importance of the intracellular Nod-like receptor (NLR) family of innate immune receptors, which are capable of sensing microbial and/or damage signals, in maintaining intestinal homeostasis and protecting against both colitis and colitis-associated tumorigenesis through a variety of mechanisms, including the promotion of intestinal barrier function and epithelial repair, the negative regulation of inflammatory responses, and the preservation of microbial eubiosis (1–3). We have previously shown that Nod1, which recognizes a peptidoglycan-related moiety found in bacteria, is important in reducing susceptibility to colitis-associated tumorigenesis in mice (1, 4–6). Nod1 is expressed ubiquitously in multiple cell types, including both epithelial cells and immune cells, and its stimulation results in the activation of both NF-κB and MAPK pathways that leads to the production of proinflammatory molecules, in particular chemokines that promote neutrophil recruitment important for host defense (7, 8).

We have previously demonstrated that Nod1 was important for maintaining intestinal homeostasis. Specifically, using a model of inflammation-associated tumorgenesis with the administration of the carcinogen azoxymethane (AOM) and colitis-inducing dextran sulfate sodium (DSS), we showed that Nod1-deficient mice were more susceptible to colitis-associated tumor that was associated with a defect in intestinal epithelial barrier function resulting in greater intestinal permeability, increased commensal-driven acute inflammatory responses, and increased epithelial proliferation (1). To further elucidate the mechanism by which Nod1 signaling reduces susceptibility to inflammation-induced tumorigenesis, we performed bone marrow chimera experiments to identify the cellular compartment important for limiting tumor development during chronic DSS-induced inflammation. In the present study, we show that the Nod1 functions primarily in the hematopoietic compartment to suppress tumorigenesis. More specifically, we demonstrate a T cell–intrinsic role for Nod1 in protecting against inflammation-associated tumorigenesis. Nod1 deficiency in T cells is associated with impaired production of IFN-γ and STAT1 activation, both of which are implicated in tumor suppression and immune surveillance, during the acute inflammatory response to DSS-induced epithelial injury. We further demonstrate that the Nod1 ligand, KF1B, can act as a costimulatory molecule and enhance IFN-γ responses to anti-CD3 stimulation of T cells. IFN-γ–deficient signaling results in increased tumorigenesis similar to that observed in Nod1-deficient mice. Finally, we show that adoptive transfer of wild-type (WT) T cells into T cell–deficient mice is sufficient to limit tumor development as compared with the adoptive transfer of Nod1-deficient T cells. The importance of IFN-γ production by T cells is further supported by the adoptive transfer of IFN-γ–deficient T cells into T cell–deficient hosts, which resulted in increased numbers of tumors compared with mice that were adoptively transferred with WT T cells. Collectively, our results suggest that Nod1 signaling promoted by the recognition of bacterial ligand in T cells augments IFN-γ production, which can reduce mice susceptibility to colitis-associated tumorigenesis.

C57BL/6J, TCRβ^−/−, and IFNγ^−/− mice were initially obtained from The Jackson Laboratory and subsequently bred in-house. Both TCRβ^−/− and IFNγ^−/− mice are in the B6 background. Nod1^−/− mice were provided by Dr. Gabriel Nunez and subsequently backcrossed a minimum of eight times against the B6 background. Nod1^−/−TCRβ^−/− mice were generated by crossing Nod1^−/− and TCRβ^−/− mice. Age- and sex-matched adult (6- to 12-wk-old) mice were used for all experiments. All animals were maintained under specific pathogen-free conditions, and animal studies were approved by the University Committee on Use and Care of Animals at the University of Michigan.

Adult mice (6 to 12 wk old) were injected i.p. with 10 mg/kg AOM (Sigma-Aldrich). On day 5, mice were treated with 2% DSS (molecular mass, 36,000–50,000 Da; MP Biomedicals) in the drinking water for 5 d followed by 16 d of regular water. After two additional cycles of DSS, mice were sacrificed for tumor counting 3 wk after the end of the third DSS cycle. For experiments involving T cell–deficient hosts, the concentration of the DSS with the second or third cycle was reduced to 1.5% to optimize mouse survival.

Bone marrow chimeras were generated as previously described (2). Recipient Nod1^−/− and WT mice were lethally irradiated with split dose irradiation to minimize injury to the intestinal epithelium (550 rad separated by 3 h for a total dose of 1100 rad). Bone marrow cells (10 × 10⁶) that had been flushed from the tibias and femurs of donor Nod1-deficient and WT mice were injected into the lethally irradiated recipient mice into the tail vein or retro-orbital sinus. Four chimeric groups were generated: WT→WT, WT→Nod1^−/−, Nod1^−/−→WT, and Nod1^−/−→Nod1^−/−. Mice were treated with 4 wk of trimethoprim-sulfamethoxazole and allowed to recover for a minimum of 8 wk prior to treatment with AOM/DSS to induce tumors.

Colon lamina propria cells were isolated as previously described (2). Briefly, colons were isolated from untreated 6- to 12-wk-old WT B6 mice and cut into small pieces in HBSS (Life Technologies) supplemented with 2.5% heat-inactivated FBS (Life Technologies) and penicillin/streptomycin (HBSS+). After three washes with magnetic stirring, the colon pieces were incubated in HBSS+ with 1 mM DTT at 37°C. Colon pieces were washed three times again and then incubated in 1 mM EDTA at 37°C with magnetic stirring for 30 min. Tissue was further digested with 400 IU/ml type III collagenase (Worthington Biochemical) and 10 μg/ml DNase I (Worthington Biochemical) for an additional 2 h at 37°C with magnetic stirring. After complete digestion of tissue, the cell suspension was filtered through a 70-μm filter before running on a 75/40% Percoll gradient to collect enriched lamina propria cells from the interface.

Spleens from donor mice were homogenized, filtered, and treated with RBC lysis buffer (BD Biosciences). T cells were purified by magnetic bead isolation (Miltenyi Biotec) or by FACS using anti-CD3 Ab (BD Biosciences, clone 145-2C11). Six to 7 million magnetic bead–purified or 3 million sorted CD3⁺ T cells were adoptively transferred by tail vein injection into WT or TCRβ^−/− mice. Approximately 2 wk after injection, T cell reconstitution was verified by flow cytometric analysis of peripheral blood using anti-TCRβ Ab (BD Biosciences), and mice were treated with AOM/DSS.

Colonic tissue or tumors from AOM/DSS-treated mice were homogenized, and total RNA was isolated using the NucleoSpin RNA kit (Macherey-Nagel). cDNA synthesis was performed using iScript (Bio-Rad), and the cDNA was then used for quantitative PCR using either SYBR Green master mix (Applied Biosystems) or TaqMan gene expression assays on the ABI 7900HT. Primer sequences are available upon request.

For measurement of T cell–derived cytokines, total lamina propria cells from day 8 AOM/DSS-treated mice were sorted for CD3⁺NK1.1⁻ T cells, which excludes NKT and invariant NKT cells, and then ex vivo stimulated with plate-bound anti-CD3 Ab (10 μg/ml) (BD Biosciences, 500-A2) or anti-CD3 and anti-CD28 (2 μg/ml) (BD Biosciences, clone 37.51), respectively. Stimulation was performed in a 96-well plate using a concentration of 2 million T cells/ml. Supernatants were collected 48 h later and cytokines were measured by ELISA.

For experiments involving stimulation with the Nod1 ligand KF1B, provided as a gift by Dr. Naohiro Inohara, T cells were isolated by magnetic bead purification of total splenocytes and then stimulated with plate-bound anti-CD3 (1 μg/ml) with or without KF1B (10 μg/ml). Supernatant was collected 24 h after stimulation and IFN-γ levels were measured by ELISA.

Lamina propria cells were prepared from WT and Nod1^−/− mice on day 8 of AOM/DSS treatment followed by intracellular cytokine staining by first incubating cells for 4 h at 37°C in the presence of the protein transporter inhibitor GolgiStop (BD Biosciences) supplemented with 100 ng/ml PMA plus 1000 ng/ml ionomycin or GolgiStop alone. Subsequently, cells were surface stained with CD3 (145-2C11 or 500A2), CD4 (RM4-5 or GK1.5), CD8 (53-6.7), and NK1.1 (PK136) (all Abs from BioLegend), followed by treatment with Cytofix/Cytoperm buffer (BD Biosciences) and incubation with fluorochrome-conjugated Abs against IFN-γ (XMG1.2, from BioLegend) or IL-17A (TC11-18H10, from BD Biosciences).

T cells magnetic bead purified from the spleens of mice treated with AOM/DSS (day 8) were stimulated ex vivo with 100 ng/ml PMA and 500 ng/ml ionomycin in RPMI 1640 with 0.5% FBS (serum starvation conditions to minimize background activation by addition of serum). At the indicated time points, cells were lysed in protein extraction buffer (150 mM NaCl, 10 mM Tris-HCl [pH 7.4], 5 mM EDTA, 0.1% Nonidet P-40, 0.5 mM DTT, 5% glycerol, and Halt’s protease and phosphatase inhibitor mixture [Pierce]). Proteins were separated by SDS-PAGE, transferred onto nitrocellulose membranes, and then immunoblotted with Abs against IκBα, p-IκBα, ERK, p-ERK, STAT1, p-STAT1 (Cell Signaling Technology), and β-actin (Santa Cruz Biotechnology).

Data are presented as mean ± SEM. Comparisons of tumor counts and cytokine expression levels between Nod1^−/− and WT mice were performed using the Student unpaired t test. A p value < 0.05 was considered statistically significant.

We have previously demonstrated that Nod1-deficient mice developed more tumors than did WT mice in the AOM/DSS model of colitis-associated tumorigenesis (Fig. 1A) (1). The increased tumorigenesis in Nod1-deficient mice was associated with increased production of proinflammatory cytokines and chemokines during the acute inflammatory response to chemically induced epithelial injury and higher levels of epithelial proliferation (1). To determine which cellular compartment (i.e., epithelial and/or hematopoietic) Nod1 functions in to mediate protection against tumor development, we generated bone marrow chimeras by lethally irradiating Nod1^−/− or WT recipient mice and injecting bone marrow from Nod1^−/− or WT donor mice into the tail vein. After verifying donor bone marrow reconstitution at ∼8 wk posttransplant (data not shown), mice were treated with AOM followed by three rounds of 2% DSS to induce tumors in the setting of chronic inflammation as previously described (Fig. 1A) (1, 9). Transplanting WT bone marrow into lethally irradiated Nod1^−/− mice was sufficient to suppress tumor development to levels similar to those observed for WT control mice that were transplanted with WT bone marrow (Fig. 1B, 1C). Conversely, WT mice that received Nod1^−/− bone marrow developed significantly more tumors compared with WT control mice (i.e., WT mice transplanted with WT bone marrow) and phenocopied Nod1^−/− mice that received Nod1^−/− bone marrow (Fig. 1B, 1C). These results strongly suggest that Nod1 signaling particularly in hematopoietic cells is important for protection against tumorigenesis.

We had previously demonstrated that during the acute inflammatory response to the first cycle of DSS-induced colitis, Nod1^−/− mice exhibited significantly higher levels of various proinflammatory mediators such as CXCL2, IL-1β, and IL-6 compared with those in WT mice specifically on days 11–13 (1–3 d after completion of DSS) (1). To determine whether earlier events predisposed mice to increased inflammatory responses, we measured the induction of several proinflammatory factors within the colon tissue of WT and Nod1^−/− mice on day 10 at the end of the first round of DSS treatment (Fig. 2A) by quantitative PCR. At this time point, we saw a striking reduction in the production of IFN-γ,whereas other cytokines such as IL-6 and IL-1β were not significantly different (Fig. 2A). As IFN-γ has been associated with antitumor activity (10–12), we also looked at the mRNA expression of IFN-γ within equally sized tumors between Nod1^−/− and WT mice at time of sacrifice 3 wk after the third and last round of DSS (around day 63) (Fig. 2B). Similar to what was observed in the colon tissue on day 10, there was significantly decreased levels of IFN-γ within tumors in Nod1^−/− mice compared with those in similary sized WT tumors. Because T cells are a major source of IFN-γ, we also measured levels of other cytokines that can be produced by T cells, such as IL-4, IL-17A, and IL-10, by quantitative PCR within the colons of AOM/DSS-treated mice on day 8 as well as in tumors, and we found no significant differences between WT and Nod1^−/− mice (Supplemental Fig. 1).

IFN-γ has been previously demonstrated to have potent antitumor activity in cancer models, and its production within the tumor microenvironment is associated with improved patient outcomes (10–12). However, IFN-γ has also been previously demonstrated to affect epithelial barrier function and contribute to increased susceptibility to colitis (13–15). To determine the role of IFN-γ in the development of tumors in the AOM/DSS model of colitis-associated colon cancer, we treated IFNγ^−/− mice with AOM/DSS and compared the number and size of tumors that developed with that in WT mice (Fig. 3). IFNγ^−/− mice developed more and larger tumors as compared with WT mice as similarly observed in Nod1^−/− mice (Fig. 3) (1), indicating that IFN-γ is important for suppressing tumorigenesis in this mouse model. These results also suggest that the impaired production of IFN-γ in Nod1-deficient mice may predispose them to increased tumorigenesis.

T cells are a major source of IFN-γ and are present in the lamina propria during DSS-induced colitis (16). To determine whether there is a defect in IFN-γ production in T cells within the lamina propria early during DSS treatment before the development of overt inflammation (day 8), we isolated colon lamina propria cells and sorted CD3⁺NK1.1⁻ T cells within the lamina propria of WT and Nod1^−/− mice and measured IFN-γ levels as well as other T cell–derived cytokines, including IL-4, IL-17, and IL-10 after ex vivo stimulation with anti-CD3 Ab (Fig. 4). We observed that after the start of the first round of DSS, only the production of IFN-γ was significantly impaired in Nod1^−/− mice (Fig. 4A). Consistently, examination of CD4⁺CD3⁺NK1.1⁻ T cells revealed increased intracellular staining for IFN-γ (Fig. 4B). IL-4 and IL-5 production by colon lamina propria CD3⁺NK1.1⁻ T cells was moderately increased in Nod1^−/− mice (Fig. 4A). The reduced IFN-γ production was not due to differences in T cell numbers within the lamina propria because Nod1^−/− mice had similar frequency and numbers of CD4⁺ and CD8⁺ T cells within the lamina propria as in WT mice (Fig. 4C). Moreover, the decreased production in IFN-γ by Nod1^−/− T cells cannot be attributed to differences in levels of T cell proliferation or apoptosis, as Nod1^−/− T cells isolated from the spleens of AOM/DSS-treated mice showed no defects in proliferation or apoptosis as assessed by CFSE staining and trypan blue staining, respectively (Supplemental Fig. 2).

IFN-γ leads to phosphorylation of STAT1, which is a known tumor suppressor that promotes antitumor immune responses as well as restricts cellular proliferation (17, 18). STAT1 has also been previously shown to be induced during AOM/DSS-induced inflammation (19). We therefore determined whether the decrease in IFN-γ production by Nod1^−/− T cells was associated with lower levels of STAT1 activation. T cells were magnetic bead purified from splenocytes of Nod1^−/− and WT mice on day 8 during the first round of DSS, ex vivo stimulated with anti-CD3 and anti-CD28 Abs, and harvested at multiple time points after stimulation. STAT1 activation was assessed by immunoblotting with phospho-specific STAT1 Abs (Fig. 4D). AOM/DSS-treated Nod1^−/− T cells exhibited lower levels of STAT1 phosphorylation compared with WT T cells. Consistently, robust activation of STAT1 by Nod1^−/− T cells was induced after administration of recombinant IFN-γ (Supplemental Fig. 3). We also investigated whether Nod1-deficient T cells isolated during the first round of DSS demonstrated impaired NF-κB or MAPK signaling, as Nod1 activation by bacteria results in upregulation of these pathways (7). As with STAT1, we also observed decreased phosphorylation of ERK and IκBα consistent with Nod1 being upstream of NF-κB and MAPK pathways, respectively.

DSS treatment of mice causes epithelial injury, which results in increased gut permeability, bacterial translocation into the mucosa, and commensal-driven inflammatory responses within the colon tissue. As suggested by Fig. 4, Nod1^−/− T cells isolated from the lamina propria during DSS treatment had impaired activation of STAT1, NF-κB, and MAPK pathways. Nod1 recognizes a peptidoglycan motif found in the bacteria cell wall. We therefore investigated whether the synthetic Nod1 agonist, KF1B, is capable of activating T cells directly or in combination with anti-CD3 stimulation to potentiate IFN-γ production (Fig. 5) (20). T cells were magnetic bead purified from WT and Nod1^−/− spleens and then stimulated in vitro with KF1B alone or together with anti-CD3. KF1B was insufficient to directly activate T cells in vitro; however, KF1B was able to synergize with anti-CD3 to significantly enhance production of IFN-γ. This effect was Nod1-dependent, as KF1B was unable to potentiate IFN-γ production in Nod1^−/− T cells with TCR activation (Fig. 5A). Similarly, magnetic bead–purified CD4⁺ and CD8⁺ T cells showed similar synergistic IFN-γ responses to KF1B and anti-CD3 stimulation (Fig. 5B, 5C). These results suggest that bacterial-derived ligands recognized by Nod1 are capable of enhancing IFN-γ responses after TCR activation by anti-CD3 Ab.

The above experiments suggest that Nod1 stimulation can increase IFN-γ production during AOM/DSS treatment and is important for protecting against the development of tumors. To determine the importance of Nod1 function in T cells specifically in vivo, magnetic bead–purified T cells from WT or Nod1^−/− mice were adoptively transferred into T cell–deficient TCRβ^−/− hosts (Fig. 6A). After approximately 2 wk to allow homeostatic proliferation of T cells, mice were treated with AOM/DSS. TCRβ^−/− mice that contained Nod1^−/− T cells developed more tumors compared with mice that had WT T cells, confirming that a T cell–intrinsic function for Nod1 is important for limiting tumors. Additionally, adoptive transfer of Nod1^−/− T cells into mice that were deficient in both Nod1 and TCRβ (Nod1^−/−TCRβ^−/−) also resulted in more tumors than in mice that received WT T cells after AOM/DSS treatment (Fig. 6B). To determine whether T cell production of IFN-γ was important for tumor suppression with the AOM/DSS model, TCRβ^−/− mice were adoptively transferred with IFNγ^−/− T cells or WT T cells followed by treatment with AOM/DSS. Similar to that observed with the adoptive transfer of Nod1-deficient T cells, TCRβ^−/− mice that were adoptively transferred with IFNγ^−/− T cells developed more tumors compared with those with WT T cells (Fig. 6C). We also observed dramatically reduced IFN-γ production in the colons of Nod1^−/−TCRβ^−/− mice with adoptively transferred Nod1^−/− CD3⁺ T cells or TCRβ^−/− mice adoptively transferred with IFNγ^−/− CD3⁺ T cells on day 10, similar to that observed in Fig. 1A, suggesting that CD3⁺ T cells are indeed a major source of IFN-γ production in these mice at this time point (Supplemental Fig. 4). Collectively, these results suggest that Nod1 functions in T cells to limit tumor development likely through the regulation of IFN-γ production.

Recent data have demonstrated that multiple members of the NLR family of intracellular immune receptors that sense microbes and damage signals are critically important for resistance to DSS-induced colitis and AOM/DSS-induced colitis-associated tumorigenesis. Mice deficient in these NLRs, specifically NLRP3, NLRP6, Nod2, and Nod1, all exhibit epithelial barrier defects with increased intestinal epithelial damage and permeability, which can result in increased bacterial translocation and heightened inflammatory responses compared with WT mice and lead to increased susceptibility to tumorigenesis (2, 3, 21–23). Based on our previous observations that Nod1^−/− mice had increased intestinal epithelial apoptosis and permeability associated with increased inflammatory responses and epithelial proliferation (1), we wanted to determine in the present study whether Nod1 functions primarily in the epithelial or hematopoietic cell compartment to mediate its protective effects against inflammation-induced tumorigenesis. Interestingly, we determined that Nod1 is important in the hematopoietic compartment, as WT bone marrow–derived cells were sufficient to rescue lethally irradiated Nod1-deficient mice from developing tumors to the same extent as WT mice. More importantly, we discovered a T cell–intrinsic role for Nod1 in protecting against tumorigenesis as suggested by the observation that Nod1-deficient CD3⁺ T cells adoptively transferred into either TCRβ^−/− or Nod1^−/−TCRβ^−/− mice resulted in increased tumors compared with mice adoptively transferred with WT T cells. Nod1 deficiency in T cells was associated with decreased production of IFN-γ and activation of STAT1 early after AOM/DSS treatment. Nod1 recognizes a specific peptidoglycan motif found in intestinal bacteria, and we demonstrate that although the Nod1 ligand KF1B is insufficient to directly activate T cells, it is nonetheless capable of acting as a costimulatory molecule and enhancing T cell activation and production of IFN-γ in response to anti-CD3 Ab stimulation (4, 6). Because DSS treatment results in epithelial damage and translocation of bacteria (24), one potential mechanism by which Nod1 represses inflammation-induced tumorigenesis is to potentiate production of IFN-γ, which has antitumor effects, by T cells through recognition of bacterial signals that arise after epithelial breach by DSS. Indeed, mice deficient in IFN-γ also develop more tumors after AOM/DSS treatment compared with those in WT mice, similar to those in Nod1^−/− mice. Furthermore, TCRβ^−/− mice adoptively transferred with IFNγ^−/− T cells also develop more tumors compared with mice adoptively transferred with WT T cells. The upregulation of IFN-γ promotes phosphorylation and activation of STAT1, which can further contribute to tumor suppression, and, consistently, phosphorylated STAT1 levels are reduced in Nod1-deficient T cells early after treatment with AOM/DSS. Although the signaling pathways governing Nod1-dependent IFNγ remain to be fully elucidated, one possibility is that bacterial stimulation of Nod1 in T cells results in activation of NF-κB and MAPK pathways, both of which have been implicated in the regulation of IFN-γ expression (25–27). Consistently, Nod1^−/− T cells exhibited reduced NF-κB and MAPK activation compared with WT T cells after AOM/DSS treatment.

A similar but distinct mechanism of tumor suppression has been proposed for NLRP3. NLRP3 is an NLR that is important for the activation of caspase-1, which leads to the secretion of the mature forms of the proinflammatory cytokines IL-1β and IL-18 (28). Nlrp3^−/− mice are more susceptible to both DSS-induced colitis and colitis-associated tumorigenesis, which was related to the defective production of IL-18 in response to DSS-induced epithelial injury (19, 22). In addition to facilitating epithelial repair and restitution, IL-18 has also been implicated in promoting antitumor activity through induction of IFN-γ (29). Indeed, Nlrp3^−/− mice had significantly reduced levels of IFN-γ within the colon tissue after AOM/DSS treatment similarly to that observed with Nod1^−/− mice in the present study, and, additionally, the reduced IFN-γ production was associated with decreased activation of STAT1 in mice deficient in caspase-1, which is downstream of NLRP3 (19). Consistently, IFNγ^−/− mice, similar to IL18^−/−, Nlrp3^−/−, and Nod1^−/− mice, also develop significantly more tumors than do WT mice after AOM/DSS treatment (30). In the case of Nod1, the mechanism of IFN-γ regulation may be through the costimulatory effects of bacterial-derived Nod1 ligand rather than through effects on IL-18 production. It remains unclear, however, whether IFN-γ may act in other ways to suppress tumorigenesis because IFN-γ has also been implicated in promoting epithelial repair and epithelial cell restitution during initial inflammatory responses while inhibiting proliferation and promoting apoptosis with chronic exposure (14, 31). However, IFN-γ can also affect tight junction assembly and promote epithelial barrier effects resulting in increased susceptibility of IFNγ^−/− mice to DSS-induced colitis (13, 15), suggesting a role other than modulation of inflammation or the epithelial barrier as the primary mechanism.

Based on examination of expression levels of T cell–related cytokines in colon and tumor tissues, we observed significant changes in IFN-γ, but not in IL-4, IL-17A, or IL-10. However, purified colon lamina propria CD3⁺NK1.1⁻ T cells from Nod1^−/− mice expressed mildly elevated IL-4 levels compared with those from WT mice. Although the focus of the present study is on Nod1-mediated IFN-γ production, which appears to be dominantly affected over other T cell–related cytokines, it remains possible that IL-4 may also contribute to increased tumorigenesis in Nod1^−/− mice. In a model of colitis-associated tumorigenesis that utilizes AOM together with trinitrobenzene sulfonic acid,which induces colitis, IL-4 was suggested to promote tumorigenesis whereas IFN-γ was protective. Additionally, IL-4 has also been implicated in exacerbating colitis and tumor growth in mice (27, 32). IL-4 can also inhibit IFN-γ production. How IL-4 contributes to tumorigenesis in AOM/DSS-treated Nod1-deficient mice remains to be determined.

Although the NLR family has been traditionally considered a part of innate immunity, there is emerging evidence that NLRs can also have T cell–intrinsic functions. Nod2, for example, has been shown to function in T cells to regulate IL-2 production and host defense against Toxoplasma gondii (33). More recently, NLRP12 was shown to be an intrinsic negative regulator of T cells such that its absence in T cells results in enhanced T cell responses and increased susceptibility of NLRP12-deficient mice to autoimmune disease (34). In the case of Nod1, our data suggest that the ligand for Nod1 can promote T cell responses after TCR activation during acute inflammatory response to AOM/DSS treatment. Similar to our results, a previous study demonstrated that Nod1 stimulation with ligand resulted in increased production of IFN-γ, particularly in both mouse and human CD8⁺ T cells, that was associated with increased activation of NK-κB and MAPK pathways (35). Interestingly, IFN-γ was also significantly enhanced in the presence of both Nod1 and TLR2 ligands during TCR activation of CD8⁺ cells, and therefore it is possible that during DSS-induced colitis, TLR signaling by translocated bacteria can further enhance Nod1-mediated IFN-γ production to limit tumorigenesis. In our study, we observe Nod1-dependent increases in IFN-γ production by T cells, but whether Nod1 is differentially important in CD4⁺ versus CD8⁺ T cells remains to be determined. Although our adoptive transfer experiments suggest that Nod1 function in conventional CD3⁺ T cells is important for IFN-γ production and tumor reduction in the AOM/DSS model, it would also be interesting to determine whether Nod1 has a similar function in other cells that can produce IFN-γ within the intestine such as the group I innate lymphoid cells. Nonetheless, our study further highlights the ability of innate immune receptors to promote T cell function and point to potential therapeutic strategies that can be used to limit colon tumor development.

We thank Drs. Gabriel Nunez, Naohiro Inohara, and Michael Shaw for helpful discussions. We also thank the Microscopy and Image Analysis Laboratory for use of microscopes and Joel Whitfield of the Immunological Monitoring Core for assistance with ELISAs.

The authors have no financial conflicts of interest.