Loss of single immunoglobulin interlukin-1 receptor-related molecule leads to enhanced colonic polyposis in Apc^min mice

Abstract

Background & Aims

Commensal bacteria can activate signaling by the toll-like and interleukin-1 receptors (TLR and IL-1R) to mediate pathogenesis of inflammatory bowel diseases and colitis-associated cancer. We investigated the role of the single immunoglobulin IL-1 receptor-related (SIGIRR) molecule, a negative regulator of TLR and IL-1R signaling, as a tumor suppressor to determine whether SIGIRR controls cell cycle progression, genetic instability, and colon tumor initiation by modulating commensal TLR signaling in the gastrointestinal tract.

Methods

We analyzed Apc^min/+/Sigirr^-/- mice for polyps, microadenomas, and anaphase bridge index. Commensal bacteria were depleted from mice with antibiotics. Akt, mTOR and β-catenin pathways were examined by immunoblotting and immunohistochemistry. Loss of heterozygosity (LOH) of Apc and expression of cytokines and proinflammatory mediators were measured by non-quantitative or quantitative PCR.

Results

Apc^min/+/Sigirr^-/- mice had increased LOH of Apc and microadenoma formation, resulting in spontaneous colonic polyposis, compared with Apc ^min/+/Sigirr^+/+ mice. The increased colonic tumorigenesis that occurred in the Apc^min/+/Sigirr^-/- mice depended on the presence of commensal bacteria in the gastrointestinal tract. Cell proliferation and chromosomal instability increased in colon crypt cells of the Apc^min/+/Sigirr^-/- mice. Akt, mTOR and their substrates were hyper-activated in colon epithelium of Apc^min/+/Sigirr^-/- mice in response to TLR or IL-1R ligands. Inhibition of the mTOR pathway by rapamycin reduced formation of microadenomas and polyps in the Apc^min/+/Sigirr^-/- mice.

Conclusions

SIGIRR acts as a tumor suppressor in the colon by inhibiting TLR-induced, mTOR-mediated cell cycle progression and genetic instability.

Introduction

The large and highly diverse populations of commensal bacteria in the mammalian colon are known to provide beneficial effects for the host, including the supply of nutrients and organic substrates ^1-4. An important prerequisite for this host-bacterial mutualism is the “microbial tolerance” of the intestinal epithelial layer, a physical and immune barrier against commensal bacteria. Toll-like receptor (TLR) and Interleukin-1 receptor (IL-1R) signaling plays a critical role in the active cross-talk between the commensal-host, maintaining tolerance to the commensal bacteria. Inappropriate activation of the TLR-IL-1R signaling by commensal bacteria contributes to the pathogenesis of inflammatory bowel diseases and elevated risk for colorectal cancer ^5-10. Importantly, murine TLR4 deficiency leads to a marked reduction of colon tumor development in a colitis-associated cancer (CAC) model ^{11, 12}, indicating the importance of TLR signaling in colon tumorigenesis. Therefore, the “microbial tolerance” is important for preventing both overt inflammation and tumorigenesis in the colon.

The “microbial tolerance” to commensal microflora is a consequence of multiple regulatory mechanisms, including modulation of TLR signaling ^{11, 13, 14}. We and others previously reported that the Single Immunoglobulin IL-1 Receptor Related molecule (SIGIRR, also known as TIR8), a negative regulator for TLR-IL-1R signaling, plays a critical role in gut homeostasis, intestinal inflammation and colitis-associated tumorigenesis ^15-21. Tumorigenesis can be divided into three mechanistic stages, initiation, promotion and progression ^{22, 23}. Using the colitis-associated cancer model, we have previously reported that SIGIRR deficiency increased inflammation-induced tumor promotion/progression. However, it remains unclear whether and how TLR signaling, in the absence of SIGIRR, impact on colon epithelium to enhance tumor initiation.

Aberrant activation of Wnt signaling represents the key initiating event for intestinal tumorigenesis, since activation of Wnt target genes dominantly drives proliferation of intestinal epithelium ²⁴. Phosphorylation of β-catenin by glycogen synthase kinase 3β (GSK3β) leads to ubiquitination and degradation of β-catenin. Activation of Wnt signaling blocks the function of the “destruction complex” consisting of tumor suppressor Adenomatous polyposis coli (Apc), adaptor AXIN and GSK3β, resulting in stabilization and nuclear accumulation of β-catenin ²⁵. Constitutive activation of Wnt signaling is directly linked to colorectal cancer. Human genetic studies show that Apc is mutated in more than 95% of the Familial Adenomatous Polyposis (FAP) patients, and more than 80% of the sporadic colon cancers ²⁶. Furthermore, 50% of colon carcinomas that retain wild-type Apc, acquire point mutations in β-catenin ²⁷. Apc^min/+ mice (containing a germ-line mutation in Apc) spontaneously develop a large number of tumors in small bowel, but very few polyps in the colon ²⁸. Loss of heterozygosity (LOH) of Apc, a genetic alteration associated with the human colon cancers, is the exclusive genetic alteration leading to the tumor initiation in Apc^min/+ mouse ^{29, 30}. Thus, Apc^min/+ mouse has been utilized as a spontaneous colon cancer model to identify genetic elements that have impact on chromosomal instability, resulting in LOH of Apc and tumor initiation ³¹.

In this study, we introduced SIGIRR deficiency into the Apc^min/+ mice and uncovered a critical role of SIGIRR in tumor initiation for the first time. SIGIRR deficiency leads to increased LOH of Apc and microadenoma formation, resulting in spontaneous colonic polyposis in Apc^min/+/Sigirr^-/- mice. Importantly, we found that mTOR and its downstream substrates, well-known regulators for translational control, were hyper-activated in SIGIRR-deficient colon epithelium in response to TLR-IL-1R ligands. Furthermore, the abrogation of mTOR pathway by rapamycin results in decreased microadenoma and polyps formation in Apc^min/+/Sigirr^-/- mice. Moreover, the increased colonic tumorigenesis in Apc^min/+/Sigirr^-/- mice is dependent on the presence of commensal bacteria in the gut, underscoring the role of dysregulated commensal bacteria-TLR signaling in colonic tumor initiation.

Materials and Methods

Mice Breeding and Genotyping

Apc^min/+ mice on C57BL/6 background were purchased from Jackson Lab (Bar Harbor, ME) and bred onto Sigirr^-/- mice (>10 generations on C57BL/6 background) to generate Apc^min/+/Sigirr^+/- and Sigirr^+/- mice. Apc^min/+/Sigirr^+/- males were then crossed to Sigirr^+/- females, generating littermates including Apc^min/+/Sigirr^+/+, Apc^min/+/Sigirr^+/-, and Apc^min/+/Sigirr^-/- mice. These mice were bred and maintained in a pathogen-free animal facility at Cleveland Clinic Foundation. All the procedures were conducted in compliance with a protocol approved by the Institutional Animal Care and Use Committee at Cleveland Clinic Foundation. Genotyping of wild type and Apc^min alleles was conducted by PCR following a protocol provided by Jackson Laboratory. Genotyping of wild type and mutated Sigirr alleles was done by southern blotting as described previously ¹⁸. All the procedures have been approved by the Institutional Animal Care and Use Committee at Cleveland Clinic Foundation.

Tumor Counts and Histology

Apc^min/+/Sigirr^+/+, Apc^min/+/Sigirr^+/-, and Apc^min/+/Sigirr^-/- littermates were euthanized at 3-4 months of age or when the first signs of morbidity appeared. The intestinal tract was removed immediately after the euthanasia, opened longitudinally and washed with PBS. The small intestine and colon were examined and tumor were counted and measured under a stereomicroscope. Colonic tissues and tumors were fixed by 10% buffered formalin overnight, and then embedded in paraffin. Paraffin-embedded specimens were cut to 5 μm sections and stained with H&E for histology analyses.

Western Blotting

Purified primary colonocytes (treated or untreated) were lysed in cell lysis buffer (50mM Tris-HCl, pH 7.5/150mM NaCl/1% Triton X-100/1mM EDTA) supplemented by protease inhibitors Complete Mini (from Roch), and 1mM PMSF/2mM Na3VO4/5mM NaF. Protein concentration was measured by BCA assay kit from Sigma. 50 μg total protein of each sample was resolved by 8-10% of SDS-PAGE, and transferred onto PVDF membrane. Western blotting was conducted using the following antibodies. Rabbit anti-p-mTOR, p-p70S6K, p-S6, p-Akt (S473), p-GSK3α/β, p-4EBP, mTOR, p70S6K, S6, Akt, GSK3α/β, 4EBP were purchased from Cell Signaling Technology, and used in 1:1000 dilution. Anti-c-Myc, p-cdk2, cdk2, cyclin E, cyclin A, cyclin D1 and actin were purchased from Santa Cruz, and diluted 1:500 for Western blotting.

Statistic analysis

ANOVA was used to analyze the tumor burden. Fisher’s exact test was used to compare percentage. Student’s T test was used for other studies and p<0.05 is considered significant.

Results

SIGIRR deficiency leads to increased colon tumorigenesis in Apc^min/+ mice

To investigate the role of SIGIRR in tumorigenesis in a spontaneous colon cancer model, Sigirr^-/- mice were bred onto the Apc^min/+ background to generate Apc^min/+/Sigirr^+/+, Apc^min/+/Sigirr^-/- and Apc^min/+/Sigirr^+/- mice. Intestinal samples from littermates at 12-16 weeks of age were examined for tumor formation. As expected, Apc^min/+/Sigirr^+/+ developed on average 42 tumors per small intestine, but only 1.3 tumors per colon on average (Fig. 1A). Importantly, Apc^min/+/Sigirr^-/- mice developed on average 4.5 polyps per colon (more than 3-fold increase as compared to Apc^min/+/Sigirr^+/+ mice) (Fig. 1A and B), although these Apc^min/+/Sigirr^-/- mice only exhibited a modest increase in small intestinal polyps (average 67 polyps per mouse) (Fig. 1A). It is also noted that colon and small intestine polyps developed in Apc^min/+/Sigirr^+/- mice were comparable to those in Apc^min/+/Sigirr^+/+ mice, implicating a haplosufficiency for SIGIRR function (Fig. 1A and B). Remarkably, about 60% polyps in Apc^min/+/Sigirr ^-/- colon were larger than 2 mm in diameter, whereas only 36% polyps in Apc^min/+/Sigirr^+/+ colon were big adenomas (>2mm in diameter) (Fig. 1C). Moreover, more large polyps (>4mm in diameter) were developed in Apc^min/+/Sigirr ^-/- small intestine than Apc^min/+/Sigirr^+/+ small intestine (Fig. 1C). Notably, the polyps developed in Apc^min/+/Sigirr^-/- mice were tubular adenomas with high degree of dysplasia (Fig. 1D).

Fig. 1.

SIGIRR deficiency results in increased intestinal tumorigenesis on Apc^min/+ background. (A) Tumor numbers detected in the colon and small intestine of Apc^min/+/Sigirr^+/+, Apc^min/+/Sigirr^+/-, and Apc^min/+/Sigirr^-/- littermates. (B) Macrophotograph of the representative colons of Apc^min/+/Sigirr^+/+, Apc^min/+/Sigirr^+/-, and Apc^min/+/Sigirr^-/- littermates. (C) Size distribution of the colon and small intestine tumors developed in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- littermates. (D) Histology of colon tumors developed in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- littermates. ANOVA (A) or Fisher’s Exact Test (C) was used for statistic analysis. *=p<0.05, **=p<0.01, N.S. = Not significant.

The critical role of commensal bacteria in tumorigenesis

The elevated colonic tumorigenesis in Apc^min/+/Sigirr^-/- mice might be due to dysregulated TLR signaling triggered by commensal bacteria derived ligands. To test this hypothesis, commensal bacteria were depleted by oral administration of a cocktail of antibiotics ¹⁵. Antibiotics treatment resulted in a significant reduction of tumor formation in the colon and small intestine of the Apc^min/+/Sigirr^-/- mice (Fig. 2A and B). In addition, the tumors developed in antibiotics treated Apc^min/+/Sigirr^-/- mice were considerably smaller than those in untreated controls (Fig. 2A). The expression of the signature TLR-responsive genes, such as KC, IL-6, and cyclin D1, were indeed substantially reduced in the mucosa and tumors of Apc^min/+/Sigirr^-/- after removal of commensal bacteria (Suppl. Fig. 1), suggesting an important link between commensal bacteria-induced TLR signaling and tumorigenesis in Apc^min/+/Sigirr^-/- colon. As a control, Apc^min/+/Sigirr^+/+ mice were also treated with antibiotics. It is important to note that depletion of microflora only had a modest impact on the intestinal tumorigenesis in Apc^min/+/Sigirr^+/+ mice (Fig. 2C). These results indicate that the elevated colonic tumorigenesis in Apc^min/+/Sigirr^-/- mice has a greater dependence on gut microflora and TLR signaling than the intestinal tumorigenesis in the Apc^min/+/Sigirr^+/+ mice.

Fig. 2.

Antibiotics-treatment ameliorates tumorigenesis in Apc^min/+/Sigirr^-/- mice. (A) The number and size distribution of colon tumors developed in Apc^min/+/Sigirr^-/- mice with or without antibiotics-treatment. (B) Small intestinal tumors developed in Apc^min/+/Sigirr^-/- and Apc^min/+/Sigirr^+/+ mice with or without antibiotics-treatment. (C) The number and size distribution of colon tumors developed in Apc^min/+/Sigirr^+/+ mice with or without antibiotics-treatment. Apc^min/+/Sigirr^+/+ or Apc^min/+/Sigirr^-/- mice at 6 weeks of age were fed with regular water or water containing 1 g/L ampicillin, 500 mg/L vancomycine, 1 g/L neomycin and 1 g/L metronidazole for 6-8 weeks. ANOVA (A and B) or Fisher’s Exact Test (C) was used for statistic analysis. *=p<0.05, **=p<0.01, N.S. = Not significant.

SIGIRR deficiency leads to increased tumor initiation in Apc^min/+ mice

The key cellular events promoting tumor initiation include uncontrolled cell proliferation, prevention of apoptosis and elevated chromosomal instability. Consistent with the role of SIGIRR in colon epithelium homeostasis ¹⁵, Apc^min/+/Sigirr^-/- mice manifested extended proliferation zones in their colonic mucosa, exhibiting elevated cell proliferation (Apc^min/+/Sigirr^-/-:15.4±1.9 vs: Apc^min/+/Sigirr^+/+: 7.3±1.5 per crypt) (Fig. 3A-B), and concomitantly less apoptotic cells per crypt (Apc^min/+/Sigirr^-/- : 2.2±0.7 vs Apc^min/+/Sigirr^+/+: 3.3±0.8 per crypt) (Fig. 3C). In Apc^min/+/Sigirr^-/- tumors, three times more cells were undergoing active cell cycling (Fig. 3A and B) while fewer cells were eliminated by apoptosis (Fig. 3C). Taken together, increased cell proliferation and decreased apoptosis probably play a critical role in accelerating colon tumor development in Apc^min/+/Sigirr^-/- mice. These findings are consistent with the previous reports that demonstrate the ability of TLR signaling in promoting cell cycle progression ^{32, 33}.

Fig. 3.

Increased cell proliferation and decreased apoptosis in Apc^min/+/Sigirr^-/- colon. (A) Immunohistochemistry staining of Ki-67 in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- colon. Magnification: 100× or 400×. (B) and (C) Quantification of Ki-67 positive cells (B) or apoptotic cells detected by in situ TUNEL assay (C) in normal crypts and tumors. Well-oriented crypts (at least 50 normal crypts), or representative tumor fields (at least 30 high power fields) were counted on 2-4 slides from each mouse and 3 pairs of mice were analyzed. Error bars represent ±SEM, and Student’s T Test was conducted. * = p<0.05, ** = p<0.01.

By examining the H&E stained sections from 6-weeks old mice, more microadenoma were detected in the colons and small intestines of Apc^min/+/Sigirr^-/- as compared to that of Apc^min/+/Sigirr^+/+ mice (Fig. 4A and Suppl. Fig. 2), demonstrating increased tumor initiation in Apc^min/+/Sigirr^-/- mice. Loss of the wild-type Apc allele is the initial genetic alteration driving intestinal tumorigenesis in Apc^min/+ mice ^{31, 34}. Chromosomal instability attributable to LOH of Apc can be assessed by the frequency of “anaphase bridges”, extended chromosome bridging between the two spindle poles during the anaphase of cell cycle. By scoring anaphase bridges index (ABI) in the normal and tumor tissues, we detected significantly increased ABI in both normal mucosa and tumors of Apc^min/+/Sigirr^-/- mice compared to that in the Apc^min/+/Sigirr^+/+ mice (Fig. 4B). In support of this, flow cytometry studies showed that DNA aneuploidy was significantly increased in Apc^min/+/Sigirr^-/- tumor cells compared to that in Apc^min/+/Sigirr^+/+ mice (Suppl. Fig. 3).

Fig. 4.

Increased tumor initiation in Apc^min/+/Sigirr^-/- colon. (A) microadenoma developed in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- colon. Swiss-rolls of colon from 5 weeks old mice were cut into 5μM consecutive sections and stained by H&E. Microadenomas were quantified over H&E stained sections from 5 pairs of mice. (B) Anaphase-bridge index was counted on H&E stained normal mucosal and colon tumors. Representative photographs show normal nuclei separation in anaphase or abnormal anaphase bridge. Magnification: 1000×. More than 300 anaphase tumor cells and 500 anaphase normal cells from each strain were examined for anaphase-bridge formation. ABI was quantified over H&E stained sections from five pairs of mice. (C) LOH of Apc was assessed by mismatched PCR ^{29, 30}. Amplified Apc alleles were digested by HindIII and then separated by 1.5% agarose gel. A 123-bp product from the Apc+ allele and a 144-bp product from the Apc^min allele were detected. LOH of Apc was detected in all the colon tumors isolated from both strains. In this study, 20 colon tumors from Apc^min/+/Sigirr^-/- mice (N=4) and 5 colon tumors from Apc^min/+/Sigirr^+/+ mice (N=4) were analyzed, and the representative data is shown. Error bars represent ±SEM, and Student’s T Test was conducted. * = p<0.05, **=p<0.01.

It is plausible that the increased chromosomal instability in SIGIRR-deficient colon tissue might contribute to the increased tumorigenesis through LOH of Apc in Apc^min/+/Sigirr^-/- mice. By examining the Apc alleles in the normal and tumor tissues of the Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- colon, we found the wild-type Apc allele was lost in all the colon tumors dissected from Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- colons (Fig. 4C). In support of this, β-catenin became stabilized and accumulated in the cell cytoplasm and nucleus in microadenomas (Suppl. Fig. 4A) and well-developed tumors (Suppl. Fig. 4B). The fact that 100% of the tumors in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- colons acquire LOH of Apc demonstrates a direct correlation between tumor incidence and LOH of Apc. Furthermore, more non-tumor colon epithelial cells in Apc^min/+/Sigirr^-/- mice showed cytoplasm and nuclear β-catenin staining (Suppl. Fig. 5), suggesting that Apc^min/+/Sigirr^-/- mice might have earlier and higher frequency of LOH of Apc in colon epithelial cells. Taken together, these results suggest that the increased chromosomal instability of the SIGIRR-deficient epithelium might contribute to a higher rate of LOH of Apc and consequent accumulation of β-catenin, promoting tumor initiation in Apc^min/+/Sigirr^-/- colon.

Hyper Akt-mTOR signaling contributes to spontaneous colonic polyposis in Apc^min/+/Sigirr^-/- mice

The mammalian target of rapamycin (mTOR) is a serine/threonine kinase that belongs to the phosphoinositide 3-kinase related kinase family. Akt-mTOR activated pathways have been linked to cell cycle progression and LOH of Apc, establishing Akt-mTOR signaling as a critical pathway driving tumor initiation in the Apc^min/+ mouse ³¹. Interestingly, we observed that SIGIRR deficiency resulted in increased constitutive activation of Akt and mTOR in the colon epithelial cells (Fig. 5A and Suppl. Fig. 6), implicating SIGIRR in the suppression of tumorigenesis through its impact on Akt-mTOR axis. Moreover, Akt and mTOR phosphorylation was further induced to a higher level by IL-1, CpG or LPS in Apc^min/+/Sigirr^-/- than that in Apc^min/+/Sigirr^+/+ colon epithelial cells (Fig. 5A), suggesting that IL-1-TLR signaling probably contribute to the aberrant activation of the Akt-mTOR pathway in Apc^min/+/Sigirr^-/- colonic crypts.

Fig. 5.

Elevated mTOR pathway critically regulates tumorigenesis in Apc^min/+/Sigirr^-/- mice. (A) Cell lysates (50μg per sample) made from primary colonocytes untreated or treated by IL-1, LPS, or CpG for indicated times, were probed by antibodies against Akt-mTOR pathway and cell cycle regulators. (B) Rapamycin treatment inhibits cell proliferation in the colon of Apc^min/+/Sigirr^-/- mice. Four weeks old mice were treated by rapamycin (Sigma, St. Louis) (5mg/kg, 5 I.P. injections per week) for two weeks. Representative Ki-67 staining images were shown. Magnification: 100×. Quantification of Ki-67-positive cells was conducted on colon sections from three treated and three untreated mice (more than 50 well-oriented crypts were counted). (C) Rapamycin treatment inhibits microadenoma formation. Four weeks old mice were treated by rapamycin (5mg/kg, 5 I.P. injections per week) for two weeks. Swiss-rolls of colon and small intestine from untreated or rapamycin-treated mice were cut into 5μM consecutive sections and stained by H&E. Microadenomas were quantified over H&E stained sections from untreated or rapamycin-treated Apc^min/+/Sigirr^-/- or Apc^min/+/Sigirr^+/+ mice (N=5 per group). Error bars represent ±SEM, and Student’s T Test was conducted. * = p<0.05, **=p<0.01.

Previous studies have shown that activation of the mTOR pathway promotes protein translation through the phosphorylation and inactivation of 4E-BP1 (a repressor for protein translation), and phosphorylation and activation of p70S6K ³⁴. We found that 4E-BP1 phosphorylation was elevated in untreated or IL-1, CpG or LPS-treated Apc^min/+/Sigirr^-/- cells than that in Apc^min/+/Sigirr^+/+ cells (Fig. 5A). Concomitantly, more phosphorylation of p70S6K and its substrate ribosomal protein S6 were detected in Apc^min/+/Sigirr^-/- cells compared to that in Apc^min/+/Sigirr^+/+ cells untreated or treated by IL-1, CpG or LPS (Fig. 5A), confirming the hyper activation of the Akt-mTOR pathway in Apc^min/+/Sigirr^-/- colonic crypts. Moreover, the mTOR pathway was also constitutively activated in Sigirr^-/- colon epithelium and further elevated at a higher level upon IL-1 or LPS stimulation compared to that in wild-type cells (Suppl. Fig. 7). These results indicate that SIGIRR critically controls mTOR activation in colonic epithelium by regulating IL-1 or commensal-TLR signaling.

The Akt-mTOR axis has been shown to promote cell cycle progression through its impact on posttranscriptional control of the key cell cycle regulators ³⁴. Cyclin D1, cyclin E and c-Myc were indeed expressed at a higher level in Apc^min/+/Sigirr^-/- crypt cells compared to that in Apc^min/+/Sigirr^+/+ cells (Fig. 5A). Furthermore, greater phosphorylation of cdk2 was detected in Apc^min/+/Sigirr^-/- crypt cells compared to that in Apc^min/+/Sigirr^+/+ cells (Fig. 5A). Apc^min/+/Sigirr^-/- crypt cells also displayed decreased expression of cyclin-dependent kinase inhibitor p27, a negative regulator for G1-S transition (Fig. 5A). It is conceivable that dysregulation of these key cell cycle regulators might promote aberrant cell proliferation, contributing to the increased LOH of Apc and consequent more tumor initiation in Apc^min/+/Sigirr^-/- crypt cells.

Notably, Ak-mTOR signaling has also been shown to converge onto the Wnt-β-catenin pathway by phosphorylating GSK3β, a key kinase in controlling the stability of β-catenin. GSK3β phosphorylation was increased in both untreated and stimulated (by IL-1, CpG or LPS) Apc^min/+/Sigirr^-/- colon crypt cells compared to that in Apc^min/+/Sigirr^+/+ cells (Fig. 5A). Interestingly, elevated GSK3β phosphorylation was observed at the base of the crypts of Apc^min/+/Sigirr^-/- mice as compared to that in Apc^min/+/Sigirr^+/+ mice (Suppl. Fig. 6). These results implicate that loss of SIGIRR might also directly impact on the β-catenin pathway through the Akt-mTOR-GSK3β axis, contributing to the increased cell proliferation and LOH of Apc in Apc^min/+/Sigirr ^-/- colon crypt cells.

Since SIGIRR-deficient epithelial cells displayed constitutive and hyper activation of mTOR in response to TLR-IL-1R ligands, we hypothesize that the Akt-mTOR activated pathway is the key molecular mechanism underlying the increased cell proliferation, LOH of Apc, and tumor initiation in Apc^min/+/Sigirr^-/- colonic epithelia. Rapamycin has been shown to specifically inhibit mTOR signaling and suppress intestinal polyp formation ³⁵. We tested whether rapamycin is able to suppress tumor formation in Apc^min/+/Sigirr^-/- colon. The Apc^min/+/Sigirr^-/- mice were treated with rapamycin (5mg/kg, 5 times/week) by I.P. injection for two weeks for microadenoma formation. As expected, rapamycin treatment effectively inhibited mTOR pathway (Suppl. Fig. 8). Remarkably, cell proliferation, β-catenin activation, microadenoma formation were greatly inhibited in Apc^min/+/Sigirr^-/- mice by rapamycin (Fig. 5B and C, and Suppl. Fig. 9-10). Moreover, rapamycin treatment also resulted in reduced microadenoma formation in Apc^min/+/Sigirr^+/+ mice, supporting a critical role for mTOR in driving tumor initiation in Apc^min/+ mice (Suppl. Fig. 9-10). Furthermore, extended rapamycin treatment for 6 weeks substantially inhibited tumor growth in Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- mice (Suppl. Fig. 11). Taken together, these results confirm that mTOR signaling plays a critical role in spontaneous intestinal polyposis in Apc^min/+/Sigirr^-/- mice.

Increased inflammatory state within Apc^min/+/Sigirr^-/- colon tumors

The fact that SIGIRR deficiency leads to increased tumor incidence and the size of tumors suggests that SIGIRR-modulated signaling probably also impacts on tumor promotion/progression. NFκB, a major component of TLR signaling, plays a critical role in tumor promotion by induction of proinflammatory signaling. In Apc^min/+/Sigirr^+/+ epithelium, activated NFκB staining was relatively weak, and exclusively induced in mature colonocytes at the top of crypts (Fig. 6A), which is consistent with a previous report documented NFκB activation in colon crypts by in vivo imaging ³⁶. In contrast, NFκB activation was detected over the whole crypt axis, including differentiated and proliferating cells in Apc^min/+/Sigirr^-/- crypts (Fig. 6A).

Fig. 6.

Elevated NFκB activation and tumor-promoting inflammation in Apc^min/+/Sigirr^-/- mice. (A) Immunohistochemistry staining of p-p65 or p-STAT3 on colon sections from Apc^min/+/Sigirr^+/+ and Apc^min/+/Sigirr^-/- mice. Representative images are shown based on similar results obtained from three pairs of mice. Magnification: 100× or 400×. (B) Induction of gene expression was examined by real-time PCR. Normal mucosal and tumors were dissected from 4 months old mice as described above. Real-time PCR was conducted on RNA samples made from 5 normal crypts and 10 tumors (about 2mm in diameter) in each group. Student’s T Test was conducted, and data was shown as mean ±SEM. * = p<0.05, **=p<0.01.

Consistent with increased activation of NFκB, a subset of NFκB target genes were up-regulated in Apc^min/+/Sigirr^-/- colon. We detected much stronger induction of inflammatory gene COX2 and cytokines (IL-6 and IL-23) in both normal tissues and tumors of Apc^min/+/Sigirr^-/- mice (Fig. 6B). In particular, IL-6 has been shown to promote tumor progression in inflammation-associated cancer models through the activation of oncogene STAT3 ²³. Interestingly, STAT3 was predominantly activated at the base of the crypts, and its activation was greatly elevated in Apc^min/+/Sigirr^-/- mice (Fig. 6A). Consistently, cyclin D1, Bcl-xL and c-myc, which have been shown to be induced by NFκB and STAT3, were up-regulated at a higher levels in the tumors of Apc^min/+/Sigirr^-/- mice (Fig. 6B), which might be responsible for the increased tumor progression in these mice.

The critical role of epithelial-derived SIGIRR in controlling colon tumorigenesis

Our results suggest that SIGIRR might provide a mechanism by which the normal intestinal epithelium guards against chronic activation and tumorigenesis in the presence of commensal flora. To test the role of epithelial-derived SIGIRR in controlling spontaneous intestinal tumorigenesis, we crossed gut epithelial specific SIGIRR-transgenic mouse (Sigirr IEC-TG, driven by Fabp promoter ¹⁵) onto Apc^min/+/Sigirr^-/- background to generate Apc^min/+/Sigirr^-/-/IEC-TG mice, restoring SIGIRR expression only in intestinal epithelium. Compared to Apc^min/+/Sigirr^-/- mice, Apc^min/+/Sigirr^-/-/IEC-TG mice had fewer tumors in their colon and small intestine (Fig. 7). Moreover, the size of the tumors formed in Apc^min/+/Sigirr^-/-/IEC-TG mice were smaller than those in Apc^min/+/Sigirr^-/- mice, demonstrating SIGIRR-regulated TLR/IL-1R signaling in epithelium plays a critical role in tumorigenesis in the Apc^min/+ model.

Fig. 7.

Epithelial-derived SIGIRR inhibits tumorigenesis in Apc^min/+/Sigirr^-/- mice. Fabp-SIGIRR transgenic mouse (IEC-TG) ¹⁵ was bred onto Apc^min/+/Sigirr^-/- to generate Apc^min/+/Sigirr^-/- and Apc^min/+/Sigirr^-/-/IEC-TGmice mice. Littermates at age of 4 month old were sacrificed to count tumors formed in colon and small intestine. ANOVA was used for statistic analysis. *=p<0.05.

Discussion

In this manuscript, we report that SIGIRR is a critical tumor suppressor that controls colonic tumorigenesis by inhibiting TLR-induced mTOR-mediated cell cycle progression and consequent genetic instability. SIGIRR deficiency leads to increased LOH of Apc and microadenoma formation, resulting in spontaneous colonic polyposis in Apc^min/+/Sigirr^-/- mice. While Akt, mTOR and their downstream substrates are hyper-activated in Apc^min/+/Sigirr^-/- colon epithelium, the inhibition of mTOR pathway by rapamycin results in decreased microadenoma formation in Apc^min/+/Sigirr^-/- mice. Importantly, the spontaneous colonic polyposis in Apc^min/+/Sigirr^-/- mice is dependent on the presence of commensal bacteria in the gut, implicating the critical role of TLR signaling in colonic tumorigenesis in these mice.

We have previously used colitis-associated cancer (CAC) model to assess the role of SIGIRR in inflammation-induced tumorigenesis ¹⁵. Using the CAC model, we reported that SIGIRR deficiency increased tumor promotion and progression. Due to lack of specific molecular markers, we failed to determine whether SIGIRR deficiency had direct impact on tumor initiation. The Apc^min/+ model is based on a defined mutation in the tumor suppressor Apc gene and does not involve chronic inflammation. Importantly, the Apc^min/+ model is a spontaneous colon cancer model mimicking the FAP syndrome, and LOH of Apc is the exclusive genetic alteration leading to the tumor initiation in Apc^min/+ mouse ^{29, 30, 37}. Therefore, introduction of a second mutation into the Apc^min/+ mice allows us to determine whether this particular genetic alteration has any impact on LOH of Apc, which is the key initiation event in intestinal tumorigenesis.

The Apc^min/+ mouse model helped to define for the first time about the critical role of SIGIRR in tumor initiation. LOH of Apc was detected in all the polyps, indicating that SIGIRR deficiency increased tumor initiation in Apc^min/+/Sigirr^-/- mice through its impact on LOH of Apc in epithelium. Consistent with this, microadenomas were developed at a higher rate in Apc^min/+/Sigirr^-/- mice, and displayed constitutive β-catenin activation. Chromosomal instability, a process directly linked to tumor initiation by promoting LOH of Apc, can occur in cells with increased cell proliferation due to dysregulated cell cycle ^{31, 38}. The anaphase-bridge index, an aberrant event commonly linked to genetic instability, was indeed enhanced in Apc^min/+/Sigirr^-/- colon epithelium. Therefore, SIGIRR can directly impact on genetic instability and tumor initiation in colon epithelium by regulating TLR-IL-1R signaling.

Another important finding of this study is the discovery of the critical molecular mechanism for SIGIRR-modulated TLR-mediated tumor initiation. SIGIRR deficiency leads to hyper TLR-IL-1R-mediated Akt-mTOR signaling, a critical pathway for cell cycle progression, LOH of Apc and consequent tumor initiation in the Apc^min/+ mouse. Recent studies have demonstrated that the Akt-mTOR pathway increases protein translation of the key cell cycle regulators, by mediating the phosphorylation of a variety of translational initiation regulatory proteins, such as 4EBP and eIF4G, directly or indirectly through p70S6K^{34, 39}. Elevated expression of cyclin D1, cyclcin E and c-myc was indeed observed in Apc^min/+/Sigirr^-/- crypt cells. Therefore, hyper-activation of mTOR and p70S6K in Apc^min/+/Sigirr^-/- colonocytes might be responsible for increased cell proliferation and higher LOH of Apc in Apc^min/+/Sigirr^-/- colon. Consistently, the abrogation of mTOR pathway by rapamycin resulted in decreased cell proliferation, less microadenoma formation in Apc^min/+/Sigirr^-/- mice. It is important to point out that intestinal microadenoma formation in Apc^min/+/Sigirr^+/+ was also inhibited by rapamycin, demonstrating that mTOR is a key signaling pathway required for intestinal tumorigenesis in Apc^min/+ mice. However, it is intriguing that depletion of microflora only had minimum impact on the intestinal tumorigenesis in Apc^min/+/Sigirr^+/+ mice, whereas antibiotic treatment greatly diminished intestinal tumorigenesis in Apc^min/+/Sigirr^-/- mice. These results indicate that the intestinal tumorigenesis in Apc^min/+/Sigirr^-/-, but not in Apc^min/+/Sigirr^+/+ mice, is dependent on gut microflora and TLR signaling. While it remains unclear how mTOR is required for intestinal tumorigenesis in Apc^min/+/Sigirr^+/+ mice, our results suggest that SIGIRR suppresses intestinal tumorigenesis at least in part through its impact on TLR-induced mTOR activation. Collectively, we identified mTOR for the first time for its role in linking TLR-IL-1R signaling to tumor initiation.

Dysregulated mTOR pathway has been linked to a broad range of human cancers, including familial and sporadic colorectal cancers ⁴⁰. Furthermore, several upstream regulators of mTOR, such as PTEN, LKB1 and TSC1/2, are genetically mutated and functionally associated with inherited cancer syndromes such as Cowden Syndrome, Peutz-Jeghers Syndrome and Harmatomatous Syndrome Tuberous Sclerosis Complex ^{39, 41}. Consistent with the effect of rapamycin in Apc^min/+/Sigirr^-/-, rapamycin and its analogues have been recently approved by FDA to treat advanced renal cell carcinoma, and shown promising signs in treating endometrial cancers and mantel cell lymphoma. Better understanding of mTOR regulation will be important for the development of this promising cancer therapy.

Taken together, our current study has clearly revealed a critical role for SIGIRR in suppressing tumor initiation in the gut epithelium. Interestingly, MyD88-deficiency on Apc^min/+ resulted in impaired TLR-induced inflammation and tumor promotion, but not tumor initiation ⁸, implicating MyD88 is dispensable in regulating tumor initiation in epithelium. While elevated NFκB signaling and inflammation mediated by immune cells primarily contribute to tumor promotion/progression in Apc^min/+/Sigirr^-/- mice, this study suggests that SIGIRR-regulated TLR-IL-1R signaling in epithelial cells might directly impacts on tumor initiation. It is important to point out that re-expression of SIGIRR only in intestinal epithelium resulted in a modest, but significant reduction of tumorigenesis in Apc^min/+/Sigirr^-/- mice, suggesting the critical role of epithelial-derived SIGIRR in modulating colon tumorigenesis. On the other hand, the Apc^min/+/Sigirr^-/-IEC-TG mice still exhibited higher tumorigenesis than Apc^min/+/ Sigirr^+/+ mice, suggesting that SIGIRR deficiency in other lineages, such as T cells, may also play a role in controlling tumorigenesis in Apc^min/+/Sigirr^-/- mice as well.