MyD88 signaling plays a critical role in the development of protective host responses to C. rodentium infection, controlling pathogen burdens, driving inflammation and promoting IEC integrity/proliferation [5], [6]. As a result, Myd88 −/− mice are highly susceptible to C. rodentium infection, carrying heavy pathogen burdens and suffering severe mucosal damage and loss of IEC barrier function (Figure 1A–1D). Since many of these MyD88 dependent protective responses involve changes in IEC function, we tested whether they reflected MyD88 signaling within the IEC. Crossing Myd88 flox mice with mice expressing the cre enzyme under the villin (IEC specific) promoter, we generated IEC-Myd88 −/− mice that were then infected with C. rodentium along with control MyD88 flox mice. Interestingly, MyD88 signaling within IEC had little impact on the host response to infection. Pathogen burdens in the IEC-Myd88 −/− mice were similar to those in control mice (Figure 1A), as were crypt heights and tissue integrity. Moreover, we noted no overt difference in IEC barrier function or inflammatory cell recruitment to the infected cecum between these two mouse strains (Figure 1B–1D). These data indicate that MyD88 signaling within IEC plays little role in driving the innate host response to C. rodentium infection.

While it is unclear why IEC play such a limited role in driving MyD88 dependent responses to C. rodentium infection and other forms of colitis [23], [24], [25], we recently showed that SIGIRR, a negative regulator of TLR and IL-1R signaling, is expressed by IEC and limits their responses to IL-1β and to most bacterial PAMPs [20]. To test the impact of SIGIRR in vivo, Sigirr −/− mice were infected with C. rodentium under co-housing conditions (WT and Sigirr −/− mice in same cage) or with WT and Sigirr −/− mice housed in different cages, and similar results were obtained. As shown in Figure 2A, infected WT mice displayed modest weight loss at day (D) 2 post-infection (pi) followed by recovery and progressive weight gain. In contrast, infected Sigirr −/− mice exhibited significantly greater (P<0.05) weight loss (∼10%) starting from D4 until D10 pi (Figure 2A). Following their euthanization, the Sigirr −/− mice showed greater macroscopic signs of infection than WT mice, with overt loss of stool content and edema seen throughout their large intestines (Figure 2B). Moreover, pus and blood were often found within the ceca of infected Sigirr −/− mice and between 60 and 80% of these mice suffered macroscopic cecal ulcers at D6 and D10 pi, a phenotype not observed in WT mice.

Based on the observed pathology, as well as the fact that baseline SIGIRR gene transcript levels in WT mice were highest in the cecum (not shown), we focused our subsequent analysis on the cecum. At both D6 and D10 pi, Sigirr −/− mice showed significantly higher cecal pathology scores compared to WT mice, reflecting increased crypt hyperplasia, edema, and greater inflammatory cell infiltration (Figure 2C and 2D, P<0.05). Specifically, we noted increased macrophage and neutrophil infiltration into the cecal mucosa of Sigirr −/− mice. Similarly, gene transcript levels for chemokines (MCP-1 and MIP-2α), antimicrobial factors (mCRAMP, β-defensin III, and Reg-3γ) and cytokines (IL-17A, TNF-α and IFN-γ) were all significantly elevated at D6 and D10 pi in Sigirr −/− mice compared to WT mice (Figure S1). Based on their exaggerated colitic responses, we expected Sigirr −/− mice would show some level of protection against infection. Instead, they carried significantly higher (10–100 fold) C. rodentium burdens in their ceca and colons compared to WT mice at both D6 and D10 pi (Figure 2E, P<0.01). Interestingly, these higher burdens reflected increased numbers of C. rodentium infecting the intestinal mucosal surface, as well as within the intestinal lumen. Notably, despite their high intestinal pathogen burdens, Sigirr −/− mice did not show increased susceptibility at systemic sites, as C. rodentium burdens in the spleens and livers of Sigirr −/− mice were similar to those in WT mice (Figure 2F).

Despite carrying heavy C. rodentium burdens and suffering exaggerated colitic damage, Sigirr −/− mice survived infection and healed their mucosal ulcers by D14 pi (Figure S2), suggesting that intestinal mucosal repair was enhanced in these mice. We assessed IEC proliferation in Sigirr −/− and WT mice (by Ki-67 staining) to examine their ability to repair mucosal damage and also measured IEC barrier integrity. Increased IEC proliferation, (as measured by the percentage of IEC in a crypt showing Ki-67 positive staining) was evident by D4 pi in the Sigirr −/− mice, but not in WT mice (Figure 3A and 3B, 47±1% vs 21±3%, P<0.01). By D6 pi, IEC proliferation in Sigirr −/− mice was significantly greater than in WT mice (Figure 3A and 3B, 75±7% vs 41±4%, P<0.01) and similar results were observed at D10 pi (Figure 3A and 3B, 60±5% vs 34±4%, P<0.01). We also assessed IEC barrier function by oral fluorescein isothiocyanate dextran (FD4) gavage, and while infection did cause an increase in serum FD4 levels in both strains, we noted no differences in FD4 levels between infected Sigirr −/− mice and WT mice (Figure 3C).

We next examined whether SIGIRR impacted on very early host responses to C. rodentium by employing the cecal ligation model of acute C. rodentium infection, allowing us to inject equal pathogen numbers into the ceca of WT and Sigirr −/− mice. As expected, similar numbers of C. rodentium were found infecting the cecal tissues of WT and Sigirr −/− mice after 10 hr of infection (WT: 1.9±0.7×10⁸ CFU/g, Sigirr −/−: 1.2±0.5×10⁸ CFU/g). Cecal tissues were collected and qPCR was used to determine transcript levels for genes encoding several antimicrobial peptides and other IEC derived factors. As shown in Figure 3D, infection of WT mice led to a modest increase in the transcription of genes encoding antimicrobial peptides (mCRAMP, Reg3γ, β-defensin III) and chemokines (MCP1 and MIP2α). In contrast, we observed strikingly elevated transcription of most genes in infected Sigirr −/− mice (P<0.05), despite the two mouse strains carrying similar pathogen burdens. These results indicate that in the absence of SIGIRR, the host intestine responds very rapidly to C. rodentium infection, and that the exaggerated colitis suffered by orally infected Sigirr −/− mice likely reflects both their lack of SIGIRR, as well as their increased pathogen burdens.

SIGIRR thus plays a critical role in suppressing C. rodentium induced colitis, but the cellular source of SIGIRR is unclear. SIGIRR is expressed by IEC, as well as by several BM-derived cell types including dendritic cells and T cells [17], [18]. To better define the cell type(s) regulated by SIGIRR during C. rodentium infection, we generated haematopoietic chimeras by BM transplantation between Sigirr −/− mice and WT mice expressing Ly5.1 on their BM-derived cells. Following infection by C. rodentium, Sigirr −/− + WT BM chimeric mice developed severe colitis, similar to that seen in Sigirr −/− mice. Moreover the ceca of infected WT + Sigirr −/− BM mice were macroscopically similar to those of WT mice, although histologically, they displayed a greater increase in crypt lengths (Figure 4B–4D) than WT mice. Pathology scoring also revealed Sigirr −/− + WT BM mice had exaggerated crypt hyperplasia, edema and immune cell infiltration when compared to WT + Sigirr −/− BM mice (Figure 4D, P<0.05). Similar to Sigirr −/− mice, Sigirr −/− + WT BM chimeras carried significantly higher intestinal C. rodentium burdens, compared to WT mice and WT + Sigirr −/− BM mice at D10 pi (Figure 4A). Furthermore, Sigirr −/− + WT BM mice exhibited hyper-proliferative IEC responses (65±5%) similar to those of Sigirr −/− mice (72±6%), as revealed by Ki-67 staining while the IEC proliferative response was similar between WT mice (39±4%) and WT + Sigirr −/− BM mice (33±3%) (Figure 4E and 4F). Thus expression of SIGIRR by non-BM derived cells plays the major role in suppressing the host inflammatory response and controlling C. rodentium burdens. Moreover, loss of SIGIRR in this compartment leads to the exaggerated colitis suffered by C. rodentium infected Sigirr −/− mice.

SIGIRR is a negative regulator of TLR and IL-1R signaling, but the identity of the (unregulated) receptor(s) that drive the exaggerated colitis and IEC proliferative responses in infected Sigirr −/− mice is unclear. To identify the receptor(s) involved, we began by removing MyD88 dependent signaling from Sigirr −/− mice. We crossbred MyD88 −/− mice and Sigirr −/− mice and infected the resulting MyD88/Sigirr −/− mice. Similar to MyD88 −/− mice, MyD88/Sigirr −/− mice became sick and required euthanization by D6-8 pi (Figure 5A). They also carried heavy C. rodentium burdens (∼10¹⁰ CFU/g) at D6 pi and suffered severe mucosal damage and increased intestinal permeability much like Myd88 −/− mice (not shown), confirming SIGIRR's actions in this model depends on MyD88 signaling.

Next, we sought to identify the specific TLRs responsible for the exaggerated colitis seen in Sigirr −/− mice. Previous studies have shown that the host response to C. rodentium infection is largely dependent on TLR2 and TLR4 [7], [10]. Therefore, we crossbred Tlr2 −/− mice and Tlr4 −/− mice with Sigirr −/− mice to generate Tlr2/Sigirr −/− mice and Tlr4/Sigirr −/− mice. Upon infecting these mice along with WT and Sigirr −/− mice, and Tlr2 −/− and Tlr4 −/− mice, all mouse strains were found to survive the infection (Figure 5A and Figure S3). As expected, all readouts were dramatically greater in the Sigirr −/− mice, as compared to WT mice. Interestingly, Tlr2 −/− and Tlr4 −/− mice developed modestly elevated pathology compared to WT mice but much less than that suffered by Sigirr −/− mice (Figure S3). Moreover, the responses in Tlr2/Sigirr −/− mice and Tlr4/Sigirr −/− mice were similar to those in the Sigirr −/− mice including pathogen burdens and pathological damage (Figure 5B and 5C). Hence, the data suggest that the exaggerated damage and increased pathogen burden observed in the Sigirr −/− mice are largely independent of TLR2 and TLR4.

In previous studies, we showed TLR2 and TLR4 play specific and critical roles in controlling IEC responses during C. rodentium infection [9]. TLR4 signaling drives IEC proliferation during infection, while TLR2 signaling promotes IEC integrity. In fact, infected Tlr2 −/− mice suffer exaggerated IEC barrier permeability, and impaired phospho-STAT3 signaling within their IEC, with these defects leading to cecal and colonic ulcers and high mortality rates [10]. When we assessed IEC responses in the Tlr2/Sigirr −/− mice and Tlr4/Sigirr −/− mice, the Tlr2/Sigirr −/− mice exhibited similar proliferative responses (74±7%, Figure 5D) to Sigirr −/− mice (75±7%) as demonstrated by Ki-67 staining. Interestingly, Tlr4/Sigirr −/− mice showed an intermediate level of IEC proliferation (58±3%), modestly but not significantly reduced compared to Sigirr −/− mice, yet still significantly greater (P<0.05) than WT mice (36±4%). This data suggests that signaling through other SIGIRR-regulated receptors drives IEC proliferation in Sigirr −/− mice, even in the absence of TLR4.

We next tested IEC barrier function in these mice (and Tlr2 −/− mice). At D6 pi, FD4 serum levels within Tlr2 −/− mice were significantly elevated compared to WT mice, confirming their impaired barrier function (Figures 5F and S3, P<0.05). Similar to WT mice and Tlr4 −/− mice, Sigirr −/− mice showed no overt barrier dysfunction, but strikingly, the concurrent loss of SIGIRR (Tlr2/Sigirr −/− mice) compensated for the barrier dysfunction seen in Tlr2 −/− mice. Furthermore, the impaired phospho STAT-3 staining seen in Tlr2 −/− mice was restored in Tlr2/Sigirr −/− mice (Figure 5E). These results indicate that loss of SIGIRR leads to IEC responses that promote mucosal integrity in a manner largely able to compensate for the defects caused by TLR2 or TLR4 deficiency, suggesting another SIGIRR regulated receptor is at play.

Aside from negatively regulating TLRs, SIGIRR also negatively regulates IL-1R and IL-18R signaling [17], [19], [26]. We tested Il-18/Sigirr −/− mice, and found no significant differences with Sigirr −/− mice in their colitic responses (not shown). To test whether IL-1R signaling contributes to the exaggerated colitic responses observed in infected Sigirr −/− mice, we first treated WT and Sigirr −/− mice with anakinra, an IL-1R antagonist [27]. While treatment with anakinra did not impact pathogen burdens (Figure 6A), it lessened the macroscopic damage that developed in Sigirr −/− mice compared to vehicle treated Sigirr −/− mice (Figure 6B and C). Histological pathology scoring revealed treatment with anakinra improved the epithelial integrity of the Sigirr −/− mice (Figure 6D–E and Figure S4). It also abrogated the exaggerated IEC proliferative response observed in Sigirr −/− mice (Figure 6F).

To further explore the potential for IL-1R signaling to contribute to the exaggerated responses seen in infected Sigirr −/− mice, we assessed the expression of its ligands IL-1α and IL-1β. Interestingly, IL-1α gene expression showed a dramatic increase in the cecal tissues of Sigirr −/− mice during infection. Moreover protein analysis (ELISA) revealed significantly increased IL-1β levels in Sigirr −/− mice compared to WT mice under both baseline and infected conditions (Figure 7A and 7B, P<0.05). Similarly, cecal crypts isolated from Sigirr −/− mice showed elevated IL-1β levels under baseline and infected conditions as measured by Western blot analysis (Figure 7C).

Continuing to address the potential role of IL-1R signaling in the response of the Sigirr −/− mice, we generated Il-1r/Sigirr −/− mice and infected them, along with Sigirr −/− and Il-1r −/− mice with C. rodentium. The Il-1r/Sigirr −/− mice rapidly lost weight during infection, requiring their euthanization by D10 pi (Figure 7D). At D6 pi, we noted that although the Sigirr −/− mice, Il-1r −/− mice and Il-1r/Sigirr −/− mice all carried similar intestinal pathogen burdens, the Il-1r/Sigirr −/− mice suffered extensive mucosal damage compared to the Sigirr −/− mice and the Il-1r −/− mice (Figure 7E–F and Figure S5). Remarkably, the protective, hyper-proliferative IEC responses observed in the Sigirr −/− mice were absent in the Il-1r/Sigirr −/− mice (Figure 7G), with almost 3 fold more IEC staining positively for Ki-67 in the Sigirr −/− mice compared to the Il-1r/Sigirr −/− mice (72±6% vs 26±5%, P<0.01). Moreover, infected Il-1r/Sigirr −/− mice suffered significantly greater intestinal permeability compared to Sigirr −/−mice as measured by serum FD4 levels (Figure 7H, P<0.01), whereas the Il-1r −/− mice showed serum FD4 levels that were intermediate between the two other mouse strains. These findings indicate that the exaggerated and protective IEC responses seen in Sigirr −/− mice are largely IL-1R dependent and IL-1R may be the key receptor regulated by SIGIRR in this model.

While Sigirr −/− mice develop exaggerated colitic and IEC proliferative/reparative responses during infection, these mice also appear more susceptible to C. rodentium colonization, carrying significantly higher pathogen burdens than mice expressing SIGIRR. To test whether Sigirr −/− mice are truly more susceptible to C. rodentium infection, we infected WT and Sigirr −/− mice with a 100 fold lower dose (LD) of C. rodentium and followed their course of infection by plating feces. Previous studies have found this low dose is unable to effectively infect WT mice [28], and indeed, LD WT mice shed only low numbers of C. rodentium (10⁴–10⁶ CFU/g of feces). In contrast, LD Sigirr −/− mice showed significantly (100×) higher C. rodentium burdens by D2 pi and their pathogen burdens continued to increase until D6 pi when they reached ∼2.6×10⁹ CFU/g of feces, similar to the burdens carried by Sigirr −/− mice given a full infectious dose (Figure 8A). At D10 pi, following tissue collection, Sigirr −/− mice were found to carry 1000 fold more C. rodentium in both cecal and colonic tissues compared to WT mice (Figure 8B, P<0.01). As expected from their pathogen burdens, Sigirr −/− mice developed severe inflammation and cecal pathology similar to that seen with a full dose infection whereas WT mice did not display any overt or histological signs of infection or pathology (Figure 8B).

Thus Sigirr −/− mice show unusual susceptibility to C. rodentium colonization and infection. We next tested whether they showed heightened susceptibility to infection by another enteric bacterial pathogen. We orally inoculated WT and Sigirr −/− mice with the enteric bacterial pathogen Salmonella enterica serovar Typhimurium. S. Typhimurium is known to poorly colonize the intestines of mice, in large part because it is unable to displace competing commensal microbes [15], [29], [30]. As a result, pretreating mice with the antibiotic streptomycin is commonly used to remove competing commensals and facilitate S. Typhimurium colonization and infection of the cecum [15]. While we found Sigirr −/− mice were more susceptible than WT mice to S. Typhimurium infection following streptomycin pretreatment (not shown), we also tested their susceptibility to S. Typhimurium in the absence of antibiotic pretreatment. As expected, oral inoculation of WT mice with a dose of 5×10⁶ CFU of ΔaroA S. Typhimurium did not lead to any significant cecal pathology or colonization by this pathogen. In contrast, the same dose led to dramatically heavier (1000 fold) S. Typhimurium burdens in the cecum and severe inflammation and pathology in the ceca of Sigirr −/− mice (Figure 8C), similar to that seen in S. Typhimurium infected mice pretreated with streptomycin. Taken together, these results show that loss of SIGIRR expression leads to dramatically enhanced susceptibility to oral infection by enteric bacterial pathogens, in a manner able to overcome the typical commensal microbe dependent resistance to such colonization.

Previous studies by our group and others found that differences in the makeup of the intestinal microbiota can affect host susceptibility to C. rodentium infection [31]. Previous sequencing studies found no overt differences between the microbiota of Sigirr −/− mice and WT mice [32], but considering the susceptibility of Sigirr −/− mice to enteric pathogens, we speculated they might instead carry either fewer total commensal microbes, or harbour a different microbial composition than WT mice. We enumerated the microbiota in the feces of these mice by fluorescence microscopy under baseline conditions, but noted no significant differences in commensal numbers between WT mice and Sigirr −/− mice (Figure 9A). Through qPCR analysis of fecal pellets, we found members of the phylum Bacteroides to be the dominant taxonomic group present within the stool of both Sigirr −/− and WT mice (∼50–60%), with smaller numbers of Firmicutes (∼20%), and other microbes making up minor components of the microbiota. We also looked at representative members of these phyla such as Bifidobacteria, Lactobacillus and Clostridium [33] and did not note any large or overt differences between the WT and Sigirr −/− mice (data not shown).

We also analyzed the above mentioned stool samples using fluorescence in situ hybridization (FISH). The FISH analysis relied on staining microbes with probes for Eubacteria as well as probes specific for γ-Proteobacteria or probes that recognize only three genera (Cytophaga, Flexibacter and Bacteroides) from the phylum Bacteroides. Microbes that stained with the Eubacteria probe, but did not stain with the other probes were placed in the “Other” category, with these microbes primarily encompassing the Firmicutes as well as minor components of the microbiota. While the makeup of the microbiota as determined by the FISH staining may be less accurate when compared to the more thorough qPCR approach (Figure 9B and 9C), both techniques were consistent in showing no statistically significant differences between uninfected Sigirr −/− and WT mice, confirming previous reports by Chan et al. [32]

Interestingly, Lupp et al. (2007), showed that C. rodentium infection causes a dramatic shift in the makeup of the gut microbiota as well as a reduction in total commensal numbers. These commensal changes were host mediated and took approximately 6 days to occur, corresponding with the time taken for C. rodentium to heavily colonize its host. We, and others have speculated these changes in the microbiota may facilitate C. rodentium colonization, as antibiotic based removal of commensals accelerates the course of C. rodentium infection [31], [33]. We quantified commensal populations within the feces of infected WT mice, and confirmed it takes approximately 6 days for the aforementioned commensal depletion (80%) to occur. In contrast, total commensal numbers in the fecal samples collected from the Sigirr −/− mice dropped dramatically by D1 pi, and this reduction remained evident until D6 pi and beyond (Figure 9D). The depletion of commensal microbes affected all relevant phyla to a roughly similar degree as measured by qPCR (data not shown). The corresponding C. rodentium burdens increased in concert (albeit slightly delayed) with the loss of commensals. Rapid depletion of total commensal microbe numbers was also detected in Sigirr −/− mice following infection with LD C. rodentium (not shown). Similarly, we also observed rapid depletion of total commensal microbe numbers by D2 post-S. Typhimurium infection (in the absence of antibiotics) (Figure 9E, *P<0.05). In contrast, this reduction in commensals was not observed in S. Typhimurium infected WT mice. Interestingly, we did not observe commensal depletion when Sigirr −/− mice were infected with an avirulent strain (ΔescN) of C. rodentium (not shown). These results suggest the impaired resistance to pathogen colonization we observed in Sigirr −/− mice reflects the ability of C. rodentium and S. Typhimurium to rapidly trigger commensal depletion in these mice, thereby opening niches within the intestine for rapid pathogen colonization.

We next addressed the basis for the rapid depletion of commensal microbes seen in infected Sigirr −/− mice. Notably, the massive commensal depletion occurs prior to overt intestinal inflammation. As outlined in Figure 3D, acute exposure to C. rodentium rapidly increased antimicrobial gene expression, likely reflecting responses by IEC. We thus examined whether the IEC of Sigirr −/− mice harbored greater bactericidal activity than WT mice towards commensal bacteria, using commensal E. coli and Lactobacilli caesi as examples. We isolated intestinal crypts under baseline condition from WT and Sigirr −/− mice; collected their crypt supernatants and tested their bactericidal activity against the aforementioned commensal microbes. After a 1 hr incubation, isolates from Sigirr −/− mice showed a killing activity of 29.9±3.5% against commensal E. coli while crypt isolates from WT mice showed only 10.6±9.9%. Similarly, isolates from Sigirr −/− mice displayed greater killing activity than WT mice against Lactobacilli caesi (38.4±2.4% versus 23.4±5.4%) a gram-positive commensal microbe (P<0.05). Therefore, crypt isolates from Sigirr −/− mice possess significantly higher bactericidal activity than those from WT mice.

We next we sought to determine whether the bactericidal activity within the crypt isolates of Sigirr −/− mice increased during infection, and whether the enteric pathogens C. rodentium and S. Typhimurium were more resistant to this bactericidal activity compared to commensal microbes. We isolated intestinal crypts from WT and Sigirr −/− mice after 24 hrs of C. rodentium infection; collected their crypt supernatants and tested their bactericidal activity against both commensal microbes and pathogenic microbes. As shown in Figure 9F, while the killing of L. casei by Sigirr −/− crypt isolates was relatively unchanged from baseline, the killing of the commensal E. coli strain significantly increased to 41.4±3.9% at day 1 pi (P<0.05). Moreover, both C. rodentium and S. Typhimurium were found to be significantly more resistant to the bactericidal activity of the crypt isolates from both WT and Sigirr −/− mice compared to the commensal microbes (P<0.05). Considering the rapid induction of antimicrobial genes seen in the cecal loop model, the heightened antimicrobial activity of Sigirr −/− mice likely explains the rapid commensal depletion seen following their infection.

Mouse strains used in this study: C57B/6 (WT), Tlr2 −/−, Tlr4 −/−, Myd88 −/−, Sigirr −/−, Il-1r/Sigirr−/−, Tlr2/Sigirr −/−, Tlr4/Sigirr −/− and Myd88/Sigirr −/− mice (8–12 weeks) were bred in house, and kept under specific pathogen free conditions at the Child and Family Research Institute. Il-1r1−/− and Myd88 flox (B6.129P2-Myd88^tm1Defr/J) and Villin cre mice were purchased from Jackson Laboratory. All double gene deficient mice were generated by crossbreeding single gene deficient mice. IEC-Myd88 −/− mice were generated by crossbreeding Myd88 flox mice with Villin-Cre mice. Mice were orally gavaged with an overnight culture of ∼2.5×10⁸ CFU of streptomycin resistant C. rodentium DBS100 or 1×10⁶ CFU of ΔaroA S. Typhimurium SL1344 (ΔaroA chosen because it does not kill Nramp1 sensitive mice) [41] and euthanized at specified time points pi. Infections were performed under both co-housing conditions (WT and mutant mice in same cage) and single housing conditions (WT and mutant in different cages). For IL-1R antagonist experiments, anakinra (Kineret) at 1 mg/mouse (or PBS as a control) was injected intraperitoneally twice daily (12 hrs apart) during the course of the infection (6 days) beginning on the day of the infection (day 0) just prior to oral gavage [42]. Body weight data are presented as the mean percentage of starting weight.

All experiments were performed according to protocols approved by the University of British Columbia's Animal Care Committee and in direct accordance with The Canadian Council on Animal Care (CCAC) guidelines. The protocols were approved by the UBC Animal Care committee (Protocols: A11-0290 and A11-0253). Mice were monitored for mortality and morbidity throughout their infection and euthanized if they showed signs of extreme distress or >15% body weight loss. All surgeries were performed under anesthesia (isofluorane and xylazine), and all efforts were made to minimize suffering.

Tissue collection and bacterial counts were performed as described previously [43]. Briefly, mice were anesthetized with isofluorane and euthanized over the course of infection, dissected and their large bowel was divided into cecum and colon to be collected in 10% neutral buffered formalin (Fisher) for histological analyses, or processed for tissue pathology assays. For viable cell counts, cecum, colon tissues and stool pellets were collected separately, and homogenized in PBS pH 7.4, with dilutions plated onto LB agar or streptomycin agar plates. Pathology was scored using a previously adapted scoring system [44]. In brief, paraffin-embedded cecal tissue sections (5 µm) that had been stained with haematoxylin and eosin were examined by two blinded observers. Tissue sections were assessed for i) submucosal edema (0- no change 1- mild 2- moderate 3- profound), ii) hyperplasia (0- no change 1-1-50% 2- 51%–100% 3 - >100%), iii) goblet cell depletion (0 no change 1 mild depletion 2 severe depletion 3- absence of goblet cells), iv) epithelial damage/integrity (0 no change 1- few cells sloughing, epithelial surface rippled 2- epithelial surface is rippled, damaged 3- epithelial surface is severely disrupted/damage, large amount of cells sloughing). For ulceration, an additional score of 1 was added for each 25% (or fraction of) of the tissue in the cross section affected ie.( a small ulcer would score as 4+1, whereas a tissue where 60% of the tissue cross section is ulcerated would be 4+3) v) mononuclear cells infiltration (per 40× field) (0- no change 1- <20 2- 20 to 50 3- >50 cells. The maximum possible score was 18.

Following euthanization of mice, tissues were immediately transferred to RNA-later (Qiagen). Total RNA was purified using QiagenRNEasy kits (Qiagen) stored according to manufacturer's instructions. cDNA was synthesized with Omniscript RT kit (Qiagen) and OligodT (Applied Biological Material Inc) followed by quantitative real-time PCR techniques. qPCR was carried out on MJ Mini-Opticon Real-Time PCR System (Bio-Rad) using IQ SYBR Green Supermix (Bio-Rad) and MIP-2, MIP3-α, β-actin, mCRAMP, Reg 3-γ, β-defensin III primers using sequences and conditions previously described [44]. Il-1α primers were generated by PrimerBank (http://pga.mgh.harvard.edu/primerbank/), 5′ GAAGACTACAGTTCTGCCATT3′, reverse 5′ GACGTTTCAGAGGTTCTCAGAG 3′ Quantification was performed using the Gene Ex Macro OM 3.0 software (Bio-Rad) where PCR efficiencies for each primer set were incorporated into the final calculations.

Immunofluorescence staining of control and infected tissues was performed using previously described procedures [43]. In brief, paraffin embedded tissues were cut (6 µm) and were deparaffinized by heating at 55–65°C for 20 mins, cleared with xylene, rehydrated through an ethanol gradient to water. Sections were blocked using blocking buffer (Goat serum in PBS containing 1% bovine serum albumin (BSA), 0.1% Triton-X100, and 0.05% Tween 20, and 0.05% sodium azide). The primary antibodies used were Ki-67 (Thermo) or phospho STAT-3 (Ab-Cam) while the secondary was AlexaFluor 568-conjugated goat anti-rabbit IgG. Tissues were mounted using ProLong Gold Antifade reagent (Invitrogen) containing DAPI for DNA staining. Sections were viewed on a Zeiss AxioImager microscope and images taken using an AxioCam HRm camera operating through AxioVision software.

Tissue were stained with Ki-67 and mounted with ProLong Gold Antifade with DAPI. Pictures were then taken for each tissue section. The number of Ki-67 positive cells and DAPI positive cells were counted for each crypt and calculated as a ratio of Ki-67 positive cells/DAPI positive cells as % positive cells. At least 6 crypts were counted from each section. The results expressed were averaged from at least 6 tissue sections.

Cecal loop infections were performed as recently described [43]. In brief, mice were anaesthetized (2.5% isofluorane and xylazine at 10 mg/kg) and following a midline abdominal incision, the cecum was exposed. The proximal colon close to the cecum was ligated twice, while the ileocecal valve prevented backflow into the ileum, thus isolating the cecum. C. rodentium cultures were diluted in PBS, and 300 µl of culture containing ∼1×10⁷ CFU were injected into the cecal loop which was returned to the abdominal cavity and the incision closed with discontinuous sutures. The mice were euthanized at 10 hr pi and tissues and stool contents were collected for CFU counts and mRNA analysis as described above.

Bone marrow was isolated from C57BL/6.Ly5.1 (WT), or Sigirr −/− mice. Four million cells were intravenously injected into lethally irradiated wild-type (WT), or Sigirr −/− recipient mice that were allowed to recover for 12 weeks. Reconstitution was confirmed by FACs analysis of peripheral blood leukocytes for Ly5.1 expression. Only mice exhibiting greater than 90% chimerism were infected. Similar methods were performed to obtain Sigirr −/− +WT BM mice.

Enumerating total microbes and FISH staining were performed as described previously [44].

At least 2 fecal pellets were collected from each animal at time points indicated. After homogenization, samples were placed in 10% Neutral Buffered Formalin to a final concentration of 3%. Samples were further diluted 1∶10 in PBS, vortexed briefly, and stored at 4°C. 2–5 µl of the 1∶10 diluted sample stored in PBS was diluted in 1 ml PBS and filtered onto Anodisc 25 filters (Whatman International Ltd) with a pore size of 0.2 µM and 2.5 cm diameter. The samples were allowed to thoroughly dry, and then were mounted on glass slides with ProLong Gold Antifade reagent containing DAPI (Molecular Probes) and viewed as above. The mean number of DAPI positive microbes counted in 3 to 6 randomly chosen fields per disc (1000×) was determined by two scorers. The total number of commensal microbes was calculated based on the mean numbers of all the counted fields and the dilution factor. The total number of commensal microbes was presented as the percentage of uninfected controls. For FISH staining, filters were passed through an ethanol gradient following filtering and samples were incubated overnight at 37°C in the dark with Texas red-conjugated EUB338 general bacterial probe (5′-GCT GCC TCC CGT AGG AGT-3′) and an AlexaFluor 488 conjugated GAM42a probe (5′-GCC TTC CCA CAT CGT TT-3′) that recognizes bacteria that belong to the γ-Proteobacter class [4], [45]. Filters were then washed in the dark with hybridization solution (0.9 M NaCL, 0.1 M TRIS pH 7.2, 30% Formamide, 0.1% SDS) for 15 min with gentle shaking. This step was repeated once with wash buffer (0.9 M NaCl, 0.1 M TRIS pH 7.2), and sections were placed in dH₂O, and then mounted using ProLong Gold Antifade reagent with DAPI (Molecular Probes) and imaged as described above.

Microbial composition was performed by qPCR as described previously [33]. Briefly, at least two fecal pellets were collected and DNA was extracted using a Qiagen DNA stool extraction kit. Following DNA extraction, 50 ng DNA/reaction was used for qPCR. qPCR was completed using 16 s rRNA group specific primers [33] to determine the relative abundance of selected bacterial phyla. The primers used were specific to the phyla Bacteroides (5′-GAG AGG AAG GTC CCC CAC-3′, 5′-CGC TAC TTG GCT GGT TCA G-3′) [46], and Firmicutes (5′-GGA GYA TGT GGT TTA ATT CGA AGC A-3′, 5′-AGC TGA CGA CAA CCA TGC AC-3′) [46], and γ-Proteobacteria (5′-TCG TCA GCT CGT GTY GTG A-3′, 5′-CGT AAG GGC CAT GAT G-3′) [47]. The relative abundance of each taxonomic group was determined by calculating the average Ct value relative to the total bacterial 16 s rRNA, as determined using the universal Eubacteria primers (5′-ACT CCT ACG GGA GGC AGC AGT-3′, 5′-ATT ACC GCG GCT GCT GGC - 3′) [48] and normalized to each primer's experimentally determined efficiency. Statistical significance was determined using a one-way ANOVA with a threshold of p<0.05.

The bactericidal capacity of cecal crypt secretions was performed as described previously [49], [50]. Cecal and Colonic crypts were isolated from naïve WT and Sigirr −/− mice and mice infected with C. rodentium for 1 day. Ceca and colons were extracted from mice following euthanization and fecal contents were removed. Tissues were washed for 30 min at room temperature on an orbital shaker with HEPES buffered HBSS containing Pen/Strep and gentamycin. Tissues were then cut into small sections and placed in a petri dish containing 10 mL of BD cell recovery solution. Tissues were incubated for 2 hours at 4°C and crypts were dislodged from the tissues. Following centrifugation, crypts were resuspended in iPIPES buffer (10 mM PIPES; pH 7.4 and 137 mM NaCl), and counted. For experiments, 1000 crypts were incubated in 45 µL of iPIPES buffer at 37°C for 30 minutes. Following this period, 10 µL of a crypt-free isolate was added to 10⁴ E. coli (commensal isolates, a gift from Dr. Ben Willing) or Lactobacili casei (a gift from Dr. Alan Stintzi) or C. rodentium or ΔaroA S. Typhimurium and incubated for 37°C for 60 minutes. The antimicrobial capacity of crypt-culture extracts was assessed by counting the overnight growth of WT and Sigirr −/− crypt-treated E. coli or Lactobacili casei, and expressed as a percentage of the growth observed in untreated cultures.

The assay was performed as previously described [9]. Uninfected or infected mice at D6 pi were gavaged with 150 µl of 80 mg/ml 4 kDa FITC-dextran (Sigma; FD4) in PBS 4 hrs prior to sacrifice. Mice were anaesthetized and blood (∼500 µl) was collected by cardiac punctures, which was added immediately to a final concentration of 3% acid-citrate dextrose (20 mM citric acid, 100 nM sodium citrate, 5 mM dextrose). Plasma was collected and fluorescence was quantified using a VictorX3 (Perkin-Elmer Life Sciences) at excitation 485 nm, emission 530 nm for 1 sec.

Tissues were collected as described above. After homogenization and centrifugation, tissue supernatants were collected at specified time points and IL-1β levels were assessed using an ELISA kit (BD Bioscience) following manufacturer's instructions.

Following isolation of intestinal crypts, cells were lysed in RIPA buffer on ice. IEC proteins (50 µg) were resolved by 12% SDS-PAGE and transferred to PVDF membranes. Blots were blocked for 1 h with 5% milk in TBST. Membranes were incubated with primary antibody in TBST overnight at 4°C and probed with the respective secondary antibody for 1 hr at room temperature. Rabbit polyclonal IL-1β (Genetex) and mouse polyclonal actin (SantaCruz) antibodies were used in this study. Membranes were then incubated with ECL Prime Western Blotting kit (GE health) and imaged using a Biorad ChemiDoc XRS (BioRad)

All results are expressed as the mean value ± SEM. Unless specified otherwise, non-parametric Mann–Whiney t-tests, or Student t tests were performed (Prism version 4.00). A P value of 0.05 or less was considered significant.

The following are the GeneIDs (Database: Entrez Gene) for each gene analyzed in this manuscript, given as gene name (official symbol GeneID #): TNF-α (Tnf GeneID: 21926); IL-1 α (Il1a GeneID: 16175); IL-1β (Il1b GeneID: 16176); IFN-γ (Ifng GeneID: 15978); IL-17A (Il17a GeneID: 16171); MCP-1 (Ccl2 GeneID: 20296); iNOS (Nos2 GeneID: 18126); mCRAMP (Camp GeneID: 12796); Mip2 (Cxcl2 GeneID: 12796); Reg-3γ (Reg3g GeneID: 19695); SIGIRR (SIGIRR GeneID: 24058)