Loss of Interleukin-10 (IL-10) Signaling Promotes IL-22-Dependent Host Defenses against Acute Clostridioides difficile Infection

Infection with the bacterial pathogen Clostridioides difficile causes severe damage to the intestinal epithelium that elicits a robust inflammatory response. Markers of intestinal inflammation accurately predict clinical disease severity.

ABSTRACT

Infection with the bacterial pathogen Clostridioides difficile causes severe damage to the intestinal epithelium that elicits a robust inflammatory response. Markers of intestinal inflammation accurately predict clinical disease, however, the extent to which host-derived proinflammatory mediators drive pathogenesis versus promote host protective mechanisms remains elusive. In this report, we employed Il10^−/− mice as a model of spontaneous colitis to examine the impact of constitutive intestinal immune activation, independent of infection, on C. difficile disease pathogenesis. Upon C. difficile challenge, Il10^−/− mice exhibited significantly decreased morbidity and mortality compared to littermate Il10 heterozygote (Il10^HET) control mice, despite a comparable C. difficile burden, innate immune response, and microbiota composition following infection. Similarly, antibody-mediated blockade of interleukin-10 (IL-10) signaling in wild-type C57BL/6 mice conveyed a survival advantage if initiated 3 weeks prior to infection. In contrast, no advantage was observed if blockade was initiated on the day of infection, suggesting that the constitutive activation of inflammatory defense pathways prior to infection mediated host protection. IL-22, a cytokine critical in mounting a protective response against C. difficile infection, was elevated in the intestine of uninfected, antibiotic-treated Il10^−/− mice, and genetic ablation of the IL-22 signaling pathway in Il10^−/− mice negated the survival advantage following C. difficile challenge. Collectively, these data demonstrate that constitutive loss of IL-10 signaling, via genetic ablation or antibody blockade, enhances IL-22-dependent host defense mechanisms to limit C. difficile pathogenesis.

INTRODUCTION

Clostridioides difficile is a leading cause of nosocomial infections in the United States. High recurrence rates, increases in community-acquired infections, and the emergence of antibiotic-resistant strains render C. difficile an urgent threat to our public health system (1–5). The manifestation of C. difficile infection is highly variable, ranging from asymptomatic colonization, diarrhea, and pseudomembranous colitis to severe cases of toxic megacolon and death (6). Disease severity is shaped by the host immune response, and patients on immunosuppressants or with autoimmune disorders are more susceptible to severe disease (7–9). Taken together, there is a need to study the host immune response to C. difficile infection to develop new therapies.

Upon intestinal colonization, C. difficile produces toxins that disrupt epithelial barrier integrity and result in the translocation of commensal bacteria into submucosal tissues. Impaired barrier integrity leads to downstream induction of a multifaceted, robust inflammatory response (10, 11). The innate immune response is essential for protection against C. difficile infection. Mice deficient in pathogen recognition receptor signaling pathways or innate immune cells exhibit increased bacterial translocation, damage to the epithelial barrier, and increased mortality following C. difficile infection (12–16). Conversely, proinflammatory mediators can simultaneously exacerbate tissue damage and promote C. difficile expansion to hinder recovery (16–18). In support of these animal studies, elevated fecal and serum proinflammatory cytokine levels are associated with increased disease severity in patients (19–21). Together, these findings highlight the complexity of the host response and demonstrate the need to fundamentally understand the timing and context of intestinal inflammation as a driver of C. difficile pathogenesis. To begin to address the contribution of the host proinflammatory immune response to promoting disease severity during C. difficile infection, elevated expression of intestinal inflammatory mediators was established in mice a priori C. difficile challenge, and the disease severity following subsequent infection was investigated.

Interleukin-10 (IL-10) is a broad immunoregulatory cytokine that negatively regulates commensal bacteria-driven immune activation at steady state (22–24). Intestinal expression of IL-10 is critical for maintaining intestinal homeostasis, as mice deficient in the Il10 gene develop microbiota-dependent spontaneous colitis characterized by chronic activation of inflammatory mediators that are also associated with C. difficile pathogenesis (25–27). Thus, Il10^−/− mice, a widely used model of intestinal immune dysregulation, offer the opportunity to decouple intestinal inflammation from infection to study the causative nature of inflammatory mediators in C. difficile pathogenesis (28).

In this report, we demonstrate that preexisting intestinal immune activation, e.g., expression of proinflammatory cytokines driven by loss of IL-10 signaling, reduces susceptibility to C. difficile infection. Host protective immunity was independent of changes in C. difficile burden, toxin production, and the microbiota. The protective capacity of IL-10-deficient immune activation was dependent on IL-22 production enhancing early host defenses against C. difficile infection.

RESULTS

IL-10 deficiency decreases susceptibility to acute C. difficile infection.

At steady state, IL-10 maintains intestinal homeostasis by negatively regulating commensal bacterium-driven expression of proinflammatory cytokines. In the context of C. difficile infection, many of these proinflammatory cytokines correlate with increased disease severity. However, it is unclear whether the inflammatory profile associated with infection emerges following C. difficile-mediated tissue damage or if it proactively drives pathology and worsens disease (20, 21). To address this question, intestinal inflammation was induced independently of C. difficile infection using the murine Il10^−/− spontaneous colitis model, and the impact of constitutive inflammation on C. difficile disease severity was examined. Cohoused Il10^−/− and littermate Il10 heterozygous mice (Il10^HET) were treated with a broad-spectrum antibiotic cocktail in their drinking water to induce susceptibility to C. difficile and mimic the microbiota dysbiosis observed in patients at high risk for contracting C. difficile. Antibiotic-treated Il10^HET mice exhibited peak disease severity within 48 hours of infection, as measured by a disease score based on weight loss, body temperature, diarrhea, and lethargy (Fig. 1A), and approximately 75% mortality rate (Fig. 1B). In contrast, Il10^−/− mice experienced reduced disease severity at 2 days postinfection (p.i.) (Fig. 1A) and were less likely to succumb to acute C. difficile infection than Il10^HET mice (Fig. 1B).

FIG 1.

Genetic ablation of Il10 results in reduced susceptibility to C. difficile infection. Antibiotic-treated Il10^−/− and Il10^HET mice were inoculated with approximately 400 spores of C. difficile (VPI 10463 strain) and monitored daily for disease. (A) Disease score and (B) survival following infection. Data shown are a combination of five independent experiments (Il10^−/−, n = 23; Il10^HET, n = 25). (C) C. difficile burden in fecal pellets at day 1 p.i. (D) C. difficile burden and (E) C. difficile toxin levels in the cecal content at day 2 p.i. **, P < 0.01. Statistical significance was calculated by a log-rank test.

Il10^−/− mice challenged with pathogenic Escherichia coli, Salmonella enterica serovar Typhimurium, Citrobacter rodentium, Toxoplasma gondii, or Candida albicans all display improved pathogen clearance via enhanced phagocytic mechanisms by innate immune cells (29–33). Thus, C. difficile burden was measured at days 1 and 2 p.i. No difference in C. difficile burden was observed in the cecal content of Il10^HET and Il10^−/− mice at days 1 (Fig. 1C) and 2 p.i. (Fig. 1D). Further, C. difficile toxin activity in the cecal content of Il10^HET and Il10^−/− mice was similar at day 2 p.i., as measured by an in vitro cell-rounding assay (Fig. 1E). Together, these data indicate that loss of IL-10 augments host immunity following C. difficile infection but does not alter the establishment of infection or production of toxins, the primary virulence factors of C. difficile.

Intestinal inflammation and the subsequent onset of spontaneous colitis in Il10^−/− mice vary between vivaria and are dependent on the microbiota (25, 26). For example, Helicobacter species are well-known colitogenic triggers in Il10^−/− mice (34, 35). Prior to cohousing, we confirmed the presence of Helicobacter spp. in the feces of our Il10^−/− mice colony but not vendor-purchased C57BL/6 mice (Fig. S1A). To test the rigor of the observed enhanced survival phenotype in C. difficile-infected Il10^−/− mice, wild-type C57BL/6 and Helicobacter-positive Il10^−/− mice were cohoused and infected with C. difficile at an independent animal facility. In agreement with our studies in Il10^HET mice, Il10^−/− mice exhibited improved survival (Fig. S1B) compared to cohoused C57BL/6 mice following C. difficile infection despite no difference in C. difficile burden (Fig. S1C) or toxin production (Fig. S1D) at days 2 and 4 p.i. These complementary experiments demonstrate the robustness of this phenotype.

Enhanced protection in Il10^−/− mice is not driven by a distinct microbiota composition.

The composition of the microbiota impacts C. difficile pathogenesis through multiple direct and indirect mechanisms (36); therefore, cohoused littermate mice were used to normalize for this variable. To test the null hypothesis that the microbiota composition between Il10^HET and Il10^−/− mice was indistinguishable, bacterial 16S rRNA marker gene profiling was conducted on cecal content from Il10^HET and Il10^−/− mice collected at day 2 after C. difficile or mock infection. Microbial community alpha diversity was not different between uninfected or infected Il10^HET and Il10^−/− mice (Fig. 2A). Comparison of 16S rRNA bacterial community profiles between Il10^HET and Il10^−/− mice by relative bacterial abundance revealed a bloom of amplicon sequence variants (ASVs) identified as C. difficile in both Il10^HET and Il10^−/− infected mice compared to uninfected mice; however, the relative abundance composition between infected groups was similar (Fig. 2B). Beta diversity comparisons between samples by unsupervised hierarchical clustering (Fig. 2C), unweighted UniFrac distances (Fig. 2D), or permutational multivariate analysis of variance (PERMANOVA) analysis (Table S1) did not support rejecting the null hypothesis that there was no microbial community level difference between Il10^HET and Il10^−/− mice on day 2 following mock infection or C. difficile infection. A linear regression model was used to identify individual ASVs that correlate with Il10^HET and Il10^−/− phenotypes. The linear model readily detected C. difficile as significantly enriched in infected mice compared to uninfected mice but failed to identify an ASV significantly different between the microbiota of infected Il10^HET and Il10^−/− mice (Fig. S2A and B).

FIG 2.

C. difficile-infected Il10^−/− mice and Il10^HET mice exhibit a similar microbiota composition. Antibiotic-treated uninfected and C. difficile-infected Il10^−/− mice and Il10^HET mice were sacrificed at day 2 p.i., and cecal content was processed for 16s rRNA bacterial gene profiling. (A) Microbial alpha diversity as determined by the Shannon diversity index. (B) Relative abundance of top 15 bacterial ASVs. Bar plot is displayed at the genus level, except for orange bars that represent an ASV aligning to C. difficile. (C) Dendrogram representation of intestinal microbial communities using unsupervised hierarchical clustering of unweighted UniFrac distances to identify similarities between samples. (D) Unweighted UniFrac principal coordinate analysis plot of 16S bacterial rRNA ASVs.

In a validation cohort, 16S rRNA marker gene profiling was conducted on fecal pellets collected from C57BL/6 and Il10^−/− mice prior to cohousing (day −64 p.i.), throughout cohousing, at the start of antibiotic treatment (day −6 p.i.), and on the day of infection (day 0 p.i.). Prior to cohousing, Il10^−/− mice exhibit a distinct microbiota (Fig. S3A, Table S2). Cohousing shifted the microbiota of C57BL/6 mice to resemble the microbiota of Il10^−/− mice, as determined by unweighted UniFrac distance analysis (Fig. S3A), relative bacterial genus abundance (Fig. S3B), unsupervised hierarchical clustering (Fig. S3C), and PERMANOVA analysis (Table S2). Antibiotic treatment between day −6 and 0 p.i. significantly reduced the alpha diversity (Fig. S3D) and shifted the microbiota of both C57BL/6 and Il10^−/− mice, but no difference between groups was observed (Table S2). To identify specific ASVs that were differentially abundant between cohoused C57BL/6 and Il10^−/− mice, a linear discriminant analysis effect size (LEfSe) comparison was conducted. Several ASVs were differentially abundant within the microbiota of C57BL/6 and Il10^−/− mice prior to cohousing (Fig. S3E). However, following cohousing and antibiotic treatment, none of these differentially abundant ASVs remained (Fig. S3E). Together, these microbial profiling data support the conclusion that the differential outcome observed in antibiotic-treated IL-10-sufficient and -deficient hosts following C. difficile infection cannot be explained by community-level differences in the microbiota.

Il10^−/− and Il10^HET mice exhibit comparable induction of innate immunity following C. difficile infection.

No differences in C. difficile colonization, toxin production, or microbiota composition were observed between infected Il10^HET and Il10^−/− mice to account for enhanced protection in Il10^−/− mice; therefore, potential immune-mediated mechanisms were assessed. Induction of IL-10 is an effective strategy employed by some enteric pathogens to dampen the host immune response to infection (29, 30, 37–39). C. difficile-derived flagellin, surface layer proteins, and toxin (TcdB) all can induce macrophages, monocytes, and dendritic cells to produce IL-10 in vitro (40–42). Indeed, C57BL/6 mice infected with C. difficile have elevated IL-10 protein in the cecal tissue at day 2 p.i. (Fig. 3A). The broad immunosuppressive functions of IL-10 include inhibiting granulocyte infiltration into mucosal tissue and limiting expression of type 1 and type 17 cytokines, components of the immune response that promote protective immunity following C. difficile infection (43–47).

FIG 3.

Il10^−/− and Il10^HET mice exhibit a comparable induction of the innate immune response following acute C. difficile infection. (A) IL-10 protein levels in the cecal tissue homogenates of antibiotic-treated uninfected and day 2 p.i. C57BL/6 mice. (B to H) Il10^−/− and Il10^HET mice were inoculated with approximately 400 spores of C. difficile (VPI 10463 strain) or mock infected and sacrificed 2 days later. (B) LCN-2 protein levels in the cecal supernatants. (C to E) Large intestine lamina propria cells were harvested and assessed by flow cytometry for (C) neutrophil (CD11b⁺, Ly6G⁺) (D) monocyte (CD11b⁺, Ly6C⁺, Ly6G⁻) (E) and eosinophil (SSC^Hi, CD11b⁺, Siglec-F⁺) recruitment. (F and G) Fold induction of (F) Ifng and Il22 and (G) IFN-γ and IL-22 effector molecules (Nos2 and Reg3g) in the colon at day 2 p.i. relative to uninfected Il10^HET mice and normalized to Hprt. (H) IFN-γ, IL-22 and (I) type 2-associated cytokine protein levels in the cecal tissue homogenate. Data shown are a combination of two independent experiments (uninfected Il10^−/−, n = 7; uninfected Il10^HET, n = 6; day 2 infected Il10^−/−, n = 8; uninfected Il10^HET, n = 7). Data shown are means ± SEM. *, P < 0.05; **, P < 0.01. Statistical significance was calculated by an unpaired t test.

First, protein levels of lipocalin-2 (LCN-2), an established marker of intestinal inflammation (48), were measured in the cecal content of antibiotic-treated uninfected and day 2 p.i. Il10^−/− and Il10^HET mice (48). LCN-2 levels increased to approximately the same concentration in both groups by day 2 p.i. (Fig. 3B). Next, to thoroughly assess the quality of the innate immune response to acute C. difficile infection, Il10^−/− and Il10^HET mice were sacrificed at day 2 p.i., and recruitment of innate immune cells and induction of proinflammatory cytokines were assessed. Both infected Il10^HET and Il10^−/− mice exhibited a robust induction of the innate immune response compared to antibiotic-treated, uninfected, control mice (Fig. 3). No statistically significant differences in the frequency (Fig. S4A to C) or total numbers of infiltrating neutrophils (Fig. 3C), monocytes (Fig. 3D), or eosinophils (Fig. 3E) were observed between Il10^−/− and Il10^HET mice at day 2 p.i. Il10^−/− and Il10^HET mice at day 2 p.i. exhibited comparably elevated gene expression of Ifng and Il22 (Fig. 3F) as well as downstream host defense genes Nos2 and Reg3g in the colon (Fig. 3G). Gamma interferon (IFN-γ) and IL-22 protein concentrations in cecal tissue homogenates were also comparable (Fig. 3H). Type 2 cytokines (IL-5, IL-13, and IL-33), associated with eosinophil activation and protection during C. difficile infection (15, 47, 49), were not significantly different in the cecum of Il10^−/− and Il10^HET mice at day 2 p.i. (Fig. 3I). Finally, no differences in proinflammatory cytokines (IL-1β, IL-6, IL-17a, IL-27, and granulocyte-macrophage colony-stimulating factor) reported to modulate C. difficile pathogenesis (16, 50–54) or chemokines (CXCL1, CXCL2, and CCL2) involved in neutrophil and monocyte recruitment were observed in the cecal tissue homogenates of Il10^−/− and Il10^HET mice at day 2 p.i. (Fig. S4D and E). Collectively, these data indicate the magnitude or quality of the innate immune response in Il10^−/− mice following C. difficile infection is not driving the attenuated disease phenotype.

Loss of IL-10 signaling prior to C. difficile infection drives immune activation in the intestine and augments protective immunity.

In contrast to the comparable immune responses observed in C. difficile-infected Il10^−/− and Il10^HET mice, antibiotic-treated, uninfected Il10^−/− mice at day 2 after mock infection had higher levels of LCN-2 in the cecal content (Fig. 3B) and increased expression of IL-22- and IFN-γ-dependent effector molecules (Fig. 3F) in the large intestine than antibiotic-treated, uninfected Il10^HET mice. Moreover, antibiotic-treated Il10^−/− mice at day 0 p.i. displayed elevated immune activation in the large intestine compared to Il10^HET mice, as determined by increased frequency (Fig. 4A) and total numbers (Fig. 4B) of infiltrating neutrophils in the large intestine as well as elevated expression of proinflammatory immune defense genes (Il22, Ifng, Reg3g, and Nos2) (Fig. 4C), in agreement with previous reports (55, 56). These results support the hypothesis that preexisting immune activation, not the magnitude of the immune response following infection, confers protective immunity in Il10^−/− mice.

FIG 4.

Loss of IL-10 signaling enhances intestinal immune activation prior to infection and decreases susceptibility to acute C. difficile infection. Antibiotic-treated uninfected and Il10^−/− mice and Il10^HET mice were sacrificed on the day of infection (prior to inoculation). (A) Frequency of neutrophils and monocytes in the large intestine lamina propria. FACS plots were gated on live, CD45⁺, non-T, non-B cells, Siglec-F^neg, CD11b⁺ cells. (B) Total number of neutrophils and monocytes in the large intestine lamina propria. Data are a combination representative of two independent experiments. Il10^−/−, n = 8; Il10^HET, n = 9. (C) Fold induction of type 1- and type 17-associated effector molecules in the colon of antibiotic-treated, uninfected Il10^−/−mice relative to antibiotic-treated, uninfected Il10^HET mice and normalized to Hprt. Data are a combination representative of three independent experiments. Il10^−/−, n = 12; Il10^HET, n = 13. Data shown are means ± SEM. (D) C57BL/6 mice were cohoused with Il10^−/− mice for 2 weeks and then were administered anti-IL10R1 or isotype control (rat IgG1) by i.p. injection weekly for 3 weeks prior to infection or received a single dose of anti-IL10R1 on the day of C. difficile infection and assessed for survival following infection. Data are a combination of two independent experiments (n = 8 per group). *, P < 0.05. Statistical significance was calculated by an unpaired t test or a log-rank test.

To determine whether the loss of IL-10 signaling prior to infection and subsequent immune activation augments protection following C. difficile infection, IL-10 signaling was selectively blocked in C57BL/6 mice starting either 3 weeks prior to infection or on the day of infection, and survival was assessed. Antibody-mediated blockade of the IL-10-specific receptor IL-10R1 (αIL10R1) administered once a week for at least 3 weeks abrogates IL-10 signaling and replicates the intestinal inflammation observed in germ line Il10^−/− mice (57). C57BL/6 mice that received weekly αIL10R1 treatment starting 3 weeks prior to infection exhibited survival comparable to that of Il10^−/− mice and significantly improved survival compared to C57BL/6 mice administered αIL10R1 at day 0 p.i. (Fig. 4D). These data support the hypothesis that IL-10 inhibits basal activation of intestinal immune defense genes prior to infection, thereby rendering the host more susceptible to C. difficile infection.

IL-22 is critical for host defense against C. difficile infection in Il10^−/− mice.

Inhibition of IL-10 signaling limits C. difficile pathogenesis only if initiated several weeks prior to infection (Fig. 4D). Further, Il10 deficiency leads to enhanced colonic il22 and ifng expression in uninfected mice (Fig. 4C). Both IL-22 and IFN-γ are critical in mounting a protective innate immune response during acute C. difficile infection (13, 14, 58). These observations suggest immune activation prior to infection promotes improved survival in Il10^−/− mice. To determine the relative contribution of these cytokine pathways to host protection in Il10^−/− mice, Il22 or Tbx21 (the gene that encodes T-bet, a master transcription factor that regulates IFN-γ production) was genetically ablated in Il10^−/− mice. Following C. difficile infection, Il10.Tbx21 double knockout (dKO) mice exhibited survival comparable to that of Il10^−/− mice, suggesting the IFN-γ pathway was dispensable for protection in Il10^−/− mice (Fig. 5A). In contrast, Il10.Il22 dKO mice were acutely susceptible to C. difficile infection (Fig. 5A). To confirm the dependence of IL-22 signaling for host protection in an IL-10-deficient setting, cohoused C57BL/6 or Il10^HET mice and Il10^−/−, Il22^−/−, Il10.Il22 dKO, and Il10r2^−/− mice (IL-10R2 is the shared receptor subunit necessary for both IL-10 and IL-22 signaling) were pretreated with antibiotics and infected with C. difficile. Genetic ablation of IL-22 signaling in an IL-10-deficient setting (Il10.Il22 dKO and Il10r2^−/− mice) led to significantly increased disease morbidity at day 2 p.i. (Fig. 5B) and mortality compared to Il10^−/− mice (Fig. 5C). Collectively, these data support the conclusion that loss of IL-10 signaling leads to activation of IL-22-dependent host defense mechanisms that limit C. difficile pathogenesis.

FIG 5.

IL-22 signaling is required for protection against C. difficile infection in Il10^−/− mice. (A) Cohoused Il10^−/−, Il10.Il22 dKO, and Il10.Tbx21 dKO mice were pretreated with antibiotics and inoculated with approximately 400 spores of C. difficile (VPI 10463 strain) and assessed for survival following infection. Survival curve is a combination of three independent experiments. Il10^−/−, n = 7; Il10.Il22 dKO, n = 12; Il10.Tbx21 dKO, n = 14. (B) Disease severity at day 2 p.i. (C) Survival curve of cohoused C57BL/6 or Il10^HET (wild-type, WT), Il10^−/−, Il10r2^−/−, Il10^−/−, Il22^−/−, and Il10.Il22 dKO mice following C. difficile infection. Data shown are a combination of four independent experiments (WT, n = 12; Il10^−/−, n = 14; Il10r2^−/−, n = 14 Il22^−/−, n = 16; Il10.Il22 dKO, n = 12). *, P < 0.05. Statistical significance was calculated by an unpaired t test or a log-rank test.

DISCUSSION

C. difficile infection induces a robust innate inflammatory response that has been extensively studied in the context of pathogenesis. Here, we employed Il10^−/− mice to decouple constitutive intestinal inflammation from C. difficile infection-induced inflammation and determine their respective roles in pathogen defense. Collectively, our data support the conclusion that the absence of IL-10 signaling elevates host defenses prior to infection, leading to reduced C. difficile pathogenesis in an IL-22-dependent manner. These results implicate a novel and deleterious role for IL-10 in dampening the IL-22 response during enteric C. difficile infection.

Previous work by Kim et al. assessed IL-10 in the context of C. difficile infection and observed more severe disease in C. difficile-infected Il10^−/− mice than in cohoused C57BL/6 mice at day 7 p.i. (59). Notably, intestinal inflammation of the Il10^−/− mice prior to infection was not increased compared to that of C57BL/6 mice in this study, suggesting immune defense mechanisms were not elevated in the Il10^−/− cohort at the time of infection. The microbiota composition, such as Helicobacter colonization status, of the Il10^−/− mice used by Kim and colleagues was not reported but could explain this lack of inflammation prior to infection. The Kim et al. study examined recovery from a mild form of C. difficile disease (10 to 15% weight loss; 100% survival in C57BL/6 mice) induced by infection with vegetative C. difficile cells (59). In contrast, the data in this report investigate the role of preexisting immune activation induced by loss of IL-10 during the acute response to a severe form of C. difficile disease (20 to 30% weight loss; 25 to 50% survival in C57BL/6 mice). Severe infection in this study was induced with C. difficile spores that mimic the natural form of exposure in the hospitalized patient population (60, 61). Together, these studies support a model where basal immune activation prior to infection can limit severe acute pathogenesis. However, prolonged Il10 deficiency during a milder form of disease tips the balance toward inflammation-driven tissue immunopathology.

Despite the protective capacity of intestinal inflammation reported in this study in the context of Il10 deficiency, expression of proinflammatory molecules does not uniformly limit C. difficile disease. Notably, Il10^−/− mice are also a widely used model of inflammatory bowel disease (IBD), a well-appreciated risk factor for C. difficile disease (8, 62). Due to its multifactorial nature, however, clinical reports that link IBD to C. difficile are not able to differentiate what features of IBD drive increased C. difficile disease severity (62). Research investigating the connection between IBD and C. difficile has employed mice treated with dextran sodium sulfate (DSS), a model of chemically induced colitis, to disentangle the role of proinflammatory immune components in pathogenesis. DSS-treated mice exhibit increased susceptibility to C. difficile infection (63), and Saleh et al. demonstrated that IL-17 competent CD4⁺ T_H17 cells activated by DSS colitis were sufficient to increase susceptibility to C. difficile infection in non-DSS-treated mice (16). This work identifies induction of IL-17 during IBD as an inflammatory pathway that promotes C. difficile disease.

In the context of the C. difficile infection, the IL-23/IL-22/IL-17 axis has a nuanced role in pathogenesis. Genetic ablation of IL-23, a cytokine upstream of IL-17, protects mice from severe C. difficile infection (17), while mice deficient in IL-22, which is also directly downstream of IL-23, are acutely susceptible to infection (13, 64). Further, wild-type mice that receive rIL-22 treatment prior to C. difficile challenge are protected from severe infection (13). Altogether, these observations, along with the results presented here, support a protective role for elevated IL-22 production. At the same time, induction of the IL-23 proinflammatory axis in the absence of IL-22, or in favor of IL-17 production, could drive more severe disease. Multiple IL-22-dependent mechanisms that mediate protection against C. difficile infection have been described. Hasegawa et al. demonstrated the induction of complement proteins via IL-22 signaling on hepatocytes is required to limit non-C. difficile bacterial translocation during severe C. difficile infection (13). In addition to this systemic role for IL-22, a more recent study demonstrated a role for IL-22 signaling in modulation of intestinal epithelial glycosylation to enable growth of bacterial consumers of succinate, a crucial metabolite for C. difficile growth (65). IL-22 also acts on intestinal epithelial cells to induce expression of genes that encode antimicrobial peptides, including RegIIIγ, lipocalin-2, and calprotectin (66–68), that limit damage of adherent or mucosa-associated commensal bacteria to the epithelium (66, 67), the latter of which has been associated with host protection against C. difficile (69). In support of this, Gunesekera et al. showed that Il10^−/− mice exhibited enriched expression of these same IL-22-dependent antimicrobial genes, all of which were uniquely lost in Il10.Il22 dKO mice (55). Thus, elevated IL-22 expression at the time of C. difficile infection, as observed in Il10^−/− mice, positions the host to limit toxin-mediated destruction of the epithelial barrier.

Immune activation in il10 -deficient hosts disrupts homeostasis at steady state but is also beneficial in the context of an acute C. difficile infection by elevating baseline defense mechanisms prior to infection. This concept of immunological tuning prior to infection has been previously observed with commensal bacteria providing tonic signaling to maintain antiviral defenses in a poised state of readiness to rapidly respond upon viral infection (70–72). Thus, the trade-off of constitutive intestinal il10 expression is diminished basal activation of immune defense genes and, therefore, a decreased capacity of the host to respond to pathogen challenge. Understanding the dynamics of this biological balancing act could help develop therapies that selectively or transiently target the protective components of immune activation in at-risk patients while avoiding deleterious side effects of prolonged inflammation.

MATERIALS AND METHODS

Mice.

Four- to 6-week-old wild-type C57BL/6J, Il10^−/−, Tbx21^−/−, and Il10rb^−/− mice were purchased from the Jackson Laboratory. Il22^−/− mice were provided by R. Flavell (Yale University). All knockout mouse strains were derived on a C57BL/6 background. All mice were bred and maintained in autoclaved cages under specific pathogen free conditions at the University of Pennsylvania. All experiments with cohoused C57BL/6 and Il10^−/− mice were done at Memorial Sloan Kettering Cancer Center. Il10Il22 dKO and Il10.Tbx21 dKO mice were generated by breeding Il10^−/− mice with Il22^−/− and Tbx21^−/− mice, respectively. The presence of Helicobacter spp., a bacterial genus sufficient to trigger early onset of intestinal inflammation in Il10^−/− mice, was confirmed by PCR in the feces of all breeder Il10^−/− mice and Il10^−/−-derived mouse strains (34, 35, 73). Our Helicobacter species-positive Il10^−/− mice began to display overt signs of colitis at around 3 to 4 months of age. Therefore, 10- to 14-week-old Il10^−/− mice were used in our studies. At this age, Il10^−/− mice exhibit increased expression of inflammatory immune genes in the intestine but do not yet display clinical manifestations of spontaneous colitis. Sex- and age-matched control mice were used in all experiments according to institutional guidelines for animal care. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania and Memorial Sloan Kettering Cancer Center.

Antibiotic pretreatment, C. difficile infection, and mouse monitoring.

Mice were cohoused for 3 weeks prior to antibiotic treatment and then supplemented with metronidazole (0.25 g/liter), neomycin (Sigma) (0.25 g/liter), and vancomycin (Novaplus) (0.25 g/liter) in drinking water for 3 days. One day following cessation of antibiotic water, mice received 200 mg of clindamycin (Sigma) by intraperitoneal (i.p.) injection. Twenty-four hours later, mice received approximately 400 C. difficile spores (VPI 10463 strain, ATCC 43255) via oral gavage. For antibody-mediated blockade experiments, mice received 1 mg of αIL10R1 antibody (clone 1B1.3A; Bio X Cell) or mouse IgG1 isotype control (clone MOPC-21; Bio X Cell) i.p. weekly starting either 3 weeks before infection or on the day of infection. After infection, mice were monitored and scored for disease severity by four parameters: weight loss (>95% of initial weight, 0; 95% to 90% initial weight, 1; 90% to 80% initial weight, 2; <80%, 3), surface body temperature (>32°C, 0; 32°C to 30.5°C, 1; 30.5°C to 29°C, 2; <29°C, 3), diarrhea severity (formed pellets, 0; loose pellets, 1; liquid discharge, 2; no pellets/caked to fur, 3), morbidity (score of 1 for each symptoms with a maximum score of 3; ruffled fur, hunched back, lethargy, and ocular discharge). Mice that exhibited severe disease, defined as a surface body temperature below 29.5°C or weight loss in excess of 30%, were humanely euthanized by CO₂ displacement.

C. difficile quantification.

Fecal pellets or cecal content were resuspended in deoxygenated phosphate-buffered saline (PBS), and 10-fold serial dilutions were plated anaerobically at 37°C on brain heart infusion agar supplemented with yeast extract, l-cysteine, d-cycloserine, cefoxitin, and taurocholic acid (CCBHIS-TA). CFU were enumerated 24 h later. Prior to infection, fecal samples from mice were cultured overnight in CCBHIS-TA liquid broth and then serially diluted and grown for 24 h on CCBHIS-TA plates to ensure that mice did not harbor endogenous C. difficile in their microbiota. Supernatants from the cecal or fecal content were obtained after centrifugation for cytotoxicity assays and LCN-2 enzyme-linked immunosorbent assay (ELISA) (Bethyl Labs).

C. difficile toxin cytotoxicity assay.

Vero cells were seeded in 96-well plates at 1 × 10⁴ cells/well and incubated for 24 h at 37°C in 5% CO₂. Cecal or fecal supernatants were added in 10-fold dilutions to the Vero cells (100 μl/well) and incubated overnight prior to removal, rinsing with PBS, and replacement with fresh media. The presence of C. difficile toxins A and B was confirmed by neutralization with antitoxin antisera (Techlab, Blacksburg, VA). The data are expressed as the log₁₀ reciprocal value of the last dilution where cell rounding was observed. Cell morphological changes were observed after 18 h using a Nikon inverted microscope. The cytopathic effect was determined as rounded cells compared to the negative-control wells.

16S rRNA sequencing.

Cecal content was collected from uninfected mice and mice infected with C. difficile at 2 days p.i. DNA was extracted using the Qiagen MagAttract power microbiome kit DNA/RNA kit (catalog no. 27500-4-EP; Qiagen) and used for rRNA sequencing and Helicobacter species PCR. Genus-specific PCR was conducted on purified bacterial DNA from feces of Il10^−/− breeder mice, and the V4-V5 region of the 16S rRNA gene was amplified from each sample using the dual indexing sequencing strategy as described previously (74). Sequencing was done on the Illumina MiSeq platform. The V4-V5 region of the 16S rRNA gene was sequenced and demultiplexed using the fqgrep tool. Data were imported into QIIME2 (v. 2020.2) (75) and denoised using the DADA2 plugin (76). For data in Fig. S3 in the supplemental material, the fqgrep tool (https://github.com/indraniel/fqgrep) was used to demultiplex the sequences, followed by denoising using the DADA2 (v. 1.14.1) (76) implementation in R (v. 3.6.3) (77). Due to quality issues on the reverse reads, only the forward reads were used for denoising. Data sets were taxonomically classified in QIIME2 using the q2-feature-classifier (78) classify-sklearn naive Bayes classifier with a newly generated classifier against Greengenes 13_8 99% operational taxonomic unit (OTU) sequences (79). Phylogenetic trees were generated using mafft (80) and the q2-phylogeny plugin (81). Data were then imported into R for further analyses with phyloseq (v. 1.30.0) (82) and visualization with ggplot2 (v. 3.3.0) (83). Unweighted UniFrac (84) dissimilarity was calculated to generate principal coordinate analysis plots and for creating dendrograms using the hclust function (stats package in R core, v. 3.6.3). Finally, a linear model was built using the lm() and padjust() functions (stats package, v. 3.6.3) as well as the Tidyverse package (v. 1.3.0) (85).

Isolation of lamina propria cells and flow cytometry.

Single-cell suspensions were obtained from the large intestine lamina propria compartment by longitudinally cutting the large intestine and washing out its contents in PBS. Intestinal tissues were incubated at 37°C under gentle agitation in stripping buffer (PBS, 5 mM EDTA, 1 mM dithiothreitol, 4% fetal calf serum, 10 μg/ml penicillin-streptomycin) for 10 min to remove epithelial cells and then for another 20 min to remove intraepithelial lymphocytes. The tissue was digested with collagenase IV at 1.5 mg/ml (500 U/ml) and DNase (20 μg/ml) in complete medium (DMEM supplemented with 10% fetal bovine serum, 10 μg/ml penicillin-streptomycin, 50 μg/ml gentamicin, 10 mM HEPES, 0.5 mM β-mercaptoethanol, 20 μg/ml l-glutamine) for 30 min at 37°C under gentle agitation. Supernatants containing the lamina propria fraction were passed through a 100-μm cell strainer and then a 40-μm cell strainer. After counting, cells were plated at 10⁶ cells per well in 96-well round-bottom plates and washed twice in PBS before incubating with a cell viability dye for 20 min at room temperature (Invitrogen AQUA dye). After Fc blockade (anti-mouse CD16/32 clone 2.4G2; BD Biosciences), cells were stained using a standard flow cytometric staining protocol with fluorescently conjugated antibodies specific to CD3ε, CD5, CD19, CD45, major histocompatibility complex class II, Siglec-F, Ly-6G, Ly6C, CD11b, and CD11c. Stained cells were kept in fluorescence-activated cell sorter (FACS) buffer at 4°C until run. Samples were acquired on an LSR-II flow cytometer (Becton, Dickinson). Data were analyzed using FlowJo version 9.9.6 software. Cell populations were calculated from total cells per colon as a percentage of live CD45⁺ cells.

Cytokine and chemokine quantification.

Cecal tissue was homogenized in tissue extraction buffer with protease inhibitors for 1 min by bead beating with steel beads. Homogenates were centrifuged at 10,000 rpm for 5 min, and supernatants were collected and stored at −80°C. Supernatants containing protein were analyzed by mouse multiplex Luminex assay (Invitrogen) at the Human Immunology Core at University of Pennsylvania. Concentrations are displayed as nanogram of analyte per gram of cecal tissue.

Tissue RNA isolation, cDNA preparation, and qRT-PCR.

RNA was isolated from proximal colon tissue using mechanical homogenization and TRIzol isolation (Invitrogen) according to the manufacturer’s instructions. cDNA was generated using QuantiTect reverse transcriptase (Qiagen). Quantitative reverse transcription-PCR (RT-PCR) was performed on cDNA using either TaqMan primers and probes or QuantiTect primers in combination with TaqMan PCR master mix (ABI) or SYBR green chemistry, and reactions were run on a RT-PCR system (QuantiStudio 6 Flex; Applied Biosystems). Gene expression is displayed as fold increase over antibiotic-pretreated, uninfected control mice and was normalized to the Hprt gene.

Statistical analysis.

Results represent means ± standard errors of the means (SEM). Statistical significance was determined by unpaired t test and log-rank test for survival curve. Statistical analyses were performed using Prism GraphPad software v6.0 (*, P < 0.05; **, P < 0.01; ***, P < 0.001).