IFN-I and IL-22 mediate protective effects of intestinal viral infection

Abstract

Products derived from bacterial members of the gut microbiota evoke immune signalling pathways of the host that promote immunity and barrier function in the intestine. How immune reactions to enteric viruses support intestinal homeostasis is unknown. We recently demonstrated that infection by murine norovirus (MNV) reverses intestinal abnormalities following depletion of bacteria, indicating that an intestinal animal virus can provide cues to the host that are typically attributed to the microbiota. Here, we elucidate mechanisms by which MNV evokes protective responses from the host. We identify an important role for the viral protein NS1/2 in establishing local replication and a type I interferon (IFN-I) response in the colon. We further show that IFN-I acts on intestinal epithelial cells to increase the proportion of CCR2-dependent macrophages and interleukin (IL)-22-producing innate lymphoid cells, which in turn promote pSTAT3 signalling in intestinal epithelial cells and protection from intestinal injury. In addition, we demonstrate that MNV provides a striking IL-22-dependent protection against early-life lethal infection by Citrobacter rodentium. These findings demonstrate novel ways in which a viral member of the microbiota fortifies the intestinal barrier during chemical injury and infectious challenges.

Figure 1:. Sequence variation in NS1/2 contributes to protection from intestinal injury.

(a) Illustration of parental and chimeric MNV strains. (b) Survival of uninfected GF, MNV-infected GF mice (CR6 and CW3) and conventional (Conv) mice following administration of 3% DSS in the drinking water for 5 days. Survival of GF mice infected with VP1 (c) or NS1/2 (d) chimeric viruses following DSS. Mice were infected for 10 days prior to DSS administration. Survival experiments with the parental strains and chimeric viruses were conducted concurrently and therefore the uninfected GF survival curve shown in (b), (c) and (d) are identical. Representative H&E images of small intestinal villi (e) and quantification of villus width (f) at 10 dpi. Scale bars represent 50μM. All groups have7 mice except for CR6 and CR6.VP1^CW3 that have 8 mice. (g) Infectious MNV in the stool as determined by plaque assay at 10 dpi. The following mice were analyzed per group: CW3, n=18; CW3.VP1^CR6, n=15; CW3.NS1/2^CR6, n=17; CR6, n=18; CR6.VP1^CW3, n=17 and CR6.NS1/2^CW3, n=18. Survival curves were analyzed using the log-rank Mantel–Cox test. Villus width and log₁₀ transformed stool titers were analyzed using ANOVA with Dunnett’s multiple comparisons test compared to GF mice or CR6-infected GF mice, respectively. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.

Figure 2:. NS1/2 sequence variation contributes to intestinal persistence, IFN-I signaling and epithelial cell proliferation following intestinal injury.

(a) OasL2, Mx2 and Isg15 mRNA expression over uninfected GF mice relative to viral log₁₀ genomes in the colon. Data represents 4 independent experiments. The following mice were analyzed per group: CW3, n=8; CR6, n=11; CW3.NS1/2^CR6, n=11 and CR6.NS1/2^CW3, n=10. (b) Viral genomes in sorted colonic IECs and LPLs. LOD=limit of detection. Data represents 2 independent experiments each with 2 mice per group. (c) OasL2 mRNA expression over uninfected GF mice relative to viral log₁₀ genomes in IECs and LPLs. Data represents 2 independent experiments each with 2 mice per group. (d) Representative images of the colon stained for Ki67 and corresponding fold change of Ki67⁺ IECs relative to uninfected GF mice. Scale bars represent 100μM. The following mice were analyzed per group: GF, n=12; CW3, n=5; CR6, n=13; CW3.NS1/2^CR6, n=8; CR6.NS1/2^CW3, n=8 and Conv, n=14. All samples were analyzed at day 5 of DSS administration. Correlation was analyzed using Pearson r with p-values and r-values as shown. Fold change of Ki67⁺ IEC/crypt was analyzed using ANOVA with Dunnett’s multiple comparisons test with p-values shown. All bars represent mean and error bars represent standard error of the mean.

Figure 3:. Protection of ABX-treated mice from intestinal injury is dependent on IL-22.

Survival of WT (a), Ifngr^−/− (b) and Il22^−/− (c) mice following DSS. Mice were treated with ABX for at least 10 days prior to CR6 infection and given DSS at 10 dpi for 6 days. Small intestinal villus width (d) and virus levels in the stool, colon and SI (e) at 10 dpi with CR6. The following ABX, ABX+CR6 and Conv mice were analyzed for villus width: WT n=5,5,4; Ifngr^−/−, n=9,11,6 and Il22^−/−, n=10,11,8. The following mice were analyzed for stool, colon and SI titer: WT, n=36,4,17; Ifngr^−/−, n=12,6,6 and Il22^−/−, n=25,11,11. (f) Representative images of pSTAT1 and pSTAT3 expression in the colon of ABX-treated WT mice. Scale bars represent 100μM. pSTAT1 (g) and pSTAT3 (h) expression score in the colon at day 6 post DSS. The following mice were analyzed for pSTAT1 and pSTAT3: WT ABX, n=9,9; WT ABX+CR6, n=10,9; Ifnar1^−/− ABX, n=4,8; Ifnar1^−/− ABX+CR6, n=5,9; Il22^−/− ABX, n=6,5 and Il22^−/− ABX+CR6, n=6,6. Proportion (i) and absolute numbers (j) of IL-22 expressing ILC (CD19⁻CD11b⁻CD90.2⁺CD3⁻TCRβ⁻) and CD4⁺ T cells (CD19⁻CD11b⁻CD90.2⁺CD3⁺TCRβ⁺CD4⁺) in the colon of ABX-treated WT and Ifnar1^−/− mice at day 6 post DSS. The following mice were analyzed for cell proportions and numbers: WT ABX, n=15,6; WT ABX+CR6, n=15,6; Ifnar1^−/− ABX, n=12,9 and Ifnar1^−/− ABX+CR6, n=14,10. Survival curves were analyzed using the log-rank Mantel–Cox test. ANOVA with Dunnett’s multiple comparisons test was used to analyze villus width and Log₁₀ transformed MNV titer compared to ABX and WT mice, respectively. A two-tailed t-test was used to analyze STAT expression. ANOVA with Tukey’s multiple comparisons test was used to analyze flow cytometry data. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.

Figure 4:. Protection from intestinal injury and IL-22 expression is associated with CCR2-dependent cells.

(a) Proportion of IL-22 expressing ILC in ABX-treated WT and ccr2^−/− mice at day 6 post DSS. The following number of mice were analyzed: WT ABX, n=8; WT ABX+CR6, n=7; ccr2^−/− ABX, n=6 and ccr2^−/− ABX+CR6, n=6. (b) Infectious virus levels in the SI, colon and stool of ABX-treated ccr2^−/− mice compared to WT mice. The following number of mice were analyzed for stool titer: WT, n=17 and ccr2^−/−, n=22; SI and colon titer: WT, n=8 and ccr2^−/−, n=13. (c) Survival of ccr2^−/− mice following DSS administration. Mice were treated with ABX for at least 10 days prior to CR6 infection and given DSS at 10 dpi for 6 days. Survival curves were analyzed using the log-rank Mantel–Cox test. Proportion of MHCII⁺ populations in ABX-treated WT and ccr2^−/− (d) and ABX-treated WT, Ifnar1^−/− and Il22^−/− (e) mice at day 6 post DSS. The following number of mice were analyzed in (d): WT ABX, n=5; WT ABX+CR6, n=5; ccr2^−/− ABX, n=3 and ccr2^−/− ABX+CR6, n=3. The following number of mice were analyzed in (e): WT ABX, n=9; WT ABX+CR6, n=9; Ifnar1^−/− ABX, n=5; Ifnar1^−/− ABX+CR6, n=7; Il22^−/− ABX, n=6 and Il22^−/− ABX+CR6, n=8. Lin⁻ =CD45⁺CD19⁻TCRβ⁻Gr-1^low. Figures (a) and (d) were analyzed using ANOVA with Tukey’s multiple comparisons test and figure (b) and (e) were analyzed using two-tailed t-test. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.

Figure 5:. IFN-I signaling in IEC is required for protection from intestinal injury.

Survival of Ifnar1^f/f (a) and Ifnar1^ΔIEC (b) mice following DSS. Mice were treated with ABX for at least 10 days prior to CR6 infection and given DSS at 10 dpi for 6 days. Survival curves were analyzed using the log-rank Mantel–Cox test. Small intestinal villus width (c) and virus levels in the SI, colon and stool (d) at 10 dpi with CR6. The following number of mice were analyzed for villus width: Ifnar1^f/f ABX, n=9; Ifnar1^f/f ABX+CR6, n=9; Ifnar1^ΔIEC ABX, n=7 and Ifnar1^ΔIEC ABX+CR6, n=9. The following number of mice were analyzed for stool titer: Ifnar1^f/f, n=71 and Ifnar1^ΔIEC, n=49; SI and colon titer: Ifnar1^f/f, n=9 and Ifnar1^ΔIEC, n=10. (e) Representative images of pSTAT3 staining in the colon and pSTAT3 expression score at day 6 post DSS. Scale bars represent 100μM. The following number of mice were analyzed for pSTAT3: Ifnar1^f/f ABX, n=8; Ifnar1^f/f ABX+CR6, n=6; Ifnar1^ΔIEC ABX, n=6 and Ifnar1^ΔIEC ABX+CR6, n=6. Figures (c) and (e) were analyzed using ANOVA with Tukey’s multiple comparisons test and figure (d) was analyzed using two-tailed t-test. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.

Figure 6:. IECs react to IFN-I stimulation and promote an IL-22 response.

Representative images of murine SI (a), murine colon (b) or human colon (c) organoids either unstimulated or stimulated with 500U/ml of IFNβ for 48 h and expression of Mx2 mRNA relative to unstimulated organoids. Scale bars represent 50μM. The following number of independent experiments were performed: (a), n=7; (b), n=3 and (c), n=3. Log₁₀ transformed Mx2 expression was analyzed by two-tailed paired t-test. (d) Phosphorylation of STAT3 at 2 h in SI organoids. Representative of 4 independent experiments. (e) Representative images of Reg3β expression and representative mean fluorescence intensity (MFI) of individual organoids of 2 independent experiments at 24 h. Scale bars represent 50μM. The following number of organoids were analyzed: unstimulated, n=34; IFNβ, n=32 and IL-22, n=40. ANOVA with Dunnett’s multiple comparisons test was used to analyze Reg3β expression. Proportion of IL-22 expressing ILC (f) and CD11b^lowCD11c⁺CD103⁻ cells (g) in ABX-treated Ifnar1^f/f and Ifnar1^ΔIEC mice at day 6 post DSS. The following number of mice were analyzed for IL-22: Ifnar1^f/f ABX, n=16; Ifnar1^f/f ABX+CR6, n=19; Ifnar1^ΔIEC ABX, n=15 and Ifnar1^ΔIEC ABX+CR6, n=15. The following number of mice were analyzed for CD11b^lowCD11c⁺CD103⁻ cells: Ifnar1^f/f ABX, n=18; Ifnar1^f/f ABX+CR6, n=20; Ifnar1^ΔIEC ABX, n=14 and Ifnar1^ΔIEC ABX+CR6, n=14. ANOVA with Tukey’s multiple comparisons test was used to analyze cell populations. (h) Gene expression analysis of IECs from MNV CR6-infected Ifnar1^f/f (n=3) and Ifnar1^ΔIEC (n=3) mice compared to uninfected mice (Ifnar1^f/f, n=3 and Ifnar1^ΔIEC, n=3) at day 6 post DSS. Circles in Venn diagram represent number of transcripts that are enriched following MNV-infection in the mice with the indicated genotypes. Pathway analysis was performed on the non-overlapping gene set and p-values determined by Ingenuity Pathway Analysis. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.

Figure 7:. MNV protects young mice from enteric bacterial infection.

Survival and weights of B6 (a) and Il22^−/− (b) mice infected with C. rodentium at 21 days old. Survival curves were analyzed using the log-rank Mantel–Cox test. For weight curves, AUC was determined for each mouse and difference between uninfected and CR6-infected groups determined by two-tailed t-test. Mean and standard error of the mean is shown for weight analysis. Bacterial titers in the stool (c), spleen (d) and liver (e) at 6 dpi with C. rodentium. The following number of mice were analyzed in (c): WT Uninfected, n=56; WT CR6, n=59; Il22^−/− Uninfected, n=48 and Il22^−/− CR6, n=47. The following number of mice were analyzed in (d): WT Uninfected, n=25; WT CR6, n=29; Il22^−/− Uninfected, n=24 and Il22^−/− CR6, n=22. The following number of mice were analyzed in (e): WT Uninfected, n=19; WT CR6, n=21; Il22^−/− Uninfected, n=23 and Il22^−/− CR6, n=23. (f) Proportion of colonic CD4⁺ and CD8⁺ T cells. (g) Proportion of colonic CD4⁺ and CD8⁺ T cells expressing IFNγ. (h) Proportion of IFNγ⁺, IL-13⁺ and IL-22⁺ cells in the colonic ILC population. The following number of mice were analyzed for cell proportions: WT Uninfected, n=12; WT CR6, n=16; Il22^−/− Uninfected, n=11 and Il22^−/− CR6, n=10. Log₁₀ transformed bacterial titers and cell proportions were analyzed using two-tailed t-tests. All bars represent mean and error bars represent standard error of the mean. All p-values are shown in the figure.