NLRP1 restricts butyrate producing commensals to exacerbate inflammatory bowel disease

Abstract

Anti-microbial signaling pathways are normally triggered by innate immune receptors when detecting pathogenic microbes to provide protective immunity. Here we show that the inflammasome sensor Nlrp1 aggravates DSS-induced experimental mouse colitis by limiting beneficial, butyrate-producing Clostridiales in the gut. The colitis-protective effects of Nlrp1 deficiency are thus reversed by vancomycin treatment, but recapitulated with butyrate supplementation in wild-type mice. Moreover, an activating mutation in Nlrp1a increases IL-18 and IFNγ production, and decreases colonic butyrate to exacerbate colitis. We also show that, in patients with ulcerative colitis, increased NLRP1 in inflamed regions of the colon is associated with increased IFN-γ. In this context, NLRP1, IL-18 or IFN-γ expression negatively correlates with the abundance of Clostridiales in human rectal mucosal biopsies. Our data identify the NLRP1 inflammasome to be a key negative regulator of protective, butyrate-producing commensals, which therefore promotes inflammatory bowel disease.

Fig. 1

Loss of NLRP1 confers protection against DSS-induced colitis. WT and Nlrp1^−/− mice were given 3% (w/v) DSS for 6 days followed by normal drinking water for 2 days and disease severity was measured according to a percentage weight loss, b colon length and c histology score for epithelial damage and inflammatory cell infiltrate of hematoxylin and eosin (H&E)-stained sections of the colon, scale bar = 1 mm. d Representative H&E-stained sections of the colon from WT and Nlrp1^−/− mice. Lethally irradiated WT and Nlrp1^−/− mice were reconstituted with WT or Nlrp1^−/− bone marrow for 12 weeks and were given 3% DSS for 6 days. Disease severity was measured by e percentage weight loss, f colon length and g histology score. Data are representative of 3 independent experiments with 3–5 mice per group. Means ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. As determined by a two-tailed, unpaired t-test. A one-way analysis of variance (ANOVA) and Tukey’s post-hoc comparisons were performed on data that involved more than two comparisons

Fig. 2

Microbiome dysbiosis in Nlrp1^−/− mice is transferred to co-housed WT mice. a WT and Nlrp1^−/− littermates were separated at weening for 6 weeks, then stool was collected and subjected to bacterial 16S analysis. Differentially abundant OTUs are presented. b WT mice were either single-housed or co-housed (CH) with Nlrp1^−/− mice for 4 weeks prior to DSS treatment and weight loss, c colon length and d histology score were measured for signs of disease. e Prior to DSS treatment fecal samples were collected for bacterial 16S analysis. Comparison of 16S sequencing showing differentially abundant OTUs in 3 pair-wise comparisons made to WT mice. Data are representative of 2 independent experiments with 3–6 mice per group. Means ± SD (a, e), means ± SEM (b–d); *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. As determined by a two-tailed, unpaired t-test

Fig. 3

Vancomycin treatment or supplementation with butyrate ablates the Nlrp1 phenotype. a Stool from WT and Nlrp1^−/− mice was harvested before and after vancomycin (50 mg/L) treatment for 4 weeks. Bacterial DNA from stool was isolated from untreated (UT) and vancomycin-treated mice, and depletion of the Coccoides group (belonging to the Clostridiales phylum) was confirmed using specific 16S primers by quantitative PCR. Results were normalized to the total bacteria present in the stool. b Vancomycin-treated mice were subjected to 3% (w/v) DSS for 6 days and disease severity was measured according to percentage weight loss and c colon length. Short chain fatty acid analysis was performed on fecal samples collected from WT and Nlrp1^−/− mice before (UT) and after vancomycin treatment. The concentration of d butyrate and e propionate was measured by gas chromatography mass spectrometry. f WT and Nlrp1^−/− mice were supplemented with 2% butyrate in drinking water ad libitum for 28 days, and then subjected to 2.5% (w/v) DSS for 6 days and disease severity measured according to percentage weight loss or g colon length. Data are representative of 3 independent experiments with 3–6 mice per group. Means ± SEM; *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001. A two-tailed unpaired t-test was used to determine statistical significance between two groups; and a two-way ANOVA with Tukey’s post-hoc comparisons was performed on data that involved more than two comparisons

Fig. 4

Hyper-activation of Nlrp1 exacerbates DSS-induced colitis independent of IL-1R signaling. Il-1r^−/−Nlrp1a^Q593P/+ and Il-1r^−/−Nlrp1a^Q593P/Q593P mice were given 3% (w/v) DSS for 6 days. a The percentage weight loss, b colon length and c histology score for epithelial damage and inflammatory cell infiltrate of H&E-stained sections of the colon were measured for disease severity. d Representative H&E-stained sections of the distal colon from Il-1r^−/−Nlrp1a^Q593P/+ and Il-1r^−/−Nlrp1a^Q593P/Q593P mice, scale bar = 2 mm. Short chain fatty acid analysis was performed on fecal samples collected from WT and Nlrp1^−/− mice before (UT) and after vancomycin treatment. The concentrations of e butyrate and f propionate were measured by gas chromatography mass spectrometry. Data are representative of 2 independent experiments with 3–5 mice per group. Means ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. As determined by a two-tailed, unpaired t-test

Fig. 5

Deletion of IL-18 removes the hyper-activated Nlrp1 DSS-induced colitis phenotype. Il-1r^−/−Il-18^−/−Nlrp1a^Q593P/+ and Il-1r^−/−Il-18^−/−Nlrp1a^Q593P/Q593P mice were given 3% (w/v) DSS for 6 days. a The percentage weight loss, b colon length and c histology score for epithelial damage and inflammatory cell infiltrate of H&E-stained sections from the colon were measured for disease severity. d Representative H&E-stained sections of the distal colon from Il-1r^−/−Il-18^−/−Nlrp1a^Q593P/+ and Il-1r^−/−Il-18^−/−Nlrp1a^Q593P/Q593P mice, scale bar = 1 mm. Data are representative of 2 independent experiments with 3–5 mice per group. Means ± SEM

Fig. 6

NLRP1 gene expression is augmented in human ulcerative colitis which positively correlates with IFNγ gene expression in inflamed regions of the colon. Microarray analyses were performed on colon biopsies from healthy controls (n = 22), and patients with ulcerative colitis or Crohn’s disease. Gene expression data expressed as relative fold change compared to healthy controls from a transverse (n = 15), cecum (n = 10), sigmoid (n = 32) and rectum (n = 29) regions of the colon from patients with ulcerative colitis and b ileum (n = 25) from patients with Crohn’s disease. Correlation between NLRP1 and IFNγ gene expression in various regions of the colon of c inflamed regions from ulcerative colitis patients. Means ± 95% asymmetric confidence interval; **p ≤ 0.01. A one-way ANOVA and Tukey’s post-hoc comparisons were performed on data that involved more than two comparisons, while a two-tailed, unpaired t-test was performed on data containing two comparisons. Pearson's correlation coefficient was used to generate the r value and the p values denoting the significance were estimated from a paired t-test