Loss of Sirt1 function improves intestinal anti-bacterial defense and protects from colitis-induced colorectal cancer

Abstract

Dysfunction of Paneth and goblet cells in the intestine contributes to inflammatory bowel disease (IBD) and colitis-associated colorectal cancer (CAC). Here, we report a role for the NAD+-dependent histone deacetylase SIRT1 in the control of anti-bacterial defense. Mice with an intestinal specific Sirt1 deficiency (Sirt1int-/-) have more Paneth and goblet cells with a consequent rearrangement of the gut microbiota. From a mechanistic point of view, the effects on mouse intestinal cell maturation are mediated by SIRT1-dependent changes in the acetylation status of SPDEF, a master regulator of Paneth and goblet cells. Our results suggest that targeting SIRT1 may be of interest in the management of IBD and CAC.

Figure 1. SIRT1 deletion increases Paneth and goblet cell number and up-regulates anti-bacterial gene expression in mice and worms.

A, Representative images of Paneth (Lysozyme⁺ cells), goblet (Periodic acid-Shiff⁺ cells), and enteroendocrine (Chromogranin A⁺ cells) cell staining. (N = 3 mice; 5–10 field per mouse, 50–100 crypt/villi per field). Bar = 50 µm. B–C, mRNA and protein levels of Paneth (Lys, Mmp7, Crypt1, Crypt4, Defa-Rs) and goblet (Klf4, Muc2) markers are increased in the proximal and distal small intestine of Sirt1^int−/− mice (N = 6–8). For RTqPCR analysis, rps12 was used as reference gene. β-Actin was used as loading control in the western blot. D, Increased expression of genes involved in pathogen defense in the sir-2.1 (ok434) C. elegans mutant (N = 6; each N means a single population of ∼500 worms). Act1 and Y45 were used as reference. E, sir-2.1 siRNA induces lys-1 and lys-7 driven GFP expression in lys-1::GFP and lys-7::GFP reporter worms. In the same figure a representative image of lys-7::GFP before and after sir-2.1 siRNA is shown (DIC, differential interference contrast). Results are expressed as mean±SEM. *P<0.05; **P<0.01; ***P<0.001.

Figure 2. Intestinal deletion of Sirt1 rearranges the gut microbiome.

A, Representative image showing a significant increase in cecum size in Sirt1^int−/− mice. B, Bactericidal capacity of small intestine (duodenum) in Sirt1^int−/− and Sirt1^L2/L2 mice was assayed by colony forming unit assay. Results are expressed as % of the control (no proteins) (N = 7). C, Distribution of microbial Phyla in colon tissue and cecum contents of Sirt1^L2/L2 and Sirt1^int−/− mice. D, Network-based analyses of cecal bacterial communities in Sirt1^L2/L2 and Sirt1^int−/− mice. Each large circle represents an animal and each smaller dot represents an OTU; Sirt1^L2/L2 (Blue) and Sirt1^int−/− (Red). E, Principal Coordinate Analysis (PCA) (PERMANOVA p = 0.012, and ANOSIM p = 0.007) showing significant separation of the microbial communities between Sirt1^L2/L2 and Sirt1^int−/− mice. F, Heatmap showing the top 20 different OTUs between Sirt1^L2/L2 and Sirt1^int−/− mice. The abundance values were log transformed and standardized and plotted within matlab. *indicates Barnesiella, Bacteroides, and Prevotella; °indicates Clostridum genus (see also Table S3 in File S1). Results are expressed as mean±SEM. *P<0.05; **P<0.01; ***P<0.001.

Figure 3. Intestinal Sirt1 deletion impacts on the development of colorectal cancer.

A–B, Schematic representation of the AOM/DSS protocol (top left panel) and representative image of colons from Sirt1^int−/− and control mice after CAC induction (Bar = 200 µm) (bottom left panel). Sirt1^int−/− mice show significantly less tumors (right panel). B, Representative picture of a colon section from a Sirt1^int−/− and control mouse after CAC induction (left panels). Tumor size (right panel). C, Colon length at the time of sacrifice (AOM/DSS). D, Percentage of body weight change. For the AOM/DSS experiment 8 mice for each genotype were used. *P<0.05; **P<0.01; ***P<0.001. E–F, Principal Coordinate Analysis (PCA) of bacterial sequences from colon tissue performed using unweighted UniFrac distance matrix. E, Sirt1^int−/− colon tissue before and after AOM treatment (PERMANOVA p = 0.003, ANOSIM p = 0.016). F, Sirt1^L2/L2 colon tissue with and without AOM treatment (PERMANOVA p = 0.067, ANOSIM p = 0.038). G–H, Most statistically significant OTUs before and after AOM in both Sirt1^L2/L2 (G), and Sirt1^int−/− (H), colon tissues; *Indicates Helicobacter and Desulfovibrio (see also Table S4 in File S1). I, PCA of bacterial sequences from colon tissue after AOM (PERMANOVA p = 0.767, ANOSIM p = 0.167).

Figure 4. Hyper-acetylation by SIRT1 stabilizes SPDEF and triggers Paneth and goblet cells maturation.

A, SPDEF target genes (Slug, uPA, Ccl6) are induced in Sirt1^int−/− intestines. GIF1 (Nrg3, Pax6, ChgA) and SOX9 (Igfbp4) target genes, as well as Spdef are not changed. B, In vitro acetylation/deacetylation assays demonstrates that p300 acetylates SPDEF and SIRT1, but not SIRT6 and SIRT7, deacetylates SPDEF. C, Nano-LC-MS/MS shows SIRT1-dependent in vitro deacetylation of AcK294 (left panel). Sequence alignment showing the evolutionary conserved K294 residue (right panel). D, SIRT1, but not SIRT1G261A, deacetylates SPDEF in HEK293 immunoprecipitates. The SPDEFK294Q mutant is not acetylated. Tubulin was used as loading control. E–F SPDEF, SPDEFK294Q, and SIRT1 were transfected in the HEK293 cell line and visualized by immunoblotting before (0 min) and after cycloheximide (CHX) treatment. HSP90 is used as loading control (left panels in E and F). The relative stability of SPDEF or SPDEFK294Q was calculated by ImageJ (right panels in E and F). G, PC3 cells were co-transfected with an E-Cadherin promoter luciferase reporter construct, wild type or SPDEFK294Q in presence or absence of SIRT1. SIRT1 represses SPDEF-dependent reporter activation. The acetylated-mimic SPDEFK294Q mutant is constitutively activate and not affected by SIRT1. H, Immunoblotting of crypt enriched fractions from Sirt1^int−/− and Sirt1^L2/L2 mice show increased SPDEF protein levels in Sirt1^int−/− intestines. β-Actin was used as loading control. Results are expressed as mean±SEM. *P<0.05; **P<0.01; ***P<0.001.