Chronic enteritis triggered by diet westernization is driven by epithelial ATG16L1-mediated autophagy

Abstract

Macroautophagy/autophagy exerts multilayered protective functions in intestinal epithelial cells (IECs) while a loss-of-function genetic variant in ATG16L1 (autophagy related 16 like 1) is associated with risk for developing Crohn disease (CD). Westernization of diet, partly characterized by excess of long-chain fatty acids, contributes to CD, and a metabolic control of intestinal inflammation is emerging. Here, we report an unexpected inflammatory function for ATG16L1-mediated autophagy in Crohn-like metabolic enteritis of mice induced by polyunsaturated fatty acid (PUFA) excess in a western diet. Dietary PUFAs induce ATG16L1-mediated conventional autophagy in IECs, which is required for PUFA-induced chemokine production and metabolic enteritis. By transcriptomic and lipidomic profiling of IECs, we demonstrate that ATG16L1 is required for PUFA-induced inflammatory stress signaling specifically mediated by TLR2 (toll-like receptor 2) and the production of arachidonic acid metabolites. Our study identifies ATG16L1-mediated autophagy in IECs as an inflammatory hub driving metabolic enteritis, which challenges the perception of protective autophagy in the context of diet westernization.Abbreviations: AA: arachidonic acid; ATG16L1: autophagy related 16 like 1; CD: Crohn disease; CXCL1: C-X-C motif chemokine ligand 1; ER: endoplasmic reticulum; GFP: green fluorescent protein; GPX4: glutathione peroxidase 4; IBD: inflammatory bowel disease; IECs: intestinal epithelial cells; PTGS2/COX2: prostaglandin-endoperoxide synthase 2; PUFA: polyunsaturated fatty acid; SDA: stearidonic acid; TLR2: toll-like receptor 2; WT: wild-type.

Keywords: ATG16L1; Crohn disease; glutathione peroxidase 4; intestinal epithelial cells; intestinal inflammation; polyunsaturated fatty acids.

Figure 1.

Dietary PUFAs induce autophagy in Gpx4 deficient intestinal epithelium. (A) Experimental model of the in vivo experiment. GFP-LC3;WT and GFP-LC3;Gpx4^+/-IEC mice were fed a PUFA-enriched western diet for three months. (B, C) Representative confocal images (B) and quantification (C) of GFP-LC3 puncta (green) of small intestinal crypts of WT and Gpx4^+/-IEC mice after three months exposure to a PUFA-enriched western diet (n = 2). DAPI (blue) indicates nuclei. Scale bar: 10 µm (top) or 5 µm (bottom). (D-F) A representative immunoblot (D) and quantification of LC3B-II:I (E) and SQSTM1 (F) of intestinal epithelial scrapings of WT and Gpx4^+/-IEC mice after three months exposure to a PUFA-enriched western diet (n ≥4). GAPDH served as the loading control. Median is shown, Mann-Whitney test for quantification in (E, F). (G) A representative immunoblot of GPX4, SOD1, GPX2, GPX1 and PRDX6 of small intestinal epithelial scrapings of WT and Gpx4^+/-IEC mice after three months exposure to a PUFA-enriched western diet (n = 5). GAPDH served as the loading control. *p < 0.05.

Figure 2.

Dietary PUFAs induce autophagy in IECs restricted by GPX4. (A, B) A representative immunoblot (A) and quantification (B) of LC3B-II:I and GPX4 of siCtrl and siGpx4 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n ≥4). GAPDH served as the loading control. (C, D) A representative confocal images (C) and quantification (D) of GFP-LC3B puncta (green) of siCtrl and siGpx4 IECs after ω-6 PUFA (AA) stimulation for 24 h. (n ≥3). DAPI (blue) indicates nuclei. Scale bar: 5 µm. Median is shown, Kruskal Wallis test with Dunn’s correction for quantification in (D). (E) A representative immunoblot of LC3B-II:I, ATG16L1 and GPX4 of siCtrl, siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n = 3). GAPDH served as the loading control. *p < 0.05, **p < 0.01, ****p < 0.0001.

Figure 3.

Epithelial-specific deletion of Atg16l1 in mice protected against metabolic enteritis induced by dietary PUFAs. (A, B) Enteritis histology score (A) and representative H&E images (B) of WT, Gpx4^+/-IEC, atg16l1^-/-IEC and Gpx4^+/-IEC;atg16l1^-/-IEC mice after three months exposure to a PUFA-enriched western diet (n = 14/14/14/13). Median is shown, each dot represents an experimental animal, Kruskal-Wallis test with Dunn’s correction (A). Scale bar: 100 µm in (B). (C-E) Representative images of MPO⁺ neutrophils (C), CD4⁺ (D) and CD8⁺ T cells (E) of intestinal sections from WT, Gpx4^+/-IEC, atg16l1^-/-IEC and Gpx4^+/-IEC;atg16l1^-/-IEC mice after three months exposure to a PUFA-enriched western diet. Scale bar: 100 µm. Black arrows indicate MPO positive cells (C), CD4 positive cells (D) and CD8 positive cells (E) in patchy inflammatory spots of Gpx4^+/-IEC mice. (F, G) Quantification of macrophages (F) and neutrophils (G) per cm small intestine of WT, Gpx4^+/-IEC, atg16l1^-/-IEC and Gpx4^+/-IEC;atg16l1^-/-IEC mice after three months exposure to a PUFA-enriched western diet (n ≥8). (H, I) Representative immunohistochemistry (H) and quantification (I) of LYZ⁺ (lysozyme⁺) cells in the small intestine of WT, Gpx4^+/-IEC, atg16l1^-/-IEC and Gpx4^+/-IEC;atg16l1^-/-IEC mice after three months exposure to a PUFA-enriched western diet (n = 5). Scale bar: 100 μm (H). Median is shown, each dot represents an experimental animal. Kruskal-Wallis test with Dunn’s correction (I). *p < 0.05, ****p < 0.0001.

Figure 4.

Rapamycin enables PUFA-induced enteritis in WT mice. (A) Experimental model of the in vivo experiment. WT and Gpx4^+/-IEC were fed a PUFA-enriched western diet for three months and treated with vehicle or rapamycin (Rapa) for the last two weeks of the experiment. (B, C) enteritis histology score (B) and representative H&E images (C) of WT and Gpx4^+/-IEC fed a PUFA WD for three months and treated with vehicle or rapamycin (Rapa) for the last two weeks of the experiment (n = 7/9/8/9). Median shown, each dot represents an experimental animal. Kruskal-Wallis test with Dunn’s correction (B). Scale bar: 100 µm (C). *p < 0.05.

Figure 5.

ATG16L1-mediated autophagy is required for PUFA-induced CXCL1 production in IECs. (A) Quantification of CXCL1 in the supernatant of siCtrl and siGpx4 of WT and atg16l1^-/- IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n = 4). (B – D) Quantification of CXCL1 in the supernatant (B), protein lysate (C) and relative Cxcl1 expression (D) of siCtrl, siGpx4 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n = 4/6). (E – G) Quantification of CXCL1 in the supernatant of siCtrl, siGpx4 and siGpx4 siAtg5 (E), siGpx4 siAtg7 (F) and siGpx4 siAtg12 (G) IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h (n > 3). Note that vehicle and siCtrl and siGpx4 controls are identical in (E – G) as experiments were performed at the same time. (H, I): Quantification of CXCL1 in the supernatant of siGpx4 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h with or without bafilomycin A₁ (H) or chloroquine (I) (n = 3). (J) Quantification of lipid peroxidation by flow cytometry of BODIPY 581/591 C11⁺ labelled siCtrl, siGpx4 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h (n = 3). (K, L) quantification of CXCL1 in the supernatant of siCtrl, siGpx4 and siGpx4 siNcoa4 (K) and siCtrl, siGpx4 and siGpx4 siRab7a (L) after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h (n ≥4). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 6.

Atg16l1 is required for PUFA-induced TLR2 activity in IECs. (A) Relative AP-1 activation in siCtrl or siATG16L1 mTLR2 reporter cell line after stimulation with vehicle, oxidized phosphatidylcholine (oxPAPC) or the lipid peroxidation by-product 4-hydroxynonenal (4-HNE) for 24 h (n = 8). (B) Enrichment plots showing significant downregulation of MAPK signaling pathway in AA and SDA stimulated IECs. siGpx4 siAtg16l1 IECs were compared to siGpx4 IECs (n = 4). Red and blue color bar: this bar corresponds to the ranked list of genes. * red: indicates genes in the MAPK signaling pathway that are upregulated in siGpx4 siAtg16l1 +AA/+SDA compared to siGpx4 +AA/+SDA. Blue: indicates genes in the MAPK signaling pathway that are downregulated in siGpx4 siAtg16l1 +AA/+SDA compared to siGpx4 +AA/+SDA. The intensity of the color reflects the degree of upregulation or downregulation. (C) A representative immunoblot of (phospho-)MAPK/JNK and (phospho-) JUN/c-Jun in siCtrl, siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n ≥4). GAPDH served as the loading control. (D) Enrichment plots showing significant downregulation of the protein processing at the endoplasmic reticulum pathway in AA and SDA stimulated IECs. siGpx4 siAtg16l1 IECs were compared to siGpx4 IECs (n = 4). Red and blue color bar: this bar corresponds to the ranked list of genes. * red: indicates genes in the protein processing at the endoplasmic reticulum pathway that are upregulated in siGpx4 siAtg16l1 +AA/+SDA compared to siGpx4 +AA/+SDA. Blue: indicates genes in the protein processing at the endoplasmic reticulum pathway that are downregulated in siGpx4 siAtg16l1 +AA/+SDA compared to siGpx4 +AA/+SDA. The intensity of the color reflects the degree of upregulation or downregulation. (E) a representative immunoblot of (phospho-)ERN1/IRE1α and HSPA5/GRP78 in siCtrl, siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 24 h. (n ≥4). GAPDH served as the loading control. (F, G) A representative immunoblot (F) and quantification by densitometry (G) of epithelial HSPA5/GRP78 relative to GAPDH of WT, Gpx4^±IEC, atg16l1^-/-IEC and Gpx4^±IEC;atg16l1^-/-IEC mice after three months exposure to a PUFA-enriched western diet (n = 5). Each dot represents one experimental animal. GAPDH served as the loading control in F. *p < 0.05, ****p < 0.0001.

Figure 7.

Pam2CSK4 induces inflammation in Gpx4^+/-IEC; atg16l1^-/-IECs mice. (A) Quantification of CXCL1 in the supernatant of siCtrl, siGpx4 and siGpx4 siAtg16l1 IECs with or without stimulation with the TLR2 agonist Pam2CSK4 for 24 h (n = 4). (B, C) Enteritis histology score (B) and representative H&E images (C) of WT and Gpx4^+/-IEC; atg16l1^-/-IECs fed a PUFA-enriched western diet for three months and treated with vehicle or Pam2CSK4 (Pam) for the last week of the experiment (n = 6/4/5/6). Scale bar: 100 µm in C. ***p < 0.001, ****p < 0.0001.

Figure 8.

Atg16l1 is required for PTGS2/COX2-mediated AA metabolism in IECs. (A, B) Heatmap showing significant altered expression of AA metabolism genes in siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after stimulation with ω-6 PUFA (AA, A) or ω-3 PUFA (SDA, B) for 8 h compared to siCtrl. AA metabolism genes are indicated on the right. (C) Ptgs2 Fold change in siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after exposure with ω-6 PUFA (AA) or ω-3 PUFA (SDA) for 8 h compared to siCtrl. (D) A representative immunoblot of PTGS2/COX2 in siCtrl, siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) exposure for 24 h. (n = 3). GAPDH served as the loading control. (E) Relative expression of Ptgs2 relative to Actb/β-actin in siGpx4, siAtg16l1 and siGpx4 siAtg16l1 IECs after ω-6 PUFA (AA) or ω-3 PUFA (SDA) stimulation for 8 h determined by qPCR (n ≥4). (F) Levels of lipid mediators in vehicle, AA or SDA exposed atg16l1^-/- IECs relative to WT IECs. The heatmap shows the percentage changes of atg16l1^-/- IECs and wt IECs compared to wt siCtrl IECs (n = 4). (g, H) Quantification of CXCL1 in the supernatant of siGpx4 IECs after ω-6 PUFA (AA, G) or ω-3 PUFA (SDA, H) stimulation for 24 h with the PTGS2/COX2 inhibitor celecoxib (0.1 µM, 1 µM, 10 µM, CEB) (n = 4). *p < 0.05, **p < 0.01.

Figure 9.

Model depicting PUFA-induced actions in Gpx4-deficient intestinal epithelium. PUFAs induce ATG16L1-mediated conventional autophagy which enables metabolic enteritis originating from Gpx4-deficient IECs. Specifically, ATG16L1 enables TLR2 activation and arachidonic acid metabolism into prostanoids. Image was created with BioRender.com.