Impact of global PTP1B deficiency on the gut barrier permeability during NASH in mice

Abstract

Objective: Non-alcoholic steatohepatitis (NASH) is characterized by a robust pro-inflammatory component at both hepatic and systemic levels together with a disease-specific gut microbiome signature. Protein tyrosine phosphatase 1 B (PTP1B) plays distinct roles in non-immune and immune cells, in the latter inhibiting pro-inflammatory signaling cascades. In this study, we have explored the role of PTP1B in the composition of gut microbiota and gut barrier dynamics in methionine and choline-deficient (MCD) diet-induced NASH in mice.

Methods: Gut features and barrier permeability were characterized in wild-type (PTP1B WT) and PTP1B-deficient knockout (PTP1B KO) mice fed a chow or methionine/choline-deficient (MCD) diet for 4 weeks. The impact of inflammation was studied in intestinal epithelial and enteroendocrine cells. The secretion of GLP-1 was evaluated in primary colonic cultures and plasma of mice.

Results: We found that a shift in the gut microbiota shape, disruption of gut barrier function, higher levels of serum bile acids, and decreased circulating glucagon-like peptide (GLP)-1 are features during NASH. Surprisingly, despite the pro-inflammatory phenotype of global PTP1B-deficient mice, they were partly protected against the alterations in gut microbiota composition during NASH and presented better gut barrier integrity and less permeability under this pathological condition. These effects concurred with higher colonic mucosal inflammation, decreased serum bile acids, and protection against the decrease in circulating GLP-1 levels during NASH compared with their WT counterparts together with increased expression of GLP-2-sensitive genes in the gut. At the molecular level, stimulation of enteroendocrine STC-1 cells with a pro-inflammatory conditioned medium (CM) from lipopolysaccharide (LPS)-stimulated macrophages triggered pro-inflammatory signaling cascades that were further exacerbated by a PTP1B inhibitor. Likewise, the pro-inflammatory CM induced GLP-1 secretion in primary colonic cultures, an effect augmented by PTP1B inhibition.

Conclusion: Altogether our results have unraveled a potential role of PTP1B in the gut-liver axis during NASH, likely mediated by increased sensitivity to GLPs, with potential therapeutic value.

Keywords: GLP-1; GLP-2; Gut microbiota; Inflammation; NASH; PTP1B.

Graphical abstract

Figure 1

Gut microbiota composition and functional changes. A) Alpha-diversity analysis for PTP1B WT and PTP1B KO mice fed a chow or MCD diet for 4 weeks. Shannon index is expressed as mean ± SEM. ^###p < 0.001 PTP1B KO chow vs PTP1B KO MCD. B) Phylogenetic β-diversity of fecal microbiota. Graph represents principal coordinate analysis (PCoA) plot based on 16 S rRNA gene sequencing of stool samples from PTP1B WT (blue symbols) and PTP1B KO (red symbols) mice fed a chow or MCD diet for 4 weeks. Differences between genotypes in this case are indicated with different clouds. Pink cloud corresponds to animals fed chow diet and green cloud surrounds animals fed a MCD diet. C) Ratio Firmicutes/Bacteroidetes (F/B) is expressed as mean ± SEM. ^∗p < 0.05 PTP1B WT chow vs PTP1B WT MCD. D-E) Predicted functional changes in gut microbiota analyzed by PICRUSt tool. Bar graphs show significant differences in KEGG pathways. MCD_1 M refers to animals fed a MCD diet for 4 weeks; Chow_ refers to animals fed a chow diet. Significant differences are shown as p-value at the right panel. n = 5–7 mice per experimental group.

Figure 2

Bile acid levels in serum samples from PTP1B WT and PTP1B KO mice fed chow or MCD diet. A-G) Mean ± SEM of bile acids concentrations in serum samples are expressed as part per billion (ppb). n = 10–15 mice per experimental group. H) Nr1h4 mRNA levels in the gut (encoding FXR protein). White dots represent mice fed a chow diet and black dots represent mice fed a MCD diet. n = 10–15 mice per experimental group. ∗p< 0.05; ^∗∗p < 0.01; PTP1B WT chow vs PTP1B WT MCD; ^#p < 0.05; ^##p < 0.01 PTP1B KO chow vs PTP1B KO MCD; ⁺p < 0.05; ⁺⁺p < 0.01 PTP1B WT MCD vs PTP1B KO MCD.

Figure 3

Alterations in the gut barrier permeability in PTP1B WT and PTP1B KO mice during NASH. A) LPS levels in serum after 4 weeks of feeding mice a chow or MCD diet. n = 6–9 mice per group. B) FITC-dextran assay in PTP1B WT and PTP1B KO mice fed a chow or MCD diet for 4 weeks. Data are expressed as percentage of FITC-dextran levels absorbed in serum samples. n = 9–16 animals per experimental group. ^∗∗∗p < 0.001 PTP1B WT chow vs PTP1B WT MCD; ^###p < 0.001 PTP1B KO chow vs PTP1B KO MCD; ⁺⁺p < 0.01 PTP1B WT MCD diet vs PTP1B KO MCD diet. C) Representative images of ZO-1 immunostaining in colon samples of PTP1B WT and PTP KO mice. Scale bar 50 μm. D) Effect of the pro-inflammatory environment on TEER values in Caco-2 shScrb and Caco-2 shPtpn1 monolayers. TEER values from Caco-2 shScrb or shPtpn1 cultured in Transwell plates treated with mCM-C or mCM-LPS are represented as percentage of baseline. Data are expressed as mean ± SEM of three independent experiments performed in triplicate. ^∗∗∗p < 0.001 Caco-2 shScb mCM-LPS vs mCM-C; ^###p < 0.001 Caco-2 shPtpn1 mCM-LPS vs mCM-C.

Figure 4

Analysis of anatomical and histological parameters in colon samples from PTP1B WT and PTP1B KO mice fed a chow or MCD diet for 4 weeks. A) Colon moisture content is expressed as percentage of water regarding total colon weight. n = 7–8 mice per experimental group. B) Colonic epithelial granuloma areas expressed as percentage of total epithelium area. n = 4 mice per experimental group. C) Representative images of H&E colonic stained sections from PTP1B WT and PTP1B KO mice. Images were taken from 5 μm colon paraffin-embedded sections stained with H&E with an Axio Scan. Z1 (Zeiss). Granuloma areas are marked with a white symbol. D) Mucin layer analysis in colonic samples from mice under NASH conditions. Images show AB (left) (n = 8–16 crypts per experimental group) and PAS (right) stained representative crypts from colonic sections that stain acidic and neutral mucins, respectively (n = 17–24 crypts per experimental group). E) Muc2 (mucin) mRNA levels are shown as fold change referred to the control PTP1B WT fed a chow diet. Data are expressed as mean ± SEM. n = 5–8 mice per experimental group. ^∗p < 0.05; ^∗∗p < 0.01 PTP1B WT chow vs PTP1B WT MCD; ^#p < 0.05; ^###p < 0.001 PTP1B KO chow vs PTP1B KO MCD.

Figure 5

mRNA expression levels of pro-inflammatory markers in colon samples of PTP1B WT and PTP1B KO mice during NASH. A-E) Cd3g (T-cell surface glycoprotein CD3 γ), GzmB (granzyme B), Ifng (interferon gamma), Nos2 (inducible nitric oxide synthase), and Ptp1n (PTP1B) mRNA levels are shown as fold change referred to the control PTP1B WT fed a chow diet. White dots correspond to mice fed a chow diet and black dots to mice fed a MCD diet during 4 weeks. Data are expressed as mean ± SEM. n = 5–10 mice per experimental group. ^∗∗∗p < 0.001 PTP1B WT chow vs PTP1B WT MCD; ^##p < 0.01 PTP1B KO chow vs PTP1B KO MCD; ⁺p < 0.05; ⁺⁺ p< 0.01; ⁺⁺⁺p < 0.001 PTP1B WT MCD vs PTP1B KO MCD.

Figure 6

Effect of PTP1B deficiency on GLP-1 release by the gut under pro-inflammatory stimuli. A) GLP-1 serum levels from C57BL/6 J mice injected with LPS (2 mg/kg body weight) as described in Materials and Methods. Black dots represent control animals and gray dots correspond to animals under an LPS acute pro-inflammatory stimulus. n = 4–7 mice per experimental group. ∗p<0.05 PTP1B WT vs PTP1B WT LPS. B) Upper panel represents GLP-1 release from colonic organoids stimulated with pCM-LPS compared with organoids treated with pCM-C and DMEM-2%FBS. n = 2 independent experiments performed in triplicate. ^∗∗ p< 0.01 DMEM vs CM-LPS. Lower panel shows colonic organoids images used in this experiment. Upper panel shows a colonic organoid with budding crypt-like domains. Bottom panel shows an organoid with mature EEC L cells. Both images correspond to the same organoid, left image under light microscopy, right image under fluorescence microscopy. L cells (yellow fluorescence YFP) are shown in the right image. C) GLP-1 levels measured in the culture media from colonic crypts cultured under pro-inflammatory conditions. Data are expressed as fold change referred to control medium (DMEM alone). n = 3 independent experiments performed in quadruplicate. ^∗∗p < 0.01; ^∗∗∗p < 0.001 data referred to basal conditions; ^###p < 0.001 referred to pCM-LPS. D) GLP-1 serum levels measured in animals under chow or MCD diet during 4 weeks. n = 5–7 mice per experimental group. ^∗∗∗p < 0.001 PTP1B WT chow vs PTP1B WT MCD; ⁺⁺p < 0.01 PTP1B WT chow vs PTP1B KO chow. E) Gcg mRNA levels (encoding proglucagon) are shown as fold difference referred to the control PTP1B WT fed a chow diet. White dots correspond to mice fed a chow diet and black dots to mice fed a MCD diet during 4 weeks. All data are expressed as mean ± SEM. n = 5–8 animals per experimental group. ^∗∗∗p < 0.001 PTP1B WT chow vs PTP1B WT MCD ; ^###p < 0.001 PTP1B KO chow vs PTP1B KO MCD.

Figure 7

Inflammatory and insulin signaling in enteroendocrine STC-1 cells stimulated with pro-inflammatory conditioned media. A) STC-1 cells were pre-incubated or not for 2 h with the PTP1B inhibitor (20 μM) before the treatment with mCM-LPS for 5, 15, or 30 min. The dishes pretreated with the PTP1B inhibitor were used throughout the experiment. At the end of the culture time, cells were lysed and protein extracts were analyzed by western blot with the corresponding antibodies. Vinculin was used as a loading control. Right panel shows representative western blots. In the left panel, graphs show quantification of three independent experiments performed in duplicate. ^∗∗p < 0.01 mCM-LPS plus PTP1B inhibitor vs mCM-LPS. B) PTP1B inhibition protected against the drop of insulin signaling in EECs stimulated with pro-inflammatory conditioned medium. STC-1 cells were pre-incubated with the PTP1B inhibitor (20 μM) before the treatment with mCM-LPS or mCM-C for 16 h. Then, insulin was added for an additional 15 min. Protein extracts were prepared, and the phosphorylation of the IR and AKT was analyzed by western blot with the antibodies against phospho-IR, phospho-AKT Ser473, phospho-AKT Thr308, total IR, and total AKT. An anti-p85-PI3K antibody was used as a loading control. Right panel shows representative western blots. In the left panel, graphs show quantification of three independent experiments. The different experimental conditions are indicated under the blots or graphs. Data are shown as mean ± SEM. ⁺ p < 0.05 mCM-LPS plus PTP1B inhibitor vs mCM-LPS.

Figure 8

mRNA expression levels of GLP-2 responsive genes in colon samples of PTP1B WT and PTP1B KO mice under NASH conditions. A-C) Lyz (lysozyme), Kfg (keratinocyte growth factor), and Vip (vasointestinal peptide) mRNA levels are shown as fold change referred to the control PTP1B WT fed a chow diet. White dots correspond to mice fed a chow diet, and black dots correspond to mice fed a MCD diet. Data are expressed as mean ± SEM. n = 4–8 animals per experimental group. ⁺p< 0.05; ⁺⁺p < 0.01 PTP1B WT chow diet vs PTP1B KO chow diet.

Supplementary Figure 1

Gut characterization in PTP1B WT and PTP1B KO mice under NASH conditions. A) Graph represents the ratio colon weight/length for body weight. White and black bars represent animals fed a chow or MCD diet, respectively. n= 5-7 mice per experimental group. B) Upper panel shows quantification of number of feces in the colon from the four experimental animal groups. Lower panel shows representative images of colon samples. n= 4-8 mice per experimental group. C) Feces humidity was measured with the moisture analyzer MB120 (Ohaus). Percentage of humidity was calculated from the difference between final dried feces and starting fresh feces. Data are expressed as mean ± SEM. n=6-11 mice per experimental group. ∗p<0.05; ∗∗p<0.01 PTP1B WT chow vs PTP1B WT MCD diet; ^##p<0.01; ^###p<0. 001 PTP1B KO chow vs PTP1B KO MCD.

Supplementary Figure 2

Effect of PTP1B inhibition in insulin signaling in EECs under proinflammatory stimulus. A) STC-1 cells were pretreated or not for 2 h with the PTP1B inhibitor (20 μM) and then stimulated with insulin for 15 min. Protein extracts were prepared, and the phosphorylation of the IR and AKT was analyzed by western blot with the antibodies against phosho-IR, phospho-AKT Ser473, phospho-AKT Thr308, total IR, and total AKT. An anti-p85-PI3K antibody was used as a loading control. Left panel shows representative western blots. In the right panel, graphs show quantification of three independent experiments. Data are shown as mean ± SEM. n=3 independent experiments. ∗p<0.05; ∗∗∗p<0.001 DMEM alone vs insulin; ⁺⁺p<0.01; ⁺⁺⁺p<0.001 insulin plus PTP1B inhibitor vs insulin. B) STC-1 cells were pre-incubated with the PTP1B inhibitor (20 μM) before the treatment with mCM-LPS or mCM-C for 16 h. Then, insulin was added for a further 15 min. Protein extracts were prepared and the phosphorylation of the IR and AKT was analyzed by western blot with the antibodies against phosho-IR, phospho-AKT Ser473, phospho-AKT Thr308, total IR, and total AKT. An anti-p85-PI3K antibody was used as a loading control. Left panel shows representative western blots. In the right panel, graphs show quantification of three independent experiments. Data are shown as mean ± SEM. ⁺⁺⁺p<0.001 mCM-LPS vs mCM-C.

Supplementary Table 1

TEER values of Caco-2 shScrb and Caco-2 shPtp1n cells under mCM-C or mCM-LPS treatment. Two stable cell lines, Caco-2 scramble (Caco-2 shScb) and Caco-2 siPtpn1 (Caco-2 silenced for Ptpn1 gene, encoding for PTP1B protein), were cultured with mCM-C or mCM-LPS. TEER values were measured after 24 and 48 h. Data are expressed as mean values ± SEM. n=3 independent experiments performed in triplicate.