Colonic inflammation triggers β cell proliferation during obesity development via a liver-to-pancreas interorgan mechanism

Abstract

Under insulin-resistant conditions, such as obesity, pancreatic β cells adaptively proliferate and secrete more insulin to prevent blood glucose elevation. We previously reported hepatic ERK activation during obesity development to stimulate a neuronal relay system, consisting of afferent splanchnic nerves from the liver and efferent vagal nerves to the pancreas, thereby triggering adaptive β cell proliferation. However, the mechanism linking obesity with the interorgan system originating in hepatic ERK activation remains unclear. Herein, we clarified that colonic inflammation promotes β cell proliferation through this interorgan system from the liver to the pancreas. First, dextran sodium sulfate (DSS) treatment induced colonic inflammation and hepatic ERK activation as well as β cell proliferation, all of which were suppressed by blockades of the neuronal relay system by several approaches. In addition, treatment with anti-lymphocyte Peyer's patch adhesion molecule-1 (anti-LPAM1) antibody suppressed β cell proliferation induced by DSS treatment. Importantly, high-fat diet (HFD) feeding also elicited colonic inflammation, and its inhibition by anti-LPAM1 antibody administration suppressed hepatic ERK activation and β cell proliferation induced by HFD. Thus, colonic inflammation triggers adaptive β cell proliferation via the interorgan mechanism originating in hepatic ERK activation. The present study revealed a potentially novel role of the gastrointestinal tract in the maintenance of β cell regulation.

Keywords: Beta cells; Endocrinology; Metabolism; Obesity.

Figure 1. Metabolic phenotypes of mice treated with DSS.

(A) Established schematic model of the neuronal relay system from the liver to the pancreas. (B) Scheme of the experimental plan for DSS administration. Mice were treated with 0% (vehicle) or 1.5% DSS (n = 5 per group). (C) Body weights at day 7 of experimental groups treated with 0% or 1.5% DSS (n = 5 per group). (D) Fasting blood glucose levels of mice at 7 days after vehicle or DSS treatment are shown (n = 5 per group). (E) Fasting insulin levels of mice at 7 days after vehicle or DSS treatment are shown (n = 5 per group). (F) The results of insulin tolerance tests after vehicle or DSS treatment are shown (n = 5 per group). Data are presented as means ± SEM. *P < 0.05 as assessed by the unpaired 2-tailed t test.

Figure 2. Colonic inflammation and colonic intestinal barrier disruption are induced in mice treated with DSS.

(A) Colonic length is shown for each experimental group (n = 5 per group). (B) Disease activity indices are shown for each experimental group (n = 5 per group). (C) Photomicrograph image of hematoxylin and eosin staining on day 7 post-DSS. Epithelial and crypt alignment of the distal colon were apparently impaired. Scale bars indicate 100 μm. (D) Photomicrograph image of the occludin immunofluorescence in the distal colon, and the relative fluorescence intensity in each experimental group was calculated (n = 5 per group). Scale bars indicate 100 μm. (E) Serum FITC-Dextran levels (40 kDa) after administration to mice in each of the experimental groups (n = 5 per group). (F) Serum LPS concentrations in the portal veins of mice in each of the experimental groups (n = 5 per group). Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by the unpaired 2-tailed t test.

Figure 3. Hepatic ERK pathway activation and β cell proliferation are enhanced in mice treated with DSS.

(A) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and phosphorylated ERK (pERK) of vehicle- or DSS-treated mice on day 7 (n = 4 per group). Band intensities were calculated as pERK intensity per ERK intensity. (B) Representative images and counted BrdU-positive β cell percentage in each group are shown (n = 10 per group). Scale bars indicate 50 μm. (C) Representative images and counted Ki67-positive β cell percentage at 5 days after vehicle or DSS administration (n = 5 per group). Scale bars indicate 50 μm. (D) Expression levels of FoxM1 gene in islets isolated from mice 5 days after vehicle or DSS administration (n = 10 per group). (E) Representative images and measured β cell masses after 7 days of vehicle or DSS treatment (n = 5 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05 as assessed by the unpaired 2-tailed t test.

Figure 4. Suppression of the hepatic ERK pathway blunts β cell proliferation in mice treated with DSS.

(A) Scheme of the experimental plan for DSS administration and adenovirus administration using LacZ or dominant-negative MEK (d/nMEK). (B) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and pERK on day 7 of each group (n = 10/12/10/12 per group). Band intensities were calculated as pERK intensity per ERK intensity. (C) Fasting blood glucose levels of mice in each group (n = 5/5/6/7 per group). (D) Representative images and counted BrdU-positive β cell percentage in each group (n = 5/5/4/5 per group). Scale bars indicate 50 μm. (E) Representative images and measured β cell masses are shown for each group (n = 15/18/20/23 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s honestly significant differences (HSD) post hoc test.

Figure 5. Blockade of the neuronal relay by pharmacologically blocking the afferent splanchnic nerve fibers blunts β cell proliferation in mice treated with DSS.

(A) Scheme of the experimental plan for DSS administration and capsaicin treatment. (B) Fasting blood glucose levels of mice in each group (n = 6/6/8/9 per group). (C) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and pERK on day 7 for each group: Vehicle-Sham, Vehicle-Cap, DSS-Sham, and DSS-Cap (n = 6/6/8/9 per group). Band intensities were calculated as pERK intensity per ERK intensity. (D) Represedntative images and counted BrdU-positive β cell percentage in each group (n = 6/6/7/9 per group). Scale bars indicate 50 μm. (E) Representative images and measured β cell masses for each group (n = 6/6/7/9 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 6. Blockade of the neuronal relay by subdiaphragmatic vagotomy blunts β cell proliferation in mice treated with DSS.

(A) Scheme of the experimental plan for DSS administration and vagotomy (Vx). (B) Fasting blood glucose levels of mice in each group (n = 4/5/6/6 per group). (C) Representative images and calculated relative intensities of liver extract immunoblottings with anti-ERK and pERK for each group: Vehicle-Sham, Vehicle-Vx, DSS-Sham, and DSS-Vx (n = 23/19/23/21 per group). Band intensities were calculated as pERK intensity per ERK intensity. (D) Representative images and counted BrdU-positive β cell percentage in each group (n = 4/4/5/5 per group). Scale bars indicate 50 μm. (E) Representative images and measured β cell masses for each group (n = 23/20/23/21 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 7. Administration of anti-LPAM1 antibody suppresses colonic inflammation in mice treated with DSS.

(A) Scheme of the experimental plan for DSS administration and antibody administration subcutaneously (s.c.). (B) Body weights at day 7 for each of the experimental groups treated with Vehicle-IgG, DSS-IgG, or DSS-LPAM1 (n = 7/7/4 per group). (C) Fasting blood glucose levels of mice in each group (n = 7/7/4 per group). (D) Colonic lengths of each of the experimental groups (n = 8/8/8 per group). (E) Disease activity indices for each experimental group (n = 7/7/4 per group). (F) Photomicrograph image of the occludin immunofluorescence in the distal colon and relative fluorescence intensity in each experimental group (n = 9/9/4 per group). Scale bars indicate 100 μm. (G) Serum FITC-Dextran levels (40 kDa) after administration to mice in each of the experimental groups (n = 7/7/4 per group). (H) Serum LPS concentrations in the portal veins of mice in each of the experimental groups (n = 8/6/5 per group). Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 8. Blockade of colonic inflammation suppresses hepatic ERK activation and β cell proliferation in mice treated with DSS.

(A) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and pERK on day 7 for each group. Band intensities were calculated as pERK intensity per ERK intensity (n = 8/8/4 per group). (B) Representative images and counted BrdU-positive β cell percentage in each group (n = 5/5/4 per group). Scale bars indicate 50 μm. (C) Representative images and measured β cell masses for each group (n = 8/8/4 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 9. Colonic inflammation was induced in mice with HFD-induced obesity.

(A) Scheme of the experimental plan for NC or HFD feeding. (B) Body weight at 4 weeks after feeding of NC or HFD in 8-week-old mice (n = 7/9 per group in NC and HFD mice). (C) Colonic lengths in each of the experimental groups (n = 7/9 per group). (D) Photomicrograph image of the occludin immunofluorescence in the distal colon and the relative fluorescence intensity in each experimental group (n = 5/5 per group). Scale bars indicate 100 μm. (E) Serum FITC-Dextran levels (4 kDa) after administration to mice in each experimental group (n = 7/9 per group). (F) Serum LPS concentrations in the portal veins of mice in each of the experimental groups (n = 3/7 per group). Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by the unpaired 2-tailed t test.

Figure 10. Hepatic ERK activation and β cell proliferation are induced in mice with HFD-induced obesity.

(A) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and pERK in each group (n = 5/6 per group). Band intensities were calculated as pERK intensity per ERK intensity. (B) Expression levels of FoxM1 and Mki67 genes in islets isolated from mice in each group (n = 4/4 per group). (C) Representative images and measured β cell masses for each group (n = 7/9 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05 as assessed by the unpaired 2-tailed t test.

Figure 11. Administration of anti-LPAM1 antibody suppresses colonic inflammation in mice with HFD-induced obesity.

(A) Scheme of the experimental plan for NC or HFD feeding and antibody administration during the last 2 weeks of NC or HFD feeding. (B) Body weight at 4 weeks after antibody-treatment in 8-week-old mice fed NC or HFD (n = 11/9/9 per group). (C) Fasting blood glucose levels of mice in each group (n = 6/5/5 per group). (D) Fasting insulin levels of mice are shown (n = 6/5/5 per group). (E) Colonic lengths of each experimental group (n = 11/9/9 per group). (F) Photomicrograph image of the immunofluorescence stain in the distal colon and the relative fluorescence intensities in each experimental group (n = 5/5/5 per group). Scale bars indicate 100 μm. (G) Serum FITC-Dextran levels (4 kDa) after administration to mice in each experimental group (n = 6/5/5 per group). (H) Serum LPS concentration in the portal vein of mice in each of the experimental groups (n = 4/3/3 per group). Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 12. Blockade of colonic inflammation suppresses hepatic ERK activation and β cell proliferation in mice with HFD-induced obesity.

(A) Representative images and calculated relative intensities of liver extract immunoblotting with anti-ERK and pERK in each group (n = 11/9/9 per group). Intensities of bands were calculated as pERK intensity per ERK intensity. (B) Expression levels of FoxM1 and Mki67 genes in islets isolated from mice in each group (n = 5/4/4 per group). (C) Representative images and counted Ki67-positive β cell percentage in each of the groups (n = 5/5/5 per group). Scale bars indicate 50 μm. (D) Representative images and measured β cell masses for each group (n = 6/5/5 per group). Scale bars indicate 200 μm. Data are presented as means ± SEM. *P < 0.05, **P < 0.01 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 13. LPS and IL-23 enhances hepatic ERK phosphorylation.

(A) Representative pixel intensity in cytokine array analyses from the portal vein blood (n = 2/3/3 per group in NC-IgG, HFD-IgG, and HFD-LPAM1, respectively). (B) Calculated relative intensities and representative images of LPS- or control-treated (PBS) Hepa1-6 extract immunoblotted with anti-ERK and pERK in each group (n = 4/3 per group). Band intensities were calculated as pERK intensity per ERK intensity. In each cohort, the band intensity at each hour is shown in comparison with 0 hour, serving as the basal intensity. (C) Calculated relative intensities and representative images of IL-23– or control-treated (PBS) Hepa1-6 extract immunoblotted with anti-ERK and pERK in each group (n = 5/5 per group). Band intensities were calculated as pERK intensity per ERK intensity. In each cohort, the band intensity at each hour is shown in comparison with 0 hour, serving as the basal intensity. Data are presented as means ± SEM. *P < 0.05 as assessed by 1-way ANOVA, followed by Tukey’s HSD post hoc test.

Figure 14. Schematic model.

The neuronal relay system from the intestine to the pancreas identified in the present study is depicted graphically.