Bile diversion to the distal small intestine has comparable metabolic benefits to bariatric surgery

Abstract

Roux-en-Y gastric bypass (RYGB) is highly effective in reversing obesity and associated diabetes. Recent observations in humans suggest a contributing role of increased circulating bile acids in mediating such effects. Here we use a diet-induced obesity (DIO) mouse model and compare metabolic remission when bile flow is diverted through a gallbladder anastomosis to jejunum, ileum or duodenum (sham control). We find that only bile diversion to the ileum results in physiologic changes similar to RYGB, including sustained improvements in weight, glucose tolerance and hepatic steatosis despite differential effects on hepatic gene expression. Circulating free fatty acids and triglycerides decrease while bile acids increase, particularly conjugated tauro-β-muricholic acid, an FXR antagonist. Activity of the hepatic FXR/FGF15 signalling axis is reduced and associated with altered gut microbiota. Thus bile diversion, independent of surgical rearrangement of the gastrointestinal tract, imparts significant weight loss accompanied by improved glucose and lipid homeostasis that are hallmarks of RYGB.

Figure 1. Biliary diversion schematic and effects on body weight, food intake and body composition.

(a) For the biliary diversion procedure, the common bile duct was ligated proximal to the pancreatic duct. The gallbladder was then anastomosed to one of the following: (1) jejunum 4 cm distal to the ligament of Treitz (GB-J), (2) ileum 4 cm proximal to the ileo-caecal valve (GB-IL) or (3) gallbladder-duodenal anastomosis (GB-D model) at the level of the ampulla of Vater. GB-D was performed without significant alteration of bile flow and functioned as a sham surgery. The RYGB procedure was performed as we previously described. Mice were fed a high fat diet (HFD) for induction of diet-induced obesity (DIO), underwent the surgical procedures and were monitored for 8 weeks post-operatively. (b) Average daily food intake, Bio-Serv F3282 (5.49 kcal g⁻¹); (c) relative change in body weight in N of 15 DIO, 15 GB-IL and 7 RYGB mice; and (d) serial body composition measures via NMR of fat and lean mass (N of 15 DIO, 12 GB-D, 11 GB-J, 15 GB-IL, 12 RYGB). Values shown are mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001 versus DIO controls by one-way analysis of variance with Dunn's post-test.

Figure 2. Biliary diversion modifies bile acid abundance and composition.

(a) Serum bile acid levels and (b) fold change in serum bile acids in mice subjected to biliary diversion (GB-D, GB-J and GB-IL) and RYGB relative to DIO controls at 8 weeks post-operative. *P<0.05, **P<0.01 versus DIO by one-way analysis of variance with post-test. Values shown are mean±s.e.m. N of 5 per group.

Figure 3. Effects of biliary diversion on glucose tolerance, insulin sensitivity and lipid metabolism.

(a) Blood glucose was measured after a 4-h period of food-restriction. (b) Experimental groups underwent intraperitoneal glucose challenge at 2, 4 and 8 weeks post-operative following biliary diversion procedures. Area under the curve (AUC) measurements between 0 and 120 min were calculated and compared among groups. (c) Fasting plasma insulin, (d) the homeostatic model assessment of insulin resistance (HOMA-IR) (e) serum cholesterol, (f) serum FFAs, (g) serum triglycerides determined at 4 weeks post-operative with GB-J, GB-IL and RYGB compared with DIO. (h,i) Insulin sensitivity was determined by the hyperinsulinemic-euglycemic clamp, where a continuous infusion of insulin (4 mU kg⁻¹ min⁻¹) was delivered with euglycemia (140 mg dl⁻¹) maintained by a variably adjusted glucose infusion. (h) Glucose infusion rates (mg glucose per kg min⁻¹) over the 120 min procedure in DIO, GB-IL and RYGB mice (N=4), (i) and the mean glucose infusion rates during the last 20 min of the clamp. (j) Faecal fat (w/w% by mass), (k) faecal cholesterol (mol per g faeces); (l) faecal FFA (mol per g faeces) and (m) faecal triglyceride (mol per g faeces) in mice under study. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 versus DIO by one-way analysis of variance with Dunn's post-test. Values shown are mean±s.e.m. (for a,b, N of 15 DIO, 17 GB-D, 18 GB-J, 23 GB-IL, 6 RYGB; for c,d; j–m, N of 5–10 per group).

Figure 4. Energy expenditure in response to bariatric procedures.

(a) Energy expenditure over a 24-h period was assessed by indirect calorimeter in DIO, GB-IL and DIO mice at 4 weeks post-operative. (b) Unadjusted energy expenditure (kcal h⁻¹). (c) ANCOVA adjusted mean energy expenditure. (d) Food intake (kcal per day) was monitored daily for 5 days in each study group. (e) The frequency of locomotor activity (pedestrian meters) as determined by beam breaks per 24-h period. N of 4 DIO, N of 6 GB-IL, N of 6 RYGB per group. Values shown are the mean±s.e.m. *P<0.05 by two-way ANCOVA.

Figure 5. Effects of biliary diversion on liver histology and hepatic gene expression.

(a) Liver steatosis as assessed by H&E staining in lean, DIO, GB-D, GB-J, GB-IL and RYGB mice at 8 weeks after surgery was significantly reduced (b) in GB-IL and RYGB mice relative to DIO. (c) Expression of FXR target genes in liver at 4 weeks post-operative. (d,e) Canonical pathways most differentially expressed between GB-IL versus DIO (d) and RYGB versus DIO (e). Expression of hepatic genes involved in (f) agranulocyte adhesion, (g) cholesterol biosynthesis, (h) eicosanoid signalling and (i) stellate cell activation. (j) Inflammation by hepatic F4/80, (k) Ki-67 and (l) caspase 3 by immunohistochemical staining. *P<0.05; **P<0.01, ***P<0.001 versus DIO by one-way analysis of variance with Dunn's post-test. Values are mean±s.e.m. (N of 5–10 per group for b–i; 4 per group for j–l). Magnification bar, 200 μm.

Figure 6. Effects of biliary diversion on ileal and hepatic gene expression.

(a) Ileum and (b) liver tissues 8 weeks post-operative were harvested for gene expression analysis using RT–PCR. *P<0.05, **P<0.01 versus DIO by one-way analysis of variance followed by two-tailed Students t-test. Values shown are mean±s.e.m. N of 3–5 per group. Bacs, bile-acid-CoA synthetase; Bat, bile acid transporter; Bsep, bile sale effluent pump; Ibabp, ileal bile acid-binding protein; Ibat, Ileal bile acid transporter; Mrp3, multi-drug resistance protein 3; Ntcp, Na-taurocholate cotransporting polypeptide; Ostα and Ostβ, organic solute transporter α and β.

Figure 7. Signalling in the ileum and liver is altered after biliary diversion.

Immunoblots of (a) ileum and (b) liver protein obtained from lean, DIO and GB-IL mice 8 weeks after surgery. Expression was normalized to GAPDH. *P<0.05, **P<0.01 by one-way analysis of variance with Dunn's post-test. ^$P<0.05, ^$$P<0.01, ^$$$P<0.001 by two-tailed Students t-test. Values shown are mean±s.e.m. N of 3–8 per group.

Figure 8. Biliary diversion procedures differentially alter the gut microbiome.

Caecal contents from DIO, GB-D, GB-J and GB-IL mice 8 weeks after surgery were subjected to 16S rRNA gene sequence analysis. Relative abundance of bacterial genera (bar chart) and phyla (pie chart) with each surgical procedure is shown. The y-axis is a weighted Unifrac analysis of the microbiota for the pooled treatment group. ‘*' indicates reference points for a given taxa (N of 5 mice per group).