Loss of Cyp8b1 Improves Glucose Homeostasis by Increasing GLP-1

Bile acids are amphiphilic molecules synthesized from cholesterol in the liver. They are physiological detergents that maintain cholesterol homeostasis (1). Bile acid synthesis is a multistep process consisting of two distinct pathways: the classical and the alternate. The classical pathway accounts for most of the bile acids produced, starting with the enzymatic conversion of cholesterol into 7α- hydroxycholesterol. This reaction is carried out by the rate-limiting enzyme of bile acid synthesis, cytochrome P450, family 7, subfamily A, polypeptide 1 (Cyp7a1). Cholic acid (CA) is the major bile acid, and cytochrome P450, family 8, subfamily B, polypeptide 1 (Cyp8b1) plays a critical role in its production. In humans, the other major bile acid is chenodeoxycholic acid (CDCA). Because CDCA is converted into muricholic acid (MCA) in rodent livers, the main bile acids found in rodents are CA and MCA instead (2). Mice with targeted disruption of Cyp8b1 (Cyp8b1^−/−) fail to produce CA and have an increased bile acid pool due to increased Cyp7a1 activity (3). The fraction of the pool consisting of CA is predominantly replaced by α- and β-muricholates in Cyp8b1^−/− mice. The resulting increase in hydrophilic bile acid species leads to a reduction in intestinal absorption and in hepatic accumulation of cholesterol (4–6). Cyp8b1^−/−× ApoE^−/− mice show reduced atherosclerotic plaques, owing to decreased levels of lipoproteins containing apolipoprotein B in the plasma, reduced hepatic cholesteryl esters, and enhanced bile acid synthesis (7). Furthermore, cholesterol-fed alloxan-induced type 1 diabetic Cyp8b1^−/− mice are protected against hypercholesterolemia and gall stones (5). These findings suggest that the absence of Cyp8b1 may be beneficial in cases of metabolic syndrome.

Bile acids also function as signaling molecules, activating nuclear (farnesoid X receptors [FXR]) and G-protein–coupled receptors (TGR5) (8). Functional activation of TGR5 with bile acids in cell lines and mice improves the secretion of GLP-1 (9,10). The incretin, GLP-1, a peptide hormone derived from differential processing of the proglucagon gene in the L cells of the small intestine, is released in response to nutrient ingestion. GLP-1 mimetic drugs improve insulin sensitivity and increase β-cell mass in insulin-resistant obese fa/fa Zucker rats (11). These beneficial effects have resulted in the development of new antidiabetic drugs mimicking GLP-1 (12).

Recent evidence suggests that colesevelam, a well-known bile acid sequestrant (BAS), suppresses hepatic glycogenolysis in a TGR5-mediated, GLP-1-dependent manner (13). Furthermore, colesevelam has been shown to preferentially bind CA and to also improve glucose tolerance in humans (1). The proposed mechanism for the modulation of GLP-1 by BAS involves increased free fatty acid (FFA) content in the ileum due to defective micellar solubilization in the jejunum after BAS treatment (14).

We hypothesized that the reduction of Cyp8b1 would lead to improved insulin secretion and glucose tolerance in mice via increased GLP-1 release. Using Cyp8b1^−/− mice, we show that reduced intestinal fat absorption resulted in increased FFA content in the ileal lumen. In addition, significantly improved glucose tolerance, insulin sensitivity, and insulin secretion due to increased plasma GLP-1 levels were observed. Furthermore, the phenotypes observed in Cyp8b1^−/− mice were lost after administration of a specific and potent GLP-1 receptor antagonist or CA, suggesting the critical involvement of elevation of GLP-1 and the absence of CA in the observed phenotype. Our data provide strong evidence that inhibition of Cyp8b1 may be a viable therapeutic target in the management of diabetes.

Cyp8b1^+/− mice were purchased from the University of California, Davis, Knockout Mouse Project (KOMP) repository (Cyp8b1^{tm1(KOMP)Vlcg}). Control and Cyp8b1^−/− mice were obtained by breeding Cyp8b1^+/− mice. Our study used female mice, aged 3–6 months. The mice had ad libitum access to water and were fed a 0.5% cholesterol diet (high cholesterol diet [HCD]; Harlan Diets, Madison, WI) for 4–6 weeks. For experiments involving fasting, mice were fasted for 4 h (7:00 a.m.–11:00 a.m.) before the procedure. Blood for all experiments was drawn from the medial saphenous vein in EDTA-coated capillary tubes. Tissue harvesting and islet isolations were performed after the mice had been fed the HCD for 6–8 weeks. All experiments were approved by the University of British Columbia Animal Care Committee.

Plasma samples were prepared for and analyzed by gas chromatography–mass spectroscopy, as previously described (15).

Blood was drawn at preestablished times from fasted mice before and after intragastric gavage of 200 μL corn oil. Plasma triglyceride (TG) content was measured using the manufacturer’s protocol (Roche, Mississauga, Ontario, Canada). For CA feeding experiments, mice were orally gavaged with 17 mg/kg CA (Sigma-Aldrich, Oakville, Ontario, Canada) or vehicle (1.5% NaHCO₃), as previously described (16), and oral fat tolerance tests (OFTT) were performed 30 min after feeding.

Fasted mice were gavaged with a bolus of a high-fat meal (Research Diets, New Brunswick, NJ) in PBS (200 μL of 200 mg/mL) and killed 1 h postgavage. Sections of the intestine (duodenum, jejunum, and ileum) were collected in 0.5 mmol/L sodium taurocholate (Sigma-Aldrich), vortexed, and centrifuged at 3,000 rpm for 15 min. Lipids were extracted using the Folch extraction method (17). FFAs were measured using an enzymatic colorimetric assay (Roche).

Fasting plasma TG and HDL cholesterol were measured from 5 μL freshly isolated plasma according to the manufacturer’s protocol (TG: Roche; HDL cholesterol: Thermo Fisher Scientific Inc., Middletown, VA).

Oral (OGTT) and intraperitoneal glucose tolerance tests (IPGTT) were performed on fasted mice by gavaging or injecting 2 g/kg glucose and measuring glucose using a glucometer (LifeScan, Burnaby, British Columbia, Canada) at predetermined time points. Insulin tolerance tests (ITT) were performed by injecting 1 unit/kg human recombinant insulin (Novo Nordisk, Mississauga, Ontario, Canada) in the fasted mice, followed by measurement of glucose as in the GTT. For CA feeding experiments, mice were orally gavaged with 17 mg/kg CA or vehicle (1.5% NaHCO₃), as previously described (16), for 3 consecutive days at 6:00 p.m., and an IPGTT was performed the following day. For the relative estimation of the effect of OGTT versus IPGTT, the change in the area under the curve was calculated as ΔAUC (AUC = AUC [IPGTT] – AUC [OGTT]), where AUC is area under the curve.

Fasted mice were gavaged with 2 g/kg glucose and blood was drawn at predetermined time points postgavage. Plasma was isolated, and total glucose-dependent insulinotropic polypeptide (GIP) and GLP-1 levels were measured using ELISA according to the manufacturer’s protocol (GIP: Millipore, Billerica, MA; GLP-1: MSD, Gaithersburg, MD). For CA feeding experiments, mice were treated as mentioned above (GTT and ITT). On the following day, blood was isolated at predetermined time points postglucose gavage to measure plasma GLP-1.

To determine insulin secretion in vivo, fasted mice were gavaged with 2 g/kg glucose, and plasma insulin was measured at different time points by ELISA (Mercodia, Uppsala, Sweden). Insulin secretion ex vivo was performed on hand-picked islets isolated after intraductal collagenase injections, as previously described (18,19), or by differential density centrifugation over Histopaque (Sigma-Aldrich). Ten islets per well were incubated in Krebs-Ringer bicarbonate buffer containing 1.67 mmol/L or 16.7 mmol/L glucose. The media were removed after 1 h, and the islets were lysed in an acid-alcohol buffer (EtOH:H₂O:HCl at 150:47:3). Insulin secretion was normalized to islet DNA content. Lysed islets were diluted 1:2,000, and islet insulin content was measured using ELISA (Mercodia).

GTT was performed in mice injected with 167 µg/kg body weight exendin-(9-39)amide (Ex-9) (Tocris Bioscience, Minneapolis, MN) or vehicle (PBS) 30 min before an oral glucose gavage. For the relative estimation of the effect of Ex-9 on GTT, the change in AUC was calculated as ΔAUC = AUC [vehicle] – AUC [Ex-9].

Total RNA was isolated from tissues using an RNA extraction kit (Qiagen, Toronto, Ontario, Canada) and from islets using the RNeasy kit (Life Technologies, Burlington, Ontario, Canada). One microgram of RNA was used to synthesize cDNA using the Superscript III First-Strand Synthesis kit (Life Technologies). Primer sequences used for measuring gene expression were as follows: Cyp7a1: Forward (5′-ACG CAC CTC GTG ATC CTC TGG G-3′), Reverse (5′-GGC TGC TTT CAT TGC TTC AGG GCT-3′); Cyp27a1: Forward (5′-GTT CGG TCT TGC CTG GGT CGG-3′), Reverse (5′-ACT TCT CCC ATC CCG GGA GCC-3′); Cyp8b1: Forward (5′-ATC GCC TGA AGC CCG TGC AG-3′), Reverse (5′-AGC TGG GGA GAG GAA GGA GTG C-3′); Glp-1: Forward (5′-GCC CAA GAT TTT GTG CAG TGG-3′), Reverse (5′-GTC CCT TCA GCA TGC CTC TC-3′); and Gapdh: Forward (5′-TGC ACC ACC AAC TGC TTA G-3′), Reverse (5′-GAT GCA GGG ATG ATG TTC-3′) was used as the housekeeping gene.

Quantitative real-time PCR was performed in an ABI Prism 7700 Sequence Detection System, using SYBR Green PCR Master Mix (Applied Biosystems, Warington, U.K.).

Formalin-fixed, paraffin-embedded sections (5 μm) from three areas of the pancreas of HCD-fed Cyp8b1^−/− and control mice were deparaffinized and rehydrated. Antigen retrieval was performed with Target Retrieval Solution (Dako, Carpinteria, CA) in a steamer for 20 min. Sections were blocked with 2% normal goat serum (Vector Laboratories, Burlington, Ontario, Canada) for 30 min and then incubated with guinea pig anti-insulin (1:200; Dako) in 0.1% BSA/PBS overnight at 4°C. Alexa 594 goat anti-guinea pig secondary antibody (1:200; Invitrogen Life Technologies, Burlington, Ontario) was applied for 1 h at room temperature. Slides were mounted using Vectashield mounting medium with DAPI (Vector Laboratories) and imaged on a BX61 microscope (Olympus, Center Valley, PA). Quantification was performed using Image-Pro software (Media Cybernetics, Bethesda, MD). β-Cell mass was measured by calculating the percentage of insulin-positive surface area from six evenly spaced sections per pancreata. The mean insulin-positive area was multiplied with the pancreatic wet weight to estimate β-cell mass.

All data are presented as mean ± SEM. As noted in the Figure legends, data were analyzed by using unpaired Student t test with two-tailed analysis or two-way ANOVA, followed by the Bonferroni post hoc test.

Cyp8b1^−/− mice display an overall increase in the bile acid pool, with complete absence of CA and relative enrichment of other bile acids (2). Therefore, we quantified plasma bile acids in the Cyp8b1^−/− mice and observed abolished CA, confirming the previously described phenotype (Fig. 1A). In addition, the Cyp8b1^−/− mice have no change in circulating CDCA levels (Fig. 1A). The unchanged CDCA levels may be explained by efficient conversion of excess CDCA to MCA, because mice lacking Cyp8b1 have increased plasma levels of α-MCA (eightfold increase; P = 0.002), β-MCA (fourfold increase; P = 0.002), and ω-MCA (twofold increase; P = 0.02) (Fig. 1A). Cyp8b1^−/− mice also have a significant increase (approximately fourfold) in plasma ursodeoxycholic acid (UDCA; P = 0.002) (Fig. 1A). Overall changes in the plasma bile acid profile were similar to those reported in the bile of Cyp8b1^−/− mice (3).

This was further corroborated by direct transcript measurements of the genes involved in bile acid biosynthesis. Quantification of hepatic mRNA revealed that Cyp7a1 gene expression was significantly higher in Cyp8b1^−/− mice compared with the controls (P = 0.005), whereas there was no change in Cyp27a1 gene expression (Fig. 1B). In addition, the Cyp8b1^−/− mice show complete knockdown of Cyp8b1 gene expression compared with control mice (P = 0.0004) (Fig. 1B). These findings are also in line with previously published data (3).

By acting as signaling molecules for FXR and TGR5, bile acids are involved in maintaining glucose metabolism and insulin sensitivity (8). Cyp8b1^−/− mice have an altered bile acid pool size (3), which may result in altered glucose homeostasis. To test our hypothesis, we first measured fasting plasma glucose levels, which showed a marked reduction in Cyp8b1^−/− mice compared with controls (P = 0.002) (Fig. 2A). Recent evidence suggests an association between increased 12α-hydroxy bile acids and insulin resistance in humans (20). Also, mice overexpressing Cyp7a1 are protected against diet-induced obesity and insulin resistance (21). Cyp8b1^−/− mice showed a reduction of ∼76% in fasting insulin levels compared with controls (P = 0.02) (Fig. 2B). OGTT results showed that Cyp8b1^−/− mice were more glucose tolerant than controls (incremental AUC; P = 0.04) (Fig. 2C and D). Because Cyp8b1^−/− mice lack 12α-hydroxy bile acids, express increased hepatic Cyp7a1 transcripts (3), and have lower fasting insulin levels, we assessed insulin sensitivity in these mice. Intraperitoneal insulin tolerance test (IPITT) results showed that Cyp8b1^−/− mice were highly insulin sensitive (ANOVA P = 0.012, AUC P = 0.0007) (Fig. 2E and F).

In contrast to insulin secretion induced by the intraperitoneal injection of glucose, increased insulin secretion after an oral glucose load is partly due to the action of incretins, an effect that is diminished in type 2 diabetes (22). To investigate the role of incretin hormones in Cyp8b1^−/− mice, we performed IPGTT and OGTT in control and Cyp8b1^−/− mice (Fig. 3A and B). Cyp8b1^−/− mice showed a greater incretin effect compared with control mice, as seen by significantly higher ΔAUC in Cyp8b1^−/− mice (P = 0.041), which represents the area between the curves for IPGTT and OGTT (Fig. 3C). These data suggest that incretin hormones may contribute to the improved phenotypes of Cyp8b1^−/− mice.

The main function of bile acids is to facilitate micelle formation and promote dietary fat absorption. We observed reduced body weight in Cyp8b1^−/− independent of dietary regimen (Supplementary Fig. 1A), primarily due to decreased adipose depot weights (Supplementary Fig. 1B). This suggests that altered bile acid composition in these mice may contribute to reduced dietary fat absorption. To determine this, we performed OFTT in Cyp8b1^−/− mice. As expected, the Cyp8b1^−/− mice showed a reduction in fat absorption (ANOVA P = 0.041, AUC P = 0.03) (Fig. 3D and E) (4). These data show that the altered bile acid profile leads to defective intestinal fat absorption in addition to the reported reduction in cholesterol absorption (3,6).

Incretin hormones are secreted in response to nutrient ingestion, and FFAs activate their release (23). Our observation of reduced fat absorption led us to hypothesize that increased FFAs may persist in the intestinal lumen of Cyp8b1^−/− mice. Thus, we quantified luminal FFA content in Cyp8b1^−/− mice gavaged with a liquefied high-fat diet and found increased FFAs in their ileal lumen compared with controls (P = 0.03) (Fig. 4A). Because FFAs are known to potentiate the release of incretins, we measured plasma levels of GIP and GLP-1 in these mice after the oral glucose load. No change was observed in the plasma total GIP levels (Fig. 4B). However, Cyp8b1^−/− mice showed a significant increase in plasma GLP-1 excursions 10 min after the oral glucose (ANOVA P = 0.04) (Fig. 4C). GLP-1 transcript levels were also higher in the ilea of these mice (P = 0.03) (Fig. 4D). These results correlate with our observation of increased FFA content in the ileal lumen of the Cyp8b1^−/− mice.

Because GLP-1 release was increased in Cyp8b1^−/− mice, we next measured insulin secretion, and it was higher 15 min after the oral glucose gavage in Cyp8b1^−/− mice (ANOVA P = 0.04) (Fig. 5A). GLP-1 stimulates transcription of the proinsulin gene and increases insulin biosynthesis (24). In line with this, insulin content was markedly higher in Cyp8b1^−/− islets (P = 0.03) (Fig. 5B). To determine if the increased islet insulin content contributed to insulin secretion, we performed ex vivo glucose-stimulated insulin secretion on freshly isolated islets. The fold increase in insulin secretion from the Cyp8b1^−/− islets in response to 16 mmol/L glucose compared with basal secretion was higher than that of controls (P = 0.04) (Fig. 5C). Cyp8b1^−/− mice, however, did not demonstrate any change in β-cell mass (Fig. 5D and E). It is essential to note that pharmacological but not physiological doses of GLP-1 increase β-cell mass in rats and diabetic mice (25,26). Our findings show that β-cell function is enhanced in mice lacking Cyp8b1, associated with an increase in circulating levels of GLP-1.

Next, we investigated if GLP-1 was the predominant contributor to the observed improvement in β-cell function of Cyp8b1^−/− mice. Ex-9 is a GLP-1 receptor antagonist that affects β-cell function and glucose metabolism in healthy subjects (27). We performed OGTT in Cyp8b1^−/− and control mice, administered with Ex-9 or vehicle. As expected, Ex-9–treated controls showed impaired glucose tolerance compared with vehicle-treated controls (Fig. 6A). Interestingly, GLP-1 receptor antagonism in the Cyp8b1^−/− mice also impaired glucose tolerance (Fig. 6B). To calculate differences in tolerance, we calculated ΔAUC as a measure of impairment induced by Ex-9. The ΔAUC in Cyp8b1^−/− mice was significantly lower than that of control mice (P = 0.01), owing to higher endogenous GLP-1 (Fig. 6C). These data suggest that GLP-1 is an important contributor to the improved glycemic phenotype of Cyp8b1^−/− mice.

To dissect the contribution of the loss of CA in manifestation of the improved phenotype, we gavaged Cyp8b1^−/− mice with CA or vehicle. Interestingly, the observed reduction in fat absorption of Cyp8b1^−/− mice was normalized upon CA administration (Fig. 7A and B). Next, we gavaged CA to Cyp8b1^−/− mice for 3 consecutive days, followed by a measurement of glucose tolerance. CA treatment of Cyp8b1^−/− mice reversed the improved glucose tolerance, normalizing the glucose levels to those of vehicle-treated controls (Fig. 7C and D). Interestingly, plasma GLP-1 excursions of the CA-treated Cyp8b1^−/− mice were normalized to those of vehicle-treated controls, after the oral glucose gavage (Fig. 7E). Together, these data suggest that the absence of CA contributes to the improved glucose tolerance in mice lacking Cyp8b1.

Mice with targeted disruption of P450 cytochrome Cyp8b1 have reduced intestinal cholesterol absorption. Contrary to the expected feedback regulation, bile acid pool size and bile acid synthesis are increased several-fold in these mice via the upregulation of hepatic Cyp7a1 (3). The upregulation of Cyp7a1 is known to be beneficial and was recently shown to be a consequence of a novel positive-feedback mechanism exerted by bile acids that antagonize FXR (28). In our study, the altered bile acids in Cyp8b1^−/− mice also lowered plasma TGs and increased the levels of HDL cholesterol (Supplementary Fig. 3A and B). These findings are in accordance with the lowering of atherosclerosis incidence in Cyp8b1^−/− concomitant to the reduction of circulating lipoproteins containing apolipoprotein B. Our study is the first to demonstrate that absence of Cyp8b1 and the consequent changes in bile acid pool result in improved glucose tolerance and β-cell function. This effect appears to be primarily mediated by increased GLP-1 secretion and subsequent improvement in insulin secretion (Fig. 8A). Our findings show that genetic ablation of 12α-hydroxylated bile acids results in improved β-cell function and insulin sensitivity. This is in agreement with a recent finding of 12α-hydroxylated bile acids being associated with insulin resistance in humans (20).

The previous observations of reduced intestinal cholesterol absorption in Cyp8b1^−/− mice prompted us to investigate intestinal fat absorption. Interestingly, Cyp8b1^−/− mice showed diminished intestinal fat absorption. Pair feeding Cyp8b1^−/− mice and controls with the HCD led to comparable body weight gain (Supplementary Fig 1C). However, improvement in oral glucose tolerance persisted even upon pair feeding (Supplementary Fig. 1D), indicating that body weight cannot completely explain improved glucose tolerance. We hypothesized that reduced fat absorption would affect luminal levels of FFAs, which are mediators of incretin release (23). The reduction in fat absorption was directly correlated with an increase in the luminal FFA content. Luminal FFA levels were elevated only in the ileum of Cyp8b1^−/− mice. These changes suggested that distal intestinal incretin, GLP-1 (produced by L cells), could be preferentially affected compared with the proximal intestinal incretin, GIP (produced by K cells). To test this, we assessed levels of GLP-1 and GIP in Cyp8b1^−/− mice after the oral glucose challenge. As expected, Cyp8b1^−/− mice exhibited higher GLP-1 release, whereas no differences in GIP levels were observed. These data indicate a potential contribution of luminal fatty acids in the OGTT phenotype of Cyp8b1^−/− mice.

Absence of CA and its derivatives, the potent activators of TGR5, leads us to speculate that the bile acid–induced TGR5-mediated GLP-1 secretion pathway may not play a major role in improving the glycemic profile of the Cyp8b1^−/− mice. However, some contribution of TGR5 in the development of the glycemic phenotype cannot be overruled because the levels of CDCA in Cyp8b1^−/− mice are increased (3). Lithocholic acid (LCA), the secondary bile acid formed from CDCA, has been shown to be the most potent activator of TGR5 (29). Moreover, an oral glucose dose is known to induce bile acid excursions (30). Levels of LCA in the colon of Cyp8b1^−/− mice were approximately threefold higher compared with controls (Supplementary Fig. 2A). To determine whether increased levels of LCA may account for the effects on glucose tolerance, we performed OGTT in mice treated acutely with LCA (15 mg/kg, 10 times lower than toxic dose) (31). We found no difference in oral glucose tolerance upon LCA treatment (Supplementary Fig. 2B). These results are in line with the report of Thomas et al. (10), who found that LCA failed to induce GLP-1 release in wild-type colonic explants but was able to do so only in colonic explants overexpressing TGR5.

Interestingly, we also found an increase in the transcript levels of GLP-1 in the ileum of Cyp8b1^−/− mice. This may be indicative of an increase in the number of GLP-1–positive L cells or an increase in the transcription of GLP-1 without a consequent increase in L-cell numbers. It will be interesting to study the effects of Cyp8b1 deficiency and the consequent changes in bile acids on the expression or differentiation of intestinal L cells.

Increased GLP-1 levels in Cyp8b1^−/− mice correlated not only with an improvement in β-cell function but also with peripheral insulin sensitivity. GLP-1 is an insulin secretagogue that improves β-cell function and promotes weight loss, thereby improving insulin sensitivity (32,33). However, Gedulin et al. (11) have shown in rats that the improvement in insulin sensitivity in response to Ex-4, a GLP-1 receptor agonist, persists even upon pair feeding. This indicates that GLP-1 improves insulin sensitivity independent of body weight. We also observed that GLP-1 receptor antagonism in Cyp8b1^−/− mice led to a complete reversal of the observed improvement in their glucose tolerance. Moreover, the effect on glucose tolerance persisted upon pair feeding. This indicates that GLP-1 plays a critical role in influencing the glycemic phenotype of Cyp8b1^−/− mice.

We found that CA administration reversed the improvements in fat and glucose tolerance in Cyp8b1^−/− mice. It must be noted that the changes in the bile acid pool of Cyp8b1^−/− mice are recapitulated in germ-free and antibiotic-treated mice, in which CA levels are diminished and cholesterol absorption is reduced (34,35). A recent study directly confirmed the similarity in bile acid composition of antibiotic-treated and Cyp8b1^−/− mice; however, regulation of blood glucose levels was not assessed (28). Interestingly, germ-free mice display improved glucose tolerance and insulin sensitivity, although the underlying mechanisms remain underexplored (36). Thus, the absence of CA likely leads to the observed improvement in the glycemic phenotype of Cyp8b1^−/− mice.

Recent evidence that murine islets express FXR and TGR5, receptors of bile acids (37,38), suggests that the altered peripheral bile acid composition of Cyp8b1^−/− mice may have a direct and important role in improving β-cell function. Thus, it would be interesting to investigate the direct role of the increased bile acid species in Cyp8b1^−/− mice (α-, β-, and ω-MCA and UDCA) in regulating glucose tolerance and β-cell function.

In conclusion, our study is the first to provide evidence that loss of CA improves glucose metabolism and insulin sensitivity in Cyp8b1^−/− mice. We show that these effects are consistent with a role for GLP-1 in the phenotype of Cyp8b1^−/− mice. It is of interest to determine if similar effects are recapitulated upon reducing CA in humans. Our data strongly demonstrate that the inhibition of Cyp8b1 has potential therapeutic implications by eliminating 12α-hydroxylated bile acids and thus increasing the level of endogenous GLP-1 in the treatment of type-2 diabetes.

Acknowledgments. The authors acknowledge the assistance of H. Mark Wang, Jason Yao, and the mouse core of the Centre for Molecular Medicine and Therapeutics for special care of the mice and the technical expertise provided by Renze Boverhof (University Medical Center Groningen, Groningen, the Netherlands) for bile acid measurements.

Funding. This work was supported by a Canadian Institutes of Health Research (CIHR)http://dx.doi.org/10.13039/501100000024 grant (MOP 106684) to M.R.H. M.R.H. is a Killam Professor of Medical Genetics and a Canada Research Chair in Human Genetics at the University of British Columbia. C.B.V. holds the Irving Barber Chair in Diabetes Research.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.K. designed and performed the experiments, analyzed data, and wrote the manuscript. J.V.P. designed and performed experiments and wrote the manuscript. W.d.H. and P.R. helped perform experiments. N.W. reviewed and edited the manuscript. A.K.G. performed the bile acid measurements. C.B.V. and R.R.S. contributed to discussion and reviewed and edited the manuscript. M.R.H. designed the experiments and reviewed and edited the manuscript. M.R.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.