Colestilan decreases weight gain by enhanced NEFA incorporation in biliary lipids and fecal lipid excretion.

Bile acid sequestrants (BASs) are cholesterol-lowering drugs that also affect hyperglycemia. The mechanism by which BASs exert these and other metabolic effects beyond cholesterol lowering remains poorly understood. The present study aimed to investigate the effects of a BAS, colestilan, on body weight, energy expenditure, and glucose and lipid metabolism and its mechanisms of action in high-fat-fed hyperlipidemic APOE*3 Leiden (E3L) transgenic mice. Mildly insulin-resistant E3L mice were fed a high-fat diet with or without 1.5% colestilan for 8 weeks. Colestilan treatment decreased body weight, visceral and subcutaneous fat, and plasma cholesterol and triglyceride levels but increased food intake. Blood glucose and plasma insulin levels were decreased, and hyperinsulinemic-euglycemic clamp analysis demonstrated improved insulin sensitivity, particularly in peripheral tissues. In addition, colestilan decreased energy expenditure and physical activity, whereas it increased the respiratory exchange ratio, indicating that colestilan induced carbohydrate catabolism. Moreover, kinetic analysis revealed that colestilan increased [(3)H]NEFA incorporation in biliary cholesterol and phospholipids and increased fecal lipid excretion. Gene expression analysis in liver, fat, and muscle supported the above findings. In summary, colestilan decreases weight gain and improves peripheral insulin sensitivity in high-fat-fed E3L mice by enhanced NEFA incorporation in biliary lipids and increased fecal lipid excretion.

periods of 3-4 days (nine samples/group), and neutral sterol and bile acid contents were analyzed by gas chromatography as described previously ( 24 ).

H]NEFA infusion and bile analysis
At weeks 9 and 10, [ 14

C]glycerol/[
3 H]NEFA infusion and bile cannulation were performed by procedures similar to those described previously ( 25,26 ). After overnight fasting, the mice were anesthetized [12.5 mg/kg midazolam (Genthon, Nijmegen, The Netherlands), 25 mg/kg fl uanisone (Janssen Pharmaceutica, Beerse, Belgium), and 0.79 mg/kg fentanyl citrate (Janssen Pharmaceutica)]. The gall bladder was cannulated, and an infusion needle was injected into the inferior vena cava. 3 H]phospholipid and total phospholipid fl ow were determined in the collected bile. Bile acid concentrations were measured using a bile acid kit (Lucron Bioproducts, Milsbeek, The Netherlands). Bile acids were isolated using Sep-Pak C18 cartridges. Subsequently, 3 H dpm incorporation in bile acids was determined. Cholesterol was extracted from bile (Bligh and Dyer method), isolated from other lipids, and quantifi ed by highperformance thin-layer chromatography (HPTLC); [ 3 H]cholesterol was determined by counting 3 H dpm in scraped HPTLC bands. Phospholipids were determined using the Phospholipids B kit (WAKO). [ 3 H]phospholipids were determined by counting 3 H dpm in scraped HPTLC bands. Levels of 14 C incorporation in bile acids, cholesterol, and phospholipids were too low to be measured, making it impossible to determine the conversion of glycerol to biliary lipids.

Fat weight and hepatic lipid content
After glycerol/NEFA infusion, abdominal (epididymal) fat, subcutaneous fat, liver, and femoral muscle tissue were harvested and weighed. The hepatic lipid content (free cholesterol, TG, and cholesteryl esters) was determined by procedures described previously ( 27 ). In brief, 10-20 mg of tissue was homogenized in PBS, the protein content was measured, and lipids were extracted and separated by HPTLC on silica gel plates. Lipid spots were stained with a coloring reagent (

Hyperinsulinemic-euglycemic clamp analysis
Different groups of mice were used for these experiments. Hyperinsulinemic -euglycemic clamp analysis was performed at week 9 as described previously ( 28 ). In brief, after overnight fasting, the mice were anesthetized [0.5 ml/kg Hypnorm (Janssen Pharmaceutica), 12.5 mg/g midazolam (Genthon)]. An infusion needle was inserted into one of the tail veins, and basal glucose parameters were determined by infusing D -[ 3 H]glucose . Thereafter, the mice were given a bolus of insulin; a hyperinsulinemic clamp was initiated by continuously infusing insulin and D -[ 3 H]glucose. Blood samples (<3 l) were collected at 10 min intervals (tail bleeding) to monitor blood glucose levels. A variable infusion of 12.5% D -glucose solution (in PBS) was initiated at time 0 and adjusted to maintain blood glucose at approximately 7.0 mmol/l. When steady-state glucose levels were reached (approximately 1 h after mildly insulin resistant ( 22,23 ). We hypothesized that colestilan induces hepatic cholesterol/bile acid/phospholipid synthesis, leading to increased glycerol/NEFA incorporation in biliary lipids. The increased glycerol/NEFA incorporation, derived from triglyceride (TG) lipolysis in adipose tissue, together with increased fecal lipid excretion would thereby contribute to decreased body weight and improved insulin sensitivity.
To investigate our hypothesis, we used infusion of radiolabeled glycerol and NEFA and determined their subsequent conversions to biliary lipids. In addition, we evaluated the effects of colestilan on fecal excretion, energy metabolism, insulin sensitivity, and analysis of gene expression relevant to lipid and glucose metabolism in liver, adipose tissue, and muscle.

Experimental animals
Heterozygous E3L transgenic male mice aged 10-14 weeks were established at The Netherlands Organization for Applied Scientifi c Research (TNO) animal facility. They were housed in a temperature-controlled room on a 12 h light-dark cycle with water and food ad libitum. Animal experiments were performed according to European Union regulation 86/609/EC, Council of Europe European Treaty Series 123 and the Dutch Experiments on Animals Act, which includes approval by the Institutional Animal Care and Use Committee of TNO. Several of the measurements and results described here were repeated in either another cohort of APOE*3 Leiden mice (plasma parameters, feces composition, hyperinsulinemic-euglycemic clamp experiments, and metabolic cage experiments) or in another diabetic mouse model KKA y mice (plasma parameters, feces composition, and the results of the hyperinsulinemic-euglycemic clamp experiments). Data acquired were in line with data presented herein.

Diets
After a run-in period on a high-fat diet (47 kCal% beef tallow, 20 kCal% protein, 33 kCal% carbohydrates; Hope Farms, Woerden, The Netherlands) for 10-12 weeks, E3L mice were divided on the basis of body weight, blood glucose, plasma cholesterol, and plasma insulin (all measured after 4 h fasting) into 2 groups (week 0). The control group was fed a high-fat diet, whereas the colestilan group was fed a high-fat diet containing 1.5% (w/w) colestilan (Mitsubishi Tanabe Pharma Corporation, Tokyo, Japan). The colestilan was added to the diet, leading to a slightly lower caloric density of the diet (4.48 vs. 4.55 kCal/g). Body weight and food intake were monitored every week.

Analysis of plasma parameters
Blood samples were collected by tail bleeding after 4 h fasting, except for clamp and fl ux analyses, where the animals were fasted overnight. Blood glucose levels were measured using the "Freestyle blood glucose measurement system" from Disetronic Medical Systems BV (Vianen, The Netherlands) according to the manufacturer's instructions. Plasma total cholesterol and TG levels were measured using the assay reagents from Roche Diagnostics (Almere, The Netherlands). Plasma insulin levels were measured using a kit from Linco Research (St. Charles, MO). Plasma NEFA was measured using the NEFA C kit from Wako (Neuss, Germany).

Fecal lipid analysis
At weeks 7 and 8 of the study period, feces were collected from three subgroups per treatment group during three different analyzed for its integrity on Aligent chips ("lab on chip" RNA 6000 Nano assay). If not suffi ciently intact, RNA was extracted again. For both groups, equal quantities of individual RNA samples were pooled and labeled in duplicate (biotin-labeled cRNA synthesis) for hybridization. Each sample was hybridized with Affymetrix mouse array MOE430A (containing probes for 22,625 genes) in duplicate. The hybridization signals were analyzed using Affymetrix software and stored in primary data fi les. Preprocessing and quality control of the microarray data was performed using the ArrayAnalysis.org pipeline (http://www.arrayanalysis. org). All samples passed the quality control . Raw signal intensities were normalized using the guanine cytosine robust multiarray averaging algorithm applied across all arrays in the dataset. Probelevel signals were summarized to gene-level intensities using the custom MNBI CDF-fi le (Entrez Gene annotation, version 16.0.0). This resulted in expression values for 12,251 genes, represented by unique Entrez Gene identifi ers. Normalized intensities were log transformed (base2) and fold changes were calculated based on the averages of the two technical replicates within each group.

Statistical analysis
All data are expressed as means ± standard errors (SEs). The signifi cance of differences was calculated using Student's t -test. P values <0.05 were considered statistically signifi cant.
initiating the insulin infusion), blood samples were collected at 20 min intervals for 1 h to determine insulin-stimulated glucose turnover.

Measurements in metabolic cages
Different groups of mice were used for this experiment. The mice were individually housed in metabolic cages (Columbus Instruments, Columbus, OH) and fed a powdered diet ad libitum. Metabolic parameters in the mice were assessed at weeks 1 and 3 for 2-3 days in metabolic cages (after 1 day acclimatization in the same cage) ( 29 ). The following variables were continuously monitored: total food intake, total physical activity, total O 2 consumption (expressed as VO 2 ml·kg Ϫ 1 ·h Ϫ 1 ), and total CO 2 production ). The respiratory exchange ratio (RER) and energy expenditure (kcal·kg Ϫ 1 ·h Ϫ 1 ) were determined from these variables.

Gene expression analysis
Different groups of mice were used for this experiment. After 8 weeks of colestilan treatment, livers, adipose tissues (perigonadal), and muscles were removed after 4 h fasting and frozen in liquid N 2 for microarray analysis . RNA was extracted using RNA-zolB (Campro Scientifi c, Veenendaal, The Netherlands), purifi ed using the RNeasy kit (Qiagen, Venlo, The Netherlands), and Fig. 1. Colestilan decreases body weight, fat mass, and plasma lipid levels and increases food intake. E3L mice in the control and colestilan groups were given a high-fat diet and high-fat diet with 1.5% colestilan, respectively, for 8 weeks. Body weight (A) and food intake (B) were measured every week, and plasma cholesterol (D) and triglyceride (E) levels were measured after weeks 4 and 8 of colestilan treatment (n = 20-21). Plasma was obtained from E3L mice that were fasted for 4 h. C: Abdominal (abd.) and subcutaneous (subc.) fat, liver, and femoral muscle were removed and weighed after weeks 9 or 10 of colestilan treatment (n = [11][12]. Data are expressed as means ± SE. * P < 0.05, ** P < 0.01 versus control by Student's t -test .
These results indicate that colestilan treatment decreases body weight and fat mass of mice despite increased caloric intake.

Colestilan decreases blood glucose and plasma insulin levels by improving peripheral insulin resistance
Because it was recently reported that colestilan improves glycemic control (in addition to its effects on plasma LDL cholesterol) in type 2 diabetic patients ( 5, 18 ), we next measured blood glucose and plasma insulin levels in E3L mice. Control E3L mice exhibited mild hyperglycemia, which was signifi cantly attenuated after colestilan treatment for 4-8 weeks ( Fig. 2A ). Plasma insulin levels in the colestilan group were signifi cantly decreased compared with the levels in the control group at weeks 4 and 8 ( Fig.  2B ), indicating that colestilan improved insulin sensitivity. To further confi rm the effect of colestilan on insulin sensitivity, we performed hyperinsulinemic-euglycemic clamp analysis after overnight fasting. Whole-body glucose disposal (i.e., the glucose infusion rate) in the colestilan group was signifi cantly higher than that in control group ( Fig. 2C ). Under hyperinsulinemic conditions, whole-body glucose uptake in the colestilan group was signifi cantly increased compared with that in the control group ( Fig. 2D ). The difference in hepatic glucose production (HGP) under basal and hyperinsulinemic conditions was expressed as the percentage of basal HGP to obtain the hepatic insulin

Colestilan decreases body weight, fat mass, and plasma lipids despite higher caloric intake
To induce mild obesity and insulin resistance, male E3L mice were fed a high-fat diet for 10 weeks. Thereafter, the same diet with or without colestilan was administered for 8 weeks. Body weight in the colestilan group was signifi cantly decreased compared with that in the control group from week 2 of treatment ( Fig. 1A ). The colestilan group exhibited signifi cantly increased food intake from week 1 or 2 ( Fig. 1B ), resulting in an on average 7.5-15% higher caloric intake and indicating that the decrease in body weight following colestilan treatment was not due to decreased food intake. Quantitative analysis of adipose tissue deposits at the end of the study revealed that colestilan treatment significantly decreased abdominal (epididymal) and subcutaneous fat ( Fig. 1C ). The liver and (femoral) muscle tissue weights did not differ between the groups ( Fig. 1C ), indicating that the decrease in body weight was primarily caused by a substantial decrease in abdominal and subcutaneous fat.
Plasma cholesterol levels in the control group remained stable during the study period (approximately 7 mmol/l) ( Fig. 1D ). As expected for BAS, colestilan signifi cantly decreased plasma cholesterol levels at weeks 4 and 8 ( Fig.  1D ) and resulted in signifi cantly decreased plasma TG levels at weeks 4 and 8 ( Fig. 1E ). Metabolic cage analysis with continuous measurements of food consumption during a period of 2-3 days confi rmed that colestilan increased food intake (data not shown). At week 1, physical activity of the mice was the same during the daytime, i.e., sleeping time; however, the colestilan-treated mice exhibited signifi cantly lower activity during the night ( Fig. 3A ). Furthermore, O 2 consumption-related energy expenditure of the colestilan-treated mice was signifi cantly lower than that of the control group ( Fig. 3B ), conceivably caused by lower physical activity of the mice. The decrease in physical activity and energy expenditure was also observed at week 3 (data not shown). RER was signifi cantly higher in the colestilan group during the night of day 6 ( Fig. 3C ); this difference became more pronounced after 3 weeks of treatment. At days 19 and 20, the mean RER of the colestilan group was signifi cantly increased at the end of the light period and at the beginning of the dark period ( Fig. 3D ).
These results reveal that the colestilan-treated mice use relatively more carbohydrates than fatty acids for energy generation, which constitutes an additional explanation for the increased glucose uptake in peripheral tissues, as observed in clamp analysis. sensitivity index. Colestilan treatment did not alter the hepatic insulin sensitivity index ( Fig. 2E ), indicating that the drug improved peripheral insulin resistance without affecting hepatic insulin sensitivity.
Collectively, these observations and the decrease in adipose tissue ( Fig. 1C ) support the contention that the colestilan-treated mice become more insulin sensitive.

Colestilan decreases energy expenditure and induces carbohydrate catabolism
The decreased body weight and increased food intake in the colestilan group suggest that colestilan increases energy expenditure. However, the current literature pertaining to this issue suggests that BAS treatment would actually result in decreased energy expenditure ( 12,17 ). No direct experimental studies to address this issue have been performed to date. To clarify this issue, we measured O 2 consumption, CO 2 production, food intake, and physical activity using metabolic cage analysis. We calculated RER, which gives information on the substrate used for energy generation, from the ratio of O 2 consumption to CO 2 production. Moreover, we calculated O 2 consumption-related energy expenditure. during colestilan treatment. Therefore, we tested the hypothesis that colestilan increased NEFA incorporation in biliary lipids by measuring the fl ow of [ 3 H]radiolabeled plasma fatty acids into bile during bile duct cannulation and by determining the mass and incorporation of [ 3 H] NEFA-derived radioactivity in biliary bile acids, cholesterol, and phospholipids. Compared with the control group, colestilan treatment tended ( P = 0.08) to induce higher bile production over 180 min ( Fig. 5G ). Total [ 3 H]NEFA-derived radioactivity in cholesterol ( Fig. 5C ) and phospholipids ( Fig. 5E ) in bile were both signifi cantly increased in the colestilan group. [ 3 H]NEFA-derived radioactivity in bile acids increased in the colestilan group; however, this increase was not signifi cant ( P = 0.14) ( Fig. 5A ). In contrast, compared with the control group, the colestilan-treated group exhibited no difference in nonlabeled bile acids ( Fig. 5B ), cholesterol ( Fig. 5D ), and phospholipids ( Fig. 5F ) in bile.
These observations indicate that colestilan increases biliary lipid synthesis using fatty acids as a substrate in the liver. Consistent with the increased use of NEFA by the liver for conversion to biliary lipids, plasma NEFA levels in the colestilan group were decreased during treatment ( Fig. 5H ).

Colestilan affects expression of genes involved in lipid and glucose metabolism
To further investigate the mechanism by which colestilan affects metabolism, we determined the expression profi le

Colestilan increases fecal lipid excretion without changing fatty acid balance
Because the decreased body weight and increased food intake could not be explained by increased energy expenditure, we also analyzed fecal excretion of fatty acids, bile acids, and neutral sterols. We hypothesized that the decrease in plasma lipids could be attributed to an increased fl ow of biliary lipids from the liver to the feces. First, we measured liver weight and hepatic lipid contents, which did not change signifi cantly, except for the free cholesterol level, which was slightly but signifi cantly increased ( Fig. 1C and Fig. 4D ).
Subsequently, we analyzed fecal lipids and found more than 7-and 3-fold increases in bile acid and neutral sterol excretion, respectively ( Fig. 4A ). Moreover, the colestilan group displayed signifi cantly increased total fecal fatty acid excretion than the control group ( Fig. 4B ). Because the colestilan-treated mice had higher food intake ( Fig. 1B ), the fatty acid balance was calculated from the difference in the total fatty acid intake and total fatty acids excreted in the feces. The fatty acid balance in these mice remained constant compared with that in the control group ( Fig. 4C ), indicating that the same amount of fat was absorbed in the control and colestilan-treated animals.

Colestilan increases the conversion of NEFA to biliary lipids
The observations of lower energy expenditure and unaltered fat absorption do not provide an explanation for the decrease in weight gain, in particular in adipose tissue, Fig. 4. Colestilan increases fecal lipid excretion without changing the fatty acid balance. Feces were collected during weeks 7-8, and fecal bile acids (A), neutral sterols (A), and fatty acids (B) were measured after lipid extraction (n = 9). Fatty acid balance was calculated from the difference between the total fatty acid intake and total fatty acid excreted in the feces (C). Hepatic lipids were measured after weeks 9 to 10 of colestilan treatment (n = 11 Ϫ 12) (D). Data are expressed as means ± SE. * P < 0.05, ** P < 0.01 versus control by Student's t -test.
for insulin-regulated glucose uptake. In contrast, in muscle the expression of Slc2a4 was decreased, suggesting that the improvement of peripheral insulin sensitivity by colestilan is due to increased glucose uptake in adipose tissue.
In general, genes involved in insulin signaling were found to be relatively unaffected. This does however not mean that colestilan did not have an effect, since regulation of the insulin signaling pathway is primarily determined by the phosphorylation state of the proteins therein.

DISCUSSION
In this paper, we demonstrated that the treatment of dyslipidemic and mildly obese and insulin-resistant E3L mice with colestilan markedly improved body weight, hyperglycemia, and dyslipidemia. Of note, colestilan decreased the amount of abdominal and subcutaneous adipose tissue and improved peripheral insulin sensitivity. Mechanistic studies revealed that the drug increased the synthesis of biliary lipids using fatty acids as a substrate in the liver.
Surprisingly, the phenotypic changes in body weight and fat deposition were accompanied by increased food intake, indicating that decreases in body weight and blood glucose levels did not result from decreased food intake. This decrease in body weight was not specifi c for E3L mice because colestilan treatment exerted a similar effect in another mouse model for obesity and insulin resistance, of 12,251 well-characterized mouse genes in liver, adipose tissue, and muscle. A selection of genes involved in lipid and glucose metabolism is depicted in supplementary Fig. I. From this selection, only those genes that were signifi cantly affected by colestilan treatment in either one or more tissues are depicted in Fig. 6 . The hepatic gene expression patterns were mostly directed and related to activation of fatty acids ( Acsl5 and Acss2 ), formation of phospholipids ( Pcyt1a , Cds2 , and Lpcat3 ), and strong enhancement of the fl ux from intermediary metabolites, starting from acetyl-CoA to cholesterol and bile acids, in line with the enhanced synthesis of phospholipids, cholesterol and bile acids. The latter was refl ected by the clear increase of genes involved in cholesterol ( Acly , Acat2 , Hmgcr , and Fdft1) and bile acid synthesis ( Cyp7a1 ) together with a decrease in expression of Nr0B2 (SHP) as major suppressor of Cyp7a1 expression. Although hepatic Ppara gene expression was strongly upregulated, this was not accompanied by stimulation of other genes involved in fatty acid activation and oxidation, most likely explained by the fact that the mRNAs were isolated after 4 h fasting whereas PPAR-␣ is activated under nutrient-defi cient conditions as an adaptive response to prolonged fasting. In contrast, the NEFA incorporation study was performed after overnight fasting, a condition known to activate fatty acid oxidation.
In adipose tissue upregulation of the expression of several genes involved in fatty acid transport ( Fabp1 , Fabp6 , and Slc27a2 ) was observed, but also an enhanced expression of Slc2a4 , the gene encoding for glucose transporter type 4 (GLUT4) , the most important protein KKA y mice (unpublished observation). Although their food intake was increased, the fatty acid balance in the colestilan-treated E3L mice remained the same as that in the control mice because colestilan profoundly increased fecal fatty acid excretion.
Colestilan improved glycemic control in the high-fatfed E3L mice, as refl ected by signifi cant reductions in blood glucose and plasma insulin levels, in line with decreased glucose and HbA1c levels (beside LDL cholesterol) observed in type 2 diabetic patients with or without hypercholesterolemia ( 2-5, 18, 30, 31 ). Hyperinsulinemic-euglycemic clamp analysis demonstrated that colestilan improved insulin sensitivity in the peripheral tissues without affecting the liver. Consistent with this fi nding, the amount of visceral and subcutaneous adipose tissues was decreased, whereas liver weight and the hepatic TG content remained unchanged, the latter in accordance with the inverse correlation between hepatic TG levels and insulin sensitivity as reported previously ( 32 ).
Multiple mechanisms have been proposed to explain the antidiabetic effects of BASs in the literature on the basis of the existing knowledge of bile acid biology ( 7,11,17 ). Cholestyramine has been recently reported to improve glycemic control in Zucker diabetic fatty rats by enhancing incretin responses such as GLP-1 and peptide YY, for which the FXR-SHP-LXR pathway was not required ( 14 ). Another study in diet-induced obese rats demonstrated that colesevelam treatment decreased plasma glucose levels by improving insulin resistance secondary to inducing GLP-1 secretion ( 15 ). In a clinical study, colestilan decreased postprandial plasma glucose concentrations and increased GLP-1 levels in patients with type 2 diabetes ( 4 ). Because we did not measure GLP-1 levels, we cannot exclude the possibility that increased GLP-1 secretion contributes to the reduction in body weight and improved metabolic control in the colestilan-treated E3L mice.
Interestingly, the benefi cial effects of colestilan on obesity and insulin sensitivity were not caused by an increase in energy expenditure. In fact, energy expenditure was found to be decreased. Simultaneously, the calculated RER in the colestilan-treated mice was significantly increased, indicating an enhanced preference for carbohydrates over fats for energy production. Remarkably, this elevated glucose catabolism occurs even under the consumption of a high-fat diet. The latter data together with improved peripheral insulin sensitivity provide an explanation for the increased whole-body glucose uptake in the colestilan-treated mice. Transcriptome analysis confi rmed these data showing that the gene encoding GLUT4 was upregulated in adipose tissue and not in muscle, suggesting increased glucose uptake in fat.
In line with our fi ndings, it has been reported that bile acids promote energy expenditure by TGR5 activation ( 12 ) and that a lowered bile acid pool size reduces energy expenditure ( 16 ). This led to the suggestion that BAS treatment would result in decreased energy expenditure ( 17 ). However, in humans, BAS treatment had no effect on energy expenditure and no correlation with plasma bile acid levels was found ( 33 ). In the present study, the decrease in energy expenditure is possibly related to the observed decrease in physical activity.
We have demonstrated for the fi rst time to our knowledge that BAS increases NEFA incorporation in biliary lipids. The results of [ 3 H]NEFA infusion experiments strongly support our working hypothesis: [ 3 H]NEFA-derived radioactivity in biliary cholesterol and phospholipids in the colestilan group was signifi cantly higher than that in the control group, whereas incorporation in biliary bile acids tended to be increased, indicating enhanced hepatic synthesis of biliary lipids using fatty acids as a substrate. Gene expression profi ling in the liver supported this mechanism by demonstrating upregulation of genes involved in activation of fatty acids ( Acsl5 and Acss2 ), formation of phospholipids ( Pcyt1a , Cds2 , and Lpcat3 ), and synthesis of cholesterol ( Acly , Acat2 , Hmgcr , and Fdft1 ) and bile acids ( Cyp7a1 ). Furthermore, these results combined with the observed decreases in adipose tissue deposits and plasma NEFA levels support the contention of colestilan-induced enhanced TG effl ux from adipose tissue. Decreased plasma NEFA levels induced by colestilan treatment may constrain the energy metabolism toward higher catabolism of carbohydrates, as refl ected by the elevated RER values. In concert with the decreased visceral obesity, this may contribute to increased insulin sensitivity.
In conclusion, we demonstrated that colestilan decreased weight gain and improved peripheral insulin sensitivity in high-fat-fed male E3L mice by enhanced NEFA incorporation in biliary lipids and increased fecal lipid excretion. Our results provide a potential mechanistic basis for the effects of BASs in humans. Fig. 6. Effect of colestilan on gene expressions in liver, adipose tissue, and muscles involved in lipid and glucose metabolism. Mice received a high-fat diet or high-fat diet with colestilan for 8 weeks. Livers, adipose tissues, and muscles were collected after 4 h fasting, total RNA was extracted and gene expression analysis was performed using Affymetrix MOE430A arrays. Data represent fold change as compared with the control group. Values in bold are considered signifi cant ( P < 0.05). Red box indicates increase, blue box decrease. Only those genes that were signifi cantly affected by colestilan treatment in either one or more tissues are shown in this fi gure. For complete data, see supplementary