Role of bile acid sequestrants in the treatment of type 2 diabetes

Abstract

Cholestyramine is a first-generation bile acid sequestrant (BAS) and antihyperlipidemic agent that currently has limited use because of its relatively weak effect on lowering low density-lipoprotein (LDL)-cholesterol (C) and poor tolerability. The current first choice drugs for hyper-LDL-cholesterolemia are 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (statins) because of their strong LDL-C lowering effects and efficacy in prevention of cardiovascular disease. However, after lowering the target levels of LDL-C in very high risk patients, combination therapy with statins and other antihyperlipidemic drugs may become more important for treatment of hyper-LDL-cholesterolemia. Second-generation BASs such as colesevelam and colestimide have a glucose-lowering effect and improved tolerance, which has led to re-evaluation of their utility in combination with statins or antidiabetic agents.

Key Words: Bile acid; Bile acid sequestrant; Type 2 diabetes

INTRODUCTION

Bile acid sequestrants (BASs) were one of the first classes of drugs to show that cholesterol-lowering therapy decreases the risk of cardiovascular disease (CAD)[1,2]. However, use of first-generation BASs such as cholestyramine and colestipol has been limited by poor tolerability and a relatively weak effect on lowering of low-density lipoprotein cholesterol (LDL-C)[1,2]. Currently, 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors or statins are the first choice for treatment of hyper-LDL-cholesterolemia based on their stronger LDL-C lowering effect and prevention of cardiovascular events[3-6]. However, co-administration of BASs with statins may produce lower LDL-C levels[7,8]. Second-generation BASs such as colesevelam (used clinically in the USA since 2000[9-11]) and colestimide (also called colestilan, and used clinically in Japan since 1999[12]) have improved tolerability. BASs also have a glucose-lowering effect[9-16], and are currently being re-evaluated for their potential use in combination with statins or antidiabetic agents.

MECHANISMS UNDERLYING THE LDL-C LOWERING EFFECT OF BILE ACID SEQUESTRANTS

Biliary excretion of cholesterol as a component of bile is an important excretion pathway for hepatic cholesterol. Conversion of cholesterol to bile acids in the liver and excretion into the intestine via the biliary duct and gall bladder also facilitates excretion of cholesterol. Over 95% of bile acids excreted in bile from the gall bladder are reabsorbed in the terminal ileum and transferred to the liver via the portal vein in a recycling pathway. BASs, which are not themselves absorbed from the gut, absorb bile acids in the intestine and inhibit enterohepatic circulation of bile acids by preventing their reabsorption. This causes a significant increase of bile acids bound to BASs in feces. The decrease in bile acids transferred to the liver via the portal vein leads to upregulation of hepatic cholesterol cytochrome P450 7 alpha1 (CYP7A1), the rate-limiting enzyme for conversion of cholesterol to bile acids, promoting compensatory conversion and, thereby resulting in a decrease of intrahepatic cholesterol. In turn, this activates the hepatic LDL receptor, which then binds circulating LDL-C and results in a decrease in the level of circulating LDL-C[17]. BASs also inhibit cholesterol absorption by preventing formation of micelles composed of bile acids in the intestinal lumen, which may also contribute to the LDL-C lowering effect.

MECHANISMS UNDERLYING THE GLUCOSE-LOWERING EFFECT OF BILE ACID SEQUESTRANTS

Many clinical studies have shown that BASs improve glycemic control in patients with type 2 diabetes[9-16]. The mechanisms underlying this effect remain unclear, but several have been proposed. Bile acids such as cholic acid (CA) and chenodeoxycholic acid (CDCA) are natural ligands for the farnesoid X receptor (FXR)[18-19], and activation of FXR in liver may increase the production of small heterodimer partner (SHP)[20], a protein that plays a central role in lipid and glucose metabolism via regulation of various downstream molecules[21,22]. The increase in SHP due to FXR activation increases glucose metabolism by inhibiting production of phosphoenolpyruvate carboxykinase (PEPCK)[22], an enzyme associated with gluconeogenesis (although conversely FXR activation has been shown to increase PEPCK activity and glucose levels[23]). FXR activation also represses glucose levels in a diabetic rat model[24]. Bile acids can also increase glucose metabolism by regulating energy homeostasis via activation of the G protein-coupled receptor 5 (TGR5)-cAMP-type 2 iodothyronine deiodinase (D2) pathway in brown adipose tissues or skeletal muscles independently of FXR[25]. These observations suggest that BASs might worsen glycemic control because of potential deactivation of hepatic FXR through a decrease of bile acids in liver, in contrast to the established beneficial effects of BASs for glucose metabolism. However, there is a report showing that BAS treatment does not change the level of total bile acids in serum, but increases the absolute level of CA as well as the CA level relative to total bile acids[26]. The relative increase in circulating CA may itself influence glucose metabolism through a decrease of glucose levels via the TGR5-cAMP-D2 pathway. However, even if this mechanism occurs it is unlikely to improve overall glycemic control because the level of CDCA relative to total bile acids may be decreased by BAS treatment, and CDCA has similar effects on TGR5 to those of CA[25].

The most plausible mechanism for the glucose-lowering effect of BASs may be associated with effects on the liver X receptor (LXR), as proposed by Bay et al[27]. LXR is a nuclear transcription factor that mainly regulates lipid metabolism, and its natural ligands are oxysterols such as 22(R)-hydroxycholesterol, 24(S)-hydroxycholesterol, and 27-hydroxycholesterol[28,29]. Reduction of bile acid flux in the portal vein by BAS treatment decreases FXR activity in liver, and this decreased FXR activity may induce an increase of LXR activity due to decreased SHP production[20]. LXR activation in liver results in improved glucose sensitivity by preventing gluconeogenesis based on inhibition of the activity of PEPCK and G6Pase[30]. LXR activation may also improve glucose metabolism by promoting expression of glucokinase and glucose transporter 4 (GLUT4) in adipocytes[31] or by promoting insulin secretion in β cells in the pancreas[32]. However, LXR activation may simultaneously increase circulating triglyceride levels by promoting production of stimulating sterol regulatory element-binding protein 1c (SREBP1c) in liver[33]. On the other hand, decreased FXR activity in liver may itself promote PEPCK or SREBP1c production via reduction of SHP[19,22]. Therefore, FXR deactivation and resulting LXR activation may have competitive effects on molecules such as PEPCK. However, it is likely that regulation of glucose or lipid metabolism by LXR activation probably overcomes the effects of FXR deactivation[34]. These mechanisms suggest that in liver both activation of FXR by FXR agonists such as CA and CDCA, and deactivation of FXR by BASs improve glycemic control, although these changes have opposite effects on TG production via SREBP1c (decreased by FXR agonists and increased by BASs).

BAS-induced secretion of glucagon-like peptide-1 (GLP-1) in the ileum may be another important mechanism. Bile acids can increase GLP-1 secretion in L cells in the intestine via TGR5[19]. However, a recent report[35] showed that colesevelam also induced the release of GLP-1 and improved plasma glucose levels and insulin resistance in the diet-induced obesity (F-DIO) rat fed a high fat/high sucrose (HF) diet. In contrast, administration of SC-435, an inhibitor of the apical sodium-dependent bile acid transporter (ASBT), did not change GLP-1, glucose levels, and insulin sensitivity in F-DIO rats fed the HF diet, compared to untreated F-DIO rats fed the same diet[35]. Addition of both colesevelam and SC-435 reduced the total concentration of bile acids in portal blood and increased CYP7A1 mRNA expression in liver, which reflects FXR deactivation[35]. These findings suggest that colesevelam can improve glycemic control by increasing GLP-1 levels in the circulation independently of an effect on FXR in liver. A clinical study also showed that colestimide decreased postprandial plasma glucose levels, but increased GLP-1 in patients with type 2 diabetes[36]. It is clear that the mechanisms underlying the glucose-lowering effect of BASs are complicated and other mechanisms may also be involved. The possible mechanisms are summarized in Figure 1.

Figure 1 Possible mechanisms of the effects of bile acids and bile acid sequestrants on glucose and lipid metabolism. Bile acids activate farnesoid X receptor (FXR) activity in liver, which leads to increased small heterodimer partner (SHP) production that inhibits phosphoenolpyruvate carboxykinase (PEPCK) and therefore inhibits gluconeogenesis in liver. Bile acids also promote energy metabolism in brown adipose tissue and skeletal muscle via the G protein-coupled receptor 5 (TGR5) activation, and increase glucagon-like peptide-1 (GLP-1) secretion via TGR5 activation in L cells in the intestine. On the other hand, bile acid sequestrants (BASs) may decrease FXR activity in liver, and the resulting decrease in SHP production may cause activation of liver X receptor (LXR) activity. LXR activation inhibits PEPCK and therefore inhibits gluconeogenesis. BASs can also increase circulating GLP-1 levels. These mechanisms may explain the glucose-lowering effects of bile acids and BASs. Because bile acids decrease sterol regulatory element-binding protein 1c (SREBP1c) production due to increased SHP, bile acids decrease triglyceride (TG) levels. In contrast, indirect LXR activation by BASs can increase SREBP1c production and circulating TG levels. In the figure, solid and dotted lines respectively show promoting and inhibitory effects. CA: cholic acid; CDCA: chenodeoxycholic acid.

EFFECTS OF BASS ON GLUCOSE AND LIPID METABOLISM IN PATIENTS WITH TYPE 2 DIABETES

The glucose-lowering effect of first-generation BASs such as cholestyramine was shown in the 1990s[13], but there was little subsequent interest in this effect for several years. In a small randomized, double-blind, crossover trial in 21 patients (20 men and 1 women) with type 2 diabetes complicated with dyslipidemia treated with glibenclamide or insulin, cholestyramine administered at 16 g/d for 6 mo significantly decreased the mean plasma glucose and LDL-C levels by 13% and 28%, respectively, compared with placebo. A decrease of glycated hemoglobin from 8.8% to 8.3% (not significant) also occurred. The glucose lowering-effect of second-generation BASs such as colesevelam and colestimide has attracted more attention. A pilot study of the effect of colesevelam on glucose levels (Glucose-Lowering effect of WelChol Study: GLOWS) showed that addition of colesevelam at 3.75 g/d in 65 patients with type 2 diabetes that was not fully controlled with sulfonylurea or metformin, alone or in combination significantly decreased HbA_1c by 0.5% from a baseline level of 7.9% and LDL-C by 11.7% compared with placebo after 12 wk[14]. Interestingly, the extent of the reduction was greater in patients with HbA_1c≥ 8.0% (1.0% reduction vs placebo). Based on the positive results for glycemic control in the GLOWS study, three phase III randomized, double-blind placebo-controlled trials with a 2 wk single blind placebo run-in were performed, in which colesevelam was added to sulfonylurea-, metformin-, and insulin-based therapy, respectively[9-11].

In a 26 wk study in 461 patients with type 2 diabetes treated with sulfonylurea as monotherapy or in combination with other oral antidiabetes drugs, colesevelam (3.75 g/d) significantly reduced HbA_1c (baseline 8.2%) by 0.54% in all patients and by 0.79% in a subgroup treated with sulfonylurea monotherapy, and reduced fasting plasma glucose (FPG) by 13.5 mg/dL in all patients, compared with placebo[9]. The reduction of HbA_1c in a subgroup with HbA_1c > 8.0% at baseline was 0.58% vs placebo. Colesevelam also significantly reduced LDL-C by 16.7% compared with placebo, while an insignificant elevation of HDL-C (+0.1%) and a significant elevation of triglyceride (TG) (+17.7%) were observed. In a 26 wk study in 316 patients with type 2 diabetes treated with metformin as monotherapy or in combination with other antidiabetes drugs, colesevelam (3.75 g/d) significantly decreased HbA_1c (baseline 8.2%) by 0.54% in all patients and by 0.47% in a subgroup treated with metformin monotherapy, and reduced FPG by 13.9 mg/dL in all patients, compared with placebo[10]. The reduction of HbA_1c in a subgroup with HbA_1c > 8.0% at baseline was 0.60% vs placebo. LDL-C was significantly reduced by 15.7%, with insignificant increases in HDL-C (+0.9%) and TG (+4.7%). In a 16 wk study in 287 patients with type 2 diabetes treated with insulin as monotherapy or in combination with other antidiabetes drugs, colesevelam (3.75 g/d) significantly decreased HbA_1c (baseline 8.3%) by 0.5% in all patients and by 0.59% in a subgroup treated with insulin monotherapy, compared with placebo[11]. The reduction of HbA_1c in a subgroup with HbA_1c > 8.0% at baseline was 0.57% vs placebo, with an insignificant reduction of FPG (-14.6%). LDL-C was significantly reduced (-12.8%), TG was significantly increased (+21.5%), and HDL-C showed a small and insignificant decrease (-0.9%).

The mean percentage compliance in the three studies with sulfonylurea, metformin, and insulin was high: 93.3%, 92.3%, and 92.7% in the colesevelam group, with similar rates in the placebo group. The results of these studies suggest that colesevelam can reduce HbA_1c levels by approximately 0.5% in all type 2 diabetic patients when added to sulfonylurea-, metformin-, or insulin-based therapy. The effect on glycemic control of colesevelam as monotherapy or in combination with other antidiabetes drugs in patients with type 2 diabetes is less clear. However, given the possible effect of BASs on GLP-1, as described above, it will be interesting to investigate the effects of coadministration of colesevelam with a dipeptidyl peptidase IV (DPPIV) inhibitor, which improves glycemic control by preventing degradation of circulating GLP-1.

An effect of colestimide on glycemic control in patients with type 2 diabetes has also been reported. Yamakawa et al randomly assigned patients with type 2 diabetes complicated by hyperlipidemia to colestimide (3.0 g/d) or pravastatin (10 mg/d) treatment groups, and investigated the effect of these drugs on lipid and glucose metabolism[15]. In both groups, 33 of 35 patients received monotherapy with oral antidiabetic drugs or insulin, or combination therapy with these agents. Colestimide and pravastatin significantly decreased LDL-C levels by 23% and 17%, respectively, but only colestimide produced a significant decrease in HbA_1c (0.9% reduction from a baseline level of 7.7%) after 3 mo of therapy. In our study of colestimide (3.0 g/d; n = 20) compared to rosuvastatin (2.5 mg/d; n = 20) in patients with type 2 diabetes complicated with hyper-LDL-cholesterolemia, HbA_1c decreased by approximately 0.6% from a baseline level of 7.9%, with a treatment difference of 0.5% between colestimide and rosuvastatin after treatment for 3 mo[16]. LDL-C was decreased by rosuvastatin and colestimide by 39% and 14%, respectively. All except 3 patients in the colestimide group had already taken other antidiabetic drugs. Kondo and Kadowaki also recently reported an effect of high dose colestimide therapy (4.5 g/d) on glycemic control in a randomized double-blind placebo-controlled study (n = 183 at the start of randomization)[12]. Colestimide was generally administered as monotherapy, except in a few patients. After 3 mo, a 0.9% reduction of HbA_1c and a decrease in FPG of 22 mg/dL were observed in the colestimide group (n = 86) compared with the placebo group (n = 86). In subgroups of patients with HbA_1c 8.0 to < 9.0% and ≥ 9.0%, colestimide decreased HbA_1c by 1.0% and 1.5%, respectively, compared to placebo. Based on these reports, the HbA_1c-lowering effect of colestimide appears to be somewhat stronger than that of colesevelam. It is important to note that the colestimide trials were all performed in Japan, and that the clinical characteristics of the patients differed from those in trials of colesevelam. A trial with direct comparison of the effects of colesevelam and colestimide on glycemic control has yet to be performed. Co-administration of colestimide with DPPIV inhibitors has also not been studied. Studies showing glucose-lowing effect of bile acid sequestrants are summarized in Table 1.

Table 1 Studies showing glucose-lowing effect of bile acid sequestrants.

	n	Duration (wk)	Therapy	Baseline (g/d)	HbA1c (%)	ΔHbA1c	ΔLDL-C	Others
Garg and Grundy	21	6	cholestyramine	16.00	no description	-0.50%a	-28.00%
Zeive et al	65	12	colesevelam	3.75	7.90%	-0.50%	-11.70%
Fonseca et al	461	26	colesevelam	3.75	8.20%	-0.54%	-16.70%	ΔCRP-11.2%
Bay et al	316	26	colesevelam	3.75	8.20%	-0.54%	-15.90%	ΔCRP-14.4%c
Goldberg et al	287	16	colesevelam	3.00	8.30%	-0.50%	-12.80%	ΔCRP-12.2%
Yamakawa et al	70	12	colestimide	3.00	7.70%	-0.90%b	-23.00%
Takebayashi et al	40	12	colestimide	3.00	7.90%	-0.50%	-14.00%	Δ8-isoPGFα-32%c
Kondo and Kadowaki	183	12	colestimide	4.50	8.00%	-0.90%	-22.50%

ΔHbA_1c (change of hemoglobin A_1c) is shown as the difference between BAS and placebo except for b.

^aGlycated hemoglobin;

^bChange from baseline. ΔLDL-C: change of low density-lipoprotein cholesterol; ΔCRP: change of C reactive protein by treatment (^cstatistical significance); Δ8-isoPGFα: change of urinary 8-iso-prostaglandin F2α (a marker of systemic oxidative stress) by treatment (^cstatistical significance).

SAFETY, TOLERABILITY, AND DRUG INTERACTIONS IN BAS TREATMENT

As mentioned above, cholestyramine, a first-generation BAS, was shown to lower LDL-C before the clinical use of statins, and was one of the first drugs to show that lowering cholesterol could decrease the risk of CAD[1,2]. However, after the appearance of statins, use of cholestyramine has been limited because of its relatively weak LDL-C lowering effect compared with statins, and because of poor compliance due to a high frequency of side effects, high dosage, requirement for suspension in water, and unpleasant taste[1,2]. Colestipol, another first-generation BAS, was better tolerated by patients than cholestyramine, but its effects were still not satisfactory[37]. The main side effect of BASs is gastrointestinal symptoms including dyspepsia, nausea, and particularly constipation, while systemic severe side effects are rare. Regarding the mechanism of constipation induced by BAS, CDCA in feces promotes secretion of water and electrolytes into the intestine by activating adenyl cyclase and increasing intracellular cAMP in colonic mucosa cells[38,39]. BAS absorbs bile acids, which causes a decrease in the level of intratubular bile acids and this may induce constipation.

Colesevelam and colestimide are second-generation BASs with reduced side effects, including constipation. For WelChol^® (colesevelam hydrochloride)[40], the frequency of adverse reactions is relatively low, with 11.0%, 8.3%, and 4.2% of patients developing constipation, dyspepsia, and nausea, respectively, compared to 7.0%, 3.5%, and 3.9%, respectively, with placebo. Myalgia was found in 2.1% of patients treated with colesevelam compared to 0.4% with placebo. The frequency of constipation with colestimide has been reported to be 3.6%[41]. Both colesevelam and colestimide can be administered as tablets (the latter is also formulated as a granulated powder). The tablet size is still somewhat large, but the drug compliance is better than that for cholestyramine. In addition, BASs can be used safely in children, in pregnant women, and in patients with liver and renal disease because they are not absorbed systemically.

Drug interactions are also an important concern in BAS administration. Cholestyramine has many drug interactions since it increases the absorption of common drugs including digoxin, diuretics, estrogens, hydrocortisone, propranolol, thyroxine, and warfarin, and may also interfere with fat-soluble vitamins such as vitamins A, D, E and K[17]. Therefore, it is recommended that other drugs are taken at least 1 h before or 4 h after cholestyramine treatment. In contrast, colesevelam does not influence the bioavailability of digoxin, fenofibrate, lovastatin, metoprolol, quinidine, valproic acid, pioglitazone, and warfarin[40]. Drugs with a known interaction with colesevelam include glyburide, levothyroxine, and oral contraceptives containing ethinyl estradiol and norethindrone, and it is recommended that these drugs should be taken at least 1 hour prior to colesevelam administration[40]. In summary, severe side effects of BASs are rare, and second-generation BASs have improved tolerability and reduced drug interactions compared to first-generation BASs.

BASS IN THE TREATMENT OF TYPE 2 DIABETES: CURRENT ROLE AND FUTURE PERSPECTIVES

The main role of BASs in treatment of diabetes appears to be as second-line drugs in combination with other antidiabetic agents. The glucose-lowering effect of BASs is moderate to mild, and BAS monotherapy for diabetes has only been examined in one study of colestimide[12]. In fact, colesevelam is only currently approved as adjunct therapy for glycemic control in type 2 diabetes in the USA (since 2008), and colestimide has yet to be approved for treatment of type 2 diabetes alone. As discussed above, BASs can decrease HbA_1c by 0.5% to 0.9%[9-16], and the glucose-lowering effect of BASs may be stronger in patients with higher baseline HbA_1c levels[9-12,14,16]. BASs rarely cause body weight gain or increase hypoglycemia[9-11], and second-generation BASs have improved drug compliance due to reduced side effects of constipation or dyspepsia[9-11,41]. Systemic severe side effects of BASs are rare. These characteristics support the utility of BASs as additional drugs for treatment of diabetes.

It is apparent that BASs are more suitable for treatment of patients with type 2 diabetes complicated with hyper-LDL cholesterolemia, rather than patients with type 2 diabetes alone, because there is evidence that reduction of LDL-C by cholestyramine decreases the risk of CAD[1,2]. However, statins are now established as the first choice treatment for hyper-LDL cholesterolemia due to their strong LDL-C lowering effect and prevention of cardiovascular events[3-6]. The beneficial effect of rosuvastatin for prevention of cardiovascular events even extends to apparently healthy men and women with baseline LDL-C levels < 130 mg/dL, but high-sensitivity C reactive protein (CRP) of ≥ 2 mg/L[5]. A beneficial effect of atorvastatin for prevention of cardiovascular events has also been shown in patients with type 2 diabetes with relatively low LDL-C levels (≤ 160 mg/dL)[4].

The target LDL-C level in very high-risk patients (those with cardiovascular disease with diabetes, cigarette smoking or factors associated with metabolic syndrome) has recently been lowered to ≤ 70 mg/dL[42]. This suggests that combination therapy will become more important in patients who cannot achieve the target levels, even with high dose statins, or cannot tolerate high dose statins because of side effects such as myalgia. In addition, a recent meta-analysis suggested that most statins, including rosuvastatin and atorvastatin, are weakly but significantly associated with new onset of type 2 diabetes[43]. This potentially deleterious effect of statins on glucose metabolism may become more apparent when they are used at a high dose[44]. Therefore, combination therapy of statins with other antihyperlipidemic drugs may be appropriate, and addition of BASs to statin therapy may be suitable, especially for type 2 diabetes with hyper-LDL-cholesterolemia. It should be noted that it is still unclear whether statins worsen glycemic control after onset of type 2 diabetes. Regarding the effects of BASs, colesevelam monotherapy can decrease LDL-C by 15% to 21%[17,40,45,46] and by a further 10% in combination with statins[8]. However, care is required with use of BASs in patients with high TG levels because of a potential TG elevation effect[47]. Anti-inflammatory and anti-oxidative stress effects of BASs have also been reported in patients with type 2 diabetes[8,16].

The above findings suggest that BASs are especially suitable for patients with type 2 diabetes complicated by hyper-LDL cholesterolemia, since these patients often fail to achieve target levels of HbA_1c and LDL-C with use of other antidiabetes drugs and statins. However, there is still no evidence to show that addition of second-generation BAS such as colesevelam and colestimide reduces the risk of cardiovascular events in these patients.

CONCLUSION

Second-generation BASs such as colesevelam and colestimide are generally well tolerated and severe systemic side effects are rare. When BAS is coadministered with antidiabetes drugs such as sulfonylurea, metformin and insulin, a reduction in HbA_1c of approximately 0.5% to 0.9% can be expected without increased hypoglycemia or weight gain. The LDL-C lowering effect of BASs is relatively mild, but coadministration of a BAS with statins is likely to produce a further decrease of LDL-C. BAS treatment may be especially beneficial for patients who have not reached target HbA_1c and LDL-C levels using other antidiabetic drugs and statins. Further studies are required to determine whether addition of a BAS in these patients will reduce mortality and the risk of cardiovascular events.