Muraglitazar, a Novel Dual (α/γ) Peroxisome Proliferator–Activated Receptor Activator, Improves Diabetes and Other Metabolic Abnormalities and Preserves β-Cell Function in db/db Mice

Muraglitazar and rosiglitazone were synthesized by BMS Medicinal Chemistry. Fenofibric acid was purchased from Sigma (St. Louis, MO).

db/db mice (C57BL/6ks lepr^−/−) and age-matched lean normal C57BL/6J or Swiss Webster mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and housed under controlled temperature (23°C) and lighting (12 h of light between 6 a.m. and 6 p.m.) with free access to water and standard mouse diet (18% protein rodent diet no. 2018; Harlan). The db/db mice were prebled, and those within a narrow range of fasted glucose levels were selected for studies to minimize variability between control and drug-treated groups. C57BL/6J mice on experimental diet were fed research diet no. 12327, which contains 40% sucrose/40% fat by calorie (Research Diets, New Brunswick, NJ) for 12 weeks before the start of the experiment and were maintained on this diet for the duration of the experiment. Mice were dosed daily by oral gavage in a vehicle composed of 20% polyethylene glycol (vol/vol), 5% N-methyl pyrrolidone, and 75% 10 mmol/l phosphate buffer, pH 7.4. Bristol-Myers Squibb study guidelines were strictly followed in the investigations.

Lean normal mice or db/db mice that were treated with vehicle or compounds, respectively, were killed, and their inguinal white adipose tissue (WAT) and liver were harvested. Total RNA was isolated from WAT or liver samples using RNeasy (Qiagen, Valencia, CA). For Northern blot analysis, 15 μg RNA was subjected to MOPS-formaldehyde gel electrophoresis. The gels were blotted to nylon membranes and hybridized to ³²P-cDNA probes according to standard procedures. The radioactivity in the hybridized bands was counted on an Instant Imager (Packard Instruments, Meridian, CT). Alternatively, SYBR-Green PCR analysis was carried out (Applied Biosystems, Foster City, CA). Oligonucleotide primers were designed using Primer Express, and RT-PCRs were carried out (primer sequences and protocol available upon request). The mRNA levels of target genes were normalized to control glyceraldehyde-3-phosphate dehydrogenase mRNA levels. WAT RNA samples from vehicle- and compound-treated mice were also analyzed by Affymetrix microarray for changes in gene expression pattern (protocol available upon request).

C57BL/6J mice that were treated with vehicle or compounds for 7 days were fasted overnight and intravenously injected with Triton-WR1339 (250 mg/kg) 1 h after the final dosing. The injection of Triton prevents the degradation of triglyceride-rich VLDL (triglyceride/VLDL) particles in plasma, resulting in an accumulation of triglycerides. The secretion rate (typically 0.16–0.20 mg · min⁻¹ · 100 g body wt⁻¹, linear for 5 h after Triton administration) was determined by calculating the amount of triglycerides accumulated 2.5 h after the Triton injection/100 g body wt. Triglyceride levels were determined using a Roche Cobas blood chemistry analyzer.

The db/db mice that were treated with vehicle or compounds for 14 days were killed, and their liver was harvested. Liver acyl coenzyme-A oxidase activity (ACO) activity [(slope of the rate of A₅₀₂ increase after addition of substrate − the slope of the background rate)/mg protein] was measured according to a published method (19).

About 50 μl tail vein blood from overnight-fasted or ad libitum–fed mice was collected in EDTA-coated tubes. Plasma glucose, triglyceride, free fatty acid (FFA), cholesterol, and HDL cholesterol levels were determined using a Roche Cobas blood chemistry analyzer; insulin, adiponectin, and corticosterone levels were determined by mouse enzyme-linked immunosorbent assay kits (Linco Research, St. Charles, MO). Corticosterone levels were assessed at 10 a.m. during regular 12-h diurnal light cycle. ED₅₀ to normalization is calculated as the midpoint of the dose-response activity curve using a four-parameter-fit equation.

The high–molecular weight (HMW), medium–molecular weight (MMW), and low–molecular weight (LMW) adiponectin complexes in db/db mouse plasma were detected according to the method described by Waki et al. (20). A total of 0.5 μl db/db mouse plasma samples were diluted (1:12) and incubated for 1 h at room temperature in reducing sample buffer (3% SDS, 50 mmol/l Tris-HCl, pH 6.8, 10% glycerol, 5% 2-mercaptoethanol, and 10 mmol/l dithiothreitol) or nonreducing sample buffer (3% SDS, 50 mmol/l Tris-HCl, pH 6.8, and 10% glycerol) and subjected to SDS-PAGE under reducing/heat-denaturing conditions (samples were heated at 95°C for 10 min) or nonreducing/nonheat-denaturing conditions, according to the standard Laemmli’s method with Criterion precast Tris-HCl 4–15% gel (Bio-Rad, Hercules, CA). For immunoblotting, proteins separated by SDS-PAGE were transferred to nitrocellulose membranes, blocked with StartingBlock (Tris-buffered saline) Blocking Buffer (Pierce, Rockford, IL), and then incubated with 1:5,000 diluted anti-mouse adiponectin globular domain monoclonal antibody (Chemicon, Temecula, CA) in Tris-buffered saline with 0.1% Tween-20 for 1 h at room temperature. After washing, the membranes were incubated with goat anti-mouse IRDye 800 (1:10,000) (Rockland, Gilbersville, PA) for 1 h at room temperature and then washed thoroughly. The membrane was scanned with the Odyssey Imaging System (Li-Cor, Lincoln, NE).

The db/db mice, which were on a 2-week dosing regimen, were fasted overnight on day 13. On day 14, an oral dose of vehicle alone or compound was given in the morning, and blood samples were collected from the tail vein for determination of baseline values (t = 0 min). The mice were then gavaged with an oral bolus of glucose (2 g/kg), and additional blood samples were collected at regular intervals (t = 15, 30, 60, and 90 min) for glucose and insulin measurement. Homeostasis model assessment index values were calculated using the following equation: (the product of the fasting insulin levels [mU/l] × fasting glucose levels [mmol/l]/22.5).

Pancreata were harvested from overnight-fasted vehicle- and drug-treated mice, placed in liquid N₂, then stored at −20°C. Pancreata were homogenized in acid-ethanol (75% ethanol, 23.5% water, and 1.5% c-HCl in 1.8 ml volume) with a polytron homogenizer. The homogenates were stored at 4°C for 28 h and then centrifuged at 1500g for 30 min at 4°C. The supernatants were diluted (1:20,000), and insulin levels were determined by enzyme-linked immunosorbent assay (21).

Liver triglyceride levels were determined using a Wako Kit (no. 997-69801). Frozen liver pieces were homogenized in saline and brought to a concentration of 0.05 mg/1 ml. Twenty microliters of the sample were solubilized with 20 μl deoxycholate (1.6% wt/vol in water), and 1 ml of the Wako reagent was added. The mixture was incubated at 37°C for 15 min, and the absorbance was read at 505 nm.

Unpaired, two-tailed Student’s t tests were performed for comparisons between compound-treated and vehicle control groups. Differences were considered significant at P < 0. 05.

As previously described, muraglitazar potently stimulates full-length human PPARγ- and PPARα-mediated reporter gene expression (EC₅₀ for PPARγ and PPARα transactivation = 0.11 and 0.32 μmol/l, respectively; 13) The ability of muraglitazar to transactivate full-length mouse PPARγ or PPARα receptor has not been determined. However, in a chimeric Gal4/mouse PPAR-mediated reporter gene assay, muraglitazar shows mouse PPARγ agonist activity at levels comparable with its human PPARγ activity (EC₅₀ for mouse PPARγ = 0.09 μmol/l for muraglitazar and 0.08 μmol/l for the PPARγ selective activator rosiglitazone; the PPARα selective activator fenofibric acid was inactive) and mouse PPARα agonist activity that is weaker than its human PPARα activity (observed EC₅₀ for mouse PPARα = 23.8 μmol/l for muraglitazar and 16.3 μmol/l for fenofibric acid; rosiglitazone was inactive). The disparity between the mouse and human PPARα activity is likely due to mouse/rodent-specific differences in the interactions of muraglitazar with several amino acid residues that are altered between mouse and human PPARα ligand binding domains (22).

The effects of muraglitazar treatment on the expression of PPAR target genes in WAT and liver were determined in db/db and normal mice. In db/db mice, muraglitazar treatment (10 mg · kg⁻¹ · day⁻¹ for 2 weeks) increases mRNA levels of fatty acid binding protein aP2, GLUT4 glucose transporter, and lipoprotein lipase in WAT and stimulates both mRNA and activity levels of ACO and suppresses apolipoprotein CIII mRNA levels in liver (Fig. 1A and B). Microarray analysis of WAT RNA from muraglitazar- or rosiglitazone (10 mg · kg⁻¹ · day⁻¹ for 7 days)-treated db/db mice shows that expression levels of genes that are implicated in 1) adipocyte differentiation, 2) insulin signaling and glucose metabolism, 3) fatty acid transport, 4) fatty acid oxidation, 5) triglyceride synthesis, and 6) energy expenditure are modulated by both muraglitazar and rosiglitazone treatment (Table 1). Both muraglitazar and rosiglitazone treatment (10 mg · kg⁻¹ · day⁻¹ for 3 days) stimulate aP2 and lipoprotein lipase mRNA levels and suppress 11β-hydroxy steroid desaturase 1 mRNA levels in normal mouse WAT (Fig. 1C). Muraglitazar, but not rosiglitazone, stimulates ACO mRNA levels in normal mouse liver (Fig. 1D). Finally, in normal mice, muraglitazar treatment (3, 10, and 30 mg · kg⁻¹ · day⁻¹ for 7 days) dose dependently inhibits triglyceride/VLDL secretion from the liver without promoting liver weight increase (Fig. 1E and F). Fenofibrate treatment (30, 50, and 100 mg · kg⁻¹ · day⁻¹) also inhibits triglyceride/VLDL secretion (Fig. 1E and F). However, this effect is accompanied by dose-dependent increases in liver weight, which is a known fibrate-induced phenomenon in rodents (23). Rosiglitazone treatment (3, 10, and 30 mg · kg⁻¹ · day⁻¹), by contrast, does not inhibit triglyceride/VLDL secretion (Fig. 1E and F). The gene expression data thus demonstrate that muraglitazar treatment results in modulation of PPAR target gene expression in WAT and liver. The PPARα agonist activity of muraglitazar may have contributed to the differences in the expression levels of various PPAR target genes in WAT and liver as well as inhibition of VLDL secretion in muraglitazar-treated compared with rosiglitazone-treated mice.

Muraglitazar was investigated in three separate studies for 1) dose-dependent lowering of fasted and fed glucose, insulin, FFA, triglyceride, and cholesterol levels in young hyperglycemic db/db mice (∼8-week-old males; 0.03–50 mg · kg⁻¹ · day⁻¹ orally for 2 weeks), 2) effect on time-dependent deterioration of glycemic control and plasma insulin levels in older db/db mice (∼12-week-old females; 30 mg · kg⁻¹ · day⁻¹ for 4 weeks), and 3) improvements in hyperglycemia and glucose tolerance and effect on pancreatic insulin content in severely hyperglycemic db/db mice (∼10-week-old females with fasting plasma glucose >500 mg/dl; 10 mg · kg⁻¹ · day⁻¹ for 2 weeks; rosiglitazone at 10 mg · kg⁻¹ · day⁻¹ was used as a positive control in the study).

In study 1, muraglitazar treatment results in dose-dependent lowering of both fasted (day 7 data shown, similar data were also obtained after 14 days) and fed (on day 15) plasma glucose, FFA, insulin, triglyceride, and cholesterol levels (Fig. 2A–E). Amelioration of hyperglycemia in the presence of reduced plasma insulin levels suggests that insulin sensitivity has been improved in muraglitazar-treated young db/db mice. As previously observed with PPARα activators in rodents, the cholesterol-lowering effect of muraglitazar is restricted to a reduction of the HDL cholesterol fraction (data not shown) (24). The ED₅₀ to normalization of glucose and triglyceride levels in fasted animals on day 14 are 0.1 and 0.2 mg · kg⁻¹ · day⁻¹, respectively, and in fed animals on day 15 are 0.5 and 1.3 mg · kg⁻¹ · day⁻¹, respectively. As observed with PPARγ activators (4), muraglitazar-treated mice (at 10 and 50 mg · kg⁻¹ · day⁻¹) experience a trend toward increased body weight gain in comparison with the vehicle-treated mice (Fig. 2F).

In study 2, the vehicle- and muraglitazar-treated db/db mice were monitored weekly for changes in fasting glucose and insulin levels. As shown in Fig. 3A and B, the vehicle-treated control db/db mice show poor glycemic control throughout the duration of the study. The vehicle-treated db/db mice also show signs of further deterioration of glycemic control (higher fasting plasma glucose levels) and some β-cell exhaustion (significant drop in plasma insulin levels) by the end of the 4-week treatment period (Fig. 3B and C). By contrast, the muraglitazar-treated mice show time-dependent improvement in glycemic control (fasting glucose levels are reduced to the levels observed in lean normal mice) and maintain reduced, but stable, plasma insulin levels during the entire 4-week period (Fig. 3A–C). Muraglitazar-treated mice also show significant improvements in other metabolic parameters such as fasted FFA (−49%), fed glucose (−60%), fasted triglyceride (−31%), and fed triglyceride (−47%) levels (Fig. 3D–F).

In study 3, at the end of the 2-week treatment period, the db/db mice were fasted overnight and, after collecting baseline (t = 0 min) plasma samples, were challenged with an oral bolus of glucose. Muraglitazar treatment results in significant reduction of baseline fasting plasma glucose (−51%), insulin (−55%), and FFA (−33%) levels as well as homeostasis model assessment index (−63%) (Table 2). When challenged with an oral bolus of glucose, muraglitazar-treated animals exhibit a reduced glucose excursion (indicating increased tolerance to glucose) and lower plasma insulin levels compared with vehicle-treated animals (Fig. 4A and B). The increased glucose tolerance, along with the concomitantly lowered insulin levels and reduced homeostasis model assessment index, indicate that insulin sensitivity has been improved in muraglitazar-treated mice. Rosiglitazone treatment also results in improved glycemic control; however, the effects are less pronounced than with muraglitazar at the same dose (Table 2, Figs. 4A and B). In addition to improvements in insulin sensitivity and glycemic control, pancreatic insulin content is increased by about fourfold by both muraglitazar and rosiglitazone treatment (Fig. 4C), which is suggestive of some preservation of β-cell function. Neither drug shows any significant impact on the weight of the pancreas in this study (Fig. 4D).

C57BL/6J mice, when maintained on a diet high in fat and sucrose (DIO-mice), develop mild hyperglycemia and high plasma triglyceride and cholesterol levels (18). Consistent with its antidiabetic and lipid-lowering effects in db/db mice, muraglitazar treatment (10 mg · kg⁻¹ · day⁻¹ for 2 weeks) lowers fasting glucose, triglyceride, and cholesterol levels of DIO-mice to the levels observed in mice on normal standard diet (Table 3).

As in some type 2 diabetic patients, db/db mice exhibit abnormally low plasma adiponectin levels and high plasma corticosterone (the murine counterpart to cortisol in humans) levels compared with age-matched normal C57BL/6J mice. Diminished adiponectin levels and elevated corticosterone levels serve as biomarkers of tissue insulin resistance and increased hepatic glucose production (20,25–30). Adiponectin exists as HMW, MMW, and LMW complexes in plasma (20,25,28). In patients and in animal models, increased levels of HMW adiponectin complex have been associated with improved insulin sensitivity (20,25,28).

In older hyperglycemic db/db mice, muraglitazar treatment (study 2) elevates plasma adiponectin levels and lowers plasma corticosterone levels to the levels observed in normal mice (Fig. 5A and B). In severely hyperglycemic db/db mice, muraglitazar treatment (study 3) elevates their plasma adiponectin levels above the levels observed in lean normal mice and significantly lowers plasma corticosterone levels (Fig. 5C and D). By comparison, rosiglitazone treatment elevates adiponectin levels to the levels observed in normal mice and lowers corticosterone to the levels comparable with muraglitazar-treated levels in this study (Fig. 5C and D). Furthermore, immunoblot analysis shows that in muraglitazar-treated db/db mice (10 mg · kg⁻¹ · day⁻¹ for 2 weeks), their plasma total adiponectin levels as well as HMW adiponectin complex levels are substantially increased compared with the vehicle-treated mice (Fig. 5E and F).

Obese patients with insulin resistance and type 2 diabetes frequently suffer from nonalcoholic fatty liver condition (31,32). Elevated lipid content in the liver has been implicated in hepatic insulin resistance, glucose overproduction, and increased VLDL synthesis and secretion (30,31). The db/db mice on a normal diet accumulate lipids (primarily triglycerides) in the liver and develop hepatic steatosis. In these mice, muraglitazar treatment (50 mg · kg⁻¹ · day⁻¹ for 2 weeks) results in significant reductions of liver triglycerides content (76 ± 3 mg/g liver tissue in muraglitazar-treated vs. 100 ± 10 mg/g liver tissue in vehicle-treated mice).

Muraglitazar is a novel dual (α/γ) PPAR activator that selectively binds to and activates human PPARγ and human PPARα (13,33–37). The in vivo pharmacological data in lean normal mice and in db/db mice demonstrate that muraglitazar modulates the expression of PPAR target genes implicated in the regulation of glucose and lipid metabolic pathways in WAT and in liver. The in vivo data also demonstrate that muraglitazar is a potent and efficacious antidiabetic and lipid-lowering agent in db/db mice. In young hyperglycemic db/db mice, muraglitazar lowers both fasted and fed glucose and triglyceride levels to the levels commonly observed in lean normal mice. In addition, muraglitazar treatment reduces fasted and fed insulin, FFA, and cholesterol levels. In older db/db mice, longer-term muraglitazar treatment prevents time-dependent deterioration of glycemic control and development of insulin deficiency. In severely hyperglycemic db/db mice, muraglitazar treatment markedly reduces fasted plasma glucose and insulin levels as well as glucose excursion. In these animals, muraglitazar also increases the insulin content in the pancreas. Muraglitazar treatment elevates the low plasma adiponectin levels, increases the HMW adiponectin complex levels, and reduces the elevated plasma corticosterone levels of db/db mice. Muraglitazar treatment also significantly lowers triglyceride content in db/db mouse liver. Finally, in DIO-mice, muraglitazar treatment normalizes diet-induced mild hyperglycemia and hyperlipidemia, which corroborates the glucose and lipid-lowering effects in db/db mice. Muraglitazar treatment did not cause hypoglycemia in mice under the experimental conditions used.

The antidiabetic and lipid-lowering effects induced by muraglitazar treatment may result from one or more of the following PPAR-mediated mechanisms: 1) improved insulin action and enhanced glucose uptake in adipose tissue and skeletal muscle, 2) increased fatty acid uptake and storage in adipose tissue, 3) reduced plasma FFA levels, 4) increased plasma total adiponectin and HMW adiponectin complex levels, 5) suppression of glucose overproduction by liver, 6) enhanced VLDL catabolism in the plasma, and 7) reduced triglycerides/VLDL synthesis/secretion in the liver (1–4,25,38–39). The HDL cholesterol lowering in muraglitazar-treated mice is most likely the result of a rodent-specific PPARα-mediated mechanism that suppresses the production of apolipoprotein A1 (a major protein component of HDL particles) in the liver (24). In humans, muraglitazar, like other human PPARα activators (e.g., fenofibrate, gemfibrozil), has demonstrated plasma HDL cholesterol–raising effects (7,8,40–42).

The trend toward increased weight gain in muraglitazar-treated db/db mice is probably due to a combination of effects including 1) enhanced adipogenesis, 2) retention of calories that would otherwise be lost due to glucosuria, and 3) water retention due to the alleviation of the glucose-driven osmotic diuresis and/or increased plasma or extracellular fluid volume (43,44). The liver triglyceride-lowering effect of muraglitazar is possibly due to reduced plasma FFA and lipid levels, which would limit fatty acid substrate availability for lipid biosynthesis in the liver. Reduced lipid content in the liver will lower hepatic insulin resistance, glucose overproduction, and increased VLDL synthesis (31,32).

In muraglitazar-treated db/db mice, the improvement in insulin sensitivity and the concomitant reduction in plasma glucose and FFA levels are anticipated to 1) lower insulin secretory demand on β-cells and 2) prevent apoptosis of β-cells, respectively. These effects may help to prevent deterioration of β-cell function, loss of pancreatic insulin content, development of insulin deficiency, and deterioration of glycemic control in muraglitazar-treated db/db mice.

The increase in both total adiponectin levels and HMW adiponectin complex levels are expected to stimulate fatty acid oxidation in liver and skeletal muscle as well as enhance insulin sensitivity and glucose uptake in skeletal muscle (20,25–28). Interestingly, an inverse correlation has been recently described between plasma adiponectin levels and the rate of incidence of myocardial infarction in men irrespective of their glycemic status (47). The reduction of corticosterone levels by muraglitazar is likely due to reduced metabolic stress and/or PPAR-mediated suppression of the 11β-hydroxy steroid desaturase 1 gene expression in WAT and liver. The reduced corticosterone levels may suppress hepatic glucose overproduction and enhance glucose uptake in peripheral tissues (29,30).

In conclusion, the novel dual (α/γ) PPAR activator muraglitazar 1) exerts potent and efficacious insulin-sensitizing antidiabetic effects, 2) prevents time-dependent deterioration of glycemic control and development of insulin deficiency, 3) increases pancreatic insulin content and, 4) improves other metabolic abnormalities such as hyperlipidemia, fatty liver, low adiponectin levels, and high corticosterone levels in db/db mice. In the clinical setting, muraglitazar treatment (0.5–20 mg for 28 days) lowers glucose, insulin, triglyceride, FFA, and apolipoprotein CIII levels and increases HDL cholesterol levels in type 2 diabetic patients (40–42). These clinical data emphasize the utility of db/db mice as a useful model for evaluating antidiabetic properties and antidyslipidemic properties of novel PPAR activators. However, the weak mouse PPARα activity, relative to its human PPARα activity, for muraglitazar may suggest that its antidyslipidemic effects are probably underrepresented in db/db mice. Finally, since type 2 diabetes patients often suffer from dyslipidemia and other metabolic abnormalities, thus putting them at high risk for cardiovascular disease, muraglitazar, with its dual (α/γ) PPAR activity, has the potential to be useful for the treatment of these patients.

We thank Drs. S. Taylor, R. Parker, J. Whaley, R. Mukherjee, J. Graham, and R. Belder for critically reading the manuscript.