Effects of the New Dual PPARα/δ Agonist GFT505 on Lipid and Glucose Homeostasis in Abdominally Obese Patients With Combined Dyslipidemia or Impaired Glucose Metabolism
OBJECTIVE We evaluated the metabolic effects and tolerability of GFT505, a novel dual peroxisome proliferator–activated receptor α/δ agonist, in abdominally obese patients with either combined dyslipidemia or prediabetes.
RESEARCH DESIGN AND METHODS The S1 study was conducted in 94 patients with combined dyslipidemia while the S2 study was conducted in 47 patients with prediabetes. Participants were randomly assigned in a double-blind manner to GFT505 at 80 mg/day or placebo for 28 (S1) or 35 (S2) days. Primary efficacy end points were changes from baseline at week 4 in both fasting plasma triglycerides and HDL cholesterol in the S1 group and 2-h glucose upon oral glucose tolerance test in the S2 group.
RESULTS In comparison with placebo, GFT505 significantly reduced fasting plasma triglycerides (S1: least squares means −16.7% [95% one-sided CI −∞ to −5.3], P = 0.005; S2: −24.8% [−∞ to −10.5], P = 0.0003) and increased HDL cholesterol (S1: 7.8% [3.0 to ∞], P = 0.004; S2: 9.3% [1.7 to ∞], P = 0.009) in both studies, whereas LDL cholesterol only decreased in S2 (−11.0% [ −∞ to −3.5], P = 0.002). In S2, GFT505 did not reduce 2-h glucose (−0.52 mmol/L [−∞ to 0.61], P = 0.18) but led to a significant decrease of homeostasis model assessment of insulin resistance (−31.4% [−∞ to 12.5], P = 0.001), fasting plasma glucose (−0.37 mmol/L [−∞ to −0.10], P = 0.01) and fructosamine (−3.6% [−∞ to −0.20], P = 0.02). GFT505 also reduced γ glutamyl transferase levels in both studies (S1: −19.9% [−∞ to −12.8], P < 0.0001; S2: −15.1% [−∞ to −1.1], P = 0.004). No specific adverse safety signals were reported during the studies.
CONCLUSIONS GFT505 may be considered a new drug candidate for the treatment of lipid and glucose disorders associated with the metabolic syndrome.
Abdominal obesity is often associated with a clustering of cardiovascular risk factors including combined dyslipidemia (i.e., high plasma triglycerides and low HDL cholesterol concentrations), hypertension, and impaired glucose tolerance, a condition often referred to as the metabolic syndrome. Metabolic syndrome predisposes to type 2 diabetes and increases the risk of cardiovascular disease (1). Insulin resistance is the suspected underlying molecular mechanism sustaining these metabolic abnormalities. Thus, novel therapies improving insulin sensitivity should be considered, even though the regulatory agencies do not yet consider the metabolic syndrome an approved indication for drug therapy.
Peroxisome proliferator–activated receptors (PPARs) are fatty acid–activated nuclear receptors that regulate an array of physiological processes (2). The PPAR nuclear receptor subfamily is composed of three members: PPARα, PPARγ, and PPARδ (also known as PPARβ). Agonists of two of these receptors are currently used therapeutically, with the hypolipidemic fibrate drugs acting as PPARα agonists and the insulin-sensitizing thiazolidinediones acting as PPARγ agonists. Currently, there are no clinically used drugs that target PPARδ. PPARδ is widely expressed and plays a critical role in mitochondrial function, muscle development, fatty acid oxidation, and insulin sensitivity (3–7). Two-week clinical studies performed in healthy volunteers (8) and moderately overweight subjects (9) showed that the synthetic PPARδ agonist GW501516 reduces fasting plasma triglyceride and increases HDL cholesterol concentrations, indicating that the targeting of PPARδ may have beneficial effects in metabolic disorders.
GFT505 and its main active circulating metabolite, GFT1007, are PPAR modulators with preferential activity on PPARα (half-maximal effective concentration: 10–20 nmol/L) and additional activity on PPARδ (half-maximal effective concentration: 100–150 nmol/L). Both GFT505 and GFT1007 undergo extensive enterohepatic cycling and are liver targeted. Phase I studies in healthy volunteers showed that GFT505 at doses from 40 to 100 mg/day induces a dose-dependent reduction in plasma triglyceride and an increase in HDL cholesterol (10).
To further verify the potential metabolic benefits of GFT505, we conducted two phase IIa studies in patients with either combined dyslipidemia or prediabetes.
RESEARCH DESIGN AND METHODS
Two prospective, randomized, parallel-group, double-blind, placebo-controlled studies were conducted with GFT505 (2-[2,6 dimethyl-4-[3-[4-(methylthio)phenyl]-3-oxo-1(E)-propenyl]phenoxyl]-2-methylpropanoic acid) in patients aged 18–75 years. In clinical study S1 (clinical trial reg. no. NCT01271751, clinicaltrials.gov), patients with combined dyslipidemia (1.69 ≤ triglyceride ≤ 6.78 mmol/L, HDL cholesterol ≤1.03 mmol/L for male subjects and HDL cholesterol ≤1.16 mmol/L for female subjects) and abdominal obesity (waist circumference ≥102 cm for male subjects and ≥88 cm for female subjects) were recruited from January 2009 to September 2009 at 36 centers in France, 6 in Romania, and 5 in Tunisia. Participants did not have previous cardiovascular disease and either were not hypertensive or were maintained at a stable dose of antihypertensive medication for at least 2 months prior to screening. Main exclusion criteria were BMI ≥40 kg/m2, >5% weight change variation within 6 months prior to screening, blood pressure >160/95 mmHg, type 1 or type 2 diabetes, known familial hypercholesterolemia or type III hyperlipoproteinemia, creatinine clearance according to the Cockroft-Gault formula <60 mL/mn, and severe hepatic dysfunction (alanine aminotransferase [ALT], aspartate aminotransferase [AST], γ glutamyl transferase [gGT] >3 N, or alkaline phosphatase [ALP] >2 N). Lipid-lowering therapy was not permitted during the study. Patients were randomly assigned in a 2:1 design to once-daily 80 mg GFT505 (four capsules before breakfast) or placebo for 4 weeks. The primary end points were mean triglyceride and HDL cholesterol change from baseline for GFT505 versus placebo after 4 weeks. The study was considered positive if both primary end points were met. Secondary end points were comparisons of GFT505 versus placebo change from baseline for additional plasma lipid parameters, glucose homeostasis parameters, inflammatory markers, and safety parameters.
In clinical study S2 (clinical trial reg. no. NCT01275469, clinicaltrials.gov), subjects with prediabetes (fasting plasma glucose [FPG] between 6.1 and 7 mmol/L, 2-h glycemia upon 75-g oral glucose tolerance test [OGTT] ≥7.8 mmol/L within 7 days before randomization) and abdominal obesity (waist circumference ≥94 cm for male subjects and ≥80 cm for female subjects) were recruited from June 2009 to December 2009 at 20 centers in France. Main exclusion criteria were identical as for S1, except for lipid parameters: triglyceride >4.57 mmol/L and/or LDL cholesterol >5.68 mmol/L. Patients using oral antidiabetic drugs within the 3 months before randomization were excluded from the study. Patients were randomly assigned equally (1:1) to once-daily GFT505 80 mg or placebo for 5 weeks. The primary end point was mean 2-h plasma glucose upon OGTT change from baseline for GFT505 versus placebo after 4 weeks. Secondary end points were comparisons of GFT505 versus placebo change from baseline for FPG, fasting plasma insulin, homeostasis model assessment of insulin resistance (HOMA-IR), lipids, physical exercise test parameters, and inflammatory markers.
The studies were conducted in accordance with the ethical guidelines of the Declaration of Helsinki and were approved by institutional review boards/independent ethics committees at participating sites. Patients provided written informed consent before enrollment.
Study visits with plasma sampling occurred at screening before the first administration, at days 0, 14, and 28 after the first administration, and 14 days after the end of the treatment period at days 42 (for S1) and 49 (for S2). In S2, a visit for exercise tests (Vo2max, respiratory quotient, and heart rate and power at maximal exercise) was scheduled after 35 days of treatment. In S2, OGTTs were performed at days 0 and 28, with blood samples withdrawn at 0, 30, 60, 120, and 180 min after an oral glucose challenge (75 g). All plasma samples were centrally dosed for efficacy and safety markers (Eurofins Medinet, Plaisir, France). Vital signs, physical examination, and adverse event assessment were performed at each visit. Electrocardiograms were performed at randomization and final visits.
For both S1 and S2, statistical analyses were conducted on an intent-to-treat basis, including all selected patients who had taken at least one unit of the study drug (GFT505 or placebo) with at least one baseline and postbaseline evaluation. For S1, a total sample size of 90 individuals (60 GFT505 and 30 placebo) was required to provide 80% power to detect mean 20 and 12% differences for triglyceride and HDL cholesterol, respectively, between GFT505 and placebo, assuming an SD of 35% for triglyceride and 20% for HDL cholesterol. For S2, a sample size of 26 individuals in each group was required to provide 80% power to detect a mean 20 mg/dL difference in 2-h glucose upon OGTT, assuming an SD of 25 mg/dL. To test the superiority of GFT505 over the placebo, tests were one sided. The α risk was set to 0.05. Because the S1 study was considered positive only if both primary end points were met, the α risk was not adjusted for multiplicity. Comparisons between GFT505 and placebo were computed using relative change of efficacy parameters as the explained variable, group as the explaining variable, and baseline values as the covariate. Least squares means were computed. Effect size was computed as difference between least squares means for the comparison of GFT505 versus placebo and expressed as least squares means ± SE along with 95% one-tailed CI. To overcome the nonnormal distribution of some efficacy parameters, we confirmed results of the global generalized linear model with Wilcoxon test. Analyses were performed using the SAS software (version 9.1; SAS Institute, Cary, NC).
In the S1 and S2 studies, 94 and 47 subjects, respectively, were randomly assigned to receive 80 mg/day GFT505 or placebo (Supplementary Fig. 1). All patients were analyzed on an intention-to-treat basis; 81 of 94 patients in S1 and all randomized patients in S2 completed the study without any major deviation. For each study, baseline demographics and biochemical characteristics were similar among the GFT505 and placebo groups (Table 1).
After 28-days’ treatment with GFT505, several lipid parameters improved significantly from baseline. Compared with placebo (Fig. 1A and B and Tables 2 and 3), GFT505 reduced fasting plasma triglyceride levels, with a relative effect size versus placebo of −16.7% (95% one-sided CI −∞ to −5.3 [S1; P = 0.005]) and −24.8% (−∞ to −10.5 [S2; P = 0.0006]). GFT505 also increased HDL cholesterol, with a relative effect size versus placebo of 7.8% (3–∞ [S1; P = 0.004]) and 9.3% (1.7–∞ [S2; P = 0.009]). In S2 only, significant reductions in total cholesterol (effect size −8.7% [−∞ to −3.7]; P = 0.0005), non-HDL cholesterol (−13.3% [−∞ to −6.9]; P = 0.0001), and LDL cholesterol (−11% [−∞ to −3.5]; P = 0.002) were obtained in response to GFT505 treatment. In both studies, the increase in HDL cholesterol in the GFT505 group was paralleled by an increase in apolipoproteins AI (S1 effect size 5.6% [2.4–∞], P = 0.002; S2 3.3% [−1.9 to ∞], nonsignificant) and AII (S1 15.5% [12.8–∞]; S2 17.6% [10.9–∞], all P < 0.0001). Apolipoprotein B (apoB) was significantly reduced following GFT505 treatment in both studies (S1 −6.6% [−∞ to −1.4], P = 0.03; S2 −14.0% [−∞ to −8.3], P < 0.0001). Consistent with the hypotriglyceridemic effect of GFT505, apoCIII, an inhibitor of lipoprotein lipase activity (11), decreased (S1 −17.9% [−∞ to −7.5], P = 0.003; S2 −20.2% [−∞ to −6.7], P = 0.0004) following GFT505 administration. The hypolipidemic effects of GFT505 were also observed after 14 days of treatment (Tables 2 and 3). In both studies, all lipid parameters returned to baseline values 2 weeks after treatment cessation (data not shown).
To specifically assess the effect of GFT505 on glucose homeostasis, the S2 study was conducted in patients with prediabetes (Fig. 1C and D and Table 3). Whereas GFT505 did not significantly reduce 2-h plasma glucose levels after the OGTT (effect size 0.52 mmol/L [95% one-sided CI −∞ to 0.61], P = 0.18), small but significant reductions in FPG (−0.37 mmol/L [−∞ to −0.10], P = 0.01) and fructosamine (−9.46 μmol/L [ −∞ to −0.60], P = 0.02) were observed in the GFT505 group. GFT505 seemed to exert an insulin-sensitizing effect, with a significant decrease in fasting insulinemia (effect size vs. placebo from baseline to end point −24.8% [−∞ to −6.5], P = 0.005) and HOMA-IR index (effect size −31.4% [−∞ to −12.5], P = 0.001). The concentration of circulating adiponectin, a PPARγ target gene (12), was not altered following GFT505 treatment. In S1, glucose homeostasis parameters were not modified by GFT505 treatment (Table 2). There were no effects of GFT505 on exercise test parameters in S2 (data not shown).
Liver function and inflammatory markers
Levels of gGT, a marker of liver dysfunction, were significantly reduced in GFT505 versus placebo patients (S1 effect size −19.9% [95% one-sided CI −∞ to −12.8], P < 0.0001; S2 −15.1% [−∞ to −1.1], P = 0.004). In addition, GFT505 significantly decreased ALT levels in S1 (−14.7% [−∞ to −3.2], P = 0.02), with no effect on AST concentrations. A consistent and highly significant reduction of ALP levels was observed following GFT505 treatment in both studies (S1 −16.4% [−∞ to −12.4]; S2 −24.5% [−∞ to −19.1], all P < 0.0001) (Tables 2 and 3 and Fig. 1D). GFT505 reduced both haptoglobin (S1 −11.7% [−∞ to −3.7], P = 0.008; S2 −15.8% [−∞ to −3.1], P = 0.008) and fibrinogen (S1 −8.8% [−∞ to −3.6], P = 0.003; S2 −10.0% [−∞ to −1.1], P = 0.01) levels, with no effect on high-sensitivity C-reactive protein concentrations.
GFT505 showed a good tolerance profile. For S1, 28 adverse events were reported by 23 patients: 13 of 63 in the GFT505 group and 10 of 31 in the placebo group. Among these adverse events, two were judged as possibly related to GFT505: one case of mild gastroenteritis and one case of moderate headache. No adverse events led to the discontinuation of GFT505, and no serious adverse events related to GFT505 were reported. For S2, 20 adverse events were reported by 11 patients: 7 of 23 in the GFT505 group and 4 of 24 in the placebo group. Six of these adverse events were considered to possibly be related to GFT505: atrial fibrillation, which was diagnosed at the week 4 visit in a 75-year-old male patient with a long history of hypertension (n = 1), back pain (n = 1), diarrhea (n = 1), gastritis (n = 1), flatulence (n = 1), and gastrointestinal hypermotility (n = 1). No abnormal evolutions of laboratory values were observed from baseline to end point. All safety parameters including ionogram, hematological parameters, creatine phosphokinase, and homocysteine (Tables 2 and 3) concentrations remained steady throughout the study without difference between treatment groups. However, a slight and reversible increase of plasma creatinine was observed in the S2 GFT505 group (5 ± 9.5 μmol/L vs. baseline; P = 0.02, t test) (Tables 2 and 3).
These two phase IIa studies provide the first clinical evidence that the use of the dual PPARα/δ agonist GFT505 improves multiple metabolic parameters in abdominally obese patients with either combined dyslipidemia or prediabetes.
After 1 month of oral treatment at a dose of 80 mg/day, GFT505 significantly improved lipid homeostasis. In both studies, GFT505 reduced plasma triglyceride concentrations (by 16.7% in S1 and 24.8% in S2) and increased HDL cholesterol (by 7.8% in S1 and 9.3% in S2). These studies confirmed the results from a previous phase I study demonstrating that GFT505 (80 mg/day) leads to a 30% reduction of plasma triglyceride and a 12% increase of HDL cholesterol in healthy volunteers (10). From a mechanistic point of view, these effects correlated with a parallel decrease of plasma ApoCIII, an inhibitor of lipoprotein lipase activity (11), and apoB, whereas GFT505 increased plasma concentrations of apolipoproteins associated with HDL cholesterol particles, apoAI and apoAII. All of these effects are similar to those of fibrates, suggesting PPARα-mediated effects (13–15). An additional role for PPARδ activation in the lipid effects of GFT505 cannot be excluded given that a 2-week treatment of moderately obese patients with a pure PPARδ synthetic agonist, GW501516, led to similar effects on plasma triglyceride and HDL cholesterol (9). Furthermore, GW501516 increases HDL cholesterol plasma concentrations and stimulates reverse cholesterol transport in obese monkeys (16).
GFT505 treatment also improved insulin sensitivity in patients with both impaired fasting glucose and impaired glucose tolerance, with a 30% decrease of HOMA-IR. In addition, slight but significant reductions of FPG and fructosamine were observed in the GFT505 group. In a comparable population of patients with impaired glucose tolerance, 10-week treatment with metformin or pioglitazone led to a quantitatively similar decrease in fasting glucose levels (~6 mg/dL), as observed in S2 with GFT505 (17). In addition, metformin improves insulin sensitivity without significantly improving 2-h glucose levels during OGTT (17). This insulin-sensitizing effect of GFT505 was probably linked to PPARδ activation for the following reasons: 1) PPARα activation by fenofibrate does not improve hepatic and peripheral insulin sensitivity in glucose clamp studies (18,19); 2) similar to GFT505, the pure PPARδ agonist GW501516 improves insulin resistance and glucose tolerance in mouse models of diabetes (4); 3) 2-week treatment with GW501516 decreases fasting plasma insulin and HOMA-IR in moderately obese men (9); and 4) there is not an increase in plasma adiponectin levels, a well-known marker of PPARγ activity. A phase II clinical trial is currently recruiting to confirm the hypoglycemic potential of GFT505 in drug-naïve patients with type 2 diabetes (clinical trial reg. no. NCT01261494, clinicaltrials.gov). In addition, the insulin-sensitizing action of GFT505 is being assessed using the hyperinsulinemic-euglycemic clamp method in patients with the metabolic syndrome (clinical trial reg. no. NCT01271777, clinicaltrials.gov).
The cardiovascular benefit of fibrate therapy remains intensively debated. Whereas the combination of fenofibrate with simvastatin failed to reduce major cardiovascular events in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study (20), results from a large meta-analysis suggested that fibrates may reduce the risk of coronary events in individuals exhibiting high cardiovascular risk and combined dyslipidemia (21). Thus, the cardiovascular consequences of the GFT505-mediated changes in lipid and glucose metabolism should be assessed in a dedicated cardiovascular outcome study.
Significant improvement in markers of liver dysfunction was observed in GFT505-treated patients, including decreases in gGT, ALT, and ALP levels. One hypothesis is that PPARδ activation may exert a beneficial effect against the development of nonalcoholic fatty liver disease (NAFLD), as suggested by studies in mice (22). Accordingly, GW501516 has been shown to simultaneously reduce gGT and liver fat content by 20% in overweight subjects (9). In contrast, the PPARα agonist clofibrate does not exert clinical benefit in patients with NAFLD (23).
Though no serious adverse events were reported, the short-term treatment period of both studies (maximum 35 days) is a limitation in assessment of the long-term safety of GFT505. Notably, GFT505 did not increase plasma homocysteine concentrations. As previously reported with fenofibrate therapy, a reversible and moderate increase in plasma creatinine levels was observed in the GFT505 group. However, recent prespecified analyses from the Fenofibrate Intervention and Event Lowering (FIELD) study indicate that this effect is rapidly reversible and suggest that fenofibrate may delay albuminuria and glomerular filtration rate impairment in patients with type 2 diabetes (24).
In conclusion, the current results position the dual PPARα/δ agonist GFT505 as a new drug candidate to improve multiple features of the metabolic syndrome, including combined dyslipidemia, type 2 diabetes, and NAFLD. The short treatment period and the use of a single dose of GFT505 are the major limitations of the present proof-of-concept studies. The efficacy and safety of GFT505 remain to be confirmed over a longer period in a dose range–finding study.
This study was funded by Genfit, Loos, France. B.C. has acted as an investigator for Boehringer Ingelheim, Eli Lilly & Co, Genfit, Johnson & Johnson, Novo Nordisk, and Roche and has received fees for consultancy, speaking, travel, or accommodation from AstraZeneca, Eli Lilly & Co., Pierre Fabre, Genfit, GlaxoSmithKline, Novartis, Novo Nordisk, Merck, sanofi-aventis, Servier, and Takeda. Y.Z. has received fees for speaking, travel, or accommodation and has acted as an investigator for AstraZeneca, Boehringer Ingelheim, Genfit, GlaxoSmithKline, Merck, Novo Nordisk, and sanofi-aventis. B.S. has received consultancy fees from Genfit and Merck. E.B. has received fees for presentations and participation to boards and has acted as an investigator for Abbott, AstraZeneca, Genfit, Merck, Novo Nordisk, Pfizer, sanofi-aventis, and Schering-Plough. No other potential conflicts of interest relevant to this article were reported.
B.C. participated in the study as an investigator, contributed to discussion, wrote the first draft, and reviewed and edited the manuscript. Y.Z. participated in the study as an investigator and reviewed and edited the manuscript. B.S. designed the study, contributed to discussion, and reviewed and edited the manuscript. E.B. participated in the study as the principal investigator, contributed to discussion, and reviewed and edited the manuscript.
Parts of this study were presented in poster form at both the 70th Scientific Sessions of the American Diabetes Association, Orlando, Florida, 25–29 June 2010 (abstract no. 677 and late-breaking abstract no. 12), and the 46th Annual Meeting of the European Association for the Study of Diabetes, Stockholm, Sweden, 20–24 September 2010 (abstract no. 880).
The authors acknowledge Pierre Clerson (Orgamétrie) for his advice on statistical analysis and Lesley Millat (Genfit) for her careful rereading of the manuscript. The authors also thank Rémy Hanf (Genfit) for his support in conducting the trials. The authors thank the investigators and contributors from each of the study sites.
This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc11-0093/-/DC1.
- Received January 14, 2011.
- Accepted May 26, 2011.
- © 2011 by the American Diabetes Association.
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.