Therapeutic Roles of Peroxisome Proliferator–Activated Receptor Agonists

Fibrates were introduced for treatment of hyperlipidemia. Trials with fibrates have shown a reduction in coronary heart disease (CHD) risk through modification of atherogenic dyslipidemia. Fibrates are now recommended for the treatment of patients with dyslipidemia who are at high risk of CHD (2).

PPARα potentiates fatty acid oxidation in the liver, heart, kidney, and skeletal muscle. Activation of PPARα leads to an increase in expression of lipoprotein lipase and apolipoprotein A-V (apoA-V) and to a decrease in hepatic apoC-III. These actions lower plasma triglycerides in chylomicrons and VLDL particles, thus liberating fatty acids, which are taken up and stored as fat in adipocytes or metabolized in skeletal muscle (3). In addition, PPARα activation increases hepatic apoA-I and -II expression, which raises HDL cholesterol levels, and promotes HDL-mediated cholesterol efflux from macrophages by inducing ATP-binding cassette A1 transporter (4). In addition to the efficacy of fibrates in the clinical management of atherogenic dyslipidemias involving reduced HDL cholesterol and elevated triacylglycerol-rich lipoprotein levels, they are effective in shifting the LDL subclass distribution toward larger particle species (5). Evidence of their beneficial effect on CVD has been obtained from several large clinical trials (6–11) (Table 1).

The Bezafibrate Coronary Atherosclerosis Intervention Trial (BECAIT) (6) investigated the effect of bezafibrate and dietary intervention on atherosclerosis in survivors of myocardial infarction (MI) who had dyslipidemia (predominantly hypertriglyceridemia). Bezafibrate slowed the progression of focal coronary atherosclerosis to a degree comparable to that achieved with statins and reduced the incidence of coronary events in young postinfarct patients. It also led to improvements in lipid profile (Table 1).

The Bezafibrate Infarction Prevention (BIP) study (7) investigated the effect of bezafibrate on fatal/nonfatal MI or sudden death in patients with stable angina or previous MI and high total cholesterol. The probability of an event was not significantly reduced, but post hoc analysis of a subgroup of patients with high baseline triglycerides revealed a significant reduction in the cumulative risk of coronary events. Bezafibrate treatment increased HDL cholesterol and lowered triglyceride levels. It should be noted that bezafibrate is an unusual member of the fibrate class in that although it acts primarily as a PPARα agonist, it also has some effect on PPARδ (12).

The Helsinki Heart Study (HHS) (8) assessed the reduction in CHD risk produced by elevating HDL cholesterol and lowering non-HDL cholesterol concentrations with gemfibrozil in asymptomatic middle-aged men with primary dyslipidemia. Marked improvements in lipid profile in the gemfibrozil group compared with placebo were associated with a significant reduction in the incidence of CHD.

The Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (VA-HIT) (9) examined the effect of gemfibrozil on nonfatal MI and coronary death in men with CHD and low HDL cholesterol. Gemfibrozil significantly reduced the risk of an event compared with placebo. The risk of stroke was also significantly reduced compared with placebo. Although LDL cholesterol levels did not differ significantly in the two groups, the reductions in coronary events and death with gemfibrozil treatment were accompanied by an increase in HDL cholesterol and reductions in total cholesterol and triglyceride levels. Although the increase in HDL cholesterol was associated with reduced risk, it did not account for all of the risk reduction, suggesting that gemfibrozil may have pleiotropic effects.

The complementary lipid-altering effects of fibrates have led to their increased use as combination therapy. However, a statin/fibrate, particularly gemfibrozil, combination may be associated with increased risk of rhabdomyolysis. In vitro studies have highlighted that gemfibrozil may increase the myotoxic effects of statins (13,14) and that use of gemfibrozil results in higher plasma concentrations of statins (15–18). Possible mechanisms for this include inhibition of statin glucuronidation (13), which may then reduce statin elimination in vivo. Inhibition of the P450 enzyme CYP2C8 by gemfibrozil (13,19) may also impede elimination of statins (13). In contrast to the findings for gemfibrozil, the use of fenofibrate in combination with statins is associated with myotoxic effects to a lesser extent. During a survey in 1998–2002 on total reports of rhabdomyolysis in patients on statin/fibrate combination therapy, fenofibrate/cerivastatin was associated with a small proportion of cases (2.3%), whereas gemfibrozil/cerivastatin accounted for the majority of reports (88%) (20). Thus, it should be recommended that gemfibrozil not be used in combination with the current generation of statins, while combinations of other fibrates with statins should be used with caution. Furthermore, it is important that the effects of new PPAR agonists on statin glucuronidation and cytochrome P450 metabolism be investigated.

Overall, studies indicate that by increasing the level of HDL cholesterol and lowering plasma triglycerides, fibrates can reduce the incidence of CHD. Currently, evidence that fibrates affect clinical outcome in patients with vulnerable plaques is predominantly associated with gemfibrozil. However, the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) trial, an ongoing prospective 5-year study in 9,795 patients with type 2 diabetes, will provide evidence as to whether fibrates are effective for primary prevention of CVD.

Patients with type 2 diabetes are at particularly high risk of atherosclerotic events. Post hoc analyses of completed trials such as the VA-HIT and the HHS suggest that patients with diabetes derive greater benefits from fibrate therapy than nondiabetic subjects (21–25). Two studies that have prospectively investigated the impact of fibrate-mediated lipoprotein modification on CVD risk in patients with type 2 diabetes, the Diabetes Atherosclerosis Intervention Study (DAIS) and the St. Mary’s, Ealing, Northwick Park Diabetes Cardiovascular Disease Prevention (SENDCAP) study, support the benefit of PPARα therapy in the management of type 2 diabetes.

The DAIS (10) was designed to assess the effects of correcting lipoprotein abnormalities using fenofibrate on coronary atherosclerosis. Patients with good glycemic control, mild lipoprotein abnormalities typical of type 2 diabetes, and at least one visible coronary lesion were randomized to micronized fenofibrate or placebo. The increase in percentage diameter stenosis and the decrease in minimum lumen diameter were significantly smaller in the fenofibrate group than in the placebo group. The reduction in angiographic progression of coronary artery disease was related, at least in part, to significant changes in total cholesterol, HDL cholesterol, LDL cholesterol, and triglyceride concentrations, even in individuals previously judged not to need treatment.

The SENDCAP study (11) was also conducted to determine whether intervention with a fibrate (combined with conventional management) could improve cardiovascular outcomes in patients with type 2 diabetes. Patients with no history of clinical CVD were randomized to bezafibrate or placebo. Compared with placebo, those treated with bezafibrate had significantly greater reductions in median serum triglycerides, total cholesterol, and total-to-HDL cholesterol ratio, and they had significantly greater increases in HDL cholesterol. Although there was no significant difference between groups in the progression of ultrasonically measured arterial disease, those treated with bezafibrate experienced a significant reduction in the combined incidence of probable ischemic change on resting electrocardiogram and documented MI.

Insulin resistance occurs long before the clinical onset of diabetes. Onset of insulin resistance is often accompanied by obesity, in particular visceral obesity. Resistance of dysfunctional fat cells to the antilipolytic effects of insulin leads to chronic elevations in plasma free fatty acid (FFA) levels. This, in turn, induces insulin resistance in the liver and skeletal muscle, resulting in reduced glucose uptake and increased gluconeogenesis. Dysfunctional fat cells also produce excessive amounts of cytokines (e.g., tumor necrosis factor-α [TNF-α], interleukin [IL]-6, and resistin) that further induce insulin resistance, inflammation, and atherosclerosis and that secrete reduced amounts of insulin-sensitizing cytokines such as adiponectin. The obesity that characterizes this pathological state is also frequently associated with nonalcoholic steatohepatitis, a progressive liver condition that carries a risk for fibrosis, cirrhosis, and end-stage liver disease and that is augmented by insulin resistance (26).

Synthetic ligands for PPARγ are of particular interest for treating patients with type 2 diabetes because they restore sensitivity to insulin. This property has been attributed to the direct effects of PPAR agonists on lipid metabolism in adipose tissue and to secondary effects on lipid and glucose metabolism in liver and skeletal muscle (27,28). PPARγ agonists promote adipocyte differentiation, and they promote FFA uptake and storage in subcutaneous adipose rather than visceral adipose tissue. This reduces FFA levels, with associated reductions in insulin resistance. In addition, activation of PPARγ is believed to increase the expression and translocation to the cell surface of the glucose transporters GLUT1 and -4, thus increasing glucose uptake into liver and skeletal muscle cells and reducing plasma glucose levels (29). PPARγ agonists may also restore insulin sensitivity by decreasing TNF-α (30) and increasing adiponectin expression (31).

The value of PPARγ agonists in the management of insulin resistance and type 2 diabetes is demonstrated by a number of important clinical trials with thiazolidinediones (32–37) (Table 2). Although troglitazone has now been withdrawn from the market because of rare but serious cases of hepatotoxicity, studies with pioglitazone and rosiglitazone indicate that this is not a class effect. Class effects associated with these agents include weight gain, edema (38), and increased lipoprotein(a) concentrations (39,40). Because of concerns about fluid retention, thiazolidinediones are not recommended for use in patients with New York Health Association class III or IV cardiac status (38). Despite these warnings, thiazolidinediones are increasingly prescribed to patients with diabetes and heart failure. A recent cohort study suggested that thiazolidinediones are not associated with risk of death from heart failure in patients with diabetes and heart failure (41). However, such patients should be monitored closely. The results of cardiovascular outcome studies with pioglitazone and rosiglitazone will establish the long-term benefits and risks of these agents.

The Troglitazone Study Group carried out a dose-ranging study of troglitazone in patients with type 2 diabetes (32). Levels of HbA_1c (A1C) fasting plasma glucose (FPG), plasma nonesterified fatty acids, triglycerides, and fasting plasma insulin were significantly lower in troglitazone-treated patients than in control subjects, and the troglitazone-treated group demonstrated increased insulin sensitivity. HDL cholesterol concentrations increased with higher doses of troglitazone (>600 mg/day).

The Troglitazone in the Prevention of Diabetes (TRIPOD) study found that troglitazone improved insulin sensitivity in two-thirds of Hispanic women with previous gestational diabetes. Troglitazone also provided protection against development of type 2 diabetes (33), thought to be associated with preservation of pancreatic β-cell function. Recent results of the Diabetes Prevention Program also showed that troglitazone reduced the incidence of diabetes, even though it was given for a limited period of time because of its discontinuation (42).

In patients with type 2 diabetes (34), pioglitazone decreased A1C and FPG levels, without changing fasting or glucose-stimulated insulin/C-peptide concentrations. Fasting plasma FFA and triglyceride concentrations were also significantly reduced. During an insulin clamp, pioglitazone significantly reduced basal endogenous glucose production, whereas insulin-stimulated total and nonoxidative glucose disposal during a second insulin clamp were significantly increased (34). Pioglitazone-induced improvements in hepatic insulin sensitivity are associated with reductions in hepatic fat content (43).

Several larger studies have confirmed the effects of pioglitazone on blood glucose and lipid profiles in patients with type 2 diabetes. In the placebo-controlled Pioglitazone 026 study, pioglitazone significantly reduced A1C, FPG, fasting insulin, C-peptide, and triglyceride concentrations, and it increased HDL cholesterol. Total cholesterol and LDL cholesterol levels were unaffected (35). Furthermore, a recent trial specifically investigating the lipid-lowering effects of thiazolidinediones found that pioglitazone was associated with significant improvements in triglycerides, HDL cholesterol, non-HDL cholesterol, and LDL particle size compared with rosiglitazone (44). The ongoing Pioglitazone Prevention of Type 2 Diabetes (PIPOD) trial (an extension of TRIPOD) is investigating whether 4 years of pioglitazone treatment offers continued protection from diabetes in women who had normal/impaired glucose tolerance or “mild diabetes” at the end of TRIPOD.

In a 3-month study in patients with type 2 diabetes, rosiglitazone improved insulin-stimulated glucose metabolism during low- and high-dosage insulin clamps and led to reductions in plasma FFA concentration and hepatic triglyceride level (36). These changes were associated with increases in extramyocellular lipid content and in the sensitivity of peripheral adipocytes to insulin. The increase in insulin sensitivity elicited by rosiglitazone is associated with improvements in the clinical and histological severity of nonalcoholic steatohepatitis (26).

In a larger study, rosiglitazone significantly reduced FPG, fructosamine, FFAs, and plasma insulin from baseline, and it increased HDL cholesterol levels. Although the total-to-HDL cholesterol ratio and triglyceride level did not change, rosiglitazone treatment increased total cholesterol and LDL cholesterol (37).

As for pioglitazone, the effect of rosiglitazone on progression of type 2 diabetes is under investigation in a number of trials. One of these, the Diabetes Reduction Approaches with Ramipril and Rosiglitazone Medications (DREAM) trial, is currently assessing whether 3 years of rosiglitazone, alone or in combination with ramipril, delays or prevents the progression of type 2 diabetes in subjects with impaired glucose tolerance.

In patients with diabetes, PPAR agonists have demonstrated several other potential benefits, in addition to their effects on insulin sensitivity and glucose homeostasis.

Over 70% of patients with type 2 diabetes have some degree of dyslipidemia (45), typically increased triglycerides and reduced HDL cholesterol levels. The U.K. Prospective Diabetes Study (UKPDS) demonstrated that lipid abnormalities were more important risk factors for CHD than A1C (Fig. 1) (46), a vital consideration given that up to 80% of patients with type 2 diabetes die from CVD complications (47). The ability of selective PPARα agonists, such as fibrates, to improve these abnormal lipid profiles is well established (48). There is also evidence that PPARγ agonists can impact positively on diabetic dyslipidemia, although currently available thiazolidinediones appear to have differential effects: whereas pioglitazone increases HDL cholesterol and lowers triglycerides (49), rosiglitazone only provides improvements in HDL cholesterol levels. Indeed, rosiglitazone has been shown to increase total cholesterol and LDL cholesterol (50). Some evidence suggests that pioglitazone lowers triglycerides by enhancing the triglyceride fractional catabolic rate rather than decreasing hepatic production (51). In addition, thiazolidinedione treatment improves postprandial lipid metabolism. Rosiglitazone, troglitazone, and pioglitazone all induce a shift in LDL phenotype from dense to large buoyant subfractions (52–54).

Type 2 diabetes is also characterized by elevated levels of inflammatory markers, such as C-reactive protein (CRP), that are also closely linked to CVD risk. Thus, increasing interest is developing in the anti-inflammatory properties of PPARα and -γ agonists, and this is discussed in more detail below.

Ongoing clinical trials are investigating the ability of dual PPARα/γ agonists to provide type 2 diabetic patients with the collective benefits of effective sustained glycemic control, lipid management, and anti-inflammatory effects. Preliminary data from animal models of obesity and diabetes are encouraging (55), and early clinical trials have shown that these agents improve both the hyperglycemia and the dyslipidemia that are characteristic of type 2 diabetes (56). Because of caution surrounding possible carcinogenicity in animals, regulatory authorities have requested 2-year toxicology data before approving initiation of clinical trials of ≥6 months. Although there is currently no evidence of carcinogenicity in humans, and the cancer risk may be species-specific, it is important to establish the safety of PPAR agonists. Some dual PPARα/γ agonists, such as ragaglitazar and MK-0767, have been withdrawn from clinical trials. However, a number of other agents are undergoing preclinical evaluation, whereas tesaglitazar (Galida) and muraglitazar have completed the 2-year toxicity studies and are currently in clinical trials to clarify their efficacy and safety.

The metabolic syndrome describes a complex of metabolic abnormalities, including obesity, diabetes, hypertension, and dyslipidemia, to which insulin resistance is central. The prevalence of this syndrome (currently 24% in the U.S.) is increasing across the globe (57). A substudy of the Botnia trial demonstrated that the effect of clustering of metabolic syndrome components on the risk of CHD morbidity in individuals with the metabolic syndrome appears greater than the relative risk associated with the individual components (58).

The data discussed in this review outline a compelling rationale for PPARα and -γ agonists as ideal agents for managing the metabolic syndrome. Thiazolidinediones improve glycemic control via their actions on insulin sensitivity, whereas fibrates improve a range of atherogenic dyslipidemias, including the dyslipidemic profile commonly associated with the metabolic syndrome. These agents also influence many other components of metabolic syndrome, including hypertension, inflammation, and vascular dysfunction and remodeling. If dual PPARα/γ agonists can bring together the beneficial effects of PPARα and -γ agonists, prospects for the management of patients with the metabolic syndrome will look brighter.

Although dyslipidemia is a central causative factor of atherosclerosis, processes such as inflammation, vascular dysfunction, and vascular remodeling are also involved. The pleiotropic effects of PPAR agonists, therefore, take on extra significance in the management of CVD.

Fibrates are known to reduce cardiovascular risk (Fig. 2) (11,25). If successful, the FIELD trial, discussed earlier, will confirm that these PPARα agonists should be used to protect type 2 diabetic patients against heart disease.

Studies to determine whether the effects of thiazolidinediones translate into long-term cardiovascular benefits are also underway. These include two randomized trials: Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROACTIVE) and Rosiglitazone Evaluated for Cardiovascular Outcomes and Regulation of Glycaemia in Diabetes (RECORD). The PROACTIVE is a double-blind, placebo-controlled trial that is assessing whether pioglitazone reduces total mortality and macrovascular morbidity in high-risk patients with type 2 diabetes. The RECORD study is evaluating the impact of open-label rosiglitazone on cardiovascular events and risk. In addition, the PPARγ Agonist for the Prevention of Late Adverse Events Following Percutaneous Coronary Revascularisation (PPAR) trial is assessing whether rosiglitazone reduces mortality, MI, and target vessel revascularization in obese individuals with hypertension or type 2 diabetes.

PPARα activation can inhibit inflammatory response genes by repressing nuclear factor-κB, signal transducer and activator of transcription, and activator protein-1 signaling. Downregulation of activator protein-1 leads to inhibition of thrombin-induced production of endothelin-1 in the arterial endothelium (59), decreased production of the inflammatory mediators IL-6 and cyclooxygenase-2 in response to IL-1, and downregulation of inducible nitric oxide (NO) synthase in macrophages (60). Inducible NO synthase is responsible for the generation of large amounts of NO, which can be toxic and proinflammatory. PPARα activation can also reduce tissue factor expression and activity in human monocytes/macrophages via inhibition of nuclear factor-κB (61,62). In the liver, PPARα activation decreases IL-1–induced CRP expression and IL-6–stimulated fibrinogen expression (63,64).

PPARα may activate a novel anti-inflammatory pathway. Fatty acids produced by LPL hydrolysis of VLDL (and to a lesser extent LDL and HDL) activate PPARα and thus reverse the increase in vascular cell adhesion molecule-1 expression induced by inflammatory stimuli such as TNF-α and lipoproteins (65). In addition to highlighting a novel anti-inflammatory pathway, these findings suggest different transcriptional responses dependent on the specific mechanism of fatty acid generation and uptake (66).

Activation of PPARγ also confers anti-inflammatory effects. In human monocytes, PPARγ is induced on exposure to oxidized LDL and is expressed at high levels in the foam cells of atherosclerotic lesions (66). Moreover, activation of PPARγ induces ATP-binding cassette A1 expression and cholesterol removal from macrophages, suggesting that PPARγ also regulates a pathway for cholesterol efflux (4). The anti-inflammatory properties of synthetic PPARγ agonists may also be explained, in part, by their ability both to inhibit production of TNF-α and other proinflammatory cytokines (30) and to increase adiponectin expression and secretion (31). In addition to increasing tissue sensitivity to insulin, adiponectin is known to inhibit inflammatory pathways.

CRP is an important marker of systemic inflammation and a powerful predictor of both diabetes and CVD risk (67,68). In a 26-week study of 357 patients with type 2 diabetes, rosiglitazone significantly reduced CRP levels (−26.8%, P < 0.01) compared with placebo; it also reduced matrix metalloproteinase-9, a useful marker of plaque stability (69). In another study, 84 nondiabetic patients with stable CHD were randomized to rosiglitazone or placebo for 12 weeks (70). Rosiglitazone significantly reduced several markers of endothelial activation and inflammation: E-selectin, von Willebrand factor, fibrinogen, and CRP. Plasminogen activator inhibitor type-1 (PAI-1) is another well-established inflammatory marker. Thiazolidinediones reduce PAI-1 protein expression in human preadipocytes and adipocytes (71). In addition, thiazolidinedione-mediated reduction in angiotensin II–induced PAI-1 expression is one suggested mechanism for a reduced incidence of microalbuminuria in diabetic nephropathy (72).

The presence of NO, at low levels, is fundamental to endothelial health, and the ability of endothelial cells to synthesize this short-lived molecule is widely used as a marker of normal function. Agonists of PPARα and -γ have been shown to increase NO expression and release, respectively, from vascular endothelial cells (73,74), demonstrating that these agents may improve endothelial function in vivo.

Vascular remodeling is now established as a key contributory factor in CVD. In vitro and in vivo studies have shown that thiazolidinediones inhibit the proliferation, hypertrophy, and migration of vascular smooth muscle cells (75–77). These agents also reduce the progression of intima-media thickness that commonly occurs in patients with atherosclerosis and, in some studies, decreases intima-media thickness per se (78,79). At least part of this action could be caused by interference with the mitogen-activated protein–kinase pathway (80).

In addition, PPARs can also exert a therapeutic effect on the microvascular complications of diabetes. In patients with diabetic nephropathy, characteristic morphologic changes such as thickening of the glomerular basement membrane, glomerular hypertrophy, and mesangial expansion lead to microalbuminuria. PPARγ ligands ameliorate microalbuminuria in patients with early diabetic nephropathy (81), reduce urinary albumin excretion in microalbuminuric patients, and decrease the serum concentration of type IV collagen (82).

PPARα and -γ ligands are in widespread clinical use for the treatment of dyslipidemia and insulin resistance, respectively. In addition, preclinical studies suggest that ligands of the more recently identified PPAR isoform PPARδ may be beneficial in patients with these conditions. PPARδ agonists have been shown to reduce levels of triglycerides and small, dense LDLs and to increase HDL cholesterol in insulin-resistant animals (83,84). These agents may therefore be of benefit in patients with diabetic dyslipidemia or the metabolic syndrome. Treatment of obese mice with a PPARδ agonist reduces obesity and insulin resistance via an effect on fatty acid oxidation in skeletal muscle (85). Although investigation of PPARδ has not yet progressed beyond the preclinical stage, these findings have led to suggestions that PPARδ may be a useful pharmacological target for treatment of obesity and insulin resistance (86). Furthermore, although the precise role of PPARδ in the inflammatory process has not been fully elucidated, this receptor may also have anti-inflammatory effects, in particular in macrophages (87).

As investigation of the PPAR family continues, it is becoming apparent that PPARs are involved in a wide range of physiological processes and, as a result, are potential targets for a number of disease processes in addition to dyslipidemia and type 2 diabetes, such as psoriasis (88) and multiple sclerosis (89).

PPARs are central to the regulation of energy homeostasis, with each isoform controlling particular aspects. As a result, agents that activate individual PPARs have different clinical effects. PPARα agonists are used in clinical practice for the treatment of dyslipidemia (in particular low HDL cholesterol and elevated triglyceride levels) and reducing cardiovascular risk, whereas PPARγ agonists are currently used for the treatment of type 2 diabetes; by reducing insulin resistance, these agents effectively ameliorate hyperglycemia. The roles played by PPARδ in control of lipid metabolism, glucose metabolism, and obesity are currently under investigation.

The finding that PPAR agonists play a role in regulating other processes, such as inflammation, vascular function, and vascular remodeling, has highlighted further potential indications for these agents, of which a few have been touched on in this article. Continued study of these agents will, no doubt, reveal new applications. Furthermore, initial studies suggest that the advent of dual PPARα/γ and pan-PPAR agonists is likely to broaden the therapeutic value of these agents.

Editorial support was provided by the Future Forum Secretariat, London, U.K. The Future Forum is sponsored by AstraZeneca.