Are peroxisome proliferator-activated receptors new therapeutic targets in diabetic and non-diabetic nephropathies?

Peroxisome proliferator-activated receptors (PPAR), members of the nuclear hormone-receptor superfamily of ligand-binding-transcription factors, are involved in the pathophysiology of the metabolic syndrome. Agonist activation of PPAR provides a new pharmacological pathway to the treatment of the metabolic syndrome and its complications. Since one of the major complications of this syndrome is nephropathy, the potential benefit of PPARα, -γ and –β/δ agonists on kidney merits examination. Moreover, numerous studies have demonstrated that, in addition to their hypolipidaemic and anti-diabetic effects, these drugs possess anti-inflammatory, anti-fibrotic and anti-proliferative properties. These data strongly suggest a potential benefit of PPAR agonists on diabetic and non-diabetic nephropathies. Herein, we describe the currently known effects of PPARα, -γ and -β/δ agonists on diabetic and non-diabetic nephropathies, and more precisely, focus on their potential positive impact on kidneys. PPARα agonists are involved in lipid-metabolism regulation primarily by increasing fatty acid β oxidation. PPARα are predominantly expressed in tissues with high peroxisomal β oxidation and mitochondrial activities (liver, heart, skeletal muscle, intestine and kidney) and, to a lesser extent, in other tissues. Pharmacological PPARα agonists, such as fibrates, are also involved in almost all steps of lipoprotein metabolism. They lower hepatic triglyceride production by increasing fatty acid β oxidation, increase the high-density lipoprotein (HDL)-cholesterol level by raising expression of apolipoproteins AI and AII, favour reverse cholesterol transfer by increasing hepatic scavenger-receptor class B type 1 (SR-B1), necessary for the uptake of HDL-cholesterol and promote HDL-mediated cholesterol efflux from macrophages located in the vascular wall by inducing ATP-binding cassette-transporter-A1 (ABCA-1). PPARα agonists also lower the triglyceride level by increasing lipoprotein lipase activity which induces lipolysis [ 1 ]. In addition to their lipid regulation, PPARα agonists have also been shown to have anti-inflammatory properties. In the kidney, PPARα are predominantly expressed in proximal tubules, the medullary thick ascending limb and, to a lesser degree, in mesangial cells [ 2 ]. During sustained starvation, PPAR activation plays a crucial role by triggering fatty acid β oxidation to maintain balanced energy production and expenditure. This phenomenon was well-illustrated by Kamijo et al . [ 3 ] who showed that sustained starvation of PPARα knockout mice increased urinary albumin excretion (UAE) because of the inability to correctly resorb albumin in renal proximal tubules. This resorption dysfunction is histologically documented by the observation of giant lysosomes containing large amounts of albumin. This failure of lysosomal enzymes to degrade albumin can be reversed by glucose administration, which restores appropriate ATP production. These observations suggest that β-oxidation activation by PPARα agonists is essential for adequate proximal tubule function during sustained starvation. Excess lipid storage in the proximal tubules favours renal dysfunction. It most commonly results from on the one hand, enhanced expression of two major transcription factors—sterol regulatory element-binding protein-1 (SREBP-1) and SREBP-2—which stimulate fatty acid and cholesterol syntheses, on the other hand decreased expression of PPARα-dependent key enzymes involved in β oxidation. In this way, several diabetic animal models with nephropathy such as Zucker diabetic fatty (ZDF) rats, db / db mice [ 4 ], SREBP-1c transgenic mice [ 5 ], age-related renal disease rat model [ 6 ] and high-fat-diet-induced obesity in C57BL/6J mice [ 7 ] are associated with excess renal lipid accumulation. This lipid accumulation, depending on the models, may induce an increased production of transforming growth factor-β (TGFβ), vascular endothelial growth factor (VEGF), connective tissue growth factor (CTGF), plasminogen-activator inhibitor-1 (PAI-1) and extracellular matrix proteins (type IV collagen and fibronectin) all of which contribute to the resulting glomerular hypertrophy, glomerulosclerosis, tubulointerstitial fibrosis, proteinuria and eventually accelerated deterioration of renal function. In non-diabetic animal, the models of renal injury, such as ischaemia–reperfusion and cisplatin-induced acute renal failure, are also characterized by excess lipid build-up in renal tubule cells secondary to fatty acid β-oxidation impairment via PPARα-activation pathways [ 8–10 ]. The increased amount of cellular lipids, in cisplatin-treated proximal tubule cells in culture, responsible for lipotoxicity, leading to apoptotic cell death, is notably prevented by prior administration of a PPARα agonist [ 8 ]. PPARα-null mice, with induced renal ischaemia–reperfusion disease, develop more severe deterioration of renal function than wild type controls [ 9 ]. Finally, prior administration of PPARα agonists, in the cisplatin-induced mouse model of renal failure and the Wistar rat model of renal ischaemia–reperfusion injury, significantly protected renal function, primarily by increasing renal fatty acid β oxidation [ 9 , 10 ]. In contrast, the nephroprotective impact of PPARα agonists was not observed when they were administered to PPARα-null mice [ 9 , 10 ]. PPARα agonists are also nephroprotective because of anti-inflammatory properties. In addition to renal excess lipids accumulation, cisplatin-induced acute renal failure is associated with increased levels of nuclear factor-κB (NF-κB)-binding activity, chemokines, pro-inflammatory cytokines and enhanced neutrophil infiltration into the corticomedullary area of kidney. Prior administration of a PPARα ligand led to significantly fewer inflammatory events caused by cisplatin [ 11 ]. In contrast, this anti-inflammatory effect of PPARα agonists in the cisplatin model was not observed in PPARα-null mice. In another rat model of renal inflammation, characterized by crescentic glomerulonephritis induced by rabbit anti-glomerular basement membrane (GBM) antibodies, high doses of the PPARα agonist bezafibrate significantly prevented glomerular proliferation, macrophage infiltration and proteinuria [ 12 ]. Cytochrome P -450 is a target for PPARα agonists. Its activation induces production of 20-HETE (hydroxyeicosatetraenoic acid) an arachidonic acid metabolite [ 13 ]. In addition to being a potent vasoconstrictor on the renal vasculature [ 14 ], 20-HETE inhibits sodium resorption in proximal tubules [ 15 ] and blocks NaCl transport in the thick ascending limb [ 16 ]. Since PPARα and cytochrome P -450 are mainly co-located in the renal proximal tubules and the thick ascending limb [ 13 ], the predominant effect of PPARα agonist on tubular function may in part explain the anti-hypertensive effect observed in some animal models such as Dahl salt-sensitive rats [ 17 ]. In humans, the PPARα agonist impact on blood pressure and natriuresis is less clear. Clinical trials using fibrates in humans for providing renal protection are primarily represented by the Diabetes Atherosclerosis Intervention Study (DAIS) and, more recently, the Fenofibrate intervention and events lowering in diabetes (FIELD) study. The DAIS extension study compared fenofibrate with placebo in type 2 diabetes (T2D) patients [ 18 ]. For the 314 diabetic patients without nephropathy who continued to take fenofibrate or placebo over a 5-year period, the respective rates of worsening albumin excretion were 8 and 18% ( P < 0.05). This benefit was predominantly attributed to blocking the progression from normal albuminuria to microalbuminuria. The FIELD study was a randomized controlled trial that included 9795 T2D patients treated for 5 years with fenofibrate or a placebo [ 19 ]. Its results indicate that the progression from normal albuminuria to microalbuminuria or from microalbuminuria to macroalbuminuria was significantly lower for the fenofibrate group than those taking the placebo 10 and 11%, respectively [ 19 ]. PPARγ are strongly expressed in adipose tissue and their activation stimulates preadipocyte differentiation. Moreover, PPARγ agonists improve insulin sensitivity and glucose homoeostasis by promoting uptake and storage of free fatty acids in adipose tissue, and by modifying adipocyte-derived signalling molecules, also called adipocytokines: by decreasing release of prodiabetic adipocytokines, including tumour necrosis factor-α (TNFα), interleukin (IL)-6, leptin and resistin and increasing the circulating concentration of the anti-diabetic adipocytokine adiponectin. However, the anti-diabetic effect of PPARγ agonists is mostly associated with a weight gain of ∼2 kg and a trend toward the onset of oedema, with fluid retention and lower blood pressure [ 20 ]. In renal tissue, PPARγ are predominantly expressed in collecting ducts and, to a lesser extent, in glomeruli, mesangial cells, proximal tubules and the renal microvasculature [ 2 ]. In ∼7% of treated patients, PPARγ agonists are associated with fluid retention, haemodilution, decreased sodium fractional excretion and diuresis, lower blood pressure, higher vascular permeability and onset of oedema [ 20 ]. Fluid retention results mainly from primary renal sodium retention induced by PPARγ agonists. Indeed, in vitro , PPARγ agonists enhance the number of luminal amiloride-sensitive epithelial Na + channels (ENaC) located on the apical membrane of a cultured cell line derived from human cortical collecting ducts [ 21 ]. These enhanced ENaC activities are abolished by a selective PPARγ antagonist. The pathophysiological relevance of PPARγ agonists in an in vivo experimental model of sodium renal retention in mice was demonstrated later by Guan et al . [ 22 ]. Early weight gain was blocked by the collecting duct-specific diuretic amiloride and was also prevented by specific PPAR -gene deletion from the collecting ducts in those mice. In the same study, the authors evaluated amiloride-sensitive Na + -channel activity in cultured inner medullary collecting duct (IMCD) cells from mice by measuring their radiolabelled sodium [ 22 Na]-flux absorption. They demonstrated that PPARγ agonists significantly increased Na + -flux absorption and that this increase was completely blocked by a specific PPARγ antagonist. However, in IMCD cells derived from PPAR gene-deleted mice, Na + -flux absorption was not stimulated by pioglitazone. It is important to retain the fact that oedema results not only from fluid retention, but also, in part, from increased capillary permeability due to lowered insulin resistance. Pertinently, insulin favours capillary permeability and sodium renal retention leading to oedema [ 23 ]. PPARγ agonists also directly increase capillary permeability by enhancing vascular VEGF production [ 24 ]. Experimental and clinical studies have shown that PPARγ agonists can significantly lower blood pressure. This effect is particularly dramatic in animal models of insulin resistance. Mechanisms by which blood pressure falls remain to be clarified. However, in several insulin-resistant animals, the blood pressure decline may be partly due to increased insulin sensitivity. Alternatively, peripheral vasodilatation might be explained by the involvement of three mediators: nitric oxide release from endothelial cells, excess syntheses of VEGF and vasodilating prostaglandin [ 24 , 25 ]. PPARγ agonists also directly induce vasodilatation secondary to inhibition of extracellular Ca 2+ uptake via calcium channels [ 26 ]. Finally, numerous observations support the direct blockage of the angiotensin II type-1 receptors by PPARγ agonists [ 27 ]. Activation of PPARγ expressed on cultured mesangial cells by pharmacological ligands, like troglitazone, or the natural ligand, 15-deoxy-delta-prostaglandin J2 (15d-PGJ2), decreases mesangial cell proliferation and expression of smooth muscle α-actin, a marker of myofibroblast activation [ 28 ]. Furthermore, PPARγ agonists inhibited the enhanced PAI-1 expression (considered a potent profibrotic factor contributing to glomerulosclerosis) by cultured mesangial cells incubated with angiotensin II [ 29 ]. Stimulation of mesangial cell proliferation by platelet-derived growth factor (PDGF) was also blocked by PPARγ agonists [ 30 ]. In addition to its anti-proliferative properties, PPARγ ligands exerted direct anti-fibrotic actions on mesangial cells by inhibiting type I collagen expression [ 31 ] and by suppressing expression of TGFβ-1-mediated smooth muscle α-actin, fibronectin and PAI-1 through an increase of hepatocyte growth factor (HGF) synthesis [ 32 ]. Exposure of opossum kidney cells, used as an in vitro model of proximal tubule cells, to low-density lipoprotein (LDL) or to albumin led to an increased production of monocyte chemotactic protein-1 (MCP-1) and TGFβ-1 which was reversed in the presence of pioglitazone [ 33 ]. Furthermore, prior addition of pioglitazone to the culture further enhanced albumin uptake by tubule cells but it was no longer associated with an exaggerated inflammatory or profibrotic cytokine response. Hence, it could be hypothesized that in vivo proteinuria would be reduced by enhanced proximal tubular uptake of albumin without potential deleterious effects on the tubulointerstitium. Exposure of the human proximal tubule cell line, HK-2 to high glucose is associated with PPARγ up-regulation. The latter is probably a protective response, as it was also associated with MCP-1 gene down-regulation and consequently, less of the inflammatory protein. This effect was reproduced by a PPARγ agonist, which further decreased AP-1 and TGFβ-1. This anti-inflammatory response was associated with anti-proliferative and proapoptotic effects [ 34 ]. Most of the results from studies where different diabetic animals model with nephropathy and proteinuria were treated with PPARγ agonists strongly suggest a renal protective effect of the PPARγ agonists, through a mechanism independent of their insulin-sensitizing action. In the Zucker fatty ( fa/fa ) rat, a prediabetic insulin-resistant syndrome model with nephropathy and proteinuria, the PPARγ agonist delays the onset of nephropathy and proteinuria and slows down the development of renal dysfunction because of the correction of metabolic abnormalities, but probably also through a direct renal effect. Indeed, proteinuria appearance in untreated Zucker fatty control rats coincided with the development of hypertension. This implies that the renal protection of the agonist PPARγ must be explained by another mechanism than lowering blood pressure [ 35 ]. Similarly, in the inbred obese ZDF rat model of severe T2D with extensive kidney damage, angiotensin-converting enzyme inhibition (ACEI) or PPARγ agonists alone significantly prevented proteinuria and impaired renal clearance, and protected against structural damage of glomerular and tubulointerstitial tissues [ 36 ]. However, treatment with PPARγ agonists was more protective than ACEI, as assessed at 6 months of age by lower proteinuria, less glomerulosclerosis and less macrophage/monocyte tubulointerstitial infiltration. Having given the wide range of PPARγ-agonist actions, numerous mechanisms to explain their superior renal protection can be envisaged. These mechanisms are extensions of PPARγ-agonist metabolic actions on glycaemia control and the lowering hyperlipidaemia, but probably also reflect a direct anti-inflammatory, anti-proliferative and anti-fibrotic action of these drugs on the kidney [ 36 ]. Finally and moreover, in the Sprague–Dawley rat model of streptozotocin-induced diabetes, troglitazone administration started a few days after streptozotocin was able to prevent glomerular hyperfiltration, albuminuria and the enhanced build-up of extracellular matrix proteins and TGFβ-1 in glomeruli, despite theabsence of effect on blood pressure or glucose levels [ 37 ]. In the different non-diabetic animal models with nephropathy such as renal ischaemia–reperfusion-induced injury in Wistar rats [ 38 ], ciclosporin A (CsA)-induced renal injury in Sprague–Dawley rats [ 39 ], renal dysfunction induced by sepsis in mice [ 40 ] and the 5/6-nephrectomy in Sprague–Dawley rats [ 41 ], prior administration of PPARγ agonists constantly generated a protective effect, as demonstrated by less interstitial inflammatory cell infiltration, interstitial fibrosis and proinflammatory mediators compared with untreated controls [ 38–41 ]. All those beneficial effects were independent of the PPARγ agonist's action on lipid and glucose homeostases. Several studies have demonstrated that PPARγ agonists can prevent proteinuria in humans. Lebovitz et al . [ 42 ] conducted a randomized trial on 493 T2D patients assigned to receive rosiglitazone or placebo for 26 weeks. They reported a statistically significant reduction of the urinary albumin/creatinine ratio (UACR) from baseline for rosiglitazone-treated patients, and that UACR was ∼30% lower in that group compared with the placebo group [ 42 ]. Except in a study where the sulfonylurea glicazide and the thiazolinedione pioglitazone significantly and similarly lowered the UACR of T2D patients treated for 12 weeks with either molecule [ 43 ], most of the studies provide superior benefits with PPARγ agonists compared with other anti-diabetic drugs on UACR. In this way, Bakris et al . [ 44 ] found a significant UACR reduction from baseline for T2D patients treated for 52 weeks with rosiglitazone, unlike the sulfonylurea glibenclamide-treated group, which achieved no beneficial effect in terms of microalbuminuria reduction. The discrepancy between the microalbuminuria outcomes could not be explained by glucose-homoeostasis control, which was similar for both the groups, but might be attributable, at least in part, to the strong relationship between UACR and blood-pressure reduction obtained only in the rosiglitazone-treated group. Hence, in addition to blood glucose control, rosiglitazone exerts a beneficial effect on microalbuminuria probably mediated, at least in part, through blood-pressure reduction. Furthermore, it cannot be excluded that the beneficial effect observed could be a consequence of other mechanisms acting directly on renal vascular protection, for example, reversal of insulin resistance (which is associated with a cluster of metabolic abnormalities including elevated PAI-1, prothrombotic and proinflammatory states …) [ 45 ]. Those results were supported by Nakaruma et al . [ 46 ], who treated 32 T2D patients for 12 months with the glibenclamide or the tiazolidinedione troglitazone and observed significant reductions of microalbuminuria only in troglitazone-treated patients. Similarly, two studies [ 47 , 48 ] compared the efficacies of troglitazone or pioglitazone, on UACR in T2D patients and also found significant UACR reductions with troglitazone after 4 weeks of or with pioglitazone after 52 weeks of but no with despite correction of The lower obtained with PPARγ agonists, like pioglitazone, could also be in part, by increased renal proximal tubule uptake of albumin [ 33 ]. In contrast, the results of the trial in pioglitazone against events were study included T2D patients at high of complications who were treated with pioglitazone or with drugs for no was the impact on and macroalbuminuria [ ]. The of diabetic nephropathy is to at least because not all patients have the same renal when they have of metabolic is a of deleterious and renal Since of diabetic nephropathy and higher rates are found T2D it is to for a between diabetic nephropathy and for example, to in the PPAR gene that could be associated with diabetic from the diabetes study [ ] that, T2D patients, those the significantly lower and to develop proteinuria less Those observations to indicate a protective effect of the in to diabetic PPARα and PPARγ agonists, fibrates and of agonists remain and are currently being their activities on cell and [ ], are involved in the pathophysiology of the metabolic more precisely, in preadipocyte and regulation of fatty acid β oxidation in skeletal In the kidney, agonists are of being involved in regulation, because in the kidney is after an and to control levels [ 52 ]. However, in addition to their metabolic actions on the agonists also nephroprotective In to the other are and expressed in all more in proximal tubules and renal medullary interstitial cells [ 2 ]. agonists a role in cell and in the of tissue This is well-illustrated by et al . [ ] in a mouse model of ischaemia–reperfusion-induced renal to epithelial cell into the tubule by and apoptotic cell death, in the proximal tubules in the of the kidney, particularly to cells that not in the of tubule and of renal function. In mice to is and renal dysfunction is compared with mice controls [ ]. Furthermore, mice with a specific agonist less severe tubule injury than untreated control mice. Finally, prior to a agonist exerts an effect on cultured human proximal tubule epithelial cells to [ ]. PPARα, and agonists provide nephroprotective effects in numerous diabetic and non-diabetic models with nephropathy, via mechanisms including of renal lipid by activation fatty acids β oxidation, anti-inflammatory, anti-fibrotic and properties. However, their nephroprotective effects in humans are lesser and only by of microalbuminuria. study results were as they to PPARγ agonist protection of the kidneys. the currently between the results of and clinical studies the efficacies of these drugs on renal disease, PPAR agonists a to prevent renal tissue damage in diabetic and non-diabetic nephropathies. studies are to and the potential renal benefits of these It would be particularly to human renal with lipid accumulation, such as diabetes, cisplatin-induced acute renal failure or age-related renal disease and to the potential nephroprotective impact of PPARα agonists on It would be also to the impact of or -γ and -β/δ agonists on renal of .