This Article Abstract Full Text (PDF) All Versions of this Article: 25/6/971    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schrijvers, B. F. Articles by Flyvbjerg, A. Search for Related Content PubMed PubMed Citation Articles by Schrijvers, B. F. Articles by Flyvbjerg, A. Pubmed/NCBI databases Compound via MeSH Substance via MeSH Medline Plus Health Information Diabetic Kidney Problems Hazardous Substances DB NITRIC OXIDE

From Hyperglycemia to Diabetic Kidney Disease: The Role of Metabolic, Hemodynamic, Intracellular Factors and Growth Factors/Cytokines

Medical Department M (B.F.S., A.F.), Medical Research Laboratories, Institute of Experimental Clinical Research, Aarhus University Hospital, DK-8000 Aarhus, Denmark; and Renal Unit (B.F.S., A.S.D.V.), Department of Internal Medicine, Gent University Hospital, B-9000 Gent, Belgium

Correspondence: Address all correspondence and requests for reprints to: Allan Flyvbjerg, M.D., D.M.Sc., Medical Department M/Medical Research Laboratories, Clinical Institute, Aarhus University Hospital, Nørrebrogade 44, DK-8000 Aarhus C, Denmark. E-mail: allan.flyvbjerg{at}dadlnet.dk

	Abstract

Top Abstract I. Introduction II. Metabolic Factors III. Hemodynamic Factors IV. Intracellular Factors V. Growth Factors and... VI. Outlook and Future... References

 
At present, diabetic kidney disease affects about 15–25% of type 1 and 30–40% of type 2 diabetic patients. Several decades of extensive research has elucidated various pathways to be implicated in the development of diabetic kidney disease. This review focuses on the metabolic factors beyond blood glucose that are involved in the pathogenesis of diabetic kidney disease, i.e., advanced glycation end-products and the aldose reductase system. Furthermore, the contribution of hemodynamic factors, the renin-angiotensin system, the endothelin system, and the nitric oxide system, as well as the prominent role of the intracellular signaling molecule protein kinase C are discussed. Finally, the respective roles of TGF-ß, GH and IGFs, vascular endothelial growth factor, and platelet-derived growth factor are covered. The complex interplay between these different pathways will be highlighted. A brief introduction to each system and description of its expression in the normal kidney is followed by in vitro, experimental, and clinical evidence addressing the role of the system in diabetic kidney disease. Finally, well-known and potential therapeutic strategies targeting each system are discussed, ending with an overall conclusion.

I. Introduction

II. Metabolic Factors

A. Advanced glycation endproducts (AGEs)

B. Aldose reductase (AR) /polyol pathway

III. Hemodynamic Factors

A. Angiotensin II (Ang II)/renin-angiotensin system (RAS)

B. Endothelin (ET)

C. Nitric oxide (NO)

IV. Intracellular Factors

A. Diacylglycerol (DAG)-protein kinase C (PKC) pathway

V. Growth Factors and Cytokines

A. Transforming growth factor ß (TGF-ß)

B. Growth hormone (GH) and insulin-like growth factors (IGFs)

C. Vascular endothelial growth factor (VEGF)

D. Platelet-derived growth factor (PDGF)

VI. Outlook and Future Perspectives

	I. Introduction

Top Abstract I. Introduction II. Metabolic Factors III. Hemodynamic Factors IV. Intracellular Factors V. Growth Factors and... VI. Outlook and Future... References

 
DIABETIC NEPHROPATHY IS one of the most serious complications of diabetes and the most common cause of end-stage renal failure in the Western world. At present, diabetic kidney disease affects about 15 to 25% of type 1 diabetic patients (1) and 30 to 40% of patients with type 2 diabetes (2, 3). Diabetic nephropathy is characterized by specific renal morphological and functional alterations. Features of early diabetic renal changes are glomerular hyperfiltration, glomerular and renal hypertrophy, increased urinary albumin excretion (UAE), increased basement membrane thickness (BMT), and mesangial expansion with the accumulation of extracellular matrix (ECM) proteins such as collagen, fibronectin, and laminin. Advanced diabetic nephropathy is characterized by proteinuria, a decline in renal function, decreasing creatinine clearance (CrCl), glomerulosclerosis, and interstitial fibrosis.

As will appear from this review, a vast majority of the studies referenced have been performed in diabetic animals. This raises the question whether diabetic rodents are good models of human disease. The most frequently used animal strain in experimental diabetes is the rat. In spontaneous and chemically [e.g., streptozotocin (STZ)] induced rat models of type 1 diabetes, as well as rat models of type 2 diabetes, early renal changes are seen: renal enlargement, glomerular hypertrophy, hyperfiltration, and an increasing UAE (4, 5). Furthermore, diabetic rats with a diabetes duration of more than 6 months develop electron microscopic (EM) changes, i.e., increased BMT (6, 7, 8). Over the past years, diabetic mice models have been used more frequently in the study of diabetic renal changes. As is the case for rats, diabetic mice present with early renal changes comparable with the changes seen in diabetic rats, i.e., renal enlargement, glomerular hypertrophy, hyperfiltration, and a progressive increase in UAE (9). In contrast to diabetic rats, diabetic mice models develop pronounced structural glomerular EM changes after approximately 2 months’ duration of diabetes with an increase in BMT and mesangial expansion (9, 10). In conclusion, various diabetic rodent models display early renal changes that have similarities to the early changes seen in human diabetes, whereas only diabetic mice present with increased BMT and mesangial expansion at an EM-level within a diabetes duration of a few months.

Over the past few decades, extensive research has elucidated several pathways that play a role in the development and/or progression of diabetic kidney disease. Beyond the role of high blood glucose, the pathophysiological role of different metabolic pathways in the development and progression of diabetic nephropathy has been studied extensively. Advanced glycation endproduct (AGE) formation as a consequence of sustained or transient hyperglycemia has been implicated. Furthermore, the potential role of the aldose reductase/polyol pathway, although controversial, will be addressed.

In addition, various vasoactive factors contribute to the development and progression of diabetic microvascular complications. These vasoactive factors include vasoconstrictors such as angiotensin (Ang) II and endothelin (ET), as well as vasodilators such as nitric oxide (NO). Besides the well-known systemic and locally hemodynamic effects, the renin-angiotensin system (RAS) and the ET system can exert nonhemodynamic effects via an autocrine or paracrine mode of action. They stimulate the proliferation of kidney cells and the expression of growth factors or cytokines, which may directly or indirectly contribute to the renal changes seen in diabetes. The controversial role of the NO system in renal hemodynamic changes in diabetes will also be addressed.

Activation of the diacylglycerol (DAG)-protein kinase C (PKC) pathway is also a well-known feature of diabetes. PKC is an important intracellular pathway that can be activated by many of the metabolic and hemodynamic factors involved in the pathogenesis of diabetic nephropathy. PKC can then be a stimulus for the initiation of several growth factors and cytokines.

Different growth factors and cytokines directly influence kidney function or exert their effects in a more indirect way by stimulating other factors. One of the first endocrine factors to be implicated in the pathogenesis of diabetes was GH (11). The attention was drawn to the possible role of GH in the development of diabetic microangiopathy (12). At about the same time, IGFs were identified in diabetic serum (13). Since then, new information has appeared with increasing pace on growth factors implicated in the pathogenesis of the renal functional and structural alterations seen in diabetes. Ample experimental and clinical studies illustrate the prominent role of TGF-ß, and during the last several years an increasing interest has developed in vascular endothelial growth factor (VEGF). Although platelet-derived growth factor (PDGF) has been studied less intensively, evidence supporting its pathophysiological role in diabetic kidney disease has been collected.

	II. Metabolic Factors

Top Abstract I. Introduction II. Metabolic Factors III. Hemodynamic Factors IV. Intracellular Factors V. Growth Factors and... VI. Outlook and Future... References

 
A. Advanced glycation endproducts (AGEs)
1. The AGE system.
When glucose and other reactive carbonyl compounds react nonenzymatically with proteins, lipids, or nucleic acids, Schiff bases and Amadori products are formed. Additional rearrangement and modification leads to the generation of diverse AGEs, such as carboxymethyl lysine (CML), pentosidine, imidazolone, and pyrraline (14, 15). AGEs interact with specific receptors (14, 15, 16). p60 (OST-48, AGE-R1), p90 (80K-H, AGE-R2), galectin-3 (AGE-R3), the macrophage scavenger receptor type II (ScR-II), and CD36 regulate the uptake and clearance of AGEs (14, 16). The best-characterized receptor is receptor for AGEs (RAGE) (14). Another molecule with AGE-binding and anti-AGE properties is the antimicrobial protein lysozyme (16). AGEs can also act in a receptor-independent way by cross-linking proteins (14, 15). AGEs alter the structure and function of intra- and extracellular molecules, increase oxidative stress, and modulate cell activation, signal transduction, and the expression of cytokines and growth factors through receptor-dependent and receptor-independent pathways (14, 15, 16).

2. Expression of AGEs in the normal kidney.
In human kidney, immunostaining for pentosidine, pyrraline, CML, and imidazolone was negative in glomeruli (17, 18, 19), but in renal tubuli, immunostaining was negative for CML and pentosidine in one study (19) but positive for pentosidine, CML, and imidazolone in others (17, 18). AGE staining was present in rodents, predominantly in endothelial cells within glomeruli and tubulointerstitium (20, 21). AGE and RAGE colocalize in the rat glomerulus (21). In rat kidney, RAGE is expressed in cortex and medulla (22) and localized primarily to glomeruli, distal tubules, and collecting ducts (21). In human kidney, a low level of RAGE is constitutively expressed in glomerular podocytes, but not in the mesangium or glomerular endothelium (19). In the mouse kidney, mRNA and protein for AGE-R1, AGE-R2, and AGE-R3 were detected as well as mRNA for RAGE and the ScR-II (23). More particularly, a distinct glomerular AGE-R1 staining was found, and AGE-R1 to -3 were also detected in cultured mouse mesangial cells (23). RAGE was detected in cultured mouse podocytes (10). Human-, rat-, and mouse-cultured mesangial cells contain binding sites for AGE-modified proteins (24, 25). AGE binding sites were also identified in proximal tubules of rat kidney, but it is unclear whether they represent one of the known AGE receptors (26).

3. In vitro evidence for AGEs effects on renal cells.
In cultured glomerular endothelial and mesangial cells, glycated albumin and AGE-rich proteins increased PKC activity (27, 28, 29), PKCß (29), TGF-ß1 levels (27, 30), and ECM expression (24, 25, 27, 29, 30). Hyperglycemia and glycated albumin appeared to exert an additive effect in glomerular endothelial cells (27). In cultured kidney epithelial cells, CML reduced proteoglycan expression (31). In a proximal tubular epithelial cell line, AGEs induced tubular-epithelial myofibroblast transdifferentiation through RAGE signaling (32).

4. Experimental evidence for a role of AGEs in diabetic kidney disease.
In vivo evidence for a role of AGEs in diabetic kidney disease comes mainly from studies in STZ-diabetic rats. BSA-AGE gold conjugate binding to mesangial matrix, BMT, and glomerular cell nuclei was more intense in renal tissue from diabetic than from control rats (33). Whether this was associated with an up-regulation of RAGE was not investigated. In STZ-diabetic rats, AGE staining was increased in the glomeruli, predominantly in glomerular endothelial cells, and tubulointerstitium (21). Renal AGE levels were increased after 3 (26), 12 (34), and 32 wk of diabetes (20, 35). Furthermore, exogenously administered AGEs induced ECM genes (36), mesangial expansion (37), and up-regulation of TGF-ß1 in nondiabetic mice and rats (36, 37). Recently, it was demonstrated in nonobese diabetic (NOD) and db/db mice, models of type 1 and type 2 diabetes, respectively, that the intake of food-derived AGEs contributes to diabetic nephropathy and that low AGE diet provides renoprotection (9). An increased number of proximal tubular AGE binding sites was found in type 1 diabetic rats together with an increase in renal AGE levels (26). Recently, in STZ-diabetic rats, RAGE was up-regulated after 4 wk of diabetes up to 12 wk, whereas AGE-R2 and AGE-R3 remained unchanged (22, 34). In NOD mice, early after the onset of diabetes, AGE-R1 expression was reduced, AGE-R2 expression was unchanged, and AGE-R3, RAGE, and ScR-II expression was increased compared with nondiabetic control mice (23). In db/db mice, glomerular RAGE staining was enhanced, especially in podocytes, compared with nondiabetic control mice (10). Diabetic transgenic mice overexpressing human RAGE developed renal and glomerular hypertrophy, increased albuminuria, mesangial expansion, advanced glomerulosclerosis, and increased serum creatinine compared with diabetic littermates lacking the RAGE transgene (38).

5. Clinical evidence for a role of AGEs in diabetic kidney disease.
In type 1 diabetic patients, increased circulating AGEs preceded the development of microvascular complications (39) and predicted the progression of the early morphological renal changes in BMT and matrix/glomerular volume fraction (40). In monocytes of type 1 diabetic patients, reduced expression of AGE-R1 was linked to elevated serum AGE levels and the presence of severe diabetic complications (41). In type 2 diabetic patients, increased serum levels of crossline, one of the structurally defined adducts of AGEs, were associated with the presence of nephropathy (42). In addition, serum concentrations of AGEs were associated with the development of diabetic microangiopathy in patients with type 2 diabetes (43). Amadori products, pentosidine, CML, pyrraline, and imidazolone were demonstrated in renal tissue of diabetic patients (17, 18). CML accumulated predominantly in the basement membrane and pentosidine primarily in the interstitium in patients with diabetic kidney disease (17, 19). The extent of CML and pentosidine immunostaining in the glomerular and tubulointerstitial compartments correlated with the severity of diabetic nephropathy (19). The distinctive pattern of CML accumulation in diabetic nephropathy differed from the more nonspecific AGE accumulation observed in nondiabetic sclerosing renal diseases. Glomeruli of patients with diabetic nephropathy demonstrated diffuse up-regulation of RAGE expression in podocytes (19).

6. Agents with effects on AGEs in diabetic kidney disease.
Various classes of drugs are able to interfere with the formation of AGEs or the cross-linking of proteins by AGEs. Aminoguanidine, pyridoxamine, 2,3-diamino-phenazine, OPB-9195, and tenilsetam inhibit AGE formation by scavenging reactive carbonyl intermediates, N-phenacetylthiazolium bromide (PTB) and ALT-711 are AGE-cross-link breakers. Furthermore, signal transduction through RAGE can be inhibited by antisense oligodeoxynucleotides (AS-ODNs), RAGE antibodies, or soluble RAGE. Drugs targeting other systems have also shown some effects on AGE-related pathways.

a. Glycated albumin antagonists.
Treatment of db/db mice with monoclonal antibodies against Amadori-modified glycated albumin (A717) attenuated the cortical overexpression of 1 (IV) collagen and fibronectin mRNAs, ameliorated mesangial matrix expansion, reduced albuminuria, and prevented the decline in renal function that developed in untreated db/db mice (44, 45). Oral administration of EXO-226 or 23CPPA, compounds that reduce circulating glycated albumin, had similar renoprotective effects in db/db mice (46, 47). The effects of 23CPPA were mediated by a reduction in glomerular TGF-ß1 (47).

b. Inhibitors of AGE formation.
Aminoguanidine retarded the development of diabetic nephropathy in long-term experimental diabetes in rats (6, 20, 35, 48). It appears that the renoprotective effects of aminoguanidine in diabetes are related to the duration and not the timing of treatment (20, 49) and are mediated by a decrease of AGE formation (21, 26, 50). In addition, aminoguanidine restored the diabetes-induced changes in lysosomal processing (51) and glomerular PKC activity (49, 52) in STZ-diabetic rats. Recently, in STZ-diabetic transgenic (mREN-2)27 rats, a model of diabetic kidney disease with hypertension and an overactive RAS, aminoguanidine ameliorated glomerulosclerosis and medullary pathology and decreased glomerular AGE immunolabeling, but had no effect on kidney weight, glomerular filtration rate (GFR), or UAE (53). Aminoguanidine had no effect on the expression of RAGE, AGE-R2, and AGE-R3 in STZ-diabetic rats. Compared with placebo, aminoguanidine modestly reduced proteinuria and marginally slowed progression of overt diabetic nephropathy in optimally treated type 1 diabetic patients (54). However, two double-blinded, placebo-controlled randomized clinical trials with aminoguanidine were performed in type 1 (I) and type 2 (II) diabetic patients with overt nephropathy. The primary endpoint, i.e., reducing the risk of doubling serum creatinine, was not achieved in ACTION I, but pimagedine reduced urinary protein excretion, serum triglycerides, and low-density lipoprotein levels. ACTION II was terminated early because of side effects and apparent lack of efficacy (55). Pyridorin (pyridoxamine dihydrochloride) (PM) inhibits the conversion of Amadori intermediates to AGEs in vitro (56). PM inhibited the increase of albuminuria and serum creatinine in STZ-diabetic rats (57). The therapeutic potential of PM is currently being investigated in humans with diabetic nephropathy (15). 2,3-Diamino-phenazine inhibited AGE accumulation in vitro (50). In STZ-diabetic rats, 2,3-diamino-phenazine normalized the renal AGE staining (50) and ameliorated collagen solubility, but had no effect on increased UAE (58). OPB-9195 effectively inhibited AGE formation and AGE-derived cross-linking in vitro. In Otsuka Long-Evans Tokushima Fatty (OLETF) rats, a model of type 2 diabetes, the administration of OPB-9195 lowered elevated serum AGE levels and UAE, attenuated glomerular AGE deposition, and prevented the progression of glomerulosclerosis (8). Furthermore, administering OPB-9195 in a RAGE transgenic mouse model prevented the features of advanced diabetic nephropathy (38). In vitro, tenilsetam restored the reduced digestibility of collagen, and in vivo, administration of tenilsetam for 16 wk suppressed the elevated levels of AGE-derived fluorescence and pyrraline in renal cortex of STZ-induced diabetic rats (59).

c. AGE-receptor blockers.
In cultured mesangial cells, anti-p60 antibodies prevented the AGE-induced increases in IGF-I, TGF-ß1, and ECM expression or production (60), and anti-p60 and anti-p90 antibodies prevented the AGE-induced increase in type IV collagen (24). In db/db mice treated with a neutralizing murine RAGE antibody for 2 months, the increases in kidney weight, glomerular volume, mesangial volume, and UAE were reduced, and the increases in CrCl and BMT were normalized (Fig. 1) (61). In db/db mice, treatment with soluble RAGE prevented the increases in VEGF and TGF-ß expression, mesangial expansion, BMT, and UAE and preserved renal function. Similarly, these diabetes-induced changes were not seen in diabetic RAGE null mice (10).

View larger version (11K): [in this window] [in a new window]   FIG. 1. BMT and total mesangial volume on d 60 in placebo-treated db/+ mice (open bars), RAGE antibody-treated db/db mice (light gray bars), placebo-treated db/db mice (black bars), and RAGE antibody-treated db/db mice (dark gray bars). Data are means + SEM, n = 6–10 in each group. *, P < 0.01 vs. db/+ mice and RAGE antibody-treated db/db mice. , P < 0.05 vs. db/+ mice and placebo-treated db/db mice. [Reproduced with permission from A. Flyvbjerg et al.: Diabetes 53:166–172, 2004 (61 ). © American Diabetes Association.]

e. Aldose reductase inhibitors (ARIs).
Epalrestat decreased the elevated levels of fructose 3-phosphate and AGEs in erythrocytes after 2 months of treatment (64), supporting the hypothesis that the polyol pathway plays a substantial role in the nonenzymatic glycation of proteins (64, 65).

f. Others.
In cultured mesangial cells, lysozyme suppressed the AGE-enhanced expression of several modulators of kidney structure and function, i.e., PDGF-B, type 1 (IV) collagen and tenascin, and normalized the AGE-suppressed expression and activity of matrix metalloproteinase-9 (MMP). In vivo, lysozyme administration to NOD and db/db mice reduced the elevated serum AGE levels, enhanced urinary AGE excretion, and decreased UAE (66). PDGF-antibody abrogated the AGE-induced increase in type IV collagen in cultured mesangial cells, indicating that this response was not mediated directly by AGEs and RAGE, but indirectly through an intermediate factor, i.e., PDGF (24). Angiotensin receptor blockers (ARBs) and angiotensin-converting enzyme (ACE) inhibitors (ACEi) lower the formation of AGEs in vitro by lowering the production of reactive carbonyl precursors (67). Ramipril attenuated the renal AGE accumulation found in STZ-diabetic rats, identifying a relationship between the RAS and the accumulation of AGEs in experimental diabetic nephropathy (34).

7. Conclusion.
In vitro and experimental data suggest that the up-regulation of AGEs and RAGE, and AGE-RAGE interaction contribute to diabetic renal changes, i.e., TGF-ß1 up-regulation, PKC activation, renal and glomerular hypertrophy, albuminuria, ECM production, mesangial expansion, and glomerulosclerosis. Reduced AGE-R1-dependent clearance may be responsible for increased circulating AGEs that are associated with the development of microvascular complications. AGEs accumulate in the kidney, especially in diabetic lesions. Inhibitors of AGE formation are promising agents because they improve functional and/or structural changes associated with diabetic kidney disease in experimental models. Despite promising preliminary effects in type 1 diabetic patients, aminoguanidine still needs to prove its effect in randomized clinical trials. Although the renoprotective effects of AGE formation inhibitors are mainly mediated through decreased AGE formation, multiple pathways are affected, i.e., cytokines such as TGF-ß1, IGF-I, PDGF-B, VEGF, PKC activity, and NO synthase (NOS). The renal effects of AGE-receptor blockers, RAGE antibodies, and AGE-cross-link breakers implicate AGE signaling pathways in the functional and structural changes of diabetic kidney disease.

B. Aldose reductase (AR)/polyol pathway
1. The AR system.
AR is a member of the aldo-keto reductases (AKR), a superfamily of nicotinamide adenine dinucleotide phosphate (NADPH)-dependent reductases with potential roles in the detoxification of aldehydes and ketones (68). The family of AR consists of 11 members, named AKR1B1 to AKR1B11 (AKR homepage, http://www.med.upenn.edu/akr). AR catalyzes the reduction of glucose to sorbitol, the first and the rate-limiting step in the polyol pathway (68). Sorbitol metabolism is also governed by sorbitol dehydrogenase (SDH), the enzyme responsible for the degradation of sorbitol to fructose using NAD⁺ (69). Sorbitol may interfere with the uptake and metabolism of myo-inositol (70). The physiological role of the AR pathway remains largely unknown. However, AR, sorbitol and myo-inositol are thought to play a role in the osmoregulation of the kidney (71). In most mammalian cells, the intracellular concentration of myo-inositol is many times higher than in the circulation by virtue of a Na⁺/myo-inositol cotransporter (72). Species- and organ-specific ARs have been implicated in hormone-regulated apoptosis (73) and hormone production (74).

2. Expression of AR in the normal kidney.
A single form of AR is expressed in human kidney (75). AR mRNA, protein, and enzyme activity were mainly found in the inner medulla (76). AR mRNA expression was pronounced in the papilla, especially in collecting duct cells, epithelial, endothelial, interstitial, and thin limb cells, and weak in the outer medulla and cortex, particularly in cells of the collecting duct, Bowman’s capsule, and the glomerular tuft (76, 77). AR protein was localized to podocytes and distal convoluted tubules (78). AR activity was detected in mesangial cells in culture (79), but not in vivo (77, 78). SDH mRNA expression was highest in the cortex, in Bowman’s capsule, glomerular tuft, collecting duct, peritubular endothelial and interstitial cells, and moderate in the papilla, in endothelial, epithelial, interstitial and collecting duct cells. In the outer medulla, it was low in proximal tubules and thick limb cells and strong in endothelial, distal tubular, and collecting duct cells (76, 77). Vascular smooth muscle cells (VSMCs) were positive for SDH, but glomeruli were all negative in control patients (77). In cultured rat mesangial cells, SDH activity was clearly detected (79). Na⁺/Myo-inositol cotransporter mRNA was detected predominantly in the papilla and inner medulla endothelial, epithelial, interstitial, and collecting duct cells; in the outer medulla collecting duct, endothelial, and thin limb cells; and in the cortex, Bowman’s capsule, glomerular tuft, collecting duct, and distal convoluted tubule cells (76).

3. In vitro evidence for AR effects on renal cells.
During hyperglycemia, glucose levels rapidly increase in tissues, such as the kidney, that are insulin-independent for glucose uptake. Excess glucose enters the polyol pathway and activates AR, but because SDH activity does not increase similarly, sorbitol accumulates (69). To explain the role of the polyol pathway in the onset of diabetic complications, different mechanisms have been proposed: accumulation of sorbitol or fructose (68, 69), myo-inositol depletion (70), or alterations in the NADPH/NADP⁺ and NADH/NAD⁺ ratios (80). In cultured rat mesangial cells, enhanced expression of the facilitative glucose transporter 1 increased AR expression and activity, polyol accumulation, and PKC levels, which may induce stimulation of matrix protein synthesis (81). In addition, in cultured mesangial cells, high glucose-induced DAG accumulation, PKC activation, and TGF-ß overproduction were mediated through polyol pathway activation (82, 83). In cultured proximal tubular cells, PKC and polyol pathway activation mediated the high glucose-induced collagen and fibronectin accumulation by decreasing their degradation (84, 85, 86). In proximal tubular cells, glucose increased sorbitol, fructose, DAG levels, PKC activity, and the expression of angiotensinogen (87), but not AR mRNA, immunoreactivity, and activity, which may be explained by a negative feedback of intracellular sorbitol accumulation on AR gene transcription (84, 88).

4. Experimental evidence for a role of AR in diabetic kidney disease.
Both short- and long-term experimental type 1 diabetes are associated with increased renal sorbitol and fructose levels (65, 89, 90, 91, 92, 93). Renal myo-inositol levels were found to be increased (92, 94) or decreased in diabetic animals (95). More specifically, renal polyol accumulation was associated with a decrease in myo-inositol levels in the kidney cortex while being unchanged in the kidney medulla (93). Renal AR level and activity were elevated in diabetic rats, but renal SDH level or activity did not change (92, 96). However, one study found a decrease in glomerular AR expression in type 1 diabetic rats after 2 wk of diabetes, whereas the expression of AR did not change in glomeruli and renal arterioles of type 2 diabetic rats (97). Na⁺K⁺-ATPase activity was reduced in glomeruli of diabetic animal models (95). In contrast, renal Na⁺K⁺-ATPase activity was unaffected in another study (93). Despite the discrepancy of these results, which may be due to variations in quantification methods, duration of diabetes, degree of hyperglycemia or animal strain, the majority of the published studies point toward activation of the polyol pathway and sorbitol accumulation in diabetes. In contrast to diabetic dogs showing nephromegaly, glomerular hypertrophy, mesangial expansion, BMT, and albuminuria, dogs fed galactose for 5 yr only developed thickening of the glomerular basement membrane (91). Kidney polyol accumulation was much more pronounced in galactose-fed dogs than in diabetic dogs, suggesting that the accumulation of polyol per se is not sufficient to produce diabetic glomerulopathy (91). Transgenic mice overexpressing human AR developed pathological changes in the kidney, i.e., thrombi of renal vessels and fibrinous deposits in Bowman’s capsules (98).

5. Clinical evidence for a role of AR in diabetic kidney disease.
Data about the polyol pathway and diabetic nephropathy in type 1 diabetic patients are rather scarce. In peripheral blood mononuclear cells obtained from type 1 diabetic patients with diabetic kidney disease, AR expression was increased compared with that in normal subjects, type 1 diabetics without nephropathy, and nondiabetic subjects with renal disease (99). Studies on AR gene polymorphisms have implicated the polyol pathway in the development of kidney complications, because the presence of the Z-2 allele or the T allele of the AR gene in type 1 diabetic subjects was associated with an increased risk for diabetic nephropathy (100, 101). Type 2 diabetic patients had higher serum and urine myo-inositol concentrations and sorbitol excretion than healthy controls (102, 103), but urinary fructose excretion was not different between the groups (103). Improvement of glycemic control reduced the urinary myo-inositol excretion (103). Serum myo-inositol concentrations were higher in patients with nephropathy than in those without nephropathy, but there were no differences in urinary myo-inositol concentrations (102). Renal AR expression was increased in type 2 diabetic patients compared with controls (77), especially in the glomerular mesangial area of patients with nephropathy (77). Reduced tubular SDH expression in diabetic patients was associated with interstitial fibrosis and thickened basement membranes (77).

6. Agents with effects on the AR system in diabetic kidney disease.
Numerous experimental and clinical studies with different ARIs have implicated the diabetes-induced polyol pathway activation in the development of diabetic retinopathy and neuropathy, but only a few studies have investigated the influence of ARIs in the diabetic kidney. Despite the large number of compounds active against AR in vitro, only two classes of ARIs can be reported regarding in vivo activity: cyclic imides (mostly spirohydantoins) and carboxylic acid derivatives with sorbinil and tolrestat being the most representative members, respectively (104). However, due to toxicity or a lack of efficacy, sorbinil, tolrestat, and ponalrestat have been withdrawn from clinical trials. In the following will be focused mainly on epalrestat, the only ARI still available (104).

a. ARIs.
Tolrestat blocked the glucose-induced increases in sorbitol, fructose, and DAG levels, PKC activity, and angiotensinogen expression in proximal tubular cells (87). Epalrestat abolished the glucose-induced increases in TGF-ß and PKC activity in cultured human mesangial cells, providing evidence for an interaction between these glucose-induced pathways (82). In type 1 diabetic rats, short-term oral administration of epalrestat prevented the increase in UAE and the reduction of anionic sites on the lamina rara externa of the glomerular basement membrane (105). In type 1 diabetic rats receiving epalrestat for 2 wk, renal cortical sorbitol concentration and the urinary enzyme excretion were not increased compared with untreated diabetic rats, suggesting that sorbitol accumulation leads to proximal tubular cell dysfunction and abnormal enzymuria (106). Long-term treatment with epalrestat ameliorated the decline of GFR and renal plasma flow (RPF) and mesangial expansion observed in placebo-treated diabetic rats, implicating the polyol pathway in functional and morphological changes in type 1 diabetes (107). In type 1 diabetes, 6 months of AR inhibition with ponalrestat or tolrestat reduced GFR in normo- and microalbuminuric patients, respectively (108, 109). However, 3 months of ponalrestat treatment had no effect on renal hemodynamics in patients with incipient nephropathy (110). In microalbuminuric type 2 diabetic patients, 5 yr treatment with epalrestat prevented the progression of incipient diabetic nephropathy (Fig. 2) (111). The development of new ARIs (112) may provide more potent and specific agents to block the polyol pathway, and future studies are warranted to elucidate their potential in the prevention of the development or progression of diabetic kidney disease.

View larger version (14K): [in this window] [in a new window]   FIG. 2. Changes in UAE of microalbuminuric patients treated with (•) and without () epalrestat. Vertical bars indicate means ± SEM. *, P < 0.01 vs. 0 yr. #, P < 0.05 vs. patients without epalrestat. [Reproduced with permission from K. Iso et al.: J Diabetes Complications 15:241–244, 2001 (111 ). © Elsevier.]

	III. Hemodynamic Factors

Top Abstract I. Introduction II. Metabolic Factors III. Hemodynamic Factors IV. Intracellular Factors V. Growth Factors and... VI. Outlook and Future... References

 
A. Angiotensin II (Ang II)/renin-angiotensin system (RAS)
1. The Ang II system.
Apart from the circulating RAS that regulates blood pressure, fluid, and electrolyte balance, many tissues, including the kidney, appear to have an independently regulated local RAS (113). Renin, produced in the juxtaglomerular cells of the kidney, converts angiotensinogen, derived primarily from hepatocytes, to inactive Ang I, which is then converted to biologically active Ang II by ACE (113). Ang II exerts hemodynamic and nonhemodynamic effects by binding to the Ang II type 1 (AT₁) or Ang II type 2 (AT₂) receptors. Most of the intrarenal actions of Ang II are mediated by AT₁ receptors that signal through activation of phospholipases and PKC and inactivation of adenylate cyclase (113). The role of AT₂ receptors in kidney has not yet been clearly defined, but involves the production of NO and prostaglandin F₂ and signal transduction through activation of guanylate cyclase, phosphatases, and potassium channels (113). AT₂ receptors might also be involved in mediating proliferation and apoptosis (114), and they counteract the effects of AT₁ receptor-mediated Ang II actions.

2. Expression of Ang II/RAS in the normal kidney.
All components of the RAS, i.e., renin, angiotensinogen, ACE, Ang II, AT₁ and AT₂ receptors are expressed in normal kidney (113). Renin mRNA was identified in cultured human mesangial cells and glomerular podocytes (115). In rat and human kidney, renin mRNA or protein expression was detected in glomerular endothelial cells and proximal and distal tubular cells (115, 116). Angiotensinogen mRNA was observed in cultured human glomerular podocytes and mesangial cells (115) and rat proximal tubule suspensions (116). In human kidney, a tubular angiotensinogen mRNA distribution was found (115). ACE mRNA was detected in cultured human mesangial cells and glomerular podocytes, in rat proximal tubules (116), and in tubules in human kidney (115). Most of the intrarenal Ang II is formed within the kidney (113). In rat kidney, AT₁ receptors were detected in afferent arterioles, arcuate arteries, vasa recta, glomerular mesangial cells and podocytes, proximal tubules, thick ascending limb, collecting ducts, and medullary interstitial cells (117). AT₂ receptors were localized in afferent arterioles, arcuate arteries, vasa recta, glomerular cells, proximal tubules, collecting ducts, and interstitial cells (117, 118, 119).

3. In vitro evidence for Ang II on renal cells.
High glucose stimulated the mRNA expression and protein synthesis of angiotensinogen in proximal tubular epithelial cells and involved activation of the polyol pathway and PKC (87). In cultured rat mesangial cells, high glucose increased Ang II production and AT₂ receptor expression and reduced the levels of aminopeptidase A, a metalloprotease that degrades Ang II, which could contribute to the increase of Ang II (120, 121). In cultured mesangial cells, the high glucose-induced inhibition of collagenase activity, reduction of MMP-2 levels, and increase in TGF-ß1 secretion, resulting in decreased matrix degradation and increased matrix accumulation, are mediated by Ang II (121). Ang II-induced ET-1 production in glomerular mesangial cells is partially PKC-dependent (122) and plays a role in the mitogenic effect of Ang II (123). Ang II induced connective tissue growth factor (CTGF) in cultured mesangial and tubular epithelial cells (124). Ang II induced apoptosis in cultured proximal tubular cells (125).

4. Experimental evidence for a role of Ang II in diabetic kidney disease.
Although the role of RAS in diabetic nephropathy is indisputable, data concerning the influence of diabetes on systemic and intrarenal RAS have been conflicting (126). In experimental type 1 and type 2 diabetes, plasma levels of RAS components have generally shown suppression of the system, and the local glomerular RAS seems to be activated, but data are variable, which may be due to the degree of hyperglycemia or the time after onset of diabetes (126). STZ-diabetic rats tended to have increased plasma and intrarenal levels of Ang II compared with control and insulin-treated rats (116). Studies investigating renin, angiotensinogen, and ACE mRNA in the proximal tubules (116) and glomeruli (119) in rats with STZ diabetes for 2 wk only showed an increase in renin mRNA in the proximal tubules. A down-regulation of cortical and proximal tubular AT₁ receptors (116) and of AT₂ receptors in all kidney regions (119) was reported, suggesting that alterations in the balance of kidney AT₁ and AT₂ receptors may contribute to Ang II-mediated diabetic glomerular injury. In diabetic (mRen-2)27 transgenic rats, diabetic renal pathology was associated with intense renin mRNA and protein in proximal tubules and juxtaglomerular cells along with overexpression of TGF-ß1 and collagen IV mRNA in glomeruli and tubules, as well as a declining GFR and UAE (127).

5. Clinical evidence for a role of Ang in diabetic kidney disease.
In type 1 diabetic patients, hyperglycemia was associated with increased plasma renin concentrations, whereas in patients with type 2 diabetes and nephropathy plasma renin levels were suppressed (126). However, measurements of circulating components of the RAS do not appear to accurately predict the state of activation of the local RAS (126). One study found stronger signals for renin, angiotensinogen, and ACE mRNA in mesangial and epithelial cells from hypertensive patients and patients with renal pathology, including some with diabetes (115). In renal biopsy specimens from type 2 diabetic patients, enhanced glomerular ACE staining was reported, especially in diabetics with glomerular nodular lesions (128). Another study reported reduced AT₁ receptor mRNA levels in kidney biopsy samples from type 2 diabetics with nephropathy when compared with controls (129). Several studies investigating polymorphisms of the ACE and angiotensinogen genes that may predispose people with diabetes to nephropathy have yielded varying outcomes. The TT genotype of the angiotensinogen M235T polymorphism has been associated with diabetic nephropathy in type 1 diabetes (130), whereas other studies in type 1 and type 2 diabetic subjects failed to find an association (131, 132, 133). The D-allele of the ACE insertion/deletion (I/D) polymorphism has been identified as a risk factor for increased progression of diabetic glomerulopathy in microalbuminuric type 1 diabetic patients (134). However, other studies in type 1 (132) and type 2 diabetic patients (133) did not support a role for ACE I/D polymorphism in the development of diabetic nephropathy. A metaanalysis of the ACE I/D polymorphism and risk of diabetic nephropathy, including 11 studies in type 1 and 10 studies in type 2 diabetic patients, suggested that the ACE I/D polymorphism contributes to the genetic susceptibility to diabetic nephropathy in Japanese but not in Caucasian type 2 diabetic patients (135). In Caucasian type 1 diabetic patients, comparison of data was complicated by differences between study populations, but a trend toward a protective effect of the II genotype on the development of increased UAE was observed (135). The risk for diabetic nephropathy in type 1 diabetic patients was similar in carriers and noncarriers of the AT₁ receptor A1166C polymorphism (132).

6. Agents with effects on the Ang II system in diabetic kidney disease.
The best known agents to target the RAS are ACEi and the newer ARBs. ACEi act in part by reducing plasma Ang II levels and increasing plasma bradykinin levels (136). Bradykinin accumulation has been associated with side effects, including cough and angioedema, but may also contribute to the antihypertensive and vasculoprotective effects of ACEi. ACEi do not provide a complete inhibition of Ang II activity, because Ang II can be formed through non-ACE pathways. ARBs might exert a more specific and complete blockade of the RAS through their inhibition of the actions of Ang II at the AT₁ receptor, but this results in higher circulating Ang II levels (137). The biological activity of Ang II or its metabolites may be more directed to other Ang II binding sites with potential adverse effects (138). In addition, ARBs raise bradykinin levels through stimulation of AT₂ receptors by elevated Ang II levels. Thus, ACEi and ARBs potentially have additive effects on both the RAS and bradykinin levels (136). The potential role of oral renin inhibitors that decrease Ang II levels in humans and of AS-ODNs designed to target renin, angiotensinogen, ACE, or AT₁ receptors in the therapy of diabetic nephropathy remains to be investigated.

a. ACEi.
ACEi consistently limited progressive renal injury in experimental models of STZ-induced diabetes (139). In Wistar fatty rats, an experimental model of type 2 diabetes with progressive kidney disease, enalapril ameliorated both existing proteinuria and the progression of proteinuria and preserved glomerular structure (5). In microalbuminuric type 1 diabetic subjects, perindopril reduced glomerular BMT and tended to reduce interstitial fibrosis (140), and enalapril prevented glomerular growth (141) and progression of glomerulopathy, indicating that ACE inhibition may influence renal structural changes (142). However, in a small cohort of normotensive type 1 diabetic patients with albuminuria and diabetic glomerulopathy, enalapril treatment did not affect renal structure (143). In normotensive normoalbuminuric type 1 diabetic patients, acute ACE inhibition by enalapril caused a decline in filtration fraction and in blood pressure but had no effect on UAE (144); 6 wk of captopril treatment had no effect on GFR and UAE (145), but treatment with enalapril for 3 months reduced UAE without effect on GFR or RPF (144). In normotensive microalbuminuric type 1 diabetic patients, UAE and blood pressure were reduced, whereas GFR was unaffected by long-term treatment with ramipril (146) or perindopril (147). In micro- and macroalbuminuric type 1 diabetic patients, acute ACE inhibition by enalapril did not change GFR, but it reduced blood pressure, UAE, filtration fraction, and renal vascular resistance, whereas RPF tended to rise (148). Six months of enalapril treatment induced similar changes except for GFR and RPF; GFR was reduced, whereas RPF remained unchanged (149). Enalapril was more effective in reducing UAE in proteinuric type 1 diabetic patients than antihypertensive patients (148). In normotensive type 1 diabetics with nephropathy treated for 8 yr with captopril, systemic blood pressure and albuminuria remained unchanged and only a minimal loss of GFR was seen when compared with patients who were not treated with captopril (150). However, captopril treatment was associated with a 50% risk reduction of the combined end points of death, dialysis, and transplantation (Fig. 3) (151). In macroalbuminuric type 1 diabetic patients with mild to moderate hypertension, treatment with enalapril reduced UAE, total protein excretion, and blood pressure, but had no effect on GFR and RPF (152). In microalbuminuric type 2 diabetic subjects, perindopril also reduced glomerular BMT and tended to reduce interstitial fibrosis (140). In micro- and macroalbuminuric type 2 diabetic patients with urinary podocyte excretion, trandolapril treatment reduced the number of podocytes excreted in the urine (153). The abnormalities in size-selective function of the glomerular barrier in type 2 diabetics with overt nephropathy were not ameliorated by ACEi or calcium channel blockade at variance to type 1 diabetes (154). Enalapril attenuated the decline in renal function and reduced the extent of albuminuria in normotensive, normo- and microalbuminuric patients with type 2 diabetes (155, 156). In hypertensive type 2 diabetic patients with normo-, micro-, or macroalbuminuria, treatment with enalapril was associated with a greater reduction in UAE than with nifedipine in the entire patient group, and especially in those with microalbuminuria. In the macroalbuminuric patients, the rate of deterioration in renal function was also attenuated by treatment with enalapril (157).

View larger version (29K): [in this window] [in a new window]   FIG. 3. Cumulative incidence of events in patients with diabetic nephropathy in the captopril and placebo groups. A, The cumulative percentage of patients with the primary end point: a doubling of the baseline serum creatinine concentration to at least 2.0 mg/dl. B, The cumulative percentage of patients who died or required dialysis or renal transplantation. The numbers at the bottom of each panel are the numbers of patients in each group at risk for the event at baseline and after each 6-month period. [Reproduced with permission from E. J. Lewis et al.: N Engl J Med 329:1456–1462, 1993 (151 ). © Massachusetts Medical Society.]

View larger version (18K): [in this window] [in a new window]   FIG. 4. Incidence of progression to diabetic nephropathy during treatment with 150 mg of irbesartan daily, 300 mg of irbesartan daily, or placebo in hypertensive patients with type 2 diabetes and persistent microalbuminuria. The difference between the placebo group and the 150-mg group was not significant (P = 0.08 by the log-rank test), but the difference between the placebo group and the 300-mg group was significant (P < 0.001 by the log-rank test). [Reproduced with permission from H.-H. Parving et al.: N Engl J Med 345:870–878, 2001 (159 ). © Massachusetts Medical Society.]

7. Conclusion.
Data regarding the renal expression of RAS components are inconsistent. Furthermore, the contribution of ACE or angiotensinogen polymorphisms to the development or progression of diabetic nephropathy remains controversial. Nevertheless, the beneficial effects of ACEi and/or ARBs on diabetic renal structural and functional changes beyond their blood pressure-lowering effect provides substantial experimental and clinical evidence for a deleterious role of RAS components in diabetic nephropathy. The combination of ACEi and ARBs seems to have an additive effect, at least on blood pressure and UAE. The effects of ACEi on renal structure and function may be partly explained by effects on growth factors and cytokines, metabolic pathways, and signaling molecules.

B. Endothelin (ET)
1. The ET system.
The ET system comprises the ETs, two receptors, and two activating peptidases. The ETs are a family of structurally and functionally related peptides, more in particular ET-1, the most potent vasoconstrictor known, ET-2, and ET-3. Preproendothelins, the ET precursors, are cleaved by endopeptidases to form inactive intermediates termed big ETs. Big ETs are then cleaved by ET-converting enzymes (ECEs) to form the final products (168, 169). In mammals, the two ET receptors, ET_A and ET_B, signal through the activation of G-proteins, phospholipase C, and PKC. ET_A receptors are involved in vasoconstriction and cell proliferation, whereas ET_B receptors binding all ETs mediate NO release and transient vasodilation (168, 169). ETs are normally not circulating hormones but act as paracrine and autocrine factors. The (patho)physiological roles of ETs in various organs have been described in detail (170). In the kidney, the ET system regulates renal hemodynamics, water and sodium homeostasis, cell proliferation, and matrix formation (168, 169, 170).

2. Expression of ET in the normal kidney.
ET-1, ET-2, and ET-3, ECE-1, and receptors ET_A and ET_B are present in the kidney (169, 171). ET-1 is found in all parts of the nephron. ET-1 is synthesized and secreted by cultured glomerular mesangial (122, 172, 173, 174), epithelial (175), and endothelial cells (176) and proximal tubular cells (177). In porcine kidney, the concentration of immunoreactive ET-1 was highest in renal inner medulla and very low in cortex (178). In rat kidney, ET-1 mRNA was demonstrated in cortical and medullary collecting ducts (179). In human kidney, mature ET and big ET-1 localized to the endothelium of glomeruli and intrarenal blood vessels (176). All three ET isoforms were detected by RT-PCR in human homogenates of renal medulla, cortex, and vessels (176). ET-3 is distributed differently along the nephron than ET-1 in rat kidney (180). ECE-1 localized to vascular endothelial and tubular epithelial cells in the cortex and medulla of human kidney (171). Both ET_A and ET_B receptor mRNAs were expressed in human mesangial cells (181), in rat glomeruli (182), and in normal human renal cortex and medulla with ET_B predominating (183).

3. In vitro evidence for ET effects on renal cells.
Various agents increase ET-1 production by mesangial cells, including hyperglycemia, PDGF, TGF-ß1, thrombin, Ang II, and ET-1 (181). ET-1 release is also increased by thrombin and bradykinin in glomerular endothelial cells (184) and by TGF-ß in collecting duct cells (185). Other factors such as shear-stress due to glomerular hyperfiltration (186) and urine flow (187) have been shown to stimulate ET-1 synthesis or release. Furthermore, AGEs induced ET-1 mRNA expression in cultured proximal tubular cells (188). High glucose modified the responses of mesangial cells to ET-1 (181). For example, hyperglycemia augmented ET-1-stimulated 1 (IV) collagen production and MAPK activity in mesangial cells, effects that are PKC dependent and associated with altered translocation of PKC and PKC (181, 189). ET-1 stimulates mesangial cell proliferation, contraction, and ECM accumulation (176).

4. Experimental evidence for a role of ET in diabetic kidney disease.
Studies examining systemic and intrarenal ET-1 levels in diabetic animals have yielded conflicting results. In STZ-diabetic rats, plasma ET-1 levels have been undetectable (190), unchanged (191), enhanced (187, 192), or suppressed (193) compared with control rats. Urinary ET-1 excretion was increased in STZ-diabetic rats, in parallel with enhanced proteinuria (139). Renal ET-1 levels were reported to be unchanged (190) or reduced (191) and glomerular ET-1 levels were enhanced (182) in STZ-diabetic vs. control rats. In early STZ-diabetic rats, no difference in glomerular ET_A receptor characteristics has been found; however, a reduction in ET_B density, as well as a reduction in glomerular Ang II receptor density has been reported (194). The mRNA levels for ET_A and ET_B did not change in STZ-diabetic rats (182). In type 2 diabetic rats, glomerular and tubulointerstitial ET-1 mRNA and protein expression was higher than in nondiabetic rats (195). The intrarenal ET system could be affected independently of the systemic ET-1 system. Discrepancies may be related to the degree of hyperglycemia, the renal localization, or varying duration of diabetes. Up-regulation of the renal ET system might favor the development of renal lesions, as suggested by in vivo studies in transgenic mice and rats. ET-1 transgenic mice with only slightly elevated tissue and plasma ET-1 concentrations developed glomerulosclerosis, interstitial fibrosis, renal cysts, and a progressive decline in GFR (196). ET-2 transgenic rats with a high renal transgene expression almost exclusively within glomeruli developed glomerulosclerosis and albuminuria without a change in GFR (197).

5. Clinical evidence for a role of ET in diabetic kidney disease.
Few clinical studies have been published on ET in diabetic kidney disease. In patients with uncomplicated type 1 diabetes, the plasma concentrations of ET-1 were increased and directly correlated to those of von Willebrand factor, and inversely correlated to plasma concentrations of fibronectin (198). Plasma ET-1 levels were higher in nonsmoking normotensive type 2 diabetic patients than in controls and higher in microalbuminuric vs. normoalbuminuric patients (199). Plasma ET-1 concentrations were increased in hypertensive type 2 diabetic patients as compared with healthy control subjects but were not different from those in normotensive patients (200). In contrast, plasma ET-1 levels were equally elevated in type 1 and type 2 diabetic patients and hypertensive nondiabetic patients compared with healthy controls, but the diabetic groups included patients with and without hypertension (201). Plasma ET-1 levels did not correlate with clinical parameters of diabetic disease progression in diabetic patients, but correlated with age and systolic blood pressure in healthy controls (201). In addition, plasma and urinary ET-1-like immunoreactivity values correlated positively with the severity of diabetic nephropathy in type 2 diabetics (202). In contrast, another study found no differences in the diurnal urinary excretion of immunoreactive ET among diabetic patients and patients with nondiabetic diseases (203).

6. Agents with effects on ET in diabetic kidney disease.
Several agents are available to inhibit the action of ET, including neutralizing antibodies against ET and ET receptors and AS-ODNs to preproendothelin-1, ECE, and the ET_A receptor. However, the potential use of these substances in the treatment of diabetic renal changes has not been explored, except for some in vitro studies. Ample experimental studies focused on the renoprotective effects of ET receptor antagonists. The ET system can be targeted indirectly by ACEi and PKC inhibitors.

a. Neutralizing antibodies.
Addition of a neutralizing anti-ET antibody to mesangial cell cultures markedly augmented the secretion and activation of MMP-2, whereas addition of exogenous ET-1 inhibited the synthesis of MMP-2. These results implicate ET as a potential factor in pathophysiological matrix turnover in the glomerulus (204).

b. AS-ODNs.
AS-ODNs targeting preproendothelin-1 mRNA were delivered into cultured human mesangial cells and inhibited ET-1 secretion and the cell proliferation. These results identify ET-1 as one of the autocrine growth factors of human mesangial cells that may be important in pathophysiological conditions characterized by mesangial proliferation (205).

c. ET receptor antagonists.
The effects of ETs can be blocked by the administration of nonselective ET receptor antagonists (bosentan, PD 142893, TAK-044, LU 224332), selective ET_A antagonists (FR139317, BQ-123, BMS-193884, LU 135252, PD156707, and EMD 94246), or ET_B antagonists (RES-701-1, BQ-788). PD 142893 reduced UAE in diabetic animals (206). In nondiabetic and STZ-diabetic (mRen-2)27 rats, oral administration of bosentan for 12 wk normalized systolic blood pressure and attenuated the diabetes-associated decline in GFR (127). Despite producing normotension, severe diabetic renal pathology was not prevented by bosentan, suggesting dissociation of ET, UAE, and hypertension from the structural injury in this diabetic model (Fig. 5) (127). Selective ET_A blockade with FR139317 has shown protective effects in experimental diabetic glomerulopathy (192). In STZ-diabetic rats, mRNA levels for various collagens, laminin, and growth factors were elevated (192) and glomerular proliferating cell nuclear antigen, c-myc, c-fos, and c-jun mRNA levels increased with progression of diabetic nephropathy (207). These changes were reduced by FR139317 (207), suggesting that ET may play a role in ECM production. Thus, ET inhibition may protect against diabetic glomerular injury, possibly by interference with growth factors and growth-related genes (192). LU 135252 prevented the Ang II-induced increase in tissue ET-1 content and functional ECE activity in rats (208). Furthermore, LU 135252 prevented renal histological alterations in STZ-diabetic rats, and decreased urinary ET-1 excretion, but had no effect on UAE (209). Both LU 135252 and LU 224332 reduced proteinuria and completely normalized the renal matrix expression of type IV collagen and fibronectin in hyperglycemic rats with STZ-induced diabetes (210).

View larger version (177K): [in this window] [in a new window]   FIG. 5. Histopathology of the kidney cortex in nondiabetic and diabetic transgenic (mREN-2)27 rats treated for 12 wk with either valsartan or bosentan. Sections are stained with periodic acid-Schiff. G, Glomerulus. A, In nondiabetic rats, there is minimal evidence of glomerulosclerosis. B, Diabetic rats have severe glomerulosclerosis, arteriolar hyalinization (asterisk). Valsartan-treated nondiabetic (C) and diabetic rats (D) and bosentan-treated nondiabetic rats (E) have minimal evidence of glomerulosclerosis. Bosentan-treated diabetic rats (F) have severe glomerulosclerosis and cortical interstitial fibrosis (arrow; magnification, x300). [Reproduced with permission from D. J. Kelly et al.: Kidney Int 57:1882–1894, 2000 (127 ). © Blackwell Publishing.]

e. PKC inhibition.
The ET-1-induced MAPK activation in mesangial cells is PKC dependent and associated with altered translocation of PKC and PKC (189). Infusion of the specific PKC inhibitor 1-(6-isoquinolinesulfonyl) piperazine reversed the down-regulation of ET receptors in association with normalization of PKC activity in STZ-diabetic rats (194).

7. Conclusion.
Various renal cell types can secrete ET-1 in vitro, which can be stimulated by glucose, Ang II, TGF-ß, and PDGF. Experimental data regarding plasma and renal ET-1 levels are conflicting, but experiments in transgenic animals clearly show that up-regulation of ET-1 or ET-2 may favor the development of structural and functional renal changes. In type 1 and type 2 diabetic patients, increased plasma ET-1 levels correlate with the severity of diabetic nephropathy. In vitro application of neutralizing ET antibodies and AS-ODNs indicates ET-1 as a potential factor in mesangial proliferation and matrix turnover. A strong argument in favor of ET-1 as a mediator of renal injury derives from studies with ET receptor antagonists in experimental diabetes. Independent of their blood pressure-lowering effect, ET receptor antagonists reduced renal ET-1 content, urinary ET-1 excretion, and the production of ECM proteins; lowered UAE; and reduced renal expression of TGF-ß and PDGF-B.

C. Nitric oxide (NO)
1. The NO system.
NO is a widespread signaling molecule that plays a major role in nearly every cellular and organ function in the body. The enzymatic formation of NO from L-arginine, generating L-citrulline as a coproduct, is catalyzed by NOS. Three mammalian NOS isoforms have been identified, neuronal (nNOS), endothelial (eNOS), and macrophage or inducible NOS (iNOS), more recently termed NOS1, NOS2, and NOS3, respectively (211). Their structure, function, and inhibition has been reviewed extensively (212). NO mediates its effects through activation of guanylate cyclase, resulting in increased levels of cyclic GMP (cGMP). In the kidney, NO is involved in the regulation of RPF, GFR, sodium excretion, extracellular fluid volume, and the maintenance of renal structural integrity (213). Endogenous inhibitors of NOS are N^G-monomethyl-L-arginine (L-NMMA) and asymmetric dimethylarginine (ADMA), substances naturally occurring in human plasma (214). ADMA is mainly cleared by the enzyme dimethylarginine-dimethylaminohydrolase (DDAH) and partly by renal secretion (215).

2. Expression of the NO system in the normal kidney.
All three NOS isoforms are present in the kidney (211). Their localization and that of alternative splice variants have been described in detail (216). The renal medulla is the main site for NO synthesis in the kidney (216). NOS1 mRNA and protein were detected abundantly in the macula densa and in some thick ascending limb cells in rat and human kidney (217, 218). In addition, NOS1 immunoreactivity has been demonstrated in the endothelium of glomerular efferent arterioles, inner medullary collecting ducts, and the parietal epithelium of Bowman’s capsule in rat kidney (217, 218). In rat and human kidney, NOS3 was typically detected in the endothelium of preglomerular, glomerular, and postglomerular vessels, whereas NOS2 was hardly detected (217, 219). In human glomeruli NOS3 mRNA was constitutively expressed, whereas NOS2 mRNA expression was barely found, and NOS1 mRNA was not detectable (220).

3. In vitro evidence for NO effects on renal cells.
Most studies in cultured renal cells have provided evidence that diabetes is a state of NO deficiency (211). For example, high glucose attenuated the detectable NO, blunted the NO responses to NOS3 agonists in human endothelial cells (221), and inhibited NO synthesis in murine mesangial cells (222). Possible mechanisms whereby high glucose decreases NO bioavailibility include NO capture by glucose (221), superoxide anion generation (223), endothelial ADMA accumulation (215), L-arginine depletion, and reduced tetrahydrobiopterin stability and availability (222). Other studies, however, in cultured mesangial cells, documented the stimulation of NO production in hyperglycemic conditions (224, 225), which may involve NOS2, PKC, and AR activation (224). NO deficiency as well as increased NO production may promote ECM accumulation in the mesangium (224, 226). The in vitro assessment of renal vascular reactivity in isolated rat kidneys and isolated renal arteries has provided more controversial evidence, suggesting increased, normal, or decreased NO synthesis and activity. Some of the discrepancies between different studies could relate to the method of preconstriction of isolated vessels (211).

4. Experimental evidence for a role of NO in diabetic kidney disease.
Research on the effect of experimental diabetes on the renal expression of NOS isoforms has also yielded conflicting results (reviewed extensively in Ref.211). At 7 d of STZ-diabetes, total cortical NOS activity was reduced and associated with decreased NOS1 mRNA and NADPH diaphorase staining in the macula densa, whereas cortical NOS3 was not different between diabetic and control rats (227). In contrast, after 2 wk of STZ-diabetes, NOS3 protein levels and immunostaining were elevated in glomerular endothelial cells and preglomerular vessels in diabetic vs. control rats (228, 229). However, another study found no differences in renal cortical expression of NOS isoforms or in diaphorase staining between control and diabetic rats (230). After 4 wk of STZ-diabetes, NOS1 immunostaining was modestly enhanced in macula densa (218), whole kidney NOS2 and NOS3 mRNA levels were unchanged (231), but NOS3 protein expression was increased, and NOS1 protein expression was unaltered in whole kidney samples (232). After 6 wk of STZ-diabetes, medullary NOS1 and NOS3 mRNA levels were elevated, immunostaining for NOS1 and NOS3 was increased in the proximal straight tubules and medullary thick ascending limb (233), and enhanced NOS3 immunostaining was also found in purified renal vascular trees (Fig. 6) (234). Data concerning the role of NO in glomerular hyperfiltration in experimental diabetes are also conflicting (reviewed extensively in Ref.211). The majority of studies in type 1 diabetic rats found glomerular hyperfiltration in the presence of increased urinary NO₂^– and NO₃^– (NOx) excretion, supporting the notion that increased NO synthesis contributes to diabetic hyperfiltration (228, 233, 234, 235, 236, 237, 238). The NO-donor glyceryl trinitrate induced renal vasodilation in control but not in diabetic rats, providing further evidence for a role of enhanced NO production and/or signaling that is not changed by additional NO (236). However, other studies suggested a defect in renal NO production and/or action in experimental diabetes. In rats with type 1 diabetes for 1 to 2 wk, increased CrCl and renal hypertrophy were accompanied by a decreased NOx excretion and unchanged plasma NO metabolites when compared with nondiabetic controls (227, 230). Reduced NO-dependent renal vasodilation (239, 240) and decreased glomerular NO-dependent cGMP generation (239, 241) were reported in diabetic rats compared with controls. The impaired glomerular NO-dependent cGMP generation may be mediated by thromboxane A₂ and PKC activation (241) or by an impaired activity of guanylate cyclase (231). NO has been generally considered as the principal mediator of endothelium-dependent vasodilation. Endothelium-dependent (i.e., NOS3-dependent) vasodilation is impaired in diabetic animal models and in humans with type 1 and 2 diabetes (242). In the renal circulation of STZ-diabetic rats in vivo, the vasodilation to intrarenal acetylcholine was reduced compared with control rats, suggesting impaired endothelium-dependent vasodilation (243). NO-dependent vasodilation was enhanced, whereas the vasodilation mediated by endothelium-derived hyperpolarizing factor was severely impaired (243).

View larger version (58K): [in this window] [in a new window]   FIG. 6. Immunocytochemical staining for eNOS in control (left) and diabetic (right) kidneys. Staining is detectable in glomerular capillaries and in afferent and efferent arterioles, visible at the vascular pole of the glomeruli. More capillary profiles are stained in the diabetic glomerulus. [Reproduced with permission from A. S. De Vriese et al.: Kidney Int 60:202–210, 2001 (234 ). © Blackwell Publishing.]