Genetic Risk Factors

The risk of nephropathy is strongly determined by genetic factors and only approximately 30% of patients with type 1 or type 2 diabetes will ultimately develop nephropathy. Genetic factors may directly influence the development of diabetic nephropathy and/or be clustered with genes influencing cardiovascular diseases. One support for genetic transmission is an experimental study showing that bone marrow–derived mesangial cell progenitors transmit diabetic nephropathy from donors with type 2 diabetes ( db/db mice) to naive, normoglycemic recipients. There is ongoing research through genomic screening and candidate gene approaches to better identify genetic loci for diabetic nephropathy susceptibility. Although some potential genes have been identified, linkage was only present in defined ethnic subpopulations and not in the majority of patients. An incomplete list of previously implicated genes is shown in Table 78.1 . The major problem with such studies is that diabetic nephropathy is a complex disease trait, meaning that several genes are likely involved and the inheritance patterns do not follow simple Mendelian rules. For example, a potential association between polymorphisms in candidate genes and the development and progression of nephropathy has been widely studied. These case-control studies are often problematic and clear guidelines for such polymorphism studies have been provided. The complexity of this genetic linkage analysis is exemplified by studies with controversial results investigating the insertion or deletion polymorphism of the angiotensin-converting enzyme (ACE) gene. Discrepancies could be explained by genetic heterogeneity and by small effects in limited cases in most studies (often <200 cases). The current opinion about the role of ACE gene polymorphism is that it may be associated with progression of disease in certain ethnic populations, but it is not a predictor of the development of diabetic nephropathy.

The introduction of Genome-Wide Association Studies (GWAS) is projected to bring new insights into the genetics of diabetic nephropathy. After scanning the entire genome, this method tries to identify single nucleotide polymorphisms (SNPs) associated with a particular disease, and thus can pinpoint areas of the genome involved in renal disease for example. So far, 2 GWAS have been conducted on patient populations with diabetic nephropathy. Pezzolesi et al. identified two loci that are significantly associated with renal disease in patients with type 1 diabetes: the FRMD3 (4.1 protein ezrin, radixin, moesin [FERM] domain containing 3) and CARS (cysteinyl-tRNA synthetase). These findings were replicated using samples from the Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) study. On the other hand, using a Japanese cohort, Maeda et al. suggest that ELMO1 (engulfment and cell motility 1 gene) and SLC12A3 (solute carrier family 12, member 3) may be associated with type 2 diabetic nephropathy. This was later confirmed by studies on African-American cohorts with type 2 diabetes as well as data from the GoKinD collection which includes Caucasians with type 1 diabetes. Because it casts a wide net, GWAS should increase our understanding of the pathology of diabetic nephropathy, and help identify novel therapeutic targets. This powerful method also promises to give clinicians and researchers a better understanding of the differences in disease progression and therapeutic responsiveness, paving the way towards a personalized medical practice.

More recently, the tools of genome-wide association studies have been applied to diabetic nephropathy. This method allows for a comprehensive genetic survey of the entire genome for chromosomal regions that are linked with a specific trait, in this case diabetic nephropathy. A genome scan for diabetic nephropathy in African-Americans identified susceptibility loci on chromosomes 3q, 7p, and 18q. Another scan in Pima Indians, a native American tribe with an alarmingly high incidence of early-onset type 2 diabetes, also identified linkage to diabetic nephropathy on chromosome 7. This powerful method, or its derivative technologies, may in the future more clearly identify the genetic risk for developing diabetic nephropathy. However, the genetic findings still need to be sussed out for their functional significance and explained in the larger pathophysiological context.

Stage I

The earliest renal manifestations in type 1 diabetes are nephromegaly and glomerular hypertrophy, which are accompanied by afferent arteriolar vasodilation, renal hyperperfusion (although some find that renal blood flow is reduced ), and glomerular hyperfiltration ( Fig. 78.1 ). Microscopically, there is thickening of the glomerular and tubular basement membranes. All these early functional and structural manifestations appear even if the patient is not destined to develop overt diabetic nephropathy. The UAE in this stage is normal (<30 mg/d or <20 µg/min) ( Fig. 78.1 ) but occasionally a transient increase in UAE (“transient microalbuminuria”) is present secondary to poor glycemic control or infection. Typically the blood pressure is below the hypertensive range (<140/90 mm Hg). The GFR is increased by 20 to 40% above normal values, with higher levels being frequently achieved when glycemic control is poor. Some studies suggest that patients with more renal hypertrophy or higher degrees of hyperfiltration (GFR >150 ml/min) are at increased risk for the future development of overt nephropathy. In patients with type 2 diabetes, this early stage is not readily discernible owing to the increased age but there may be modest degrees of nephromegaly and hyperfiltration.

Stage II

Incipient or latent nephropathy is defined by the appearance of microalbuminuria (UAE of 30–300 mg/day or 20–200 µg/min). Without intervention, UAE increases at the rate of 10% to 20% per year and is almost always accompanied by a steady rise in blood pressure. Hypertension (>140/90 mm Hg) is typically diagnosed one to two years after the appearance of microalbuminuria in type 1 diabetes. Microalbuminuria rarely develops before five years of disease duration (median, 10 years) in type 1 diabetes, but in type 2 diabetes microalbuminuria may be present at the time of diagnosis of hyperglycemia in up to 20% of patients with as many as 40% of patients having elevated blood pressure as well. This may be due to the fact that many of these patients have had impaired glucose tolerance for years before actually being diagnosed with diabetes, or it may reflect another disease causing increased UAE as part of the metabolic syndrome, most notably essential hypertension. Other possible causes of increased UAE in this population include renovascular hypertension, morbid obesity, and sleep apnea. Even during this clinically “silent” phase of the disease, there may be significant mesangial matrix expansion or diffuse glomerulosclerosis, further thickening of the glomerular and tubular basement membranes, and some degree of podocyte loss. The GFR may remain elevated or may decrease to within the “normal” range (100–120 ml/min) ( Fig. 78.1 ). Longitudinal studies have shown that the patient with microalbuminuria is at greater risk for the development of overt proteinuria (UAE >300 mg/d) compared with patients with normoalbuminuria. For example, type 1 diabetic patients with persistent microalbuminuria have approximately an 80% chance of developing established nephropathy within the ensuing five to seven years. In contrast, the percentage of patients with type 2 diabetes who progress to overt proteinuria is significantly lower, in part because of excess mortality due to cardiovascular events. Thus, the predictive value of microalbuminuria in patients with type 2 diabetes is much less clear than in type 1 patients.

Stage III

This stage is characterized by the development of overt proteinuria (total protein excretion >500 mg/day) or macroalbuminuria (UAE >300 mg/d) ( Fig. 78.1 ). In type 1 patients this occurs after an average of 15 years of diabetes. Hypertension is almost always present, and the worse the blood pressure control is, the more rapidly the GFR declines ( Fig. 78.4 ). If a renal biopsy were to be performed, the glomeruli would typically demonstrate diffuse glomerulosclerosis and/or nodular glomerulosclerosis, further podocyte loss with focal areas of foot process effacement, arteriolar hyalinosis in both the afferent and efferent arterioles, and variable degrees of tubulointerstitial fibrosis. Structure–function correlations indicate a highly significant inverse correlation between declining GFR and mesangial expansion. The progressive expansion of the glomerular mesangium causes a reduction in the glomerular filtering surface area. Nephron loss due to tubulointerstitial fibrosis is another major cause of the reduction in GFR. In the untreated patient, the GFR falls at a rate of about 1 ml/min/month, but this rate of fall can vary significantly from patient to patient. As the GFR falls from previously supernormal levels, the serum creatinine may remain in the normal range or be slightly elevated and is therefore not a reliable indicator of the magnitude of disease progression.

Stage IV

After approximately five years of overt nephropathy, untreated patients progress to advanced nephropathy, as characterized by nephrotic-range proteinuria (>3.5 g/d), worsened hypertension that becomes difficult to control, and a progressive decline in GFR ( Fig. 78.1 ). In fact, diabetic nephropathy is the most common cause of the nephrotic syndrome in the adult population. Parenchymal and vascular lesions become more severe. The rate of decline in GFR is steady over a period of months but is variable from patient to patient and depends on the degree of elevation of blood pressure as well as the amount of UAE.

Stage V

The final stage is progressive renal failure reaching ESRD, with the GFR declining to 15 ml/min or lower. The patient may soon need renal replacement therapy to control uremic symptoms or excessive volume expansion. The average time for progression to ESRD from the time of diagnosis of diabetes is about 20 to 25 years, with a more rapid course developing in patients with uncontrolled hypertension and/or heavy proteinuria. Many patients, especially those with type 2 diabetes, never reach ESRD because of the severely increased risk of cardiovascular mortality in this population.

Early Hemodynamic Alterations in the Glomerulus

Altered renal hemodynamics is an early characteristic feature of diabetes in humans as well as animal models. It is widely held that glomerular capillary hypertension in diabetes is the major hemodynamic alteration that contributes to progressive glomerular injury. The increase in glomerular capillary pressure is accompanied by increased glomerular blood flow, which is caused by afferent arteriolar dilation but with little or no dilation of the efferent arteriole. An imbalance of a variety of vasoactive and growth factors including the renin-angiotensin-aldosterone system (RAAS), atrial natriuretic peptide, insulin-like growth factor-1, endothelin, prostanoids, and eicosanoids has been implicated in diabetic hyperfiltration, but evidence strongly implicates the NO system as the main mediator for afferent arteriolar dilation. Three major theories have been put forward to account for the hemodynamic changes in the glomerulus: (1) a primary alteration in vascular function, (2) a primary alteration in tubular function, and (3) a primary growth of the total filtration surface area (part of the hypertrophic response), mediated by endothelial cell proliferation, capillary elongation, and new capillary formation.

It has been well characterized that diabetes is associated with impaired autoregulation at the afferent arteriolar level. This vascular theory suggests that vascular smooth cells, mesangial cells, and endothelial cells are primarily responding to a combination of high-glucose concentrations, local autacoids, and systemic signals to alter the normal autoregulatory response to the prevailing systemic pressure. Studies supporting a primary vascular role include impaired calcium transients in afferent arterioles from diabetic rats. Numerous groups have found a persistent altered responsiveness to vasoconstrictors of vascular smooth muscle cells and mesangial cells obtained from diabetic rats. Some of these studies, using mesangial and vascular smooth muscle cells cultured in high glucose or obtained from diabetic rats, have identified protein kinase C (PKC), reactive oxygen species (ROS), and TGF-β to be important mediators of vascular dysfunction.

A case for tubuloglomerular feedback as the initiating factor has been put forward by Blantz and coworkers. These investigators have convincingly demonstrated that increased uptake of glucose and sodium in proximal tubular segments may limit sodium delivery to the macula densa, thus inhibiting tubuloglomerular feedback and preventing constriction of the afferent arteriole. Presumably the enzyme ornithine decarboxylase plays an important role in this pathway. Studies in diabetic animals and patients with salt loading lend further evidence to this hypothesis. NaCl restriction would cause a decrease in GFR in the normal situation; however, in animals with longstanding diabetes and in diabetic humans, salt restriction causes a surprising increase in GFR. This salt paradox could be explained by the further decrease in salt delivery to the macula densa and further afferent arteriolar dilation. However, several studies report unimpaired tubuloglomerular feedback in diabetes; one group suggested that enhanced tubuloglomerular feedback may mitigate the increase in GFR. It is likely that much of the discrepant results are accounted for by variations in diabetes induction, degree of hyperglycemia, weight loss, insulin levels, and duration of diabetes. Without a standardized experimental approach it is virtually impossible to interpret the various results.

Regardless of the cause of hemodynamic alterations in diabetes, progressive renal injury eventually ensues. One potential scenario to explain this outcome has been put forward by Kriz and coworkers. With increased glomerular capillary pressure, there is stretching of mesangial cells with loss of tethering to the GBM. Loss of tethering may be contributed by altered integrin expression and growth factor production. This would lead to a ballooning of the capillary loop and denuding of the GBM on the epithelial side. Alterations in GBM composition may also play a role. Dropout of podocytes may ensue due to abnormal stretching of podocytes and loss of adherence to the GBM. Furthermore, abnormal stretch may stimulate TGF-β production by mesangial cells, leading to a sclerotic response. This scenario would fit with the glomerular volume increase noted in experimental and human diabetes as well as the diffuse mesangial matrix expansion and podocyte dropout.

Enhanced tubular transport of solutes and water is correlated with glomerular hyperfiltration. The elevation in the glomerular transcapillary hydraulic pressure gradient as well as the increase in glomerular plasma flow leads to an increase in GFR. This, in turn, enhances the colloid osmotic pressure in postglomerular capillaries, which can facilitate the reabsorption of water and sodium in proximal tubules. These processes provide a mechanistic link between enhanced tubular transport and the primary abnormality of glomerular hyperfiltration. As an alternative explanation, a primary abnormality in sodium reabsorption has been linked to glomerular hyperfiltration. This explanation suggests that an increase in reabsorption of sodium chloride in the proximal tubule or the loop of Henle leads to an increase in GFR by an intact macula densa mechanism. Diabetes-induced hypertrophy of tubules that mediate stimulated sodium chloride reabsorption could be pivotal in this process, again connecting the structural changes with the hemodynamic adaptation in diabetes. Therefore, both mechanisms could explain the increase in tubular reabsorption that occurs in diabetic nephropathy.

Role of Glucose Uptake and Metabolism

Extracellular hyperglycemia may define the diabetic state but it is the high intracellular glucose concentration that appears to be the critical metabolic abnormality that promotes pathological changes in diabetic nephropathy. Renal cells do not require insulin for glucose uptake but instead rely on a family of transmembrane proteins to facilitate glucose transport across the cell membrane. Glucose diffuses down its concentration gradient, and in diabetes, renal intracellular glucose levels rise in proportion to the degree of hyperglycemia. The most widely expressed glucose transporter is the GLUT-1 isoform, a high-affinity, low-capacity, facilitative transporter that typically would be saturated at or near physiological glucose concentrations. GLUT-1 is normally found in the glomerulus and the tubules, but diabetes changes its distribution and cellular expression. While most tissues downregulate GLUT-1 expression in the face of hyperglycemia to protect cells from excess glucose transport and metabolism, some renal cells such as mesangial cells actually upregulate GLUT-1 gene transcription and protein translation when cultured in high-glucose media. This positive feedback is an apparent maladaptive response in mesangial cells, because the cells can readily take up more glucose from a diabetic environment, resulting in high intracellular glucose concentration that then initiates signaling cascades, which are integral to the pathophysiology of diabetic kidney disease. In fact, mesangial cells that overexpress GLUT-1 demonstrate avid uptake of glucose even when cultured in normal concentrations of glucose; the cells also behave like wild-type mesangial cells incubated in high glucose in that they increase the synthesis of extracellular matrix proteins such as collagens and fibronectin. Interestingly, several relevant factors in the diabetic state in addition to hyperglycemia can upregulate GLUT-1 expression on the surface of mesangial cells. These factors, which include TGF-β, Ang II, and shear stress, can then stimulate net glucose uptake and intracellular metabolism and therefore promote glucotoxicity.

The Polyol Pathway

Metabolic pathways for intracellular metabolism of glucose are myriad. Activation of the polyol pathway has been shown to be important in the development of some complications of diabetes mellitus such as cataracts, retinopathy, and neuropathy, but its involvement in the development of nephropathy remains much less established. Many of the early, but not late, features of diabetic renal disease can be attributed to activation of the polyol pathway. Aldose reductase, the first and rate-limiting enzyme in this pathway, catalyzes the NADPH-dependent reduction of hexose or pentose sugars to their corresponding sugar alcohols, or polyols. Tissues that do not require insulin for glucose uptake (kidney, lens, retina, and peripheral nerves) become subject to relatively greater loads of intracellular glucose. Increased oxidation of sorbitol to fructose via fructose dehydrogenase is coupled to reduction of NAD ⁺ to NADH, and a more reduced cytosolic ratio of NADH/NAD ⁺ may result in abnormalities of cellular function, including myo-inositol depletion, ROS generation, PKC stimulation, and even TGF-β production. Glomerular hyperfiltration in diabetic rats can be ameliorated by the administration of sorbinil, an aldose reductase inhibitor. In a small randomized trial of type 1 normoalbuminuric diabetic subjects, ponalrestat (a carboxylate-containing aldose reductase inhibitor) reduced renal hyperfiltration. Urinary albumin excretion in diabetic rats has also been reduced by aldose reductase inhibitor therapy. Feeding sorbinil to diabetic rats has been reported to decrease the width of the GBM, and the stimulation by high glucose of collagen type IV expression in cultured proximal tubular cells can be abolished with sorbinil, although cellular hypertrophy is not prevented. However, the long-term effects of polyol-pathway inhibition on renal extracellular matrix accumulation have not been uniformly favorable. The aldose reductase inhibitor statil did not reduce the magnitude of glomerulosclerosis in diabetic rats after seven months despite normalization of erythrocyte sorbitol levels with this drug.

The sorbitol pathway may also produce metabolites capable of nonenzymatically glycating intracellular proteins (see subsequent sections), which can be further phosphorylated via novel pathways that are activated in diabetes mellitus. Sorbitol-3-phosphate, fructose-3-phosphate, 3-deoxyglucosone, and other unidentified metabolites are increased in erythrocytes of diabetic subjects and could participate in protein glycation and cross-linking. Such studies provide an intriguing pathogenetic link between the products of the polyol pathway (or other pathways of glucose metabolism) and the reactions of nonenzymatic glycation.

The Hexosamine Biosynthetic Pathway

Fructose-6-phosphate can be converted to glucosamine-6-phosphate, with glutamine donating its amido group to become glutamate. The final products of the pathway are uridine diphosphate (UDP)-N-acetyl-glucosamine (GlcNAc) and other nucleotide hexosamines. The amination of fructose-6-phosphate is rate limiting and is catalyzed by glutamine: fructose-6-phosphate amidotransferase (GFAT). Interestingly, high glucose, superoxide production, and Ang II can all stimulate GFAT activity in endothelial or mesangial cell cultures. Also, the high-glucose-induced stimulation of TGF-β and fibronectin that is observed in mesangial cells is partly mediated by activation of the hexosamine pathway because treatment of the cells with the GFAT inhibitor azaserine or with GFAT antisense oligonucleotides diminishes this stimulation. The mechanism by which increased flux through the hexosamine pathway induces gene transcription is likely through covalent modification of key transcription factors and signaling molecules. For instance, hyperglycemia-induced mitochondrial superoxide overproduction activates the hexosamine pathway, which stimulates Sp1 activity by post-translational O-linked glycosylation at critical serine or threonine sites; this activation of the Sp1 transcription factor then leads to induction of key prosclerotic factors including TGF-β1 and plasminogen activator inhibitor-1 (PAI-1). The hexosamine pathway can also regulate vascular function by inhibiting endothelial NO synthase (eNOS) via O-linked glycosylation modification of a key signaling enzyme, the Akt/PKB.

Signaling through Protein Kinase C

The signaling pathways that mediate some of the effects of hyperglycemia may involve activation of one or more isoforms of PKC. Based on a series of ¹⁴ C-radiolabeling experiments, the relationship between hyperglycemia and PKC activation has been ascribed to an increased intracellular metabolism of glucose that promotes the de novo formation of diacylglycerol, the major endogenous activator of the classic and novel isoforms of PKC. Interestingly, diacylglycerol synthesis may also be linked to the glucose-induced increase in the ratio of NADH/NAD ⁺ that derives from the polyol pathway. PKC activity, measured by the cytosol-to-membrane shift of various classical PKC isoforms, is increased in mesangial and endothelial cells incubated in high-glucose media. Glomeruli from diabetic rodents also display elevated PKC activity and cytosol-to-membrane translocation of diacylglycerol-sensitive PKC isoforms. Increased activity of PKC in glomerular mesangial cells may lead to an increase in extracellular matrix expression, such as fibronectin, laminin and type IV collagen.

Specific blockade of one particular PKC isoform in vivo has been made possible with the engineering of a safe and orally active PKC-β inhibitor Ruboxistaurin (LY333531). In diabetic rats this compound ameliorates the renal hemodynamic abnormalities, decreases the proteinuria, and prevents the glomerular overexpression of TGF-β and the matrix components, fibronectin and type IV collagen. A randomized, double-blind, placebo-controlled, pilot study conducted on patients with type 2 diabetes receiving renin-angiotensin system inhibitors showed that the addition of Ruboxistaurin reduced albuminuria relative to the patients’ baseline. However, when comparing the treatment and placebo arms of the trial, the changes in the albumin-to-creatinine ratios were not significant. Other PKC isoforms, particularly PKC-α, may also mediate the renal lesions in diabetes, but this possibility will have to await the design of selective isoform inhibitors. The actual molecular targets downstream of PKC activation remain unclear, but they somehow stimulate other signal transduction pathways (such as MAPKs, PI3Kinases and various growth factors). For instance, markedly elevated ambient glucose levels in mesangial cells can activate p38, a MAPK family member, but they do not significantly affect JNK activity. One member of MAPK family, ERK, is markedly activated in mesangial cells cultured under high-glucose conditions and in glomeruli of diabetic rats, and this activation is mediated through a PKC-dependent mechanism. The activated ERK participates in the induction of TGF-β and the production of extracellular matrix.

Oxidative Stress as a Common Initiator

Generation of ROS has received much attention with regard to the initiation of diabetic microvascular complications, including kidney disease. As espoused by Brownlee, increased ROS generation is a consequence of persistent exposure to high glucose and this process involves increased mitochondrial oxidation as well as decreased scavenging of ROS. Glycolysis of intracellular glucose results in production of NADH and pyruvate. Mitochondrial NADH and FADH ₂ provide energy for ATP production through oxidative phosphorylation via the electron transport chain. Electron transfer through the mitochondrial complexes generates ATP synthase via a proton gradient. With increased electron flux, a high-proton gradient, and a high electrochemical potential difference, superoxide generation is enhanced. In response to elevations in extracellular glucose, aortic endothelial cells generate intracellular ROS primarily via mitochondrial NADH and FADH ₂ that are generated by the tricarboxylic acid cycle from cytosolic pyruvate. Subsequent superoxide generation initiates DNA strand breaks leading to poly (ADP-ribose) polymerase (PARP)–mediated glyceraldehyde phosphate dehydrogenase (GAPDH) inhibition. Inhibition of GAPDH may then mediate PKC activation, hexosamine formation, advanced glycation end products (AGEs) formation, and NF-κB activation. Studies in human mesangial cells have shown a host of similar changes related to mitochondrial ROS generation. An interesting finding is that high-glucose stimulation of mitochondrial ROS mediates COX-2 upregulation in human mesangial cells.

Oxidative stress has been clearly demonstrated in several animal models of diabetic kidney disease. Markers for this state are often reflected in decrements of glutathione or increased urinary 8-isoprostane. Inhibition of ROS with α-lipoic acid or via overexpression of a cytosolic superoxide dismutase prevents diabetic kidney disease in rodent models. Inasmuch as the accumulated data show that mitochondrial generation of ROS occurs in endothelial cells, this mechanism likely plays an important role in the generalized endothelial dysfunction of diabetes, including the development of albuminuria. Apart from mitochondrial ROS, a membrane-bound NADPH oxidase may potentially contribute to high-glucose-induced ROS production. Studies have identified a novel form of NADPH oxidase, NOX-4, in the kidney, which is upregulated in animal models of diabetic kidney disease. The role of NOX-4 in diabetic kidney disease is starting to be uncovered. Administration of antisense oligonucleotides inhibiting NOX-4 have been shown to decrease glomerular hypertrophy, and resulted in decreased fibronectin expression in renal cortex and notably in the glomeruli. Moreover, NADPH oxidases have been localized to podocyte, and the NOX-1 and NOX-4 isoforms are upregulated in hyperglycemic milieu both in vitro and in vivo . The accumulation of ROS from mitochondrial and non-mitochondrial pathways in the podocyte builds up into apoptotic signals such as the p38 MAPK pathway. Accordingly oxidative stress could be contributing to the pathological changes in diabetic nephropathy, including podocyte loss and albuminuria. The subcellular source of ROS in diabetic kidney disease remains to be clarified. Also, the role of inhibition of discrete pathways of ROS generation is an exciting area of research that can help dictate future therapies for diabetic nephropathy.

Nonenzymatic Glycation of Proteins

Glycated proteins arise from a condensation reaction, driven by the ambient glucose concentration, in which a free sugar covalently attaches to a protein at reactive NH ₂ groups. Glycation proceeds through the formation of a labile Schiff base adduct, which then undergoes an intramolecular Amadori rearrangement to become a stable glucose-modified protein. The reaction occurs slowly, and the degree and duration of hyperglycemia influence the amount of glycated protein. Amadori-modified proteins may further evolve through a series of spontaneous rearrangement, dehydration, and polymerization reactions to become advanced glycation end products (AGEs). Nonenzymatic glycation can be mediated by a large variety of carbohydrates including glucose, and by methylglyoxal and 3-deoxyglucosone. In fact, methylglyoxal and 3-deoxyglucosone are far more reactive than glucose, forming adducts with amino groups of proteins, nucleic acid, and phospholipids up to several thousand times more readily than glucose. Although the chemical reactions involved are nonenzymatic, the production and detoxification of AGE precursors is actually controlled by enzymatic mechanisms. Thus the accelerated formation of glycated proteins in diabetes may result from increased production and/or decreased degradation. A large body of research has been devoted to understanding the role of glycated proteins in the pathogenesis of diabetic complications.

Renin-Angiotensin-Aldosterone System

Data from animal studies suggest that local angiotensin-generating systems exist in many tissues including the kidney, and these systems may operate independently of the systemic RAAS. Micropuncture and microdialysis experiments in normal rats have demonstrated that proximal tubular and interstitial fluid contain Ang II concentrations in the nanomolar compared with picomolar range in the systemic circulation. These studies suggest that tubular cells possess a RAAS and can produce Ang II. Significant activation of the proximal tubular RAAS is presumed to occur in the diabetic state. Studies in early STZ-induced diabetes have shown enhanced expression of renin mRNA, which contributes to this local RAAS activation. It is hypothesized that the increased local production of Ang II could contribute to tubulointerstitial injury in diabetes.

Of particular importance is the observation that systemic application of an ACE inhibitor results in almost complete inhibition of systemic Ang II formation but has only little effect on intrarenal Ang II production. Also of interest is that intact Ang II is intracellularly present in endosomes and is derived from receptor-mediated endocytosis that follows Ang II binding to its putative receptor. This could be an important mechanism because it has been demonstrated in certain cells that Ang II may directly go into the nucleus and could regulate gene transcription.

Ang II is metabolized by peptidases such as aminopeptidase A (APA) into angiotensin III and further into angiotensin IV (Ang IV). Ang IV binds to a specific receptor named AT4. This receptor is widely expressed in the kidney including endothelial, proximal tubular, and distal convoluted tubular cells. Ang IV stimulates PAI-1 expression in proximal tubular and endothelial cells through AT4 receptors. Since PAI-1 reduces extracellular matrix turnover, Ang IV may induce renal fibrosis independently of activation of AT1 and AT2 receptors. Moreover, Ang IV-generating enzymes are upregulated in conditions with presumed high levels of Ang II in the kidney such as diabetic nephropathy, likely shifting more Ang II into the degradation pathway to Ang IV.

Drugs interfering with the RAAS are a mainstay of therapy in preventing the progression of diabetic nephropathy as first reported by Zatz and coworkers in experimentally-induced diabetes. Although initially considered to act solely through normalization of systemic and glomerular hypertension, it is now clear that inhibition of the RAAS has many effects, including antifibrotic and anti-inflammatory mechanisms. In fact, Ang II itself induces in renal cells many proinflammatory and profibrogenic cytokines, chemokines, and growth factors. High glucose stimulates expression of renin and angiotensinogen in mesangial and tubular cells. This stimulation results in an increase in local Ang II concentrations that may, in turn, through autocrine and paracrine pathways, induce a whole battery of different cytokines and growth factors. Experimental studies indicate that high-glucose–mediated generation of ROS is important in the upregulation of angiotensinogen in proximal tubular cells.

Inhibition of the RAAS significantly reduces proteinuria in diabetic nephropathy as compared with other antihypertensive classes of medications. Treatment with ACE inhibitors or AT1-receptor antagonists attenuate podocyte foot process broadening in STZ-diabetic rats. An AT1-receptor antagonist, but not the calcium channel blocker amlodipine, normalized the reduced nephrin expression in podocytes from spontaneously hypertensive rats with superimposed STZ-induced diabetes. Thus, a local increase in Ang II leads to suppression of nephrin expression in podocytes, and thereby can enhance the ultrafiltration of proteins. Podocytes express AT1 and probably AT2 receptors and could respond to stimulation with Ang II. Transgenic rats with targeted overexpression of the AT1-receptor to podocytes showed pseudocysts in podocytes, followed by foot process effacement and local detachments. These changes subsequently progressed to focal segmental glomerulosclerosis.

Podocytes are also a likely site for Ang II generation. It has been previously shown that high ambient glucose concentration induces Ang II formation in podocytes through upregulation of angiotensinogen expression. Furthermore, proteinuria and the transit of proteins through the ultrafiltration barrier likely activate Ang II formation in podocytes. Finally, mechanical stretch could increase Ang II generation in podocytes. Interestingly, Ang II formation as a consequence of mechanical stress appears to be independent of ACE. In this regard, it has been demonstrated that chymase, an Ang II forming enzyme not inhibited by ACE inhibitors, is upregulated in glomeruli of patients with nephropathy due to type 2 diabetes. This observation suggests that local glomerular formation of Ang II in the diabetic state may be partly independent of ACE and may not be abolished by ACE inhibition alone.

Activation of the RAAS may play an important role in macrophage recruitment. Biopsy studies from patients with diabetic nephropathy and investigations in various animal models have revealed the presence of inflammatory cells, especially mononuclear cell infiltrates, in the glomerular and tubulointerstitial compartments. Monocyte chemoattractant protein-1 (MCP-1) is an important chemokine for macrophages/monocytes, and high glucose has been shown to stimulate MCP-1 expression in mesangial cells. In fact, increases in tubular expression of MCP-1 and RANTES, members of the C-C chemokine subfamily with chemoattractant properties for monocytes/macrophages, have been found adjacent to infiltrating immune cells in renal biopsy specimens from patients with diabetic nephropathy. The MCP-1 increase could be relevant in another way that does not involve macrophages. MCP-1 directly stimulates matrix expression in mesangial cells and increases cellular motility in podocytes, having the associated effect of making a podocyte monolayer more permeable to albumin. Extrapolated to diabetic nephropathy, these MCP-1 effects may abet the process of mesangial matrix expansion and diabetic albuminuria. The proinflammatory transcription factor NF-κB has been detected mainly in tubular cells in biopsy specimens from 11 patients with type 2 diabetes and overt nephropathy, indicating that proteinuria may have contributed to this activation. Activation of the RAAS may play an important role in this process. In animal models of kidney disease there is upregulation of NF-κB in tubular cells that is inhibited by inhibitors of the RAAS. Ang II, through an effect mediated by AT2 and not AT1 receptors, increases the mRNA and protein levels of RANTES in cultured glomerular endothelial cells. Intraperitoneal infusion of Ang II into naive rats for 4 days significantly stimulates glomerular influx of macrophages/monocytes as well as the expression of RANTES mRNA and protein in the glomeruli, mainly in endothelial cells and small capillaries. Ang II also stimulates the chemoattractant cytokines MCP-1 and osteopontin in mesangial and tubular cells, which may play a role in glomerular and tubular inflammation associated with proteinuria.

Aldosterone, working independently from the RAAS, may be involved in the development of diabetic nephropathy. The aldosterone antagonist spironolactone attenuates the enhanced TGF-β1 expression and the increased collagen deposition in rats three weeks after STZ administration. In a preliminary study, Schjoedt and colleagues observed that an increased plasma aldosterone level during long-term treatment with an AT1-receptor blocker is associated with a decline in GFR in patients with nephropathy due to type 1 diabetes. These data demonstrate that aldosterone contributes to the progression of diabetic nephropathy despite blockade of the AT1 receptor. Spironolactone decreases proteinuria in patients with type 2 diabetes and early nephropathy. The more selective aldosterone antagonist, eplerenone, also decreases the microalbuminuria in patients with type 2 diabetes independent of its antihypertensive effect. In addition to direct effects on renal cells, aldosterone also potentiates the Ang II-mediated signal processes such as MAPK activation, indicating that Ang II and aldosterone act in concert.

In summary, there are many deleterious effects of the RAAS in the pathophysiology of diabetic nephropathy. Early treatment with drugs that interfere with RAAS activation, especially ACE inhibitors and AT1 receptor blockers, is a major mainstay in the clinical management of the diabetic patient who is at risk of developing nephropathy.

TGF-β is the Common, Downstream Mediator of Diabetic Nephropathy

The cytokine TGF-β has risen to prominence in the past several years as the principal mediator of progressive renal diseases including diabetic nephropathy. TGF-β has been shown to be the final common pathway or mediator that leads to the hypertrophic and prosclerotic changes in diabetic kidney disease, thus fulfilling all of Koch’s postulates as a chief etiologic agent of the disease. TGF-β stimulates the synthesis of key extracellular matrix molecules including type I collagen, type IV collagen, fibronectin, and laminin. TGF-β also decreases matrix degradation by inhibiting proteases as well as activating inhibitors of those proteases (e.g., PAI-1). Additionally, TGF-β promotes cell–matrix interactions by upregulating integrins, the cell-surface receptors for matrix.

The biologic effects of TGF-β in kidney cells, which include cell hypertrophy and stimulation of extracellular matrix production, closely resemble those of hyperglycemia. Almost all of the molecular mediators and intracellular signaling pathways that have been identified in diabetic kidney injury have also been found to stimulate the renal TGF-β activity as an intermediary step (see also Table 78.4 ). These include high-glucose concentration; early (Amadori) as well as AGE-modified proteins; oxidative stress and overproduction of superoxide by the mitochondrial electron-transport chain; cyclical stretch/relaxation of mesangial cells in culture; de novo synthesis of diacylglycerol and PKC activation; stimulation of MAPK, such as ERK; glucosamine overproduction; and high levels of vasoactive substances such as Ang II, endothelin, and thromboxane.

Studies in animal models of both type 1 and type 2 diabetes further implicate TGF-β as an important mediator of diabetic kidney disease. TGF-β1 mRNA and protein levels are increased in both the glomerular and tubular compartments of rat and mouse models of diabetes. Of particular significance is the finding of concomitant upregulation of the renal TGF-β type II signaling receptor in tubules and glomeruli. In almost all renal cell types studied in tissue culture (except perhaps for podocytes), high ambient glucose has been shown to upregulate the expression and bioactivity of TGF-β; these include proximal tubular epithelial cells, glomerular mesangial and endothelial cells, and interstitial fibroblasts. In all renal cells studied, including podocytes, high glucose also upregulates the TGF-β type II receptor.

Human studies have also corroborated the important role that the renal TGF-β system plays in the pathogenesis of diabetic nephropathy. Glomerular TGF-β1 mRNA is markedly increased and there is net production of TGF-β1 protein by the kidney in diabetic patients. Interestingly, therapy with an ACE inhibitor protects the kidney through a host of mechanisms that also include the lowering of TGF-β1 production. In fact, a decrease in the circulating TGF-β1 level predicts a better long-term preservation of the GFR.

Blocking the activity of the renal TGF-β system in diabetic animals has provided proof-of-concept that the development of renal disease is due to overactivity of this system in diabetes. Neutralizing monoclonal antibodies to TGF-β prevent the glomerular hypertrophy and attenuate the increase in TGF-β1, α1(IV) collagen, and fibronectin mRNAs in STZ-induced diabetes in mice. Also, the antibody therapy in the db/db mouse, a model of type 2 diabetes, prevents the mesangial matrix expansion (but without affecting the albuminuria). This latter finding may suggest that diabetic proteinuria, unlike the hypertrophic and prosclerotic renal manifestations of diabetes, is mediated by increased podocyte-derived VEGF rather than by TGF-β.

Hypoxia and Diabetic Nephropathy

Clinical studies have provided evidence that even mild anemia (hemoglobin <13.8 g/dl) increases risk for progression in patients with type 2 diabetes and nephropathy. Moreover, treating anemia early in renal failure with erythropoietin slows the decline of renal function. In practice, however, the benefit of erythropoiesis-stimulating agents has not held up to the scrutiny of several large, prospective, randomized controlled trials such as TREAT which showed no renal benefit and no delay in the progression to ESRD with darbepoetin alfa. The exact mechanisms by which anemia increases the risk for progression of diabetic nephropathy are incompletely understood, but a few suggestions could be made. Anemia likely causes renal hypoxia. It has been described in experimental models of chronic renal injury that hypoxia is an important factor aggravating interstitial fibrosis, partly by the induction of factors such as TGF-β and VEGF. The transdifferentiation of tubular epithelial cells into fibroblasts, an important process contributing to tubular atrophy and interstitial fibrosis, is stimulated by cellular hypoxia and mediated, appropriately enough, by hypoxia-inducible factor-1 (HIF-1), a transcription factor that plays a role in regulating genes critical to the adaptation to low oxygen. Certain growth factors and cytokines are stimulated by HIF-1 which itself can be increased by Ang II via activation of AT2 receptors, leading to the suppression of SM-20, a dihydrogenase involved in hydroxylation of HIF-1α. Since hydroxylated HIF-1α is rapidly degraded through the proteasome, the Ang II-mediated decrease in hydroxylation leads to stabilization of this important transcription factor. Thus, hypoxia plus Ang II may induce growth factors and signaling pathways with potentially deleterious effects in diabetic nephropathy. On the other hand, erythropoietin application may have additional effects in addition to correcting anemia, and the mobilization of potential progenitor cells by this treatment has become the focus of active research. Certainly, more experiments are necessary to elucidate how anemia may contribute to the development and progression of diabetic nephropathy. Lastly, any well-meaning attempt to correct anemia for the sake of diabetic nephropathy has to be weighed against the not-inconsiderable risk of stroke, hypertension, other cardiovascular event, and possibly increased mortality that seem to arise from the overzealous use of erythropoiesis-stimulating agents.