Type 1 Diabetes

For simplicity, the course of renal involvement in type 1 diabetes can be roughly divided into five stages (Table 28-1). Stage 1, present at diagnosis, is that of renal hypertrophy-hyperfunction. At this stage, patients at risk and not at risk of DN currently cannot be clearly separated. Genetic factors associated with predisposition to or protection from DN (see Genetic Predisposition) could, in the future, add to prediction of risk during this period. Although some studies^14–17 and a recent meta-analysis¹⁸ suggest that the presence of GFR above the normal range (glomerular hyperfiltration) is a risk factor for DN, this remains controversial and does not fit with data showing that more than 65% of young type 1 diabetic subjects have increased GFR.¹⁹ Most studies addressing this issue evaluated cohorts of patients with average diabetes duration of a decade or so. In an inception cohort study of patients with adult-onset type 1 diabetes within 6 months of diabetes diagnosis, higher than normal levels of albumin excretion rate (AER), blood pressure, hemoglobin A_1c, male sex, and shorter stature were significant independent (albeit imprecise) predictors of microalbuminuria development over 18 years of follow-up, while GFR was noncontributory. Stage 2 is characterized by the presence of glomerular lesions in patients with normal albumin excretion rate and normal blood pressure levels. Preliminary studies suggest that normoalbuminuric patients with more severe glomerular lesions are at increased risk of progression.²⁰ Further, patients with type 1 diabetes, normoalbuminuric at baseline, with greater increase in DN lesions over the subsequent 5 years, have greater increases in AER over that time (Kukla, Caramori, and Mauer, unpublished observations). Decreased normal nocturnal blood pressure decline (dipping) and increased nocturnal systolic blood pressure in 24-hour blood pressure monitoring may be an early indicator of DN risk, preceding the development of persistent microalbuminuria.²¹ The presence of persistent microalbuminuria with at least 2 of 3 consecutive values between 20-200 µg/min for 30-300 mg/g of creative defines stage 3. This typically occurs after 5 or more years of type 1 diabetes, but microalbuminuria may be present earlier, particularly during adolescence and in patients with poor glycemic control and high-normal blood pressure levels, and may be more frequent and less prognostic in daytime urine samples due to postural proteinuria, especially in adolescents (Mauer, unpublished observations). Compared to normoalbuminuric patients, patients with persistent microalbuminuria are at three- to fourfold increased risk of progression to proteinuria and ESRD.²² Current studies indicate that about 20% to 45% of microalbuminuric type 1 diabetic patients will progress to proteinuria after about 10 years of follow-up, while 20% to 25% will return to normoalbuminuric levels, and the rest will remain microalbuminuric, at least over the subsequent 6 to 10 years.^22–26 Of note, studies conducted 2 to 3 decades ago estimated that about 80% of type 1 diabetic microalbuminuric patients would progress to proteinuria over the next 10 years.^27–29 The differences in progression rates may be related to overestimation of risk in these earlier small post hoc studies, differing definitions of microalbuminuria, improved prognosis due to advancements in treatment, or some combination of these factors. Alternatively, the natural history of DN may have changed over the past decades. At stage 3, glomerular lesions are generally more serious than in the previous ones. GFR is usually normal and stable or slowly declining during this stage, and recent data suggest that the subset with microalbuminuria and declining GFR may be at particularly high risk of progression.³⁰ Blood pressure levels tend to increase, and hypertension is not uncommon. Some patients also have hypercholesterolemia and hypertriglyceridemia, along with increased fibrinogen, von Willebrand factor, and prorenin levels. Other microvascular (diabetic retinopathy and neuropathy) and macrovascular (coronary artery disease and stroke) complications are also more common. Stage 4 is defined by the presence of overt proteinuria [AER > 200 µg/min or 300 mg/24 hour or albumin/creatinine ratio (ACR) > 300 mg/g]. Proteinuria typically develops only after 10 to 20 years of type 1 diabetes. Patients with proteinuria at shorter type 1 diabetes duration deserve a clinical renal biopsy. However, microalbuminuria or proteinuria may be present at diagnosis in patients with type 2 diabetes where onset of diabetes may go undetected for many years or because the increased urinary protein is of another pathogenesis in patients with minimal or atypical DN lesions. GFR is often reduced, and hypertension and dyslipidemia are very common. Retinopathy and peripheral and autonomic neuropathy are present in most patients, and if retinopathy is absent, this should call the diagnosis of DN into question. The risk for cardiovascular events is extremely high, and asymptomatic myocardial ischemia is frequent. Without treatment, GFR declines about 14 mL/min/yr in type 1 diabetic patients with proteinuria and hypertension. Progression to ESRD (stage 5) will occur 5 to 15 years after the development of proteinuria.

Type 2 Diabetes

As already noted, diabetes duration is usually not precisely known. At diagnosis, about 10% of type 2 diabetic patients have nephropathy, 20% retinopathy, 70% hypertension, and 60% dyslipidemia. Interestingly, in studies of Pima Indians where type 2 diabetes onset is much more precisely known and the age of onset is much younger than in most other type 2 diabetes cohorts, the course of diabetic kidney disease is very similar to that of type 1 diabetic patients. Overall, in patients with type 1 and type 2 diabetes, the rate of progression from microalbuminuria to proteinuria is similar to that of type 1 diabetic patients, around 30% after 10 years of follow-up.^31,32 Thus, the Steno 2 study reported that 31% of initially microalbuminuric type 2 diabetic patients progressed to proteinuria, while the same proportion (31%) of patients became normoalbuminuric, and 38% remained microalbuminuric after about 8 years of follow-up.³² Moreover, those who were normoalbuminuric at follow-up had a lower rate of GFR decline (2.3 mL/min/yr) compared to patients who remained microalbuminuric (3.7 mL/min/yr; P = 0.03) or who progressed to proteinuria (5.4 mL/min/yr; P < 0.001).³² GFR decline may be more variable in type 2 than in type 1 diabetes. Microalbuminuric type 2 diabetic patients with GFR decline usually have more advanced typical diabetic glomerulopathy lesions and worse metabolic control,³³ while those with mild or atypical lesions have no significant GFR decline over nearly 5 years of follow-up. The above categories are general, and progression is highly variable and often nonlinear. The expression and natural history of these overlapping stages may be influenced by complex genetic, environmental, and treatment interactions that may greatly affect progression and outcome. Thus, the scheme presented here is only a useful general guide, but it cannot be considered an accurate map of the course of individual patients.

For example, large long-term clinical trials have demonstrated that improved blood glucose^34,35 and blood pressure control, perhaps especially with renin-angiotensin system blockers,^12,36,37 slows the progression of DN in proteinuric patients with already substantially reduced GFR. Indeed, the natural history of DN may have changed after the results of these trials were available to the scientific community and to patients. This has led to the publication of revised treatment guidelines by several entities recommending stricter glycemic and blood pressure control for patients with diabetes. Although in the past it was believed that once DN was present the decline of renal function was inexorable, it is now clear that this may not always be the case.^12,36,37 Finally, the demonstration that prolonged euglycemia leads to reversal of established DN lesions in patients with type 1 diabetes³⁸ contradicts the long-held belief that these lesions are irreversible.

Structural-Functional Relationships in Diabetic Nephropathy

Mesangial expansion is the major cause of GFR loss in type 1 diabetic patients.⁵¹ Increase in Vv(Mes/glom) is strongly inversely correlated with filtration surface (Fig. 28-10) per glomerulus which in turn is strongly directly correlated with GFR in type 1 diabetes.⁶⁵ Mesangial fractional volume is also directly correlated with AER^51,52 (Fig. 28-11A and B) and systemic blood pressure.^51,66 The relationships of GBM width with these clinical manifestations of diabetic kidney disease are also important but are somewhat weaker than those for mesangial changes.^51,52 However, mesangial and GBM changes together explain most of the AER variability in type 1 diabetes, ranging from normoalbuminuria to proteinuria.⁵²

As mentioned, decreased podocyte number and detachment are related to albuminuria. Increases in podocyte foot process width also correlate with AER increases in type 1 diabetic patients.61,67,68 If podocyte number or shape changes are early predictors of DN risk,⁶⁹ this would support an important pathogenetic role for this key glomerular cell in this disease.

Although glomerular capillary filtration surface is directly correlated with GFR in type 1 diabetes,65,70,71 linear regression models only partially explain GFR variability in these patients.⁵² Global sclerosis⁵⁸ and interstitial expansion⁴¹ are additional independent predictors of GFR loss. However, the conclusion that the interstitium is more important than glomerular changes in diabetes came from studies where most patients were in advanced stages of kidney failure.^72–74 At these advanced stages, when serum creatinine is already clearly elevated, especially if above 2 mg/dL, interstitial changes are common to most chronic renal diseases and are not specific to diabetes. During most of the natural history of DN, glomerular parameters are more important determinants of renal dysfunction. Moreover, as already discussed, early interstitial expansion in type 1 diabetes is mainly due to expansion of its cellular component, and increased interstitial fibrillar (scar) collagen is primarily seen in patients whose GFR is reduced,⁵⁶ while glomerular ECM changes are due to accumulation of basement membrane components. Thus, the interstitial and glomerular changes of diabetes probably have different pathogenetic mechanisms.

Through many years of the natural history of DN, lesions develop without detectable clinical or laboratory abnormalities. Often, when microalbuminuria initially manifests, lesions are far advanced. Once proteinuria is present in type 1 diabetes, GFR loss typically progresses relatively rapidly toward ESRD. This nonlinear clinical course is best reflected by nonlinear analyses of structural-functional relationships.⁵⁰ Using piecewise linear regression instead of simple linear models, glomerular structural variables alone explained 95% of variability in AER. These glomerular structures, mesangial fractional volume, GBM width and filtration surface density, however, explained less than 80% of GFR variability. This increased to more than 90% with the addition of measures of glomerular tubular junction abnormalities and interstitial expansion.⁵⁰

In summary, most of the renal functional abnormalities in type 1 diabetes are explained by diabetic glomerulopathy. Structure is highly variable in patients without functional abnormalities. Structural-functional relationships are largely driven by more advanced lesions.

Microalbuminuria and Renal Structure

Persistent microalbuminuria antedates clinical nephropathy, whereas normoalbuminuria in longstanding type 1 diabetic patients predicts lower nephropathy risk. However, the relationship of renal structural changes to low levels of albuminuria (i.e., normal to microalbuminuria) is far from simple. As a group, normoalbuminuric longstanding type 1 diabetic patients have diabetic glomerulopathy lesions.^52,75 Their structural measurements range from normal to pathology overlapping in severity with microalbuminuric and proteinuric patients (see Fig. 28-11B; Fig. 28-12B).^52,75 Increased GBM width predicts progression to microalbuminuria and proteinuria.^75,76 Microalbuminuric patients have on average even more severe lesions, with few type 1 diabetic patients having renal structural measures still within the normal range (see Fig. 28-11B and Fig. 28-12B).^52,75 Hypertension and reduced GFR are also more frequent in these patients, so microalbuminuria is a marker of more advanced lesions and other functional disturbances.^52,75 However, reduced GFR may be present in normoalbuminuric longstanding type 1 diabetic patients. This is more frequent in females with retinopathy and/or hypertension and is associated with advanced glomerulopathy lesions.^23,77,78 Microalbuminuria may not be the first indicator of DN, and careful GFR and blood pressure measurements are needed, especially in female patients with the described characteristics.

Normal identical twins of type 1 diabetic patients have normal renal structure.⁴² Diabetic twins have greater GBM and TBM width and mesangial fractional volume than their nondiabetic twin, albeit sometimes still within the normal range despite many years of diabetes.⁴² Over time, all type 1 diabetic patients appear to develop some structural changes of DN, but the rate may be so slow that the lesions would be undetectable except by comparison with their nondiabetic twin and would not lead to clinical disease. There is also striking variability in the rate at which lesions develop in kidneys transplanted into type 1 diabetic patients who all had ESRD secondary to DN.⁷⁵ This cannot be fully explained by glycemia and suggests genetically determined renal tissue susceptibility.⁷⁵

Interestingly, studies of type 1 diabetic transplant recipients indicate that having a single kidney does not appear to accelerate the rate of development of DN lesions, arguing against reduced nephron number as a risk factor.⁷⁹ In fact, proteinuric diabetic patients without advanced renal failure have normal numbers of glomeruli.⁸⁰ However, reduced glomerular number could be associated with faster GFR decline once overt DN develops.

Contrasts in Nephropathy Lesions Between Type 1 and Type 2 Diabetes

There have been fewer studies of renal pathology and structural-functional relationships in type 2 diabetes, even though it accounts for more than 80% of ESRD among diabetic patients. A Danish study found that Caucasian proteinuric type 2 diabetic patients have similar structural changes when compared to proteinuric type 1 diabetic patients, and the severity of these changes predicted the subsequent rate of GFR decline.⁸¹ However, the authors also noted more variability in glomerular structure in these patients than in type 1 patients, with some type 2 proteinuric patients having minimal or no diabetic glomerulopathy.⁸¹ One study from Northern Italy of type 2 diabetic patients biopsied for clinical reasons⁸² found typical DN lesions in a third of patients; an increase of globally sclerosed glomeruli, severe tubulointerstitial lesions, and minimal or no diabetic glomerulopathy in another third; while the rest showed changes of diabetic glomerulopathy plus changes of other diseases such as proliferative glomerulonephritis and similar changes.⁸² In another Danish study, most proteinuric type 2 diabetic patients had diabetic changes,⁸³ but about a fourth of those studied had nondiabetic glomerulopathies, including “minimal lesions,” glomerulonephritis, or mixed diabetic and other changes. In this study, all patients with proteinuria and diabetic retinopathy had classical DN lesions, but less than half of those without retinopathy had DN.⁸³ A British study found similar results.⁸⁴ However, these high rates of diseases other than or superimposed upon DN are almost certainly because the patients in these studies had clinical indications for kidney biopsies, often including atypical clinical courses. In another study from Northern Italy, when renal biopsies in type 2 diabetes were performed solely for research purposes, definable renal diseases other than secondary to diabetes were in fact uncommon.⁸⁵ However, only about a third of type 2 microalbuminuric patients in this research study had findings typical of diabetic glomerulopathy; approximately a third had minimal abnormalities, and about 40% had varying combinations of global glomerulosclerotic, vascular, and tubulointerstitial lesions out of proportion to their absent to relatively mild diabetic glomerulopathy lesions.⁸⁵

Structural-Functional Relationships in Type 2 Diabetes

Although initially reported to have renal structural-functional relationships that were similar to type 1 patients,⁸⁶ a more recent study in Japanese type 2 diabetic patients indicated greater heterogeneity.⁸⁷ A Danish study found fewer glomerular lesions and higher GFR levels in type 2 versus type 1 patients with similar AER.⁸¹ Much larger glomerular volumes were found in the type 2 patients, and perhaps this preserved filtration surface.⁸¹ Nonetheless, the proteinuria in these type 2 diabetic patients was at least in part unexplained. Mesangial fractional volume increased progressively with albuminuria and global glomerular sclerosis correlated inversely with GFR in Pima Indian type 2 diabetic patients.⁶¹ Also, glomerular podocyte loss was related to proteinuria (but not microalbuminuria) in these patients.

The less precise correlation between glomerular structure and renal function in type 2 diabetes is probably due to these more varied patterns of renal injury.⁸⁵ This is of prognostic significance, since patients with more typical diabetic glomerulopathy lesions are more likely to have progressive GFR loss.^33,83

In summary, renal structural changes in type 2 diabetes are more heterogeneous, and diabetic glomerulopathy lesions are on average less severe in type 2 than in type 1 diabetic patients with similar albuminuria levels. Approximately 40% of the type 2 diabetic patients show atypical renal injury patterns. These atypical patterns are associated with higher body mass index and less diabetic retinopathy.⁸⁵ Thus, the atypical manifestations of renal injury in type 2 diabetes could be related to obesity, hypertension, hyperlipidemia, accelerated atherosclerosis, and aging interacting with the effects of hyperglycemia. The markedly increased risk of ESRD in certain type 2 diabetic populations (e.g., African American, American Indian, Hispanic) could represent variability in the renal consequence of one or more of these pathogenetic influences (e.g., there are differences in the renal structural consequences of hypertension in African American and Caucasian patients,⁸⁸) genetic susceptibility to DN, or both. Further cross-sectional and longitudinal studies in type 2 diabetic patients are required before these complexities can be better understood.

Other Renal Disorders in Diabetic Patients

Based on the above discussion, it is strongly recommended that type 1 diabetic patients with proteinuria and less than 10 years duration of diabetes and type 2 diabetic patients with proteinuria and no retinopathy should be referred to a nephrologist and be fully evaluated for other renal diseases. In these cases, renal biopsy should be strongly considered both for diagnosis and prognosis purposes.

Diabetic Nephropathy Lesions are Reversible

Despite rapid reversal of mesangial expansion in rats cured of diabetes by islet transplantation,⁸⁹ there was no improvement in DN lesions in the native kidneys of type 1 diabetic patients normoglycemic for 5 years after successful pancreas transplantation.⁹⁰ However, after 10 years of normoglycemia, there was dramatic healing of diabetic renal lesions.⁹¹ GBM and TBM width and mesangial fractional volume were lower at 10 years versus baseline and 5 years, often with these structural parameters having returned to the normal range (Fig. 28-13A and B).⁹¹ The reduction in mesangial expansion primarily resulted from the disappearance of excess mesangial matrix material (see Fig. 28-13C and D). Light microscopy revealed remarkable glomerular architectural remodeling, often including the complete disappearance of Kimmelstiel-Wilson nodules (Fig. 28-14A to C) and increased patency of glomerular capillaries.⁹¹ It is not clear why reversal of lesions was not seen in the first 5 years. Perhaps there was cellular (epigenetic) memory for the diabetic state. Alternatively, glycation of ECM may have slowed the degradation process. Unlike the normal situation where throughout adult life, glomerular ECM content remains quite constant,⁹² this healing process requires ECM removal to exceed production. Thus, the cellular recognition of an abnormal structural environment and the cellular machinery to initiate and sustain the healing and remodeling processes must exist. Remodeling and healing in the tubulointerstitium also occurred in these patients.⁹³ There was an estimated decrease in total amount of cortical interstitial collagen and disappearance of atrophic tubules, the latter presumably through reabsorption.⁹³ If it becomes possible to stimulate these healing processes over the processes of injury, DN could be delayed or prevented.

Familial Studies of Blood Pressure, Cardiovascular Disease, and Diabetic Nephropathy

Several studies showed associations between DN and predisposition to hypertension and cardiovascular disease.107–112 Higher blood pressure levels and increased prevalence of hypertension have been described in nondiabetic parents of proteinuric and microalbuminuric type 1^107–109 and type 2¹¹¹ diabetic patients. Diabetic patients with advanced nephropathy also had higher mean arterial blood pressures during adolescence,¹⁰⁹ and prediabetic blood pressure levels predicted AER after diabetes onset.¹¹⁰ The prevalence of cardiovascular disease and cardiovascular death was greater in the parents of type 1 diabetic patients with nephropathy,¹¹² and DN patients who had suffered a cardiovascular event more often had a positive family history of cardiovascular disease. This indicates that familial predisposition to cardiovascular disease increases both the risk of nephropathy and the risk of cardiovascular disease in type 1 diabetic patients with nephropathy, linking the pathogenesis of DN to factors also favoring the development of atherosclerosis.^107,108,113 Moreover, AER and blood pressure levels were linked and heritable in Caucasian families with type 2 diabetes.¹¹⁴

Sodium/Lithium and Sodium/Hydrogen Countertransport

Sodium/lithium (Na⁺/Li⁺) countertransport activity, a genetically regulated system linked to hypertension and cardiovascular disease in nondiabetic persons,¹¹⁵ was increased in microalbuminuric and proteinuric type 1^108,116,117 but not in type 2¹¹⁸ diabetic patients. Increased Na⁺/Li⁺ countertransport activity has also been described in parents of proteinuric type 1 diabetic patients¹¹⁹ and is similar in identical twin pairs discordant for type 1 diabetes¹²⁰ where both twins had values higher than age-matched nondiabetic controls.¹²⁰ Raised erythrocyte Na⁺/Li⁺ countertransport activity was associated with a fourfold increased risk of development of persistent microalbuminuria in patients with type 1 diabetes.¹²¹

Na⁺/Li⁺ countertransport has parallels with the sodium/hydrogen (Na⁺/H⁺) antiport exchange system involved in the regulation of intracellular pH, cell volume and growth, and proximal tubular sodium reabsorption.¹²² The Na⁺/H⁺ antiport activity is increased in leukocytes¹²³ and cultured skin fibroblasts^124–126 of type 1 and type 2 diabetic patients with microalbuminuria and proteinuria. Also, since cultured skin fibroblast Na⁺/H⁺ antiport activity was highly correlated among sibling pairs concordant for type 1 diabetes and for DN lesions,¹²⁷ and independent of environmental factors, it was suggested that these cultured skin fibroblast abnormalities are at least in part genetically regulated.

Diabetic Nephropathy Genes

As already noted, genetic predisposition to DN has been strongly suggested in studies of siblings concordant for type 197,98,100 and type 2^101,103 diabetes. There are ongoing searches for genetic loci related to DN susceptibility through genomic scanning and candidate gene approaches, although neither approach has yet yielded definitive results,^128–130 probably because multiple genes are involved.^128,129 Sib-pair studies in Pima Indian families with type 2 diabetes showed linkage between four chromosomal regions (chromosomes 3, 7, 9, and 20) and DN, the strongest one in chromosome 7.¹³¹ Previous studies in Caucasian populations have shown linkage between DN and chromosome 7q in type 2 diabetic patients¹³² and chromosome 3q in type 1 diabetic patients.¹³³ Segregation analysis studies in Caucasian families with type 2 diabetes also suggest a major gene effect for AER.¹³⁴ In the Family Investigation of Nephropathy and Diabetes (FIND) study, a genomewide scan demonstrated multiple chromosomal regions linked to estimated GFR in multiethnic families ascertained by a proband with DN,¹³⁵ but these may be related more to DN progression than to DN risk.

Genetic polymorphisms in candidate genes have been evaluated in several studies. Associations with DN risk have also been reported for polymorphisms in the human inducible nitric oxide synthase,¹³⁶ ectonucleotide pyrophosphatase/phosphodiesterase-1,¹³⁷ heparan sulfate proteoglycan,¹³⁸ paraoxonase 2,¹³⁹ matrix metalloproteinase-9 (MMP-9),¹⁴⁰ transforming growth factor-β1 (TGFB1),¹⁴¹ protein kinase C (PKCB),¹⁴² and haptoglobin¹⁴³ genes. Studies of polymorphisms in endothelial nitric oxide synthase (eNOS),^144–146 aldose-reductase,^147–149 plasminogen activator inhibitor 1,^150,151 interleukin 1,^152,153 apolipoprotein E,^154–157 glucose transporter-1 (GLUT1),^158,159 and methylenetetrahydrofolate reductase^160,161 genes showed conflicting results. Many other candidate genes have been studies.^70–82 A leucine repeat of the carnosinase (CNDP1) gene was associated with DN in both type 1 and type 2 diabetic patients.^162,163 Also, multiple variations in the superoxide dismutase 1 (SOD1) gene were significantly associated with development of persistent microalbuminuria and severe nephropathy in type 1 diabetic patients enrolled in the Diabetes Control and Complications Trial/Epidemiology of Diabetes and Its Complications (DCCT/EDIC) study.¹⁶⁴

One particularly extensively studied set of genes are those related to the renin-angiotensin system. Polymorphism in the angiotensin converting enzyme (ACE) gene, consisting of an insertion or a deletion (I/D) of a 287-base-pair sequence that determines most of the interindividual variance in the ACE activity, has also been associated with rate of GFR decline in nondiabetic renal diseases.¹⁶⁵ It remains unclear if this ACE polymorphism is important in the genesis of DN,^166–170 although a follow-up study supports this view.¹⁷¹ I/D ACE polymorphism may be involved in the progression of DN and the response to ACE inhibitor therapy.^172–174 Normoalbuminuric and microalbuminuric type 1 diabetic patients with the II genotype appears to have the fastest increase in AER¹⁷⁴ and the best response to ACE inhibitor therapy.^173,174 On the other hand, D was the risk allele for development of DN after 6 years of follow-up of Caucasian type 1 diabetic patients.¹⁷¹ The, DD genotype was also more frequent in normoalbuminuric or microalbuminuric Japanese type 2 diabetic patients who doubled AER or changed AER class (progressors) than in patients who did not progress after 10 years of follow-up.¹⁷² Although progression was not be associated with ACE genotype in Caucasian normoalbuminuric or microalbuminuric type 2 diabetic patients,¹⁷⁵ patients with II genotype (n = 11) did not decrease GFR after 9 years of follow-up. Moreover, presence of the D allele in Caucasian type 2 diabetic patients was related to the presence of more severe DN lesions.¹⁷⁶ Other studies did not observe an association between this ACE gene polymorphism and DN^{129,151,177–186} or rate of GFR decline¹⁸⁷ in type 1 or type 2 diabetic patients. One meta-analysis concluded that the D allele was significantly associated with DN in a dominant model (DD + ID versus II) in both type 1 and type 2 diabetic patients.¹⁸⁸ Another found this association to be significant in Japanese type 2 diabetic patients but not in Caucasian type 1 or type 2 diabetic patients.¹⁸⁹

Association with DN risk in type 1 or type 2 diabetic patients and polymorphisms in the angiotensinogen gene was found in some^190,191 but not in all^{181,182,192,193} studies. One study observed association between the T-allele of the angiotensinogen gene and elevated AER only when interaction with the D-allele of the ACE I/D polymorphism was considered.¹⁹⁴ Interactions between ACE and angiotensinogen gene polymorphism have been previously described in cross-sectional¹⁶⁶ but not confirmed in longitudinal¹⁷¹ studies. A meta-analysis did not find association between angiotensinogen gene polymorphism and DN.¹⁹⁵ Two large studies did not find a role for angiotensin II type 1 receptor polymorphism in DN,^196,197 while another found this polymorphism to significantly contribute to DN risk in type 1 diabetic patients only when accompanied by poor glycemic control.¹⁹⁸

In summary, the genetic determinants of DN risk are not fully understood, and it is likely that several genes with moderate or small effects interacting with the metabolic abnormalities of diabetes may be responsible for the genetic susceptibility to DN. The failure to find reproducible DN-associated candidate genes may be related to inadequate number of subjects, different ascertainment strategies, real genetic differences among populations, selection bias, poor study design, weak genetic contribution, wrong candidate selection, or combinations thereof. Hopefully, the ongoing genomewide studies will clarify this still-confusing area.

Glycemic Control

Although important modulating factors may exist, DN is secondary to the long-term metabolic aberrations found in diabetes and exposure to elevated glucose levels appears central to the expression of this disorder. Studies in type 1 and type 2 diabetes proved that improved glycemic control could reduce the development of DN. The DCCT demonstrated that the risk of microalbuminuria in type 1 diabetic patients was decreased by strict glycemic control.¹⁹⁹ Importantly, although the DCCT did not show a benefit of 6.3 years of improved glycemic control in the progression from microalbuminuria to proteinuria, additional follow-up of these patients showed significant reduction in the rates of progression to proteinuria and ESRD.²⁰⁰ The United Kingdom Prospective Diabetes Study (UKPDS) also demonstrated a decreased incidence of DN after 10 years of intensive glycemic control in recently diagnosed type 2 diabetic patients.³⁵ Moreover, a 5-year randomized clinical trial in 48 type 1 diabetic kidney transplant recipients demonstrated that the development of one of the earliest diabetic renal lesions, increased mesangial matrix fractional volume per glomeruli (Vv[MM/glom]), was prevented by strict glycemic control.²⁰¹ Also, intensive insulin treatment for 26 to 34 months decreased the rate of GBM thickening and increase in Vv(MM/Mes) and matrix star volume in a controlled trial in 18 microalbuminuric type 1 diabetic patients.²⁰² Finally, regression of established diabetic glomerular lesions has been demonstrated in the native kidneys of 8 type 1 diabetic patients with 10 years of normoglycemia induced by successful pancreas transplantation (see Fig. 28-14 A to C).^90,203 These studies strongly suggest that hyperglycemia is not only necessary for DN lesions to develop, but it is also necessary to sustain established lesions. Removal of hyperglycemia allows reparative mechanisms to be expressed that ultimately result in healing of the original diabetic glomerular injuries. Studies evaluating GLUT1 in animals reinforce the pathogenic role of glycemia on the development of DN. High glucose can increase GLUT1 expression and glucose transport in rat mesangial cells,²⁰⁴ effects mediated by TGF-β.²⁰⁵ Rat mesangial cells transduced to overexpress human GLUT1 protein in normal glucose concentration showed increased glucose uptake and ECM synthesis.²⁰⁶ The mechanisms involved in the glomerular injury caused by hyperglycemia will be discussed in detail later.

Hemodynamic Mechanisms

Berkman and Rifkin²⁰⁷ described a diabetic patient with unilateral renal artery stenosis and marked diabetic lesions in the kidney exposed both to hypertension and diabetes, while the contralateral kidney, protected from hypertension by the narrowed renal artery, had only ischemic changes. A model of unilateral renal artery stenosis in diabetic rats²⁰⁸ proposed to confirm this human observation, although hypertensive and diabetic changes were not differentiated in this study. Insulin-treated diabetic rats have glomerular hyperfiltration explainable by increased single nephron GFR due to increased single nephron blood flow and, in the Münich-Wistar rat strain, increased glomerular capillary pressure.²⁰⁹ Thus, alterations in intraglomerular hemodynamics could influence the rate at which diabetic lesions develop. Some studies reported that glomerular hyperfiltration was a risk factor for the development of microalbuminuria.^16,210 However, there is still controversy as to whether increased GFR is an independent predictor of progression.^15,211

Several mediators of diabetes-induced hyperfiltration have been proposed. Increased renal and urinary kallikrein levels were associated with increased GFR and renal plasma flow in moderately hyperglycemic diabetic rats,²¹² which were normalized by a bradykinin B2 receptor antagonist.²¹³ Urinary kallikrein excretion was also increased in type 1 diabetic patients with hyperfiltration.²¹⁴ The genesis of glomerular hyperfiltration has also been associated with increased nitric oxide (NO) generation or action by a glucose-dependent mechanism.^215,216 The vasodilatation caused by increased NO could enhance the permeability to macromolecules, leading to albuminuria.²¹⁷ Furthermore, the enhanced NO synthesis by endothelial nitric oxide synthase (eNOS) in afferent arterioles and glomerular endothelial cells in diabetic rats suggests a pivotal role of NO in preferential afferent arteriolar dilatation, glomerular enlargement, and functional glomerular hyperfiltration in the early stages of DN.²¹⁸ Treatment with a nonselective NOS inhibitor normalized GFR and plasma renal flow levels in diabetic rats.²¹⁹ NO activity was shown to be increased in microalbuminuric compared to normoalbuminuric type 1 diabetic patients or nondiabetic controls and related to GFR and AER.²²⁰ Other studies suggest that hyperfiltration may result from an increased proximal tubular fluid reabsorption independent of any primary malfunction of the glomerular microvasculature.²²¹

However, one cannot explain the genesis of DN on hyperfiltration alone. Reduction in nephron mass in rats by uninephrectomy²²² produces glomerular hemodynamic perturbations that are similar to diabetes²⁰⁹ but does not produce the lesions²²³ that are classic for DN in animals.^224,225 The central lesion associated with hyperfiltration in rats, focal segmental glomerular sclerosis,²²⁶ is not an important lesion in human or animal diabetes (see the section Kidney Structure in Diabetes). Reduced glomerular number at birth has been proposed as a risk factor for progressive renal disease and hypertension.²²⁷ One study suggested that reduced glomerular number was associated with increased DN risk in type 2 diabetic patients.²²⁸ This association was not confirmed in studies with type 1 diabetic patients. Glomerular number, estimated during autopsy, was decreased in type 1 diabetic patients only if ESRD was present.⁸⁰ No differences in glomerular number were found between controls and type 1 diabetic patients with established DN but not ESRD.⁸⁰ Moreover, there was no relationship between glomerular number (estimated by morphometric techniques in kidney biopsy tissue) and functional or structural parameters in type 1 diabetic patients with a wide AER range.²²⁹ Also, reduced nephron mass, such as in patients with uninephrectomy, has not been documented to produce DN lesions in humans. More than that, DN lesions may not develop faster in type 1 diabetic kidney transplant recipients (single-kidney diabetic patients) than in type 1 diabetic patients with their native kidneys.^79,230 These studies suggest that glomerular number is probably not a crucial variable in the development of DN, although this variable may be important in progression once clinical findings of DN are present.

Lowering glomerular capillary pressure by long-term ACE inhibitor therapy prevented the development of focal segmental glomerular sclerosis²³¹ but did not prevent the GBM widening and increase in Vv(Mes/glom) that occurs in the first 6 months of diabetes in rats.²³² Thus, renin-angiotensin system blockage may not prevent the development of specific early DN lesions in diabetic animals.²³² ACE inhibitor therapy prevented increase of GBM width and perhaps changed Mes composition but had no effect on Vv(Mes/glom) in microalbuminuric type 1 diabetic patients with established diabetic lesions.²³³ Cortical interstitial expansion was also prevented in type 1²³⁴ and type 2²³⁵ diabetic patients with already increased AER. These effects may not be specific for ACE inhibition, since β-blocker therapy in these studies showed similar effects on blood pressure and lesions.^233,234 The influences of renin-angiotensin system blockage on clinical progression of DN are discussed in the Treatment section. Taken together, these studies could support the hypothesis that hemodynamic abnormalities may be more important in influencing the progression of established DN lesions than in serving as the genesis of these structural changes.

Although we are arguing that a case has not been made which currently allows acceptance of the hypothesis that hemodynamic abnormalities are responsible for the genesis of the early lesions of DN, it is nonetheless possible that manipulations that can affect glomerular hemodynamics might also affect the development of early diabetic renal lesions through other than hemodynamic influences. Drugs affecting the renin-angiotensin system might also impact ECM dynamics236–239 and thereby renal structure in diabetes. However, the mechanisms of actions of these agents appear complex, are incompletely understood, and may in part be independent of the renin-angiotensin system. Also, TGF-β expression may be regulated in parallel with renin.^240,241 Thus, glomerular hyperfiltration could promote ECM accumulation by increases in the expression of TGF-β,²⁴² and this might be modeled in vitro by the mechanical stretching of mesangial cells.²⁴³

Growth Factors

This subject was previously reviewed in detail by Flyvbjerg²⁴⁷ and Chiarelli²⁴⁸ and, using a similar organizational scheme, it is summarized here.

Protein Kinase C

The PKC enzyme regulating Na⁺/K⁺-ATPase²⁹⁹ has a role in regulating cell proliferation, vascular contractility and permeability, and basement membrane syntheses. PKC can be activated by diacylglycerol³⁰⁰ and thus by high glucose (Fig. 28-15). High glucose–induced PKC activation in glomerular cells in vitro is followed by increased TGF-β³⁰¹ and MAPK³⁰² activity. Increased type IV collagen expression in mesangial cells in high glucose states is prevented by nonspecific PKC inhibitors,³⁰³ and treatment of PKC agonists can increase type IV collagen expression.³⁰⁴ Treatment with a PKCβ inhibitor blocks glomerular TGF-β1 mRNA, MM, GFR, and AER increase in diabetic rodents.^305,306 In another study, PKCβ inhibition prevented glomerular hypertrophy and albuminuria without affecting the TGF-β axis in diabetic rats,³⁰⁷ arguing that PKC renal effects in diabetes may at least in part be independent of the TGF-β axis.

Cytokines

As discussed earlier in the section on hemodynamics, there are in vitro and animal model data linking the pathogenesis of DN to the renin-angiotensin system. More than the other pathways described in this section, the renin-angiotensin system has been studied in regard to DN in humans.

Oxidative Stress

Oxidative stress is increased in diabetes.³⁵¹ The debate has been whether oxidative stress has a primary role in the pathogenesis of DN.³⁵² Skin fibroblasts from type 1 diabetic patients with DN do not show the expected increase in mRNA expression and activity for catalase and glutathione peroxidase, two antioxidant enzymes, when exposed to a high glucose condition in vitro.³⁵³ Catalase converts hydrogen peroxide (H₂O₂) to oxygen and water, whereas glutathione peroxidase reduces peroxide and superoxide (O₂⁻) levels. Glutathione peroxidase requires a high cellular level of reduced glutathione to be effective. Lipid peroxidation, induced by exposure to high glucose, was also increased in cells of type 1 diabetic patients with DN. These findings suggest that increased oxidative stress in type 1 diabetic patients with DN is at least in part associated with a decreased response of antioxidant enzymes to high glucose. Increased oxidative stress was also demonstrated in families of type 1 diabetic patients,³⁵⁴ suggesting that an abnormal redox state could even precede diabetes onset. Interestingly, erythrocyte glutathione content is correlated to Na⁺/H⁺ exchanger (NHE) activity,³⁵⁵ linking oxidative stress to an ion-transport system associated with DN risk.

There is an association between oxidative stress, NO production, and endothelial dysfunction.^356,357 Endothelium-derived NO is a potent vasodilator that also has antiatherogenic properties, including inhibition of endothelial platelet and leukocyte adhesion and of smooth muscle cell migration.³⁵⁸ Endogenous NO produced by glomerular capillary endothelial cells may modulate TGF-β production by these cells and also by mesangial cells.³⁵⁷ Endothelial NO synthesis can be stimulated by shear stress and pulsatile vessel stretching,³⁵⁹ hypoxia,³⁶⁰ and by agonists such as acetylcholine and bradykinin.³⁶¹ Exposure to high glucose concentrations increases eNOS gene expression and NO release, with a concomitant increase in O₂° production.³⁶² O₂° inactivates and reacts with NO to form peroxynitrite,³⁶² a potent oxidant, leading to endothelial dysfunction.³⁶³ Normalization of NO-mediated vasorelaxation in high glucose conditions by superoxide dismutase,³⁶⁴ a scavenger that converts O₂° into hydrogen peroxide, further adds to this association. NO suppressed TGF-β activity and abolished TGF-β1 mRNA and collagen synthesis increase in mesangial cells grown under high glucose conditions. Also, in vitro stimulation of inducible nitric oxide synthase (iNOS) activity reduced collagen and fibronectin accumulation in rat mesangial cells.³⁶⁵ NO may also modulate MMP-2 activity in these cells.³⁶⁶ Decreased basal and stimulated NO release and reduced NO bioavailability and vascular smooth muscle cell responsiveness have been shown in diabetes. Hyperglycemia may cause these abnormalities by several mechanisms, including increase in advanced glycation and end-product (AGE) production, polyol pathway and PKC activities, and oxidative stress, as further described later. AGEs can quench NO both in vitro and in vivo,³⁶⁷ decreasing its bioavailability and impairing NO’s antiproliferative effects on mesangial cells.³⁶⁸ Interestingly, aminoguanidine, an AGE crosslink inhibitor, also inhibits NO²¹⁵ and restores endothelium-dependent relaxation in diabetic rats.³⁶⁷ Increased aldose reductase activity in chronic hyperglycemia causes reduced NADPH, an essential cofactor for NOS, and could thus lead to reduced NO production.³⁵⁶ Increased PKC could also lead to impaired endothelium-dependent relaxation.³⁶⁹

Thus, the increase in intracellular reactive oxygen species (ROS) induced by hyperglycemia may arise from multiple pathways,³⁷⁰ including (1) glucose autoxidation; (2) H₂O₂ generated from the oxidation of enediols formed from Amadori products; and (3) O₂⁻ formed by the mitochondrial oxidation of NADH to NAD⁺ or during prostaglandin generation (prostaglandin synthase utilizes NADH [and NADPH], generating O₂⁻). Both H₂O₂ and hydroxyl radicals (^·OH) may be derived from O₂⁻. Mitochondrial superoxide overproduction stimulates aldose reductase activity, thus initiating the de novo synthesis of diacylglycerol, activating PKC. Reactive oxygen species can also initiate intracellular AGE formation (see Fig. 28-15). Inhibitors of the mitochondrial electron transport system and manganese superoxide dismutase, an enzyme that leads to reactive oxygen species scavenging in the mitochondrial matrix, inhibit reactive oxygen species production secondary to hyperglycemia. These agents also prevent high glucose–induced PKC and NF-κB activation, methylglyoxal-derived AGEs formation, and sorbitol accumulation through the polyol pathway. This strongly suggests that reactive oxygen species are not only necessary for the damage caused by hyperglycemia, they can also be the link between the several pathophysiologic hypotheses for DN. Recent studies suggest that the high glucose–induced sustained increase in NF-κB p65 unit expression in vitro after cells are transferred to a normal glucose medium is caused by epigenetic modifications. These epigenetic changes are caused by increased generation of methylglyoxal due to hyperglycemia-induced formation of ROS by the mitochondrial electron transport system. Methylglyoxal, the most reactive dicarbonyl AGE-intermediate in cross-linking of proteins, can in turn regenerate reactive oxygen species in the course of glycation reactions.³⁷¹ In vitro studies demonstrated that vitamin E, an antioxidant, blocks the high glucose–induced increase in PKC, TGF-β bioactivity, collagen, and fibronectin synthesis in mesangial cells.^301,372 Taurine, another antioxidant, also inhibits PKC activation,³⁰¹ TGF-β increase,³⁰¹ and collagen accumulation³⁷³ in rat mesangial cells exposed to high glucose.³⁰¹ In contrast, vitamin E did not alter the TGF-β induced increase in matrix protein synthesis in mesangial cells.³⁰¹ Treatment of diabetic rats with antioxidants shows conflicting results. Diabetic rats treated with selenium and vitamin E³⁷⁴ or nitecapone, another antioxidant,³⁷⁵ showed reduction of AER, hyperfiltration, and glomerular lesions. Intraperitoneal injection of vitamin E also prevented glomerular increase of diacylglycerol content, PKC activity, and filtration rate and fraction, and improved AER in diabetic rats.³⁷⁶ However, another study observed less proteinuria, glomerular hypertrophy, glomerulosclerosis, tubulointerstitial fibrosis, and preserved glomerular type IV collagen staining in taurine treated animals, whereas vitamin E–treated animals had extremely high mortality and worse renal lesions.³⁷⁷ In combination with vitamin C, vitamin E administered for 3 months reduced AER in a randomized, double-blind, placebo-controlled clinical trial in type 2 diabetic patients.³⁷⁸ Further studies evaluating the long-term effects of vitamin E supplementation on renal function of patients with diabetes are needed.

Glycosylation Products

Nonenzymatic reactions between reducing sugars, such as glucose, and free amino groups, lipids, or nucleic acids, known as Maillard reactions, are one of the major metabolic effects of diabetes.^379,380

Many proteins, including hemoglobin,³⁸¹ albumin,³⁸⁰ low-density lipoproteins,³⁸² erythrocyte membrane proteins,³⁸³ and crystalline lens,³⁸⁴ undergo nonenzymatic glycosylation in diabetes, leading to altered physiochemical properties of these molecules. This early glycation process proceeds through the labile Schiff base adducts formation and intramolecular Amadori rearrangement to become a stable glucose-modified protein. Amadori products comprise the majority of plasmatic glucose-modified proteins, and receptors for some of those modified proteins have been defined.³⁸⁵ Further modifications of Maillard reaction products lead to inter- and intramolecular crosslinks and formation of AGEs. These reactions are complex, and the reader is referred to several excellent reviews for greater detail.^386,387

Renal cells grown in high glucose have up-regulation of the TGF-β255,388,389 and PKC^304,390 systems and exhibit ECM overproduction.^303,391 Amadori-glycated proteins have similar effects, even without high-glucose media. Glomerular epithelial cells exposed to glycated fetal bovine serum show increased laminin and GBM antigens, unchanged type I collagen and fibronectin, and decreased collagenase activity.³⁹² Glycated albumin and high glucose together cause an even greater in vitro increase in TGF-β1, TGF-β type II receptor, α1(IV) collagen, and fibronectin mRNA than either manipulation alone.^393,394 Moreover, antibody to glycated albumin prevented the increase in fibronectin and type IV collagen expression in rodent endothelial and mesangial glomerular cells exposed to glycated albumin.^379,393,395 Antiglycated albumin antibody administration lowers plasma glycated albumin concentration, reduces AER, and improves glomerular mesangial pathology in db/db diabetic mice.^396,397 Treatment with an inhibitor of glycated albumin formation (EXO-226) normalized glycated albumin plasma concentration, reduced AER, prevented GFR decline, and reduced MM accumulation and α1(IV) collagen cortical mRNA expression in db/db mice despite persistent hyperglycemia.³⁹⁸ Blockage of overall PKC or PKC-β activity prevented the glycated albumin-induced increase in type IV collagen production,^305,306 consistent with the idea that PKC, particularly the β isoform, plays a role in the glycated-albumin induced ECM accumulation. Type 1 diabetic patients showed higher plasma levels of glycated albumin than nondiabetic controls, and type 1 diabetic patients with DN showed higher plasma levels of Amadori albumin than normoalbuminuric patients.³⁹⁹ However, studies examining the role of Amadori-glycated proteins in the progression of diabetic renal disease have reached conflicting conclusions.^400–402

Accumulation of glycation products may be a major contributor to the development of diabetic complications.403–405 Increased AGEs can stimulate the synthesis of various growth factors, including IGF-I and TGF-β. Binding of AGE-proteins to its receptors on cell surfaces induces an intracellular oxidative response in vitro, characterized by increased NF-κB.⁴⁰⁶

A hallmark of the glomerular changes caused by diabetes is mesangial expansion, mainly due to ECM increase. Mesangial cells exposed to high glucose when the matrix is being made or degraded show reduced rates of mesangium degradation,⁴⁰⁷ favoring ECM accumulation.⁴⁰⁸ Human and rat mesangial cells exposed to glycated bovine serum albumin showed increased IGF-1, IGF-2, and TGF-β₁ mRNAs, IGF-1 and TGF-β proteins, and ECM (laminin, type IV collagen, and fibronectin) proteins and mRNAs. These effects were prevented by aminoguanidine or by an AGE-receptor 1 antagonist.⁴⁰⁹ Both high glucose and glycated bovine serum albumin can increase AGE-receptors expression in rat mesangial cells.⁴¹⁰ Induction of diabetes increases mRNA expression and immunohistochemistry staining of AGE-receptors in rat glomeruli.⁴¹⁰ Further, intraperitoneal AGE-modified mouse albumin increases glomerular TGF-β1 and ECM components mRNA⁴¹¹ and causes glomerular hypertrophy. These effects are reduced by aminoguanidine.⁴¹¹ Oral treatment with aminoguanidine also normalized IGF-1 and IGFBP renal mRNA expression in diabetic rats.⁴¹² Aminoguanidine also diminished albuminuria and the increase in collagen-related fluorescence in the aorta, glomeruli, and renal tubules of diabetic rats.⁴⁰⁵ In this study, aminoguanidine prevented the increase in Vv(Mes/glom) but did not prevent GBM thickening. However, another study found aminoguanidine to ameliorate the GBM thickening but not the Vv(Mes/glom) increase in diabetic rats.⁴¹³ In similar studies (M.W. Steffes, M. Mauer, unpublished data), aminoguanidine failed to reverse GBM thickening and mesangial expansion in severe longstanding diabetic rats. Novel inhibitors of advanced glycation have been shown to exhibit renoprotective effects similar to aminoguanidine.^414,415 OPB-9195 reduced renal TGF-β1, VEGF, and type IV collagen mRNA expression, glomerular TGF-β1, VEGF, type IV collagen, and AGE immunoreactivity, albuminuria, and progressive glomerulosclerosis, without changing glycemia, in an animal model of type 2 diabetes.^416,417 Moreover, serum AGE levels were independent predictors of progression of early structural glomerular lesions (Vv[MM/glom]) in microalbuminuric type 1 diabetic patients.⁴¹⁸ Pimagedine, a second-generation AGE inhibitor, reduced urinary protein excretion in proteinuric type 1 diabetic patients and also decreased the chances of doubling serum creatinine in patients with an initial creatinine lower than 1.5 mg/dL.⁴¹⁹ In a randomized, double-blind, placebo-controlled study, pimagedine significantly reduced proteinuria and the rate of GFR decline in type 1 diabetic patients with nephropathy and retinopathy at baseline.⁴²⁰ Also, fewer pimagedine-treated patients experienced a three-step or greater progression of the retinopathy (Early Treatment of Diabetic Retinopathy Study) score.⁴²⁰ However, some patients receiving high-dose pimagedine developed glomerulonephritis.⁴²⁰

Further studies of glycation’s role in pathogenesis and treatment strategies for DN are warranted. However, the fact that glycation is a nonenzymatic process dependent on the duration and magnitude of glycemia currently leaves unexplained why only about half of patients with very poor glycemic control develop clinical DN.⁹⁴

Increased Activity of Aldose Reductase

Aldose reductase, the enzyme that catalyses the reduction of glucose to sorbitol in the polyol pathway (see Fig. 28-15), has been associated with DN and other microvascular complications. Aldose reductase, not present in all mammalian cells, is present in all the target tissues of diabetic complications. These tissues include kidney, lens, retinal capillary wall pericytes, vascular endothelium, and peripheral nerve (Schwann cells). Increased activity of aldose reductase leads to accumulation of sorbitol, which is further converted to fructose by the sorbitol dehydrogenase enzyme, using NAD⁺ as substrate. The NAD⁺/NADH ratio decreases, and the conversion of glyceraldehyde 3-phosphate to 1,3 bisphosphoglycerate is blocked, leaving more substrate (glyceraldehyde 3-phosphate) for the synthesis of α-glycerol phosphate, a diacylglycerol precursor. Diacylglycerol is a PKC activator that, as discussed, could regulate ECM synthesis and removal. Also, increased activity of aldose reductase consumes NADPH, leading to decreased or depleted glutathione. As already noted, glutathione is used by glutathione peroxidase to reduce peroxides or O₂°, yielding oxidized glutathione. Glutathione can also act as a detoxificant for carbonyl compounds.³⁵² Hypertonicity stimulates aldose reductase gene expression, and kinases (p38-MAPK and ERKs) appear to be involved in the regulation of this response.⁴²¹ Both glucose and H₂O₂ activate p38-MAPK in vitro, suggesting that increased glucose may contribute to increased aldose reductase expression independently of hyperosmotic conditions. PKC also has a role in controlling aldose reductase gene transcription.⁴²² Moreover, reactive oxygen species decrease NO content, which activates aldose reductase. Increased aldose reductase activity may also contribute to impaired renal autoregulation.⁴²³ An aldose reductase inhibitor prevented type IV collagen and fibronectin accumulation in human renal proximal tubular cells exposed to high glucose.⁴²⁴ Treatment of diabetic animals with aldose reductase inhibitors have generated divergent results.^425–427 One study showed prevention of renal and glomerular hypertrophy but not GBM thickening,⁴²⁸ while another reported prevention of mesangial expansion.⁴²⁹ Tolrestat, an aldose reductase inhibitor, prevented increased GFR, AER, and glomerular hypertrophy and attenuated the accumulation of basement membrane–like material in diabetic rats, with no effect on mesangium expansion.⁴³⁰ Peripheral blood mononuclear cell aldose reductase mRNA levels are increased in patients with type 1 diabetes and DN but not in diabetic patients without nephropathy.⁴³¹ Aldose reductase inhibitors have a partial^432–434 or no^435,436 effect in ameliorating renal microvascular complications in human studies.

The Steno Hypothesis

According to the Steno hypothesis, albuminuria reflects widespread vascular damage.⁴³⁷ Hyperglycemia, by decreasing the synthesis of and reducing sulfation levels can lead to partial depletion of heparan sulfate,⁴³⁸ the main glycosaminoglycan component of glomerular basement membranes.⁴³⁹ Reduced renal heparan sulfate can decrease the glomerular capillary wall electrostatic charge barrier.⁴⁴⁰ Abnormal glomerular charge permselectivity has been observed in diabetic rats⁴⁴¹ and humans,⁴⁴² suggesting that loss of glomerular capillary-wall proteoglycans may be responsible for the initial increase in the excretion of negatively charged albumin.^439,443 Changes in heparan sulfate have also been associated with other diabetic complications.^444,445 Administration of a modified heparin glycosaminoglycan preparation to mesangial cells exposed to high glucose in vitro caused inhibition of TGF-β1 overexpression, probably through inhibition of the glucose-induced PKC activation.⁴⁴⁶ The changes in the balance of α1(IV) collagen, MMP-2, and TIMP-2 in human and murine cells exposed to high glucose were partially reversed by heparin supplementation.⁴⁴⁷ Animal studies have shown that subcutaneous administration of low molecular weight heparin, perhaps by inducing heparin sulfate synthesis, prevented increased AER, GBM, and α1(IV) collagen expression and deposition in diabetic rats.^448–450 Reduction in AER was also reported after treatment with subcutaneous low molecular weight heparin or similar drugs in microalbuminuric and proteinuric type 1^451–454 and type 2⁴⁵⁵ diabetic patients, but the results are still controversial.⁴⁵⁶ Danaparoid, a mixture of sulfated glycosaminoglycan consisting mainly of heparan sulfate, not only decreased AER⁴⁵³ but also reduced retinal hard exudates in proteinuric type 1 diabetic patients.⁴⁵⁷ These studies suggest that decreased glycosaminoglycans can have a pathogenetic role in the development of diabetic complications. However, it is also clear from earlier discussions (see Kidney Structure in Diabetes section) that important DN lesions can develop in patients with normal glomerular permselectivity. Thus, measurably abnormal glomerular permselectivity is not necessary for the genesis of DN lesions.

In summary, several mechanisms have been proposed to explain the development of renal complications in diabetes. It appears that multiple pathways interact and are involved in a process that is probably genetically regulated. Recent advances in molecular biology, genetics and epigenetics will hopefully bring new insights as to the mechanisms involved in the genesis of DN. These advances will allow early identification of patients who are at risk of—or safe from—DN and will hopefully lead to specific preventive strategies based on the understanding of the pathophysiologic bases of DN. Moreover, the demonstration that DN lesions are reversible could lead to treatment aimed at stimulating the healing capacities of the kidney, with the goal of restoring lost function or preventing further loss.