structural-functional correlations of diabetic nephropathy

Kidney International, Vol. 45 (1994), pp. 612—622

NEPHROLOGY FORUM

Structural-functional correlations of diabetic nephropathy

Principal discussant: S. MICHAEL MAUER

University of Minnesota Medical School, Minneapolis, Minnesota

Patient 1. A 30-year-old woman had had type-I (insulin-dependent)diabetes mellitus (IDDM) for 27 years. At the time of evaluation, sheweighed 50 kg, was 157 cm tall, and had a blood pressure of 114/80 mmHg. She had background retinopathy and no clinical neuropathy. Herurine was positive for protein by dipstick and the albumin excretion ratewas 2500 /tg/min. Her hemoglobin A1 was 8.9%; serum creatinine, 1.2mg/dl; and creatinine clearance, 79 mI/mm/i .73 m2. As part of the studyof renal structure and function in siblings concordant for IDDM, sheunderwent a percutaneous renal biopsy.

Light microscopy revealed that 30% of the glomeruli were globallysclerosed. The others all had moderately severe diffuse mesangialexpansion. Many afferent and efferent glomerular arterioles showedvarying amounts of hyalinosis, with some vessel walls completelyreplaced by the waxy, homogeneous, pink hyaline material. Milddiffuse interstitial expansion was evident, and patchy areas of moremarked interstitial fibrosis were found, mainly in association withatrophic tubules or sclerotic glomeruli. Electron microscopic morpho-metric analysis showed the mesangial volume fraction (that is, fractionof glomerular volume occupied by mesangium [VvMesJ) to be 0.40(upper limit of normal, 0.22). The VvMes cells measured 0.11 andVvMes matrix 0.26; the ratio of mesangial matrix to total mesangiumwas 0.65 (normal ratio 0.5). The surface density of the peripheralglomerular basement membrane (that is, the filtration surface per unit oftuft area {SvPGBM]) was 0.087 (lower limit of normal, 0.095) and theGBM width was 661 nm (upper limit of normal, 450 nm).

Patient 2. The 28-year-old sister of Patient 1 developed IDDM at age5. She had no retinopathy or neuropathy. She weighed 60 kg, was 164cm tall, and had a blood pressure of 102/64 mm Hg. The urine was

negative for protein by dipstick; urine albumin excretion rate was 75g/min. Her hemoglobin A1 was 14%; serum creatinine, 0.9 mg/dl; andcreatinine clearance, 71 ml/min/l .73 m2.

Light microscopic evaluation of her renal biopsy specimen wassimilar to that of her sister. Morphometric analysis revealed a VvMes of0.45, a ratio of mesangial matrix to total mesangium of 0.63, anSvPGBM of 0.83, and a GBM width of 712 nm.

Discussion

DR. S. MICHAEL MAUER (Co-Director, Pediatric Nephrol-ogy, Department of Pediatrics, and Professor of Pediatrics,University of Minnesota Medical School, Minneapolis, Minne-sota): Diabetic nephropathy represents the single most impor-tant cause of renal failure in adults in the Western world and isresponsible for 25% of all new cases of uremia [1]. Overall,approximately 35% to 45% of patients with long-standing insu-lin-dependent diabetes mellitus (IDDM) ultimately will developnephropathy, defined as dipstick-positive proteinuria, hyper-tension, and falling GFR [2]. The peak incidence for thedevelopment of clinical renal disease occurs after about 16years of IDDM but does not reach zero even after 40 years [2].Patients with diabetic nephropathy account for virtually all theincreased mortality associated with type-I diabetes [2]; thosewho escape nephropathy have a mortality rate that differs littlefrom that of the nondiabetic population [2]. It is clear thatclinical diabetic nephropathy is the consequence of the ad-vanced development of a constellation of specific lesions uniqueto the diabetic state. These advanced lesions represent the finalstages in a continuum of progression of renal structural abnor-malities. Animal and human studies support the general conceptthat the factors involved in the genesis of the early structuralconsequence of renal injury might be different from the factorsdriving the progression of advanced renal structural changes; inthis Forum, I will focus primarily on the earlier aspects ofdiabetic nephropathy, because an understanding of thesephases of the natural history of this disorder (genesis) is morelikely to promote the development of specific prevention strat-egies than is knowledge about the later stages of the disease(progression). In addition, I will attempt to develop the conceptof susceptibility to diabetic nephropathy; the lack of uniformityin the renal structural response to the diabetic state appears torepresent a critical variable determining the development ofserious diabetic renal disease.

Presentation of this Forum is made possible by grants from MerckSharp & Dohme International; Dialysis Clinic, Incorporated; MarionMerrell Dow, Incorporated; and Mead-Johnson Pharmaceuticals.

© 1994 by the International Society of Nephrology

Pathologic findings

The early pioneering work of Østerby demonstrated that theglomerulus, measured morphometrically, is enlarged to somedegree at the onset of IDDM but otherwise appears normal [3].

612

EditorsJORDAN J. COHENJOHN T. HARRINGTONNIC0LA0s E. MADIAS

Managing EditorCHERYL J. ZUSMAN

State University of New York at Stony Brookand

Tufts University School of Medicine

Case presentations

Nephrology Forum: Diabetic nephropathy 613

Our studies of identical twins discordant for IDDM confirmedthis finding [41. In these latter studies the nondiabetic twins,presumably identical to their twins with IDDM except for thepresence of diabetes, all had normal glomerular structure. Thusthe diabetic state itself, rather than some independent factor(s),appears to be a prerequisite for the expression of diabeticglomerulopathy. Østerby demonstrated that GBM thickeningcan be detected 2 years after the onset of IDDM; mesangialexpansion can be quantitated after 3.5 to 5 years of diabetes [3].Ultimately, in patients developing serious renal disease, thedominant pathologic process is most often diffuse mesangialexpansion [5, 6]. The Kimmeistiel-Wilson nodular lesion is themost marked expression of pathologic mesangial expansion indiabetes. While specific for diabetes, only a minority of patientswho develop overt diabetic nephropathy have nodular glomer-ulosclerosis [7].

Other common lesions in the kidney of patients with diabetesinclude afferent and efferent glomerular arteriolar hyalinosisand exudative hyaline lesions (fibrin caps) in the subendothelialspace and under the parietal layer of the epithelium alongBowman's capsule (capsular drops) [8]. The latter might beconsequent to the adhesions of tufts containing capsular dropsto the parietal layer of glomerular epithelium, followed by GBMrupture, extrusion of the fibrin cap, and subsequent retractionand remodeling of the glomerular tuft. The hyaline lesions in theglomerular arterioles can be among the earliest renal changes inIDDM and could be important in promoting glomerular scarring[9]. Capillary loops can sometimes develop microaneurysms,which appear to be the forerunners of Kimmeistiel-Wilsonnodules [10]. Epithelial and endothelial cell structures aregenerally well preserved until late in the course of nephropathy[11, 12]. Diffuse interstitial expansion can occur early in dia-betic nephropathy [5, 13], but advanced tubulointerstitialchanges are seen primarily as the disease progresses to its laterstages [5, 141.

Structural-functional relationshipsAccurate morphometric techniques have allowed quantita-

tion of the structural changes of diabetic nephropathy [15, 16].This has facilitated correlation of these changes with renalfunction.

Glomerular and tubular basement membranes. Measurementof GBM width using precise approaches such as the orthogonalintercept method [15] demonstrate thickening in almost allpatients with long-standing IDDM [3, 5, 17]. In identical twinsdiscordant for IDDM, GBM width is always greater in thediabetic twin, even if still in the normal range (Fig. 1). Thisthickening, detectable after 2 years of diabetes [3, 7], is gener-ally uniform both within and between glomeruli, and primarilyinvolves the lamina densa. Initial immunofluorescent studies byFalk et a! that showed increased staining for normal GBMconstituents [18] suggest that increased production, decreasedturnover, or both result in the accumulation of GBM material.More recent studies by Kim et a! showed that some GBMconstituents, such as the cs3(IV)NC and a4(IV)NC domains oftype-IV collagen, have normal or increased staining intensityalong the central aspect of the GBM that persists into advancedstages of the disease, but cd(IV) and a2(IV)NC domains oftype-TV collagen are decreased in staining intensity along theinner (endothelial) aspect of the GBM [19]. These studies

Fig. 1. Glomerular (GBM) and tubular basement membrane (TBM)widths in nondiabetic (S) and type-I diabetic (0) identical twins. Twinpairs are connected by lines. Note that the GBM width of nondiabetictwin is always in normal range (shaded areas) and, whether in thenormal range or above, that of the diabetic twin is always greater thanthat of the nondiabetic twin. The same is true of TBM width, but normalvalues are not yet established. (From Ref. 4.)

confirm the structural heterogeneity of the GBM and indicatethat changes in extracellular matrix (ECM) that occur in dia-betic nephropathy are highly site specific. Preliminary quanti-tative immunogold electron microscopic studies by Zhu andcoworkers describe decreased density of type-IV collagen inthe al(IV) and a2(IV)NC domain distribution along the inner(endothelial) aspects of the GBM in patients with more ad-vanced lesions [20]; this work confirms the immunofluorescentmicroscopy observations. Mesangial matrix density also tendedto be decreased. Makino and colleagues, using scanning elec-tron microscopy, showed mesangial interposition in advanceddiabetic nephropathy, and offer a possible explanation for thedecrease in al(IV) and a2(IV) staining at both the mesangialand GBM sites in advanced disease [211. Vernier et a! foundthat histochemically identified subepithelial GBM charge-sitedensity, corresponding to heparan sulfate proteoglycan, wasreduced in more advanced cases of proteinuric diabetic patientsbut was normal at the earlier stages and with the milder lesionsof normo- or microalbuminuric patients [221. Makino et a! alsofound decreased heparan sulfate staining by immunogold elec-tron microscopy in advanced cases of diabetic nephropathy[21]. In parallel, epithelial cell structure measured either as footprocess width or as the more sensitive parameter of filtrationslit-length density is normal in patients with IDDM until high-level microalbuminuria or overt proteinuria is present [23]. Inother words, only when advanced structural changes of diabe-tes, including GBM and mesangial expansion, are present arethe earliest epithelial cell structural changes detectable [23].

In summary, early GBM thickening is not associated withspecific or yet-identified ECM compositional changes, but late

E

1000

800

600

400

200

0

Glomerular Tubular

2000

1600

1200

800

400

01

7

614 NephrologyForum: Diabetic nephropathy

lesions are associated with increases in some and decreases inother ECM components. Thus these late changes might moreclosely relate to the progression as opposed to the genesis ofthe earlier GBM widening.

In diabetic patients, immunofluorescence microscopy dis-closes increased localization of plasma proteins in GBM as wellas in other renal basement membranes, including tubular base-ment membranes (TBM) and Bowman's capsule [24, 25]. Evi-dence suggests that this increased localization is related t thecharge of the circulating molecules, as anionic proteins bindpreferentially. This binding is relatively irreversible, however,and requires harsh treatment with acids to release the boundprotein [25]. Binding of these proteins could relate to decreasednet negative charge of the GBM, perhaps secondary to de-creased heparan sulfate content.

The GBM width correlates relatively poorly with mesangialvolume fraction and with the functional parameters of GFR andhypertension [5]. However, GBM width correlates better withurinary albumin excretion (UAE) [5] and with lamina raraexterna charge-site density [221 (Fig. 2); thus GBM width mightbe interrelated with alterations in glomerular permselectivity indiabetes. The pathogenesis of this relationship is not known.

Because it can be precisely measured, increased GBM widthis a sensitive marker of the diabetic state [3, 8, 171. As Imentioned, in pairs of identical twins discordant for 10DM,only diabetic twins showed GBM thickening; thus the diabeticstate is a prerequisite for expression of this abnormality [4].Tubular basement membrane width also is increased in patientswith IDDM [4]. This finding suggests that pressures and flows inthe glomerular capillaries (normally much higher than thosepresent in peritubular capillaries) are not prerequisites for thedevelopment of renal basement membrane thickening in IDDM.

Mesangium. Shortly after the onset of IDDM, the fraction of

the glomerular tuft occupied by mesangium is normal, withroughly equal proportions of cells and matrix. Mter 2 to 3years, matrix begins to accumulate [12, 261, and mesangialexpansion out of proportion to any increase in glomerularvolume (increased mesangial volume fraction or VvMes) be-comes quite marked after 10 to 30 years of diabetes in thosepatients destined to develop clinical diabetic renal disease [5].Approximately two-thirds of this mesangial expansion is due tomatrix accumulation [27] and, initially, the expanding me-sangium appears to contain increased amounts of the usualmesangial extracellular matrix antigens [18]. In contrast, ad-vanced mesangial expansion (diffuse glomerulosclerosis) is as-sociated with decreased intensity of al(IV) and a2(IV)NCdomain [19] and decreased merosin (a laminin-like molecule)staining intensity (Kim Y, personal communication). Further,advanced expansion is accompanied by increasing deposition ofscar collagens (types I and II). Even in relatively early lesions,however, there might already be a subtle decrease in the densityof the classical collagen chains in mesangial matrix, as assessedby quantitative immunogold electron microscopy [20]. As thesefindings indicate, the nature of the ECM molecules that accu-mulate during the development of diabetic nephropathy lesionsis not well understood.

Approximately one-third of mesangial expansion in diabetesis due to cellular expansion. Normal mesangial cell antigens,including smooth muscle antigens, are initially increased inimmunofluorescent staining intensity [28]. It is unclear whethermesangial cell number, cell size, or both are increased at thevarious stages of diabetic nephropathy. Although the totalnumber of cell nuclei per glomerulus is increased [29], glomerulitend to be larger in diabetic patients, and this increase in cellnuclei per glomerulus could be proportional to that enlarge-ment. Also, pathologists using light or electron microscopicroutine sectioning strategies tend to overestimate the number ofmesangial cells by misclassifying endothelial cells as mesangial.Thus, although mesangial cell proliferation is a possibility indiabetes, it is far from proven. Despite the frequent acceptanceof this possibility as an established fact, the appropriate studiesto determine whether mesangial cell proliferation actually oc-curs in diabetes have not been done. Most of the mesangialexpansion occurring in diabetes is due to an increase in relativeand absolute mesangial matrix volume [27]. In fact, the propor-tion of the mesangium that is matrix is increased even inpatients with IDDM whose mesangial volume fraction is in thenormal range [271 (Fig. 3).

The mesangium and glomerular filtration rate (GFR). Mesan-gial expansion is closely associated with renal function indiabetic nephropathy [5]. A strong inverse correlation existsbetween mesangial volume fraction and GFR. This relationshipprobably results from the expanding mesangium compromisingthe structure of contiguous glomerular capillaries. We defineglomerular filtration surface as the trilaminar GBM, endothelialcell, and epithelial cell interface, with no intervening me-sangium. As the relative volume of the glomerulus occupied bymesangium increases, the relative peripheral capillary wallfiltration surface decreases (Fig. 4). Substantial support for thisconcept comes from the work of Ellis et al [30], confirmed byØsterby and coworkers [311, showing that peripheral capillary

D

D

0

0

E

a

800

700

600

500

400

300

0*

0a

I • I • I • I • I • I • I • I•18 10 12 14 16 18 20 22 24

LRE, sitespermicron

Fig. 2. Relationship of charge-site density in lamina rara externa(LRE) and GBM width in type-I diabetic patients. (From Ref. 22.)


0.2

0.00.0 0.2 0.4 0.6 0.8 1.0

Mesangium VvFig. 3. Ratio of mesangial matrix to total mesangial volume (matrixand cell) to mesangial volume fraction (Vv) in type-I diabetic patients.The normal ratio is 0.5. Note that almost all diabetic patients are abovethis value regardless of whether or not mesangium Vv is increased(above 0.22). (From Ref. 27.)

filtration surface per glomerulus is directly related to GFR (Fig.5).

The mesangium and hypertension. Virtually all patients withmarked mesangial expansion (specifically, a mesangial volumefraction > 37%, when mesangium is defined excluding mesan-gial GBM) have overt proteinuria. Seventy-five percent of thepatients with these advanced mesangial lesions have hyperten-sion [5], but as Patient I illustrates, 25% are normotensive.Severe mesangial expansion is thus closely associated with allthe manifestations of overt nephropathy, decreased GFR, pro-teinuria, and hypertension.

As I noted, 25% of patients with overt nephropathy andmarked mesangial expansion do not have hypertension. Fur-ther, hypertension can occur in patients with IDDM who havelesser degrees of mesangial expansion. Hypertension in apatient with IDDM is most frequently associated with increasedurinary albumin excretion (UAE) and well-established glomer-ular lesions. Hypertension does not appear to be more commonthan in the general population in patients with IDDM who donot have other structural or functional features of advanceddiabetic nephropathy.

The mesangium and proteinuria. How is albuminuria relatedto the structural lesions of diabetic nephropathy? Using mesan-gial expansion as the key structural parameter [5, 6], we foundthat patients with long-standing IDDM and entirely normalrenal function (as indicated by normal UAE, GFR, and bloodpressure) have mesangial volume fractions ranging from normalto levels bordering on those regularly associated with overtnephropathy (Group 1 in Fig. 6) [32]. Patients with microalbu-minuria (increased UAE but below detectability by dipstick),normal GFR, and normal blood pressure (UAE usually <45mg!24 hrs or 30 g/min) have mesangial volume fractionscompletely overlapping with those in patients whose renalfunction is normal (Group 2 in Fig. 6) [32]. Patients with IDDM

who have microalbuminuria and either hypertension, decreasedGFR, or both (UAE usually >45 mg124 hrs) have more ad-vanced mesangial expansion (Group 3 in Fig. 6) [32]. Values forUAE between the upper limit of normal (22 mg/24 hr or 15g/min) and 45 mg!24 hr (30 g/min) do not always predict the

later development of overt nephropathy. Thus normal UAEdoes not preclude the presence of important lesions, whilemicroalbuminuria in the lower ranges does not necessarilypredict structural abnormalities. However, if other manifesta-tions of overt nephropathy, including hypertension or de-creased GFR, are present, microalbuminuria is usually in thehigher ranges and generally indicates more advanced structuralinjury [32]. This relationship is well illustrated by today'ssecond patient; she had microalbuminuria (UAE of 108 mg/24hr or 75 sg/min), reduced GFR, and advanced glomerularlesions.

In summary, microalbuminuria in the range regularly "pre-dictive" of the later development of proteinuria is, in fact, amarker of well-established diabetic glomerulopathy [32, 33].Nonetheless, measurement of UAE is the most reliable testavailable for identifying patients at high risk of progression to

1.0

0.8

c,)0.6

0.4

.•.

0.16 o

00.15

00.14 0

0.13 0

00.12

C.)

' 0.11 0U) 0

0.10° a

a 0 00 0

o o.og 0., 00

a)0.08 0

0.07 0 0

Cl) 000.06 0 o00.05

0

0

0.04

0.03

0.02k

0

10 20 30 40 50 60 70 80

% Total mesangium

Fig. 4. Relationship between the surface density (Sv) of the peripheralcapillary filtration surface and the total % mesangium or mesangialvolume fraction (Vv). (From Ref. 5.)

0 0Do

P 0DO El0

PDU El

a 0 DDaD 0na 0•

B00DO

616 Nephrology Forum: Diabetic nephropathv

0.0Group I Group II Group III

Fig. 6. Volume fraction of the mesangium (VvMes) in the studygroups. The shaded area represents the mean 2 SD in normalsubjects. The broken line represents the level above which overtnephropathy regularly occurs (Group Ill vs. Group I, two-tailed P =0.002; Group III vs. Group II, two-tailed P = 0.007 by Mann-WhitneyU test). Most patients in Group II had UAE <45 mg/24 hr (30 sglmin);most in Group III had UAE > 45 mg/24 hr. (From Ref. 32.)

overt nephropathy at a time when the disease might still beamenable to therapeutic intervention.

Glomerular arterioles. Hyaline degenerative changes caninvolve both the afferent and efferent glomerular arterioles indiabetes. This involvement ranges in severity from smallamounts of PAS-positive translucent material under the elasticmembrane of the arterioles to virtually complete replacement ofthe smooth muscle of the vessels by this material. Hyaline inthese vessel walls, along the parietal layer of Bowman's capsule(capsular drop), and in the glomerular subendothelial space(fibrin cap), appears to contain the same materials: immunoglo-bulins, fibrinogen, and complement, as well as other plasmaproteins [34]. The pathogenesis of these lesions is unknown. Werecently found that the frequency of the more advanced gb-merular arteriolar lesions, defined as complete replacement ofsmooth muscle cells by hyaline, correlated directly with greaternumbers of globally sclerotic glomeruli; this relationship isillustrated by the patients presented here [91. This observationsupports the hypothesis that global glomerular sclerosis indiabetic nephropathy might derive in part from vascular pathol-ogy. This view is suggested by studies indicating that thecortical distribution of global glomerular sclerosis in patientswith IDDM fits a pattern consistent with a vasculo-occlusiveprocess. An additional mechanism of glomerular sclerosis indiabetes might be extreme mesangial expansion with conse-quent capillary closure [29].

Interstitium. Interstitial fibrosis is a concomitant of diabeticnephropathy. Bader and colleagues showed an inverse relation-ship between morphometric measures of the severity of inter-stitial fibrosis and GFR [131. In our studies of structural-functional relationships, we found a significant correlation amongseveral structural variables, including mesangial volume fraction,arteriolar hyalinosis, percentage of glomerular sclerosis, and cor-tical interstitial volume fraction [35]. All four variables correlatedwith GFR and UAE. Mesangial volume fraction and corticalinterstitial volume fraction were each partially independent vari-ables relating to these functional parameters.

SusceptibilityThe patients presented today illustrate one of the central

themes of this Forum. Both these siblings, the first with overtproteinuria and the second with microalbuminuria, have ad-vanced diabetic nephropathology, including light microscopicfindings of arteriolar hyalinosis, global glomeruloscierosis, andinterstitial expansion, as well as electron microscopic morpho-metric findings of similarly increased GBM width and increasedmesangial volume fraction, and decreased filtration surfacedensity. Thus these sisters have remarkable concordance ofglomerular, vascular, and tubulointerstitial lesions of diabetes.Evidence is developing to support the view that the rate ofprogression of the lesions of diabetic nephropathy that arecritical in determining clinical outcome is only partially deter-mined by glycemic control, and is highly influenced by suscep-tibility factors that might be genetic, environmental, or both.

Along these lines Seaquist et al examined nephropathy risk inIDDM multiplex families (in which more than one sibling hasdiabetes). The proband (the first of the siblings to developdiabetes) had a duration of diabetes of at least 10 years and thesibling a duration of at least 7 years [36]. Of the siblings ofprobands free of diabetic nephropathy, 17% had a UAE above45 mg124 hr (30 gImin). In contrast, 82% of the siblings ofprobands with end-stage renal disease had end-stage renaldisease or a UAE above 45 mg/24 hr, and only 17% werenormal.

Borch-Johnsen and colleagues found a higher prevalence ofovert nephropathy in siblings of probands with nephropathy(33%) than in siblings of normoalbuminuric probands (10%;P < 0.04) [37]. Both studies were unable to separate geneticsusceptibility from similarities due to shared environment. Inthe latter study, hemoglobin A1C loosely but significantlycorrelated in the probands and siblings, but environmental andgenetic factors that could potentially affect glycemic controlcould not be isolated.

These studies of familial predisposition to renal disease in

SS

S

See SS

S

S

SS •I

ISS

0.4

0.3

0.2

0.1

IS

E

E

Co

210

190

170

E 1503 130

: 70

50

30

0 --

0 50 100 150 200 250 300

Capillary filtration surface area glomerulus, x 1Opm2

Fig. 5. Direct correlation between glomerular filtration rate estimatedby creatinine clearance and capillary filtration surface area per gb-merulus. (From Ref. 30.)

.5IS


patients with IDDM are supported by observations in twogenerations of Pima Indians with type-Il (non-insulin-depen-dent) diabetes mellitus (NIDDM) [381. In these studies, afteradjustment for other risk factors, proteinuria occurred among14% of the offspring if neither parent had proteinuria, in 23% ifat least one parent had proteinuria, and in 46% if both parentshad diabetes and proteinuria. Preliminary studies by Fioretto etal, from which the two patients presented here were taken [39],indicate that among IDDM sibling pairs, the concordance innephropathy incidence is probably based on a concordance innephropathic lesions. These ongoing studies reveal less vari-ance between siblings than among all patients for structuralparameters, including GBM width and mesangial fractionalvolume. Thus, after correction for duration of IDDM, the effectof sibship on lesions was still significant [39]. These data showthat the strong familial concordance for nephropathy risk oper-ates through a concordance in nephropathic lesions, but thesestudies do not determine whether this shared risk is due tocommon genetic or environmental influences.

The risk of diabetic nephropathy among Israeli Jews ofnon-Ashkenazi origin is higher than that in Ashkenazi Jews[40]. This 1991 study showed that glycemic control and lowersocioeconomic status significantly and independently corre-lated with albuminuria. However, even after the authors ad-justed for these variables, patients of non-Ashkenazi origin stillhad an almost threefold greater risk of microalbuminuria. Also,the odds ratio for risk of nephropathy increased as the quartileof glycemic control worsened [40]; risk factors for nephropathythat might be inherited thus might be far more likely to beexpressed in patients with poor glycemic control and lowsocioeconomic status. A similar point was made by Krolewskiet al, who suggested that the increased risk for diabetic ne-phropathy in patients with a genetic background of essentialhypertension was most clearly expressed in patients with poorglycemic control [41].

If a familial and ethnic propensity to diabetic nephropathy ispartly based on genetic risk factors, then a search for candidategenes is logical. Such efforts have begun. Some authors [41, 42]but not all [43] have found that parents of diabetic offspring withnephropathy have higher levels of arterial blood pressure thando parents whose diabetic offspring do not have diabeticnephropathy. Familial clustering of cardiovascular disease oc-curs in diabetic patients with nephropathy. Cardiovasculardisease was twice as common a cause of death among parents ofdiabetic patients with nephropathy than among parents ofdiabetic patients without nephropathy. Among diabetics withnephropathy, those who had a cardiovascular event were 6times more likely to have a familial history of cardiovasculardisease than were those who had no such event [441.

Hypertension has been implicated in the natural history ofboth cardiovascular and renal disease of diabetes mellitus. Thelevel of activity of the red blood cell sodium/lithium counter-transport system (Na/LP CTT) has a high degree of heritabil-ity [451. Increased Na/Li CTT activity is associated with anincrease in blood pressure [46]. Two independent studies foundthat the activity of this cation exchanger was elevated ininsulin-dependent diabetic patients with diabetic nephropathy[41, 47]. Another study found that although the Na/Li CTTactivity was higher in diabetic patients, no difference existedbetween those with and without nephropathy [48]. A study in

England found that red cell Na/Li CTT was elevated inparents whose diabetic offspring had nephropathy comparedwith parents whose diabetic offspring had escaped this compli-cation [49]. This finding was not confirmed in a Danish study,however [43]. The question of whether genes regulating red cellNa/Li Cr1' can confer risk of, or protection from, diabeticnephropathy has not yet been answered.

Red blood cell Na/Li CTF activity has similarities with aphysiologic ubiquitous cation exchanger—the sodium/hydrogenantiport (Na/H). This cell membrane exchanger is involvedin the regulation of intracellular pH, cell volume and growth,and sodium reabsorption in the proximal renal tubule [50—52].Increased Na/H antiport activity has been described inleukocytes and red blood cells from humans with either essen-tial hypertension [53] or diabetic nephropathy [54]. Also, instudies of cultured skin fibroblasts from patients with IDDM,Na/H antiport activity was increased in those with nephrop-athy compared with those without nephropathy [55—57]. Cur-rent evidence suggests that Na/H exchange is mediated bydifferent Na/H exchanger subtypes and that the same sub-type of amiloride-sensitive exchanger (NHE-1) can be found infibroblasts, in mesangial cells, and in the basolateral site ofrenal tubular epithelial cells [58, 59]. Thus, changes in mem-brane cation transport can be present in a variety of cell typesin patients with diabetic nephropathy. A clear genetic linkbetween increased Na/H antiport activity and diabetic ne-phropathy risk is far from established, however. Extracellularmatrix molecules such as fibronectin can interact with integrinsto activate fibroblast Na/H antiport activity [60], so it isconceivable that the abnormalities of this cation transporter incultured skin fibroblasts of patients with IDDM and nephropa-thy are secondary to other cellular perturbations.

Doria and colleagues presented preliminary data suggestingthat insulin receptor gene polymorphism is associated withnephropathy risk [611, but that patients with microalbuminunaand overt proteinuria depart from Hardy-Weinberg equilibriumin opposite directions [611. Although these interesting prelimi-nary data require confirmation, they do illustrate the potentialpower of molecular biology in unraveling the genetics of sus-ceptibility to diabetic nephropathy.

It is also noteworthy that the siblings presented here evi-denced discordance between the severity of the renal, ocular,and neurologic complications of diabetes. Although frequentlyconcordant, such discordant cases make it clear that pathoge-netic factors specific to the organs at risk of diabetic complica-tions must be operative. In summary, these two sisters withIDDM illustrate the constellation of the specific and uniquelesions that represent diabetic nephropathology and depict theinterrelationships of renal pathology and function in this dis-ease. They also exemplify the growing interest in unraveling thegenetic mechanisms that might be important in the pathogenesisof this disorder.

Questions and answers

Da. NICOLAOS E. MADIAS (Chief, Division of Nephrology,New England Medical Center, Boston, Massachusetts): Wouldyou please speculate on what role glycemia might play in thegenesis of the pathologic abnormalities of diabetic nephropa-thy?

618 Nephrology Forum: Diabetic nephropathy

DR. MAUER: The advanced structural abnormalities of dia-betic nephropathy seem to require an increase in glucose to bemanifested. A few convincing cases in the literature suggestthat the abnormalities in glycemia required to produce theselesions are very subtle [621. We've all heard about, read about,and sometimes see patients with minor abnormalities in glucosetolerance or just-discovered type-IT diabetes who have ad-vanced renal or retinal disease. Glycemia is clearly a risk factorbut, as data from the Joslin Clinic and other centers have shown[631, even the best-controlled patients are at risk, although atlow risk. In developing our hypothetical constructs, we need tokeep in mind the huge variability among the structural abnor-malities among the end organs at risk in the same individual aswell as the variability in susceptibility among individuals. Wealso must keep in mind that there are factors within families thatdetermine how this susceptibility works. Thus, although glyce-mia-related abnormalities may be important, they do not ex-plain the whole story. For example, the process of alteration ofmolecules by glycosylation is presumably non-enzymatic andrelated to the duration and magnitude of the hyperglycemia. It'sthus unlikely that glycosylation per se could explain everything.At best it must be a partial explanation or must work throughintermediary steps that are themselves genetically regulated.Similarly, the aldose reductase story, very much related toglycemic control, could explain part but not all of the clinicalphenomena of diabetic complications. We need to better under-stand the factors regulating the balance between matrix produc-tion and matrix breakdown. Evidence in animals suggests thatoverproduction might not be the key variable [64]. Matrixbreakdown in diabetes has received too little attention. Thefactors that could be influencing the rate at which matrixaccumulates could include hemodynamic factors, growth fac-tors, and other variables known to influence the rate at whichmatrix is made and broken down. Very little specific informa-tion is available. This is a wide open field for new work.

DR. MADIAS: I am interested in the changes in GBM collagenthat you described. I realize that these are still preliminaryobservations, but how specific are those changes for diabeticnephropathy as compared with other nephropathies?

DR. MAUER: In terms of immunochemical analysis by thelight microscopic method, they seem quite specific for diabetes[191. The other disease that Dr. Youngki Kim in our group haslooked at carefully is membranous nephropathy, and there areclear differences from diabetes [651. Not all of the new markersfor matrix proteins have been explored in all diseases, so it'svery early to be certain that the changes in diabetes are unique,but currently this appears to be the case. In terms of electronmicroscopic immunochemical studies, even less is known aboutother renal diseases. Currently in our studies we are dividingdiabetic patients into "slow-" and "fast-track" for the devel-opment of diabetic nephropathy lesions. The question beingasked is whether early on specific matrix abnormalities couldindicate whether an individual is at risk or safe from the laterdevelopment of progressive diabetic renal injury. The answer isnot yet available.

DR. JoHN T. HARRINGTON (Chief of Medicine, Newton-Wellesley Hospital, Newton, Massachusetts): I'm interested inyour experience with serial biopsies in individual patients. Inwhat kind of patients have you done such studies, and what isthe rate of growth of the matrix portion of the mesangium?

DR. MAUER: To use serial biopsies to look at whether the rateof development of lesions is linear in a given case, one needs 3spaced biopsies. Two biopsies will not tell us whether thelesions are linear over time, Our only experience so far is in therenal allograft. It is easier to obtain biopsy tissue from thetransplanted kidney, and the baseline biopsy representing akidney never exposed to the diabetic state can be performed atthe time of transplant surgery. Our findings support the viewthat the lesions develop linearly over time [66]. Our studies alsodemonstrate marked variability in the rate of development ofmatrix lesions between different individuals (unpublished data).There are no carefully done studies at 3 time intervals in thenative kidney, however.

DR. MADIAs: There are now diabetic patients with microal-buminuria who have been treated for several years with ACEinhibitors and who have shown stabilization of the level ofmicroalbuminuria or return of albumin excretion to the normalrange. Does any information from baseline and followup biop-sies on such patients allow structural-functional correlation?

Da. MAUER: Not that I know of.DR. ANDREW S. LEVEY (Division of Nephrology, New En-

gland Medical Center): One potential way of separating glomer-ular changes due to diabetes from changes due to nephropathyis to examine insulin-dependent individuals who have had 20 to40 years of diabetes but who have not developed nephropathy.Have you examined the ultrastructural features of glomerulifrom these patients?

DR. MAUER: We find the longer the duration of diabetes, theless is the correlation between duration and severity of thelesion. What drives the relationship between duration andseverity is the fact that serious lesions are uncommon in the first5 to 10 years. For example, with 30 years of diabetes, lesionsrange from end-stage renal disease to virtually none. We'vestudied 2 sisters—one with 55 years and the other with 45 yearsof diabetes—whose kidney structures were still in the normalrange. Beyond 30 years of IDDM, one sees fewer individualswith advanced lesions, because the people who are going todevelop nephropathy already have done so.

DR. LEVEY: It is certainly very interesting that some patientswith structural evidence of diabetic nephropathy have not yetmanifested abnormal urinary albumin excretion after 30 to 40years.

Da. RICHARD LAFAYETTE (Division of Nephrology, NewEngland Medical Center): In your discussion regarding thepathogenesis of albuminuria, you suggested that mesangialexpansion leads to changes in epithelial cell function andmembrane charge. Does the loss of membrane charge densityoccur only in glomeruli with significant mesangial expansion, ordoes it occur diffusely?

DR. MAUER: The glomeruli are remarkably uniform in dia-betic nephropathy. Thus, morphometric analysis of 3 glomerulican show a correlation with GFR or with any other functionalmeasures. We measured charge site density only in the laminarara externa; we don't know what's going on in the other sitessuch as the mesangium and other areas of the GBM. Thechanges in the lamina rara externa charge-site density and inepithelial cell structure seem to develop in parallel with risingalbumin excretion rates. Since lesions are generally well devel-oped when UAE is substantially increased, these GBM charge-site and epithelial changes occur relatively late in the process.


The structural parameter that is strikingly disturbed, with lessoverlap in patients who have versus those who do not havehigh-level microalbuminuria than any other structural parame-ter, is the reduction of relative filtration surface. One possibilityis that this loss of filtration surface provides a signal to theglomerulus to alter its homeostasis, perhaps its internal hemo-dynamics. There are other possibilities including growth fac-tors, etc. However, it is possible that hemodynamic alterationsresulting from established lesions and reduction in filtrationsurface could lead to changes in charge-site density and epithe-hal cell structure and pore size. These latter changes could thenbe either a consequence or a cause of altered glomerularpermeability. These thoughts are consistent with the idea thatimproved glomerular hemodynamics induced by ACE inhibi-tors reduce glomerular permeability in proteinuric [671 andmicroalbuminuric [68] diabetic patients.

DR. JACK LEAHY (Endocrine Division, New England MedicalCenter): I have two questions that might be naive, but they areclinically based. Your work compares structure-function rela-tionships. Clinically, we rarely perform biopsies on individualswho have diabetes. In the second case that you presented, thepatient had microalbuminuria with already markedly abnormalrenal structure. Also, you describe people who didn't have anyproteinuria at all but who appear to have major structuraldamage from diabetes. What role, if any, do biopsies play indiagnosing or managing diabetic renal disease?

DR. MAUER: That is a difficult question. In part the answerdepends on what we can offer the patient as a consequence ofour examination. First, diabetologists tend to overestimate theseriousness of the renal biopsy in terms of discomfort or risk.Given current technology, it's a minor procedure. Second, letme argue that from a nephrologist's point of view, the dominantreasons for performing a renal biopsy are diagnosis and prog-nosis. Only 20% to 25% of the time do the biopsies lead to atherapeutic decision that would not have been made without thebiopsy. Third, other diseases such as systemic lupus erythema-tosus (SLE) have a far lower risk of producing end-stage renalfailure. Further, renal disease in SLE is less associated withexcess mortality compared with renal disease in the type-I ortype-TI diabetic patients. Yet patients with SLE undergo renalbiopsy much more frequently than do people with diabetes.Performing biopsies in diseases such as SLE has taught usmuch about the natural history of the disease, and sometimeswe make treatment decisions based on these biopsies. Why dowe not biopsy the diabetic patients? Diabetic patients don'tdevelop clinical manifestations until late in the disease. As youpoint out, the current evidence indicates that serious underlyingstructural injury is often present before functional tests candetect the damage. Unfortunately, in diabetic patients we donot have the advantage of using markers of ongoing injury suchas microhematuria. In my view, the biopsy is not used enoughin diabetic patients. At the very least, if we could detectpatients developing serious lesions, we could offer the patientimproved glycemic control. Certainly we should use relativelyminor manifestations of renal disease, such as increasing albu-min secretion (microalbuminuria) or changing blood pressure(even if it's normal but increasing for that patient), as indicatorsof serious underlying events. However, as I have pointed out,some patients do not manifest these findings until structuralchanges are relatively far advanced. In fact, some patients may

have a decline in GFR before microalbuminuria is manifest [69].This tends to argue that we should be using the renal biopsymore frequently since functional markers are imprecise. Oneimplication of your question is that without functional abnor-malities, the structural lesions might not be important. How-ever, serial biopsy studies (Fioretto P, unpublished data) indi-cate that the lesions in normoalbuminuric patients withstructural abnormalities are progressive. The number-one causeof renal failure in the Western world is diabetes, and we do havetherapeutic alternatives. We need to apply those alternatives tothe patients at risk and avoid their use in patients not at risk. Wecurrently don't have any way of selecting patients very early intheir course of developing important renal lesions except for therenal biopsy, and we can't change the natural history of diseaseto one that is more to our liking, that is, one with earlierfunctional warnings of underlying structural changes. I arguefor a much closer relationship between diabetologists andnephrologists. In the past, nephrologists have been restricted toa functional evaluation of diabetic nephropathy even though,like the eye, the function only deteriorates when it's too late.Imagine asking ophthalmologists to limit themselves to func-tional eye tests for diabetic retinopathy, precluding their look-ing directly at the retinal vasculature that is being damaged.Under these circumstances, would laser treatment of prolifer-ative retinopathy have developed?

DR. LEAHY: I have great difficulty in understanding renaldisease in type-I versus type-Il diabetes. Clearly, these patientsmanifest many similarities, but some striking dissimiiaritiesalso exist. How do you view the pathogenesis of the differentforms of diabetes?

DR. MAUER: Much less is known about type-Il for a numberof reasons. The risk of renal disease in type-TI diabetes isprobably not very different from that in type-I, but the age ofonset is so much later that many patients die of other causes inthe time frame required for them to develop renal disease. InPima Indians who develop type-IT diabetes at an earlier age, therelationship between duration and risk of terminal uremia isapproximately the same as in type-I. In a couple of studies, 25%of the proteinuria in type-Il diabetics results from another renaldisease [70, 71]. But the chance of finding other than diabeticrenal disease in a biopsy specimen from a proteinuric patientwho has had tyne-I diabetes for 10 years is only about 4%. Thusif your patient has had IDDM for 10 years and has proteinuria,you can have 95% certainty that diabetes is the cause of theproteinuria. In type-TI, your certainty is down to 75%. If thetype-TI patient has retinopathy, the risk of error in diagnosis ismuch lower [70]. Gambara and coworkers in Bergamo recentlypublished the description of a blending between glomerosclero-sis and diabetic nephropathy in type-IT patients—not surprisingin that age population [72]. I would also speculate that thiscombination of lesions might be why black people with type-ITdiabetes have a higher risk of end-stage renal disease comparedwith whites, whereas in black people with type-I diabetes, thatis, the younger population, there is little or no increased risk[73]. I imagine that the combination of hypertensive and dia-betic nephropathy in the type-Il population is responsible forthe increase in risk among black people with type-TI diabetes.Thus, in the type-Il patient, the causes of renal failure tend tobe more multifactorial: a higher incidence of other renal dis-eases, more hypertensive nephropathy, and more effects of


aging in addition to diabetic nephropathy. However, the major-ity of type-IT patients developing clinical renal disease have"garden variety" diabetic nephropathy.

DR. MADrAS: You mentioned that substantial evidence indi-cates that IDDM patients with microalbuminuria or clinicalproteinuria feature an overactive Na/H exchanger in variouscells. Indeed, Dr. Viberti and colleagues have shown that innormotensive, normo- or microalbuminuric IDDM patients,increased red blood cell Na/Li CTT activity was associatedwith reduced insulin sensitivity, left-ventricular hypertrophy,and lipid disturbances [74]. I am particularly fascinated by thefindings of a study by Can et al in normotensive and normoal-buminuric IDDM patients: 33% of these patients had increasedred blood cell Na/Li Cr1' activity and were the very samepatients who featured hyperfiltration and increased filtrationfraction compared with IDDM patients without elevated activ-ity versus control subjects [75]. The authors speculated thatincreased activity of this transporter might identify IDDMpatients destined to develop essential hypertension and diabeticnephropathy; this concept has potentially tremendous clinicalimplications. Do you have any data on the activity of theNa/H exchanger in that subgroup of IDDM patients youmentioned who have normal albumin excretion but elevatedblood pressure? Also, is there any information correlatingactivity of the Na/H exchanger with renal histology innormoalbuminuric individuals?

DR. MAUER: Not that I know of. Regarding the hypertensivenormoalbuniinui-jc patients, Dr. Paola Fioretto and the group inPadova led by Dr. Romano Nosadini has shown that diabetichypertensive patients with normal red blood cell Na/Li CTTactivity tend to have normal albumin execretion rates [76].Also, work from the Padova group confirmed that Na/Hantiport and red blood cell Na/Li CTT are functional modesof the same transporter in type-I diabetic patients [77]. How-ever, studies have not been designed specifically to answer thisquestion, and hypertensive normoalbuminuric patients withIDDM are relatively rare. There is currently inadequate workdone to answer your question regarding Na'7H exchangeractivity and renal histology in normoalbuminuric patients. Dr.Lurbe in Dr. Batlle's group in Chicago working with Dr.Fioretto and our group showed that normoalbuminuric type-Idiabetic patients have normal cultured fibroblast Na/H ex-changer activity, and had, as expected, less serious renallesions than did a proteinuric group of patients [78]. However,at present there is insufficient information to determine whetherraised Na/H exchanger activity in normoalbuminuric pa-tients predicts worse underlying nephropathology or a worseprognosis.

DR. ANDREW KING (Division of Nephrology, New EnglandMedical Center): Several inherited factors have been postulatedto increase the risk of progression of renal disease. Recently, ithas been postulated that nephron number is one such factor[79]. In this regard, does the rate of progression in a diabeticpatient predict the rate of progression alter transplantation inthe same patient?

DR. MAUER: No, it does not. We looked at the regressionanalysis of the number of years of type-I diabetes to end-stagerenal failure and the rate of development of changes of diabetesin the renal allograft, and there was no relationship whatsoever[66].

DR. JosE C. TEIxEut DA SILVA (Fellow in Nephrology,Boston University Medical Center, Boston): A recent publica-tion by Hogan et al shows that advanced glycosylation endproducts quenching nitric oxide decrease the antiproliferativeeffect of the latter on mesangial cells [80]. What would be theimpact on the progression of diabetic nephrosclerosis? Is pro-liferation of mesangial cells a factor of mesangial matrix expan-sion and sclerosis?

DR. MAUER: Let me make a couple of points in this regard.We have no clear evidence yet that mesangial cell proliferationis a component of diabetic nephropathy. The bulk of mesangialexpansion in diabetes is matrix and not cell [27]. Cell numberhas never been carefully enumerated; it's extremely difficult todo, even at the electron microscopic level. On a given section,a skilled individual will mistake endothelial cells for mesangialcells 20% to 30% of the time. You have to do serial sections toenumerate mesangial cells; otherwise it's not an accurate mea-sure. So there is no clear evidence yet that mesangial cellproliferation out of proportion to the increase in glomerularvolume is a component of diabetic nephropathy. Second, I can'tcomment much on the relationship between the nitric oxide andglycosylation except to say that we have renal physiologic datain animals suggesting that the nitric oxide system might beoveractive rather than underactive in the diabetic rat kidney[81]. We need to do a lot more work to determine the role of thenitric oxide system in the kidney in diabetes both in terms ofrenal hemodynamics and in terms of cell proliferation andmatrix dynamics.

DR. HARRINGTON: If you take a kidney from a diabeticanimal and transplant it into a nondiabetic animal, does thedisease resolve and, if so, over what time period?

DR. MAUER: We did that back in 1975. Using light micro-scopic studies, we put normal kidneys into diabetic rats andfound that the lesions developed. We then took kidneys thathad been in the diabetic environment for 6 or 7 months andtransplanted them into normal rats, and we found regression ofthe lesions within 2 months [82].

DR. MADIAs: What kind of rats did you use?DR. MAUER: These were highly inbred Lewis rats, so rejec-

tion of the kidney was avoided. However, when patients with15 to 20 years of diabetes and well-established lesions are curedof diabetes with a successful pancreas transplant, the renallesions do not improve over the next 5 years [83]. Whether thismeans early lesions are reversible while late lesions are not, orthat rat and human lesions are inherently different, is notknown.

Acknowledgments

The work described in this Forum was supported in part by NationalInstitutes of Health grants AM 13083, AM 20742, and RROO400, and bythe Juvenile Diabetes Foundation.

Reprint requests to Dr. S. M. Mauer, Department of Pediatrics, Box491 UMHC, University of Minnesota Medical School, 420 DelawareStreet S.E., Minneapolis, Minnesota 55455, USA

References

1. Ro TKS, HIRSH S, AVRAM MM, FRIEDMAN EA: Prevalence ofdiabetic nephropathy in Brooklyn, in Diabetic Renal-Retinal Syn-drome, edited by FRIEDMAN EA, L'ESPERANCE FA JR, New York,Grune and Stratton, 1980, pp 205—207


2. ANDERSEN AR, CHRISTIANSEN JS, ANDERSEN JK, KREINER S,DECKERT T: Diabetic nephropathy in type 1 (insulin dependent)diabetes: An epidemiological study. Diabetologia 25:496—501, 1983

3. Ø5TERBy R: Early phases in the development of diabetic glomeru-loscierosis. Acta Med Scan 574 (suppl 1): 1—80, 1975

4. STEFFES MW, SUTHERLAND DER, GOETZ FC, RICH SS, MAUERSM: Studies of kidney and muscle biopsy specimens from identicaltwins discordant for type I diabetes mellitus. N Engi J Med312:1282—1287, 1985

5. MAUER SM, STEFFES MW, ELLIS EN, SUTHERLAND DER,BROWN DM, GOETZ FC: Structural-functional relationships indiabetic nephropathy. J Clin Invest 74:1143—1155, 1984

6. STEFFES MW, STERBY R, CHAVERS B, MAUER SM: Mesangialexpansion as a central mechanism for loss of kidney function indiabetic patients. Diabetes 38:1077—1081, 1989

7. GELLMAN DD, PIELkNI CL, SOOTHILL JF, MUEHRCKE RC, MADU-RUS W, KAaI RM: Structure and function in diabetic nephropathy.The importance of diffuse glomerulosclerosis. Diabetes 8:251—256,1959

8. MAUER SM, STEFFES MW, BROWN DM: The kidney in diabetes.Am J Med 70:603—612, 1981

9. HARRIS RD, STEFFES MW, BILous RW, SUTHERLAND DER,MAUER SM: Global glomerular sclerosis and glomerular arteriolarhyalinosis in insulin dependent diabetes. Kidney Int 40:107—114,1991

10. SAIro Y, KIDA H, TAKEDA 5, YOSHIMURA M, YOKOYAMA H,K0sHIN0 Y, HATTORI N: Mesangiolysis in diabetic glomeruli: Itsrole in the formation of nodular lesions. Kidney mt 34:389—396,1988

11. FARQUHAR MG, HOPPER J, MOON HD: Diabetic glomerulosclero-sis: Electron and light microscopic studies. Am J Pathol 35:721—753, 1959

12. ØSTERBY R: The number of glomerular cells and substructures inearly juvenile diabetes. Acta Pathol Microbiol Scand [A] 80:785—800, 1972

13. BADER R, BADER H, GRUND KE, MACKENSEN-HEAH S, CHRIST 5,BONLE A: Structure and function of the kidney in diabetic glomer-ulosclerosis: Correlations between morphologic and functionalparameters. Pathol Res Pract 167:204—216, 1980

14. CHURG J, DACI-IS 5: Diabetic renal disease: Arteriosclerosis andglomerulosclerosis. Pathol Annu 1:148—171, 1966

15. JENSEN EB, GUNDERSEN HJG, 05TERI3Y R: Determination ofmembrane thickness distribution from orthogonal intercepts. JMicrosc 115:19—33, 1979

16. WEIBEL ER: Stereological Methods: Practical Methods for Biolog-ical Morphometrj. London, Academic Press, 1979

17. ØSTERBY R: Morphometric studies of the peripheral glomerularbasement membrane in early juvenile diabetes. II. Development ofinitial basement membrane thickening. Diabetologia 8:84—92, 1972

18. FALK RJ, SCHEINMAN JI, MAUER SM, MICHAEL AF: Polyantigenicexpansion of basement membrane constituents in diabetic nephrop-athy. Diabetes 32:34—39, 1983

19. KIM YK, KLEPPEL MM, BUTKOWSKI R, MAUER SM, WIESLANDERJ, MICHAEL AF: Differential expression of basement membranecollagen chains in diabetic nephropathy. Am J Pathol 138:413—420,1991

20. ZHU D, KIM Y, STEFFES MW, MAUER SM: Glomerular (G)distribution of extracellular marix (ECM) components in insulin-dependent diabetes (D) by high resolution quantitative immuno-chemistry (HRQI) (abstract). JAm Soc Nephrol 3:770, 1992

21. MAKINO H, YAMASAKI Y, HARAMOTO T, SHIKATA K, HIRONAKAK, OTA Z, KANWAR YS: Ultrastructural changes of extracellularmatrices in diabetic nephropathy revealed by high resolution scan-ning and immunoelectron microscopy. Lab Invest 68:45—55, 1993

22. VERNIER RL, STEFFES MW, SissoN-RosS 5, MAUER SM: Heparansulfate proteoglycan in the glomerular basement membrane in typeI diabetes mellitus. Kidney mt 41:1070—1080, 1992

23. ELLIS EN, STEFFES MW, CHAVERS B, MAUER SM: Observationsof glomerular epithelial cell structure in patients with type Idiabetes mellitus. Kidney mt 32:736—741, 1987

24. MILLER K, MICHAEL AF: Immunopathology of renal extracellularmembranes in diabetes mellitus. Specificity of tubular basementmembrane immunofluorescence. Diabetes 25:701—708, 1976

25. MICHAEL AF, BROWN D: Increased concentration of albumin inkidney basement membrane in diabetes mellitus. Diabetes 30:843—846, 1981

26. ØSTERBY R: A quantitative electron microscopic study of mesangialregions in glomeruli from patients with short-term juvenile diabetesmellitus. Lab Invest 29:99-1 10, 1973

27. STEFFES MW, BILOUS RW, SUTHERLAND DER, MAUER SM: Celland matrix components of the glomerular mesangium in type Idiabetes. Diabetes 41:679—684, 1992

28. SCHEINMAN JF, FISH AJ, MICHAEL AF: The immunohistopathol-ogy of glomerular antigens: The GBM, collagen, and actomyosinantigens in normal and diseased kidneys. J Clin Invest 54:11,414—11,454, 1974

29. GUNDERSEN HJG, STERBY R: Glomerular size and structure indiabetes mellitus. II. Late abnormalities. Diabetologia 13:43—48,1977

30. ELLIS EN, STEFFES MW, GOETZ FC, SUTHERLAND DER, SIM-MONs RL, NAJARIAN JS: Glomerular filtration surface in type Idiabetes mellitus. Kidney Int 29:889—894, 1986

31. ØSTERBY R, PARVING H-H, NYBERG G, HOMMEL E, JORGENSENHE, LOKKEGAARD H, SVALANDER C: A strong correlation betweenglomerular filtration rate and filtration surface in diabetic nephrop-athy. Diabetologia 31:265—270, 1988

32. CHAVERS BM, BIL0uS RW, ELLIS EN, STEFFES MW, MAUER SM:Glomerular lesions and urinary albumin excretion in type I diabeteswithout overt proteinuria. N EngI J Med 320:966—970, 1989

33. STEFFEs MW, CHAVERS BM, BIL0uS RW, MAUER SM: Thepredictive value of microalbuminuria. Am J Kidney Dis 13:25—28,1989

34. BURKHOLDER PM: Immunohistopathologic study of localizedplasma protein and fixation of guinea pig complement in renallesions of diabetic glomerulosclerosis. Diabetes 14:755—770, 1965

35. LANE PH, STEFFES MW, FIORETTO P, MAUER SM: Renal intersti-tial expansion in insulin-dependent diabetes mellitus. Kidney Int43:661—667, 1993

36. SEAQUIST ER, GOETZ FC, RICH 5, BAaosA J: Familial clusteringof diabetic kidney disease: Evidence for genetic susceptibility todiabetic nephropathy. N Engi J Med 320:1161—1165, 1989

37. BORCH-JOHNSEN K, NØRGAARD K, HOMMEL E, MATHIESEN ER,JENSEN JS, DECKERT T, PARVING H-H: Is diabetic nephropathy aninherited complication? Kidney Int 41:719—722, 1992

38. PETTITT DJ, SAAD MF, BENNETT PH, NELSON RG, KNOWLERWC: Familial predisposition to renal disease in two generations ofPima Indians with Type 2 (non-insulin-dependent) diabetes melli-tus. Diabetologia 33:438—443, 1990

39. FIORETTO P, STEFFES MW, BARBOSA J, SPRAFKA M, MAUER SM:Glomerular structure in siblings with insulin dependent diabetesmellitus (IDDM) (abstract). JAm Soc Nephrol 2:289, 1991

40. KALTER-LEIBOVICI 0, VAN DYK Di, LEIBOVICI L, LOYA N,ERMAN A, KREMER I, BONER G, ROSENFELD JB, KARP M, LAR0NZ: Risk factors for development of diabetic nephropathy andretinopathy in Jewish IDDM patients. Diabetes 40:204—210, 1991

41. KROLEWSKI AS, CANESSA M, WARRAM JH, LAFFEL LMB, CHRIST-LIEB AR, KNOWLER WC, RAND LI: Predisposition to hypertensionand susceptibility to renal disease in insulin dependent diabetesmellitus. N Engi J Med 318: 140-145, 1988

42. VIBERTI GC, KEEN H, WISEMAN MJ: Raised arterial pressure inparents of proteinuric insulin-dependent diabetics. Br Med J 295:515—517, 1987

43. JENSEN iS, MATHISEN ER, NORGAARD K, HOMMEL E, BORCH-JOHNSEN K, FUNDER J, BRAHM J, PARVING HH, DECKERT T:Increased blood pressure and erythrocyte sodium/lithium counter-transport activity are not inherited in diabetic nephropathy. Diabe-tologia 33:619-624, 1990

44. EARLE K, WALKER J, HILL C, VIBERTI GC: Familial clustering ofcardiovascular disease in patients with insulin-dependent diabetesin nephropathy. N EngI J Med 326:673—677, 1992

45. WEDER AB, FITZPATRICK MA, TORRETTI BA, HINDERLITER AL,EGAN BM, JULIUS 5: Red blood cell lithium-sodium countertrans-port as a genetic marker in essential hypertension. J Hypertens 4(suppl 6):S373—S375, 1986

46. CANESSA M, ADRAGNA N, SOLOMON HS, CONNOLLY B, TOSTESONDC: Increased sodium-lithium countertransport in red blood cells


of patients with essential hypertension. N Engi J Med 3110:772—776, 1980

47. MANGILIR, BENDING JJ, SCOTT G, Li LK, GUPTA A, VIBERTI GC:Increased sodium-lithium countertransport activity in red cells ofpatients with insulin-dependent diabetes and nephropathy. N EngIJ Med 318:146—150, 1988

48. ELVING LD, WETZELS JFM, Dn PONT JJHHM, BERDEN JHM Isincreased erythrocyte sodium-lithium countertransport a usefulmarker for diabetic nephropathy? Kidney mt 41:862—871, 1992

49. WALKER JD, TARIQ T, VIBERTI GC: Sodium-lithium countertrans-port activity in red cells of patients with insulin-dependent diabetesand nephropathy and their parents. Br Med J 302:635—638, 1990

50. ARONSON PS: Mechanisms of active H secretion in the proximaltubule. Am J Physiol 254:F647—F659, 1983

51. PARIS S, POUYSSEGUR J: Growth factors activate the sodium!hydrogen antiport in quiescent fibroblasts by increasing its affinityfor intracellular Ht J Biol Chem 259:10,989—10,994, 1984

52. OWEN NE: Effect of TPA on ion fluxes and DNA synthesis invascular smooth muscle cells. J Cell Biol 101:454—459, 1985

53. No LL, DUDLEY C, BOMFORD J, HAWLEY D: Leucocyte intracel-lular pH and Na-H antiport activity in human hypertension. JHypertens 7:471—475, 1989

54. No LL, SIMMONS D, FRIGII V, JAuuDo MC, BOMFORD J: Effect ofprotein kinase C modulators on the leucocyte Na/H antiport inType I diabetic subjects with albuminuria. Diabetologia 33:278—284, 1990

55. TREVISAN R, Lu. K, MESSENT J, TARIQ T, EARLE K, WALKER JD,VIBERTI GC: Na/H antiport activity and cell growth in culturedskin fibroblasts of IDDM patients with nephropathy. Diabetes41:1239—1246, 1992

56. DAVIES JE, No LL, KAFOED-ENEVOLDSEN AK, LI LK, EARLEKA, TREVISAN R, VIBERTI GC: Intracellular pH and Na/Hantiport activity of cultured skin fibroblasts from diabetics. Kidneymt 42:1184-1 190, 1990

57. LURBE A, FI0RLTT0 P, LAPOINTE M, MAIJER M, BATLLE DC:Enhanced proliferation in cultured skin fibroblasts from patientswith Type I diabetes and nephropathy. Role of Na/H antiport(abstract). JAm Soc Nephrol 3:761, 1992

58. ORLowsKI J, KANDASAMY RA, SHULL GE: Molecular cloning ofputative members of the NaJH exchanger gene family. J Biol Chem267:9331—9339, 1992

59. CLARK JD, LIMBIRD LE: Na/H exchanger subtypes: A predic-tive review. Am J Physiol 261:C945—C953, 1991

60. SCHWARTZ MA, LECHENE C, INGEER DE: Insoluble fibronectinactivates the Na/H antiporter by clustering and immobilizing inte-grin a5/31, independent of cell shape. Proc NatI Acad Sci USA88:7849—7853, 1991

61. DORIA A, WARRAM JH, KROLEWSKI AS: Insulin receptor genepolymorphism is associated with the development of overt protein-uria in IDDM (abstract). JAm Soc Nephrol 3:757, 1992

62. TCHOBROUTSKY G: Prevention and treatment of diabetic nephrop-athy, in Advances in Nephrology (vol 8), edited by HAMBURGER J,CROSNIER J, GRUNFELD J-P, MAXWELL MH, London, Year BookMedical Publishers, 1979, pp 63—68

63. KROLEWSKI AS, WARRAM JH, CHRISTLIEB AR, BUSICK EJ, KAHNCR: The changing natural history of nephropathy in type I diabetes.Am J Med 78:785—794, 1985

64. IHM C-G, LEE GSL, NAST CC, ARTISHEVSKY A, GUILLERMO R,LEVIN PS, GLASSOCK RL, ADLER SG: Early increased renalprocollagen nl(IV) mRNA levels in streptozotocin induced diabe-tes. Kidney mt 41:768—777, 1992

65. KIM Y, BUTKOWSKI R, BURKE B, KLEPPEL MM, CRossoN J, KATZA, MICHAEL AF: Differential expression of basement membranecollagen in membranous nephropathy. Am J Pathol 139:1381—1388,1991

66. MAUER SM, GOETZ FC, MCHUGH LE, SUTHERLAND DER, BAR-BOSA J, NAJARIAN JS, STEFFES MW: Long-term study of normal

kidneys transplanted into patients with type I diabetes. Diabetes38:516—523, 1989

67. MORELLI E, LOON N, MEYER J, PETERS W, MYERS BD: Effects ofconverting enzyme inhibition on barrier function in diabetic glom-erulopathy. Diabetes 39:76—82, 1990

68. WEGMANN JB, GIsH HERRON K, CHONKO AM, MACDOUGALLML, MOORE WV: Effect of angiotensin-converting enzyme inhibi-tion on renal function and albuminuria in normotensive type Idiabetic patients. Diabetes 41:62—67, 1992

69. LANE PH, STEFFES MW, MAUER SM: Glomerular structure inIDDM women with low glomerular filtration rate and normalurinary albumin excretion. Diabetes 41:581—586, 1992

70. PARVING H-H, GALL M-A, SKØTT P, JØRGENSEN HE, LØKEGAARDH, JØRGENSEN S, NIKLSON B, LARSEN 5: Prevalence and causes ofalbuminuria in non-insulin dependent diabetic patients. Kidney mt41 :758-762, 1992

71. AMOAII E, GLICKMAN JL, MALCHOFF CD, STEERGILL BC, KAISERDL, BOLTON WK: Clinical identification of nondiabetic renaldisease in diabetic patients with type I and type II disease present-ing with renal dysfunction. Am J Nephrol 8:204—211, 1988

72. GAMBARA V. MECCA G, REMUZZI G, BERTANI T: Heterogeneousnature of renal lesions in type II diabetes. J Am Soc Nephrol3:1467—1473, 1993

73. COWIE CC, PORT FK, WOLFE RA, SAVAGE PJ, MOLL PP. HAW-THORNE UM: Disparities in incidence of diabetic end-stage renaldisease according to race and type of diabetes. N Engi J Med321:1076-1079, 1989

74. LOPES DE FARIA JB, JONES SL, MACDONALD F, CHAMBERS J,MATTOCK MB, VIBERTI G: Sodium-lithium counter-transport ac-tivity and insulin resistance in normotensive IDDM patients. Dia-betes 41:610—615, 1992

75. C.&iut S, MBANYA IC, THOMAS T, KEAVEY P, TAYLOR R, ALBERTIKGMM, WILKINSON R: Increase in glomerular filtration rate inpatients with insulin-dependent diabetes and elevated erythrocytesodium-lithium countertransport. N EngI J Med 322:500—505, 1990

76. FIORETTO P, SAMBATARO M, CIPOLLINA MG, DUNER E, GI0RATOC, MOROCUTTI A, MOLLO F, BEN OP. CARRARO A, SACERDOTI D,BROCCO E, OPOCHER G, NOSADIN R: Impaired response to angio-tensin II in type I (insulin-dependent) diabetes mellitus. Role ofprostaglandins and sodium-lithium countertransport activity. Dia-betologia 34:595—603, 1991

77. SEMPLICINI A, MOZZATO M-G, SAMA B, NOSADINI R, FioanrroP, TREVISAN R, PESSINA A, CREPALDI G, DAL PALU C: Na/H andLi/Na exchange in red blood cells of normotensive and hyperten-sive patients with insulin dependent diabetes mellitus (IDDM). AmJHypertens 2:174—177, 1989

78. LURBE A, FloRErro P, LAPOINTE M, MAyER M, BATLLE DC:Enhanced proliferation in cultured skin fibroblasts from patientswith type I diabetes and nephropathy. (abstract). J Am SocNephrol3:761, 1992

79. BRENNER BM, COHEN RA, MILFORD EL: In renal transplantation,one size may not fit all. JAm Soc Nephrol 3:162—169, 1992

80. HOGAN M, CERAMI A, BUCALA R: Advanced glycosylation end-products block the antiproliferative effect of nitric oxide. Role inthe vascular and renal complications of diabetes mellitus. J ClinInvest 90:1110—1115, 1992

81. TOLINS JP, BROWN DM, R.&u L, SHULTZ P, MAUER SM: Abnormalhemodynamic response to renal hyperperfusion in diabetic rats:Role of nitric oxide. Am J Physiol (in press)

82. LEE SC, MAUER SM, BROWN DM, SUTHERLAND DER, MICHAELAF, NAJARIAN JS: Renal transplantation in diabetes mellitus inrats. J Exp Med 139:739—800, 1974

83. FIoa.al-ro P, MAUER SM, SUTHERLAND DER, BILOUS RW, GOETZFC, STEFFES MW: Effects of solitary pancreas transplantation onrenal structure in insulin-dependent diabetic patients with their ownkidneys. Lancet (in press)

structural-functional correlations of diabetic nephropathy

Documents