Kidney International, Vol. 43 (1993), pp. 381—386

Structural-functional relationships in type I mesangiocapillaryglomerulonephriti s

MASUJI HATTORI, YOUNGKI KIM, MICHAEL W. STEFFES, and S. MICHAEL MAUER

Department of Pediatrics, Hyogo College of Medicine, Hyogo, Japan; and from the Departments of Pediatrics and Laboratory Medicine andPathology, University of Minnesota Medical School, Minneapolis, Minnesota, USA

Structural-functional relationships in type I mesangiocapillary glomer-ulonephritis. We quantitated glomerular and cortical interstitial struc-tures in nine type I mesangiocapillary glomerulonephritis (MCGN)patients aged 6 to 20 years whose creatinine clearance (Car) was 10 to129 ml/minJl.73 rn,2 as compared to age-matched normal controls.Mean glomerular volume and mesangial volume fraction [Vv(mes/glom)] were increased and the percentage of the capillary endothelialcircumference which was defined as filtration surface was decreased intype I MCGN patients. Vv(mes/glom) was inversely related to filtrationsurface area per glomerulus (r = —0.73, P < 0.05) and directly tovolume density of cortical interstitium [Vv(int/cortex)] (r = +0.90; P <0.01). Vv(mes/glom) (r = —0.87; P < 0.01), filtration surface area perglomerulus (r = +0.83; P < 0.01) and Vv(int/cortex) (r = —0.86; P <0.01) were correlated with Car. Thus, in type I MCGN, measures bothof glomerular and of cortical interstitial structure are highly correlatedwith glomerular function.

Type I mesangiocapillary glomerulonephritis (MCGN) is aprimary glomerular disease with distinct morphologic patternsand hypocomplementemia. Structurally, there is marked me-sangial expansion, with a tendency to glomerular lobulation.Mesangial interposition between the endothelial layer and theglomerular basement membrane leads to thickening of capillarywalls and the characteristic "double contour" or tramtrackappearance. Many authors have reported that the long-termprognosis of type I MCGN is generally unfavorable, though ina given case it may be variable [1, 2]. Recently Schmitt et al [3]reported that within five years after biopsy, 49% of the patientshad died or needed dialysis treatment; after 10 years thisincreased to 64%. Clinically, nephrotic syndrome, hyperten-sion, and elevated serum creatinine values at the time of biopsyare of prognostic significance. Structurally, lobular glomerulo-nephritis, crescent formation and tubulointerstitial changeshave been related to the prognosis for this disease [4, 5].Further, it has been argued that interstitial but not glomerularstructural changes are related to functional alterations inMCGN [6, 7]. However, careful quantitation of both glomerularand interstitial structure and relating of these measures to renalfunction have not been done for MCGN as has been reportedfor diabetes [8]. Thus, we analyzed the renal biopsies of nine

Received for publication May 19, 1992and in revised form August 17, 1992Accepted for publication August 17, 1992

© 1993 by the International Society of Nephrology

381

patients with type I MCGN by morphometry, and the findingswere compared to age-matched normal kidney donors.

Methods

Patients

Nine (5 male) patients with type I MCGN were evaluated(Table 1). The diagnosis of type I MCGN was made using thefollowing criteria: (1) light microscopy, diffuse mesangial pro-liferation and thickening of glomerular capillary walls withdouble contours or tramtrack appearance; (2) electron micros-copy, subendothelial deposits and mesangial interposition, oc-casional subepithelial deposits; and (3) immunofluorescencemicroscopy, predominant C3 and properdin in a peripherallobular pattern, sometimes with IgG.

At the time of renal biopsy the nine patients were between 6and 20 years of age [11.1 4.6 years ( SD)]. Patientsnumbered 1, 8 and 9 were found to have abnormal urineanalysis on routine examination; patient #7 was found to havean abnormal urine analysis after complaints of abdominal pain;the remaining four patients presented with edema. Before renalbiopsy each patient had one or two 24-hour urine collections forcreatinine clearance (Car) and urinary protein excretion usingstandard laboratory procedures. Blood pressure data werebased on repeated determinations available over at least severaldays of the biopsy admission, and hypertension was defined bythe criteria of the Joint National Committee on the Detection,Evaluation and Treatment of High Blood Pressure [9] and theTask Force of Blood Pressure Control in Children [10]. Apercutaneous renal biopsy was performed for clinical indica-tions after obtaining informed consent from each patient or theresponsible parent. Biopsy was performed using the Vim-Silverman needle.

At the time of biopsy three patients had hypertension, fourimpaired renal function and six (all but patients number I and 2)had hypocomplementemia (Table 1). However, none of the fivepatients whose Cr exceeded 80 ml/min/l .73 m2 had hyperten-sion (Table 1). Renal biopsy was performed 16 days to 16months (average 4.5 months) from the first detection of thedisease; six cases (67%) were biopsied within the first twomonths of disease detection.

Control biopsies were obtained at the time of donation forkidney transplantation from five (4 male, 1 female) normal agedmatched people ages 6 to 20 years (13.6 6.6 years).

382 Hattori er a!: Structure and function in type I MCGN

Table 1. Characteristics of patients with type I MCGN

Case no.

Age atbiopyyears Sex

C atbiopsymi/mini1.73 m2

Urinaryprotein

excretiongidaylm2 Hypertension

Time offollow-up

since biopsymonths

Current cimica! status

C.mi/mini1.73 m2

Urinaryprotein

excretiorgidayim

1

23456789

207989

11146

16

MFFFMFMMM

1023346382

102119128129

3.372.502.082.831.840.150.472.822.32

YesNoYesYesNoNoNoNoNo

64203655

107593

CRF88327573

136137143154

TxND1.000.071.630.370.590.821.03

NMeanSD

911.14.6

976.746.3

92.041.08

925.227.2

8104.843.7

70.790.51

Abbreviations are: Car, creatinine clearance; CRF, chronic renal failure; Tx, renal transplantation; ND, not done.

Tissue processingA portion of the tissue was fixed in Zenker's solution,

embedded in paraffin, and sections stained with periodic acidSchiff (PAS) were used for interstitial and glomerular volumemeasurements.

Tissue for electron microscopy was fixed, embedded andsectioned as described elsewhere in detail [8]. Randomly se-lected centermost, nonscierosed glomeruli at least one tubulardiameter from the edge of the tissue were thin sectioned [8].Sclerotic glomeruli were excluded. At least three glomeruli foreach subject were photographed at magnifications of approxi-mately x3,900, xll,000 and x22,000. The x3,900 micrographswere placed together to form a montage of the entire glomerularcross section [11]. A calibration grid was photographed witheach montage to determine final magnification [8].

Morphometric analysisMean glomerular volume (GV) was determined on the light

microscopic sections at an approximate magnification of x 200using the method of Hirose et al [12]. Only sections at least 150m apart were used in order to minimize remeasuring the sameglomerulus. All biopsies but two had at least 30 giomeruli forevaluation. Case #3 had 14 and case #8 had 15 glomeruli. Thenumber of glomeruli averaged 56 44 (range 14 to 169).

Volume density of mesangium [Vv(mes/glom)] was measuredon the montage taken at x3900. A double lattice square gridwith equally spaced coarse points 6 cm apart, and equallyspaced fine points 3 cm apart was used so that each coarse pointdefined four fine points. Fine points were counted to determinethe number of points falling on mesangium (PM), whereascoarse points were counted to determine points falling on theglomerular area (PG). Since each coarse point defined four finepoints, then

PM ________________________________________Vv(meslglom)

= (m /pm ) [8, 13]r3 X 's

Then on photographs of an approximate magnification ofx 11,000 fine points were counted to determine the number of

points falling on the mesangial cell (PMC) and matrix (PMM)respectively. And

PMC 3VV(mes ceIl/glom) = VV(mes/glom) X . (sm /pm )

rf + PMM

PMMVV(mes matrix/glom) = V(mes/gIom) x (sm3/j.m)

MC + PMM

Mesangial volume per glomerulus was calculated by multiplying average glomerular volume times Vv(mes/glom) [141.

Surface density of the peripheral glomerular basement membrane (PGBM) [SV(pGBglQm)l and the mesangial capillar)lumenal surface [Sv(mes lumen/glom)] were determined b)counting interceptions (I) between the interface of interest anthe lines on the grid [15, 16]. Then,

2x1Sy = 2 X 1L =

60000(m3im3)

XPGmagnification

Where 'L = number of intercepts divided by the length of thgrid lines overlying the reference space, the glomerular tuft.

The PGBM was defined as the trilaminar surface of thepithelial cell, GBM and endothelial cell of the peripheracapillary wall, as previously described [14]. The mesangiacapillary lumenal surface represents the interface of endotheliacells with mesangial matrix or cellular components.

Together PGBM and mesangial capillary lumenal surfacrepresent the endothelial circumference of the glomerular capillary. Thus % of endothelial circumference occupied by PGBIvcan be determined.

PGBM% of endothelial circumference

Sv(PGBM/glom) x 100Sv(pGBM/glom) + Sv(mes lumen/glom)

And PGBM surface per glomerulus was then calculated bmultiplying average glomerular volume times SV(POBM/gIQm).

Hattori et a!: Structure and function in type I MCGN 383

Table 2. Morphometric data in type I MCGN and normal humans

Mean Mesangialglomerular volume!

volume glom.

Vv(mesangium!

glom.)

Vv(mes.cell!

glom.)

Vv (mes.matrix!glom.)

PGBM % ofendothelial

circumferencePGBM surface

area/glom.Xj

Sv Vv(PGBM/ (interstitium!glom.) cortex) GBM

widthnmCase no. x106 .un3 3/3 im34un3

1

23456789

2.9 1.62.3 1.33.6 1.92.4 0.94.7 2.02.1 0.72.4 0.62.1 0.74.0 1.6

0.560.560.530.390.420.360.240.330.41

0.330.360,340.260.270.230.170.230.26

0.230.200.190.130.150.130.070.110.15

30.018.414.734.213.536.850.351.535.4

34.616.217.996.856.094.8

216.3147.7115.4

0.01 0.480.01 0.330.01 0.250.04 0.220.01 0.210.05 0.130.09 0.110.07 0.180.03 0.17

338241320293377271308332449

MeanSDCV

Normal valuesMeanSDCVP

2.9 1.30.92 0.540.31 0.43

1.0 0.20.12 0.050.12 0.20

<0.001 <0.01

0.420.110.26

0.220.050.21

<0.01

0.270.060.23

0.120.030.22

<0.001

0.150.050.33

0.080.030.34

<0.02

31.614.10.44

62.43.40.05

<0.001

88.466.00.75

110190.17NS

0.03 0.230.03 0.110.86 0.49

0.12 0.130.02 0.030.19 0.26

<0.001 NS

325610.19

342600.18NS

Abbreviations are: mes, mesangial; gb m., glomerulu5; PGBM, peripheral gIomerular basement membrane.

Volume density of the cortical interstitium [Vv(int/cortex)]was determined on the light microscopic sections at an approx-imate magnification of x 170 by point counting images projectedonto a white surface with a projection microscope. Fine pointswere counted to determine the number of points falling on theinterstitium (P1) defined as points falling other than on glomer-uli, tubules and vessels larger than capillaries, whereas coarsepoints were counted to determine points falling on the renalcortex Since each coarse point defined four fine points,then

P1Vy(interstitiumjconex) =

Prc X (jsm3jsm3)

Mean glomerular basement membrane (GBM) width wasdetermined using the orthogonal intercept method [8, 17, 18].

StatisticsRelationships between renal structural and functional values

were examined by regression analyses (method of leastsquares). Comparisons of group data were performed by thetwo-tailed Student's t-test. Results are expressed as meanstandard deviation. Values of P < 0.05 were considered to bestatistically significant.

Results

Patient characteristicsRenal function among the nine MCGN type I patients varied

widely. Cr ranged from 10 to 129 mI/minll .73 m2, urinaryprotein excretion (UPE) from 0.15 to 3.37 g/day/m2 and UPE/mICcr from 2 to 423 rg/ml (Table 1). Three patients werehypertensive; all three had CCr less than 80 ml/min/l.73 m2.Structural-functional relationships were performed on baselinebiopsies prior to the institution of anti-inflammatory treatment.Followed for 3 to 64 months, four patients have improved, four

are stable and one with Cr of 10 at the time of study hasrequired renal transplantation. Earlier, patients were treatedwith daily prednisone and azathioprine. More recently, alter-nate day prednisone has been used. Given the small number ofpatients, the multiplicity of treatment schedules and the wideclinical range among the patients at baseline, no conclusionsregarding the effectiveness of these therapies could be derivedfrom the current study.

Structure in type I MCGN versuscontrols

Mean glomerular volume (GV) was markedly increased in allpatients, being 2 to 4.5 times the mean GV for the controlsubjects (Table 2). Whereas GV in the MCGN type I groupaveraged about three times that of the normals, mesangialvolume per glomerulus was more than five times that of thenormals. This reflected in an approximate doubling of theVv(mes/glom) in the patients (Table 2). Vv(mes cell/glom) wasgreater in the MCGN patients than in the controls (P < 0.001)and Vv(mes matrix/glom) was also increased (P < 0.02; Table2). In fact, on average about 2/3 of the increase in Vv(mes!glom)was due to increased Vv(mes celL/glom), and 1/3 due to in-creased Vv(mes matrix/glom).

The PGBM % of the endothelial circumference was de-creased in the Type I MCGN patients. However, because ofglomerular enlargement PGBM surface area per glomerulus wasnot significantly reduced in the patients compared to the con-trols (Table 2). Vv(int/cortex) varied widely among the patientsand was not significantly different from controls where widevariations were also seen. There was no significant difference inGBM width.

Structural-functional correlations in type I MCGN

Vv(mes/glom) correlated inversely with C. (r = —0.87; P <0.01; Fig. 1) and PGBM surface per glomerulus correlateddirectly with Cr (r = +0.83; P <0.01; Fig. 2). In addition, C.

384 Hattori et a!: Structure and function in type I MCGN

0.6m

Fig. 1. Relationship between Vv(mesangium/glom) and creatinineclearance (r = —0.87, P < 0.01).

BB

Fig. 2. Relationship between PGBM surface area per glomerulus andcreatinine clearance (r = +0.83, P <0.01).

0 20 40 60 80 100 120 140 160Creatinine clearance

mI/mm/i. 73 in2

Fig. 3. Relationship between Vv(interstitium/cortex) and creatinineclearance. Note log scale for Vv(interstitium/cortex). (r = —0.87, P <0.01)

0 100 200 300

alterations with the functional consequences of the disease.Confirming other studies [4, 5], patients with type I MCGN hada marked increase in glomerular volume to three times the meanglomerular volume as well as a near doubling of volume densityof mesangium per glomerulus. Thus, on average, the absolutevolume of the mesangium per glomerulus was increased morethan sixfold. This was mainly due to increased mesangial cellvolume whereas expansion of mesangial matrix was less impor-tant. This is opposite to what occurs in diabetic nephropathywhere mesangial expansion is due to increased mesangialmatrix [19]. Finally, compared to normal glomeruli, the per-centage of the glomerular capillary lumenal endothelial circum-ference which is peripheral GBM, was diminished in type I

B

E0

E

B

0.5

0.4

0.3

0.2

U

0

B1)

>>

0

U

0

0 20 40 60 80 100 120 140 160Creatinine clearance

mI/mm/i. 73 m

0

B

U)

E0

U)

0

300

200

100

0

a

B

BB

B

0

0 20 40 60 80 100 120 140 160Creatinine clearance

mI/min/i.73 m2

B

B

2.5

'a

2.0E

-< 1.0Cs0)

0.5.

0.0

PGBM surface area/glomerulusx103 pm2

Fig. 4. Relationship between mesangial volume per glomerulus andPGBM surface area per glomerulus (r = —0.73, P < 0.05).

fB

and log Vv(int/cortex) were inversely correlated (r = —0.87; P<0.01; Fig. 3). There was no significant relationship betweenCcr and mesangial volume per glomerulus (r = —0.49, NS) orGBM width (r = +0.41, NS).

There was an inverse relationship between mesangial volumeper glomentlus and PGBM surface area per glomerulus (r =—0.73; P < 0.05; Fig. 4). There were significant relationshipsbetween Vv(int/cortex) and Vv(mes/glom) (r = +0.91; P <0.01; Fig. 5) and SV(PGBM/gIom) (r = —0.76; P <0.05; Fig. 6).

Discussion

This is the first study providing detailed data regardingglomerular structural alterations and structural-functional rela-tionships in Type I MCGN. These studies are important in thatthey provide strong evidence linking glomerular structural

MCGN patients. This decrease was probably consequent tomesangial expansion [8] as well as to mesangial interposition.

Creatinine clearance in type I MCGN patients was highlycorrelated, inversely with Vv(mes/glom) and directly withPGBM surface area per glomerulus. Creatinine clearance was,as has previously been reported [6, 7], also correlated with theextent of the interstitial changes [6, 71. Thus, in type I MCGNbiopsies there are correlations between GFR and both theglomerular lesions and cortical interstitial lesions. These cor-relative findings cannot be given direct causal interpretationsand it is not possible to conclude with certainty whether it is theglomerular or the interstitial lesions or both which are thecrucial renal functional determinants in type I MCGN. How-

phenomenon. Also, it should be recognized that measurementsof cortical interstitial volume fraction are relatively crude anddo not per se indicate whether the changes represent abnormal-ities in interstitial matrix, alterations in interrelationships be-tween matrix, peritubular capillaries and interstitial lymphaticstructures [231, changes in interstitial cell number or type orabnormalities in interstitial water content. Changes in each ofthese interstitial components may well have different functionalconsequences. It would therefore be of great interest to studypure interstitial renal injury to determine whether interstitialalterations of the magnitude associated with those seen inglomerular disorders such as diabetes or MCGN are, by them-selves, associated with similar degrees of renal dysfunction.

In summary, renal functional disturbances are linked tomesangial expansion and especially to mesangial cellular expan-sion in MCGN type I. Efforts to understand the pathogenesis ofthis expansion might lead to new treatment options for thisimportant form of chronic glomerular injury.

Acknowledgments

We thank Ms. Cindy Dawis for her typing of this manuscript. Weappreciate the technical assistance of John Basgen, Tom Groppoli, andSusan Sisson-Ross. This work was presented at the 24th AnnualMeeting of The American Society of Nephrology in Baltimore, Mary-land, November 17—20, 1991.

Hattori et a!: Structure and function in type I MCGN 385

0.5U ever, it can no longer be accepted that one cannot predict

glomerular function from examining glomerular structure inprimary glomerular disorders. Morphometric analyses in dia-

0.4 betic nephropathy [8, 13], membranous nephropathy [20] andnow MCGN type I indicate that, carefully quantitated, glomer-ular structure is at least as informative as interstitial structure in

U predicting renal filtration function.0.3 In each of these three glomerular disorders important direct

correlations have emerged between measures of glomerularu filtration surface and GFR, providing a ready explanation for

the GFR alterations observed. On the other hand the mecha-> 02 El

nism whereby increased cortical interstitial volume may beEl influencing GFR is less clear. A significant inverse correlation

U has been reported between crude measures of kidney function

0 1 ________________________________________ (serum creatinine) and an estimate of interstitial capillary nu-

0 2 0 3 0 0 0 07 merical density in patients with a variety of glomerulopathies[21]. However, it is not possible to determine from these studies

Vv (mesangium/glom) whether the changes in interstitial capillaries are related toum34um3 interstitial pathology or represent downstream alterations con-

sequent to decreased glomerular capillary circulation. Since inFig. 5. Relationship between Vv(interstitium/cortex) and Vv(me- both diabetes [8, 22] and MCGN type I, the volume fraction ofsangium/glom) (r = +0.91, P < 0.01). . . .

cortex which is interstitium and the seventy of glomerularlesions are highly directly correlated, it is possible that theinterstitial changes in these disorders represent a secondary

0

U

U U

0

xa0

EE. u.

C

>>

0.00 0.02 0.04 0.06 0.08 0.10

Sv (PGBM/glom.)j.tm 3/pm3

Fig. 6. Relationship between Vv(interstitium/cortex) and Sv(PGBM/gloin). Note log scale for Vv(interstitium/cortex) (r —0.76, P < 0.05).

U

Reprint requests to S. Michael Mauer, M.D., Department of Pedi-atrics, University of Minnesota, Box 491, 515 Delaware Street S.E.,Minneapolis, Minnesota 55455, USA.

References

1. CHURO J, SOBIN LH: Diffuse mesangiocapillary glomerulonephri-tis, in Rena! Disease. Classification and Atlas of GlomerularDiseases, Tokyo, New York, Igaku-Shoin, 1982, pp. 83—110

386 Hattori et al: Structure and function in type I MCGN

2. CAMERON JS, OGG CS, TURNER DR, WELLER RO, WHITE RH,GLASGOW EF, PETERS DK, MARTIN A: Mesangiocapillary glomer-ulonephritis and persistent hypocomplementemia, in Glomerulone-

phritis, edited by KINCAID-SMITH P, MATHEW TH, LOVELL BEC-KER E, New York, John Wiley & Sons, 1973, pp. 541—556

3. SCHMITT H, BOHLE A, REINEKE T, MAYER-EICHBERGER D, VOGLW: Long-term prognosis of membranoproliferative glomerulone-phritis type I. Nephron 55:242—250, 1990

4. WHITE RHR: Mesangiocapillary glomerulonephntis, in PediatricKidney Disease, edited by EDELMANN CM JR. Boston, LittleBrown and Company, 1992, pp. 1307—1324

5. WEST CD, MCADAMS AJ: Membranoproliferative glomerulonephri-tis, in Pediatric Nephrology (2nd ed), edited by HOLIDAY MA,BARRATTTM, VERNIER RL, Baltimore, Williams & Wilkins, 1987,pp. 420—431

6. BOHLE A, BADER R, GRUND KE, MACKENSEN 5, NEUNHOEFFERJ: Serum creatinine concentration and renal interstitial volume.Analysis of correlations in endocapillary (acute) glomerulonephritisand in moderately severe mesangioproliferative glomerulonephri-tis. Virchow Arch 375:87—96, 1977

7. BOHLE A, MACKENSEN-HAEN S, V GIsE H: Significance of tubu-lointerstitial changes in the renal cortex for the excretory functionand concentration ability of the kidney. A morphometric compari-son. Am J Nephrol 7:421—433, 1987

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

9. The 1980 Report of the Joint National Committee on Detection,Evaluation and Treament of High Blood Pressure. U.S. Depart-ment of Health and Human Services, NIH Publication No 81-1088

10. Report of Task Force on Blood Pressure Control in Children.Pediatrics 59(Suppl):797—820, 1977

11. ØSTERBY R: Early phases in the development of diabetic glomeru-lopathy. Acta Med Scand (Suppl) 574:13—80, 1974

12. HIROSE K, ØSTERBY R, NOZAMA M, GUNDERSEN HJG: Develop-

ment of glomerular lesions in experimental long-term diabetes inthe rat. Kidney mt 21:689—695, 1982

13. ELLIs EN, BASGEN JM, MAUER SM, STEFFES MW: Kidney biopsytechnique and evaluation in diabetes mellitus, in Methods ofDiabetes Research (vol. 2), edited by LARNER J, CLARKE WL,POHL SL, New York, John Wiley and Sons, 1986, pp. 633—647

14. ELLIS EN, STEFFES MW, GOETZ FC, SUTHERLAND DER, MAUERSM: Glomerular filtration surface in type I diabetes mellitus.Kidney mt 29:889—894, 1986

15. WEIBEL ER: Stereologic Methods, in Practical Methods for Bio-logical Morphometry, London, Academic Press, 1979, pp. 30—39

16. ØSTERBY R, GUNDERSEN HJG: Fast accumulation of basementmembrane material and the rate of morphologic changes in acuteexperimental diabetic glomerular hypertrophy. Diabetologia 18:493—500, 1980

17. JENSEN EB, GUNDERSEN HJG, ØSTERBY R: Determination ofmembrane thickness distribution from orthogonal intercepts. JMicroscopy 115:19—33, 1979

18. GUNDERSEN HJG, SEEFELDT T, ØSTERBY R: Glomerular epithelialfoot processes in normal man and rats. Cell Tissue Res 205:147—155, 1980

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

20. SHEMESH 0, Ross JC, DEEN WM, GRANT GW, MYERS BD: Natureof the glomerular capillary injury in human membranous glomeru-lopathy. J Clin Invest 77:868—877, 1986

21. B0HLE A, v GIsE H, MACKENSEN-HAEN S, STARK-JAKOB B: Theobliteration of the post glomerular capillaries and its influence onthe function of both glomeruli and tubules: Functional interpreta-tion of morphologic findings. Klin Worchenschr 59:1043—1051, 1981

22. LANE P, STEFFES MW, MAUER SM: Relationship of mesangial andinterstitial structure to renal function in insulin dependent diabetesmellitus. (abstract) JAm Soc Nephrol 2:293, 1991

23. LEMLEY Ky, Koiz W: Anatomy of the renal interstitium. Kidneymt 39:370—381, 1991