JOURNAL OF PATHOLOGY, VOL.
167: 49-58 (1992)
DIABETIC GLOMERULOSCLEROSISIMMUNOGOLD ULTRASTRUCTURAL STUDIES ON THE GLOMERULAR DISTRIBUTION OF TYPE IV COLLAGEN AND HEPARAN SULPHATE PROTEOGLYCAN DAVID WOODROW*, JILL MOSS',
IAN SHORE* AND ROBERT G . S P I R O ~
*Department of Histopathology, Charing Cross and Westminster Medical School, London. U.K.; t Elliott P. J o s h Research Laboratory, Harvard Medical School, Boston, U.S.A. Received 25 September 1991 Accepted26 November 1991
SUMMARY We have undertaken an ultrastructural immunogold investigation of the distribution of type IV collagen and heparan sulphate proteoglycan (HSPG) in glomeruli from the kidneys of one normal control and three patients with diabetes mellitus and proteinuria. The sample included both diffuse and nodular diabetic glomcrulosclerosis. In the control and diabetic kidneys, the type IV collagen was present predominantly on the endothelial aspect of the glomerular basement membrane (GBM), and by contrast the HSPG was found mainly on the epithelial side. In the mesangium in both control and diabetic glomeruli, type IV collagen was found predominantly in the central regions, while HSPG was mostly restricted to the region beneath the epithelial cells. Consequently, where there is a marked increase in mesangial matrix with nodule formation in diabetics there is a corresponding increase in the amount of type IV collagen but not of HSPG. Although the three diabetic patients were proteinuric, thc HSPG was not decreased in the thickened GBMs. KEY
WORDS-Immunogold, ultrastructure, glomerular basement membrane, diabetes melhtus, heparan sulphatc proteoglycan, type IV collagen. INTRODUCTION
The glomerular lesions of diabetes mellitus include a n increase in thickness of the G B M together with a diffuse increase in the mesangial matrix which in some cases is accentuated to form conspicuous mesangial nodules in the glomerulithe Kimmelstiel-Wilson lesion.' Immunohistochemical studies indicate that the extracellular matrix of diabetic glomeruli contains the normal G B M constituents type TV collagen, laminin, fibronectin, and HSPG,24 while biochemical analysis of the extracted glomerular matrix has shown a n increase in the amount of type TV and type VI collagen and a decrease in the Addressee for correspondence: Dr D . F. Woodrow, Department of Histopathology. Charing Cross and Westminster Medical School, Fulham Palace Road, London W6 8RF, U.K.
0022-341 7/92/050049-10 $05.00 0 1992 by John Wiley & Sons, Ltd.
amount of laminin and H S P G when expressed as a percentage of the total dry weight of the g I o m e r ~ l i . ~The - ~ more recent studies on diabetic glomeruli have also stressed that the matrix components are not equally distributed throughout the extracellular matrix and that alterations in the mesan ial matrix may occur independently of the G B M .f4.7,* The unequal distribution of G B M components within the human glomerulus has also been demonstrated in immune-complex glomerulonephritis and in control kidneys using the ultrastructural immunogold Type IV collagen is found distributed predominantly in the mesangium and on the endothelial aspect of the G B M , while the H S P G is localized mainly on the epithelial aspect of the G B M with very little within the central regions of mesangial matrix. H u m a n diabetic glomeruli
50
D. WOODROW ET AL.
have not been studied by this method but ultrastructural immunogold investigations of the glomeruli of streptozotocin-induced diabetes in rats have demonstrated an accentuation of type IV collagen on the endothelial aspect of the GBM following induction of the lesion.13 Since alterations in extracellular matrix are an important feature of diabetic glomerulosclerosis, we have used the ultrastructural immunogold technique to study the distribution of type IV collagen and HSPG in early and late diabetes with particular reference to the nodular lesions, and have compared these findings with those from normal glomeruli.
MATERIALS AND METHODS Needle biopsies were obtained from three patients with diabetes mellitus. The first patient was an insulin-dependent 23-year-old female who showed thickened glomerular basement membranes at biopsy. The second patient was a non-insulindependent 54-year-old male with diffuse glomerulosclerosis, and the last patient was an insulindependent female aged 53 years with nodular diabetic glomerulosclerosis. Additional clinical information and routine renal biopsy findings are outlined below. Patient I Female aged 23 years; insulin dependent; diagnosed 2 years before biopsy; normotensive. Renal biopsy was performed because of a transient episode of proteinuria and decreased renal function. The cause was not discovered and renal function returned to normal. There was no further proteinuria. Renal biopsy-Nine glomeruli. There was a slight increase in mesangial matrix in otherwise normal glomeruli (Fig. la). Some tubules had thickened basement membranes. The interstitium was normal. The arterioles were within normal limits and there was focal fibroelastic intimal hyperplasia of an artery. Immune studies were negative. Electron microscop~-Capillary loops were patent and lined by normal endothelium. Many GBMs were evenly thickened (up to 650nm) and there was segmental foot process effacement of epithelial cells. Mesangial regions were normal.
Patient 2 Male aged 54 years; non-insulin-dependent diabetic for 27 years. Proteinuria was consistently about 1 .O g/24 h. Hypertensive. Serum creatinine: 90pmol/l. No retinopathy. Treated with Glibenclamide and Metformin. Renal biopsy-Six glomeruli. One glomerulus was hyalinized while the remainder showed a diffuse increase in mesangial matrix (Fig. lb). Occasional tubules were occluded by intraluminal casts and the interstitium was normal. There was focal hyaline change in the arterioles and fibroelastic intimal hyperplasia of the arteries. Electron microscopy-Capillary loops were patent and lined by normal endothelium. GBMs were evenly thickened (up to 900 nm) and there was segmental foot process effacement of the epithelial cells. The mesangial matrix was focally increased. Patient 3 Female aged 53 years; insulin dependent for 27 years. Hypertensive with proteinuria consistently about 1.0 g/24 h. Serum creatinine: 150-190 ,umol/l. Renal biopsy-Twenty-five glomeruli. Three glomeruli were sclerosed while in the remainder there was diffuse thickening of the capillary loops, together with a diffuse and prominent increase in the mesangial matrix. In some glomeruli the matrix formed nodular aggregates (Fig. Ic). Fibrin cap lesions and capsular drops were also present. There were foci of tubule atrophy and many tubules had thickened basement membranes. The interstitium was fibrosed. The arterioles showed severe hyaline degenerative change which in some glomeruli was seen in both the afferent and the efferent vessels. The arteries were within normal limits. Electron microscopy-Part of the glomerulus showed patent capillary loops lined by normal endothelium. GBMs were evenly thickened (up to 860 nm) and there was a large increase in mesangial matrix. The remainder of the glomerulus was sclerosed. Control patient A biopsy of morphologically normal kidney was obtained from a female, aged 45 years, who had a nephrectomy for a renal adenocarcinoma. By
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
51
Fig. I-Light micrographs showing glomeruli from (a) patient 1, where there is a slight increase in mesangial matrix; (b) patient 2: diffuse diabetic glomerulosclerosis, where a diffuse increase in mesangial matrix is seen (PAS); and (c) patient 3: nodular diabetic glomerulosclerosis, where thickened capillary loops, increased mesangial matrix, and nodules are all present
52
D. WOODROW ET A L .
electron microscopy, GBMs measured up to 290 nm in thickness. Immunological reagents Rabbit anti-human type IV collagen was obtained from the Institut Pasteur de Lyon, France (1041 1). The antigen was extracted from human placenta and was purified before immunization as previously d e ~ c r i b e d .The ' ~ specificity of this antibody had been determined by radioimmunoassay as indicated by the manufacturer, and cross-reactivity with types I, 11, and I11 collagens, and fibronectin was less than 0.5 per cent. In a normal human control kidney by immunofluorescence, the type IV collagen was localized in a linear pattern along the GBM and within the mesangium. It was also present in the Bowman's capsule basement membrane, the tubule basement membranes, and the matrix around the smooth muscle cells of the arterioles, as reported previ0us1y.l~ Rabbit anti-bovine HSPG specificity was verified by solid-phase radioimmunoassays and by immunoblotting of electrophoretically separated antigens." By indirect immunofluorescence on the normal kidney, the HSPG stained the capillary loops of glomeruli in a linear pattern and the mesangial regions were less intensely stained. All tubule basement membranes stained and also the capillary membranes and the matrix around the smooth muscle cells of the arteries, as has been previously reported. l 7 For indirect immunofluoresence, FITC-labelled sheep anti-rabbit immunoglobulin was obtained from Wellcome Research Laboratories, U.K. Goat anti-rabbit IgG conjugated to lOnm gold was obtained from Cambio Ltd.. U.K.
ethanol; infiltrated with Lowicryl K4M resin; and embedded in Lowicryl resin under ultraviolet light at low temperature.' The immunogold technique was carried out at room temperature as previously described.' Ultrathin sections were preincubated in an ovalbumin solution, followed by primary antiserum, washes in buffer, and then gold-labelled secondary antibody. The sections were stained with uranyl acetate and Reynold's lead citrate. Controls consisted of omitting the primary antiserum, or replacing the primary antiserum with normal rabbit serum or immunoglobulin fraction. RESULTS Control kidney Type IV collagen was localized throughout the GBM but predominantly on the endothelial aspect (Fig. 2a). In the mesangium, the gold particles were distributed diffusely in the matrix but were less abundant in the peripheral subepithelial zones. In contrast, the HSPG was found throughout the GBM but more prominently on the epithelial aspect (Fig. 3a). In the mesangium, HSPG was most conspicuous in the subepithelial regions with few particles seen within the central regions of the matrix (Fig. 3b). Patient I The distribution of type IV collagen and HSPG in the thickened GBM and mesangium was essentially the same as the control kidney and there was no obvious decrease in the level of gold staining for HSPG on the epithelial aspect of the GBM (Figs 4a and 4b).
Light microscopy The renal biopsies were fixed in Duboscq-Brazil solution and were routinely processed and stained as previously described.Ix Immunoelectron microscopy Fresh tissue was placed in 4 per cent paraformaldehyde in 0.1 M phosphate buffer, pH 7.2; dissected into 1 mm'; left for 3 h at 4°C; and stored in 1 per cent paraformaldehyde at 4°C for 1 week. The samples were processed to Lowicryl K4M resin (Taab Laboratories Ltd., U.K.). Briefly, the tissue was dehydrated through ascending grades of
Patient 2 Type IV collagen was seen across the entire thickened GBM although there was still more present on the endothelial aspect (Fig. 2b). The expanded areas of mesangial matrix of the diabetic glomerulosclerosis were diffusely and prominently stained. This pattern of staining was in sharp contrast to the HSPG, where, as in the control kidney, only very infrequent gold particles were found in the central regions of the mesangium. (Fig. 5a). The HSPG was also found in all of the capillary loops, where it was distributed particularly on the epithelial aspect of the GBM (Fig. 5b) and was not
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
53
Fig. 2-Electron micrographs ofglomeruli to show thedistribution of type IVcollagen (a) predominantly on the endothelial aspect of the GBM of the control kidney, (b) across the thickened GBM of patient 2, and (c) predominantly in the non-peripheral region of a mesangial nodule of patient 3. (a) x 29 000; (b) x 29 000; (c) x 23 000. (U: urinary space)
54
D. WOODROW ET A L .
Fig. 3-Electron micrographs of thecontrol kidney showing the distribution of HSPG (a) predominantly on the epithelial aspect of the G B M and (b) mainly in subepithelial regions o f t h e mesangium. (a) x 29000; (b) x 29 000. (U: urinary space)
diminished in amount compared with the control kidney.
Patient 3 Type IV collagen was abundant throughout the increased mesangial regions and the nodules, being less evident at the periphery (Fig. 2c). It was again present predominantly on the endothelial aspect of the thickened GBMs but often also seen distributed throughout the entire thickness. Nodules and central regions of increased mesangium contained minimal amounts of HSPG, while all peripheral capillary loops contained HSPG, particularly on the epithelial aspect (Figs 6a and 6b). In none of the loops was the level of staining less conspicuous than in the control glomerulus. In all of the four biopsies examined, no localization of type IV collagen or HSPG was seen in the epithelial, endothelial, or mesangial cells.
DISCUSSION This ultrastructural immunogold study shows that both type 1V collagen and HSPG can be localized in paraformaldehyde-fixed, Lowicrylembedded kidneys, and that there is a marked difference in the distribution of the two components. In the GBM, type IV collagen is present predominantly on the endothelial aspect, although in some of the diabetic capillary loops it extends across the full thickness of the membrane. This distribution is similar to that reported in immunofluorescence studies on diabetic kidneys'-4 and previous ultrastructural immunogold investigations on normal GBM9." and patients with immune-complex g l o m e r u l ~ n e p h r i t i s .It ~ ~is' ~also the pattern found in rats with streptozotocin-induced diabetes meIIitus. I' Type IV collagen is also found in the mesangial matrix and is diffusely and evenly distributed in
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
55
Fig. 4-Electron micrographs of a glomerulus from patient 1 to show the predominantly subepithelial distribution of HSPG in (a) the GBM and (b) the mesangium. (a) x 25 000; (b) x 23 000. (U: urinary space)
Fig. 5-Electron micrographs of a glomerulus from patient 2 to show the Predominantly subepithelial distribution of HSPG in (a) the mesangium and (b) the GBM. (a) x 20 000; (b) x 20 000. (U: urinary space)
56
D. WOODROW ET A L .
Fig. &Electron micrographs of a glomerulus from patient 3 to show the predominantly subepithelial distribution ofHSPG in (a) the mesangium and (b) the GBM. (a) x 23 000; (b) x 27 000. (U: urinary space)
central regions both in normal glomeruli and in the expanded diabetic niesangial matrix and nodules, indicating that there is an increase in the amount of type IV collagen present. This reflects the increased amount of type IV collagen found from biochemical analyses and immunofluorescence studies of diabetic g l o r n e r ~ l i . ' ~ ~ ~ ~ In all of the glomeruli examined, the HSPG is predominantly located on the epithelial aspect of the GBM and mesangium with very little found in the central areas of the normal mesangium or the grossly expanded diabetic mesangial matrix and nodules. This distribution is similar to that previously observed in normal kidney and in cases of human glomerulonephritis,"." and contrasts markedly with the type IV collagen findings. There is also no diminution of staining for HSPG around the periphery of the capillary loops in any of the diabetic patients with proteinuria, a finding already reported in human glomerulonephritis.' The only other ultrastructural studies on the distribution of HSPG using antibodies have been carried out in rats, where in aminonucleoside nephrosis there was no alteration in the distribution or amount of glomerular HSPG following the development of the nephrotic syndrome." In this study the HSPG was found to be regularly spaced along
the lamina rarae, whereas other workers showed a different arrangement with the HSPG diffusely distributed throughout the GBM.'" The fact that HSPG is not diminished while type IV collagen is increased in the diabetic glomerulus together with their different distributions demonstrates the need to consider variations in the composition of the matrix in different regions of the glomerulus when investigating the response of the glomerulus to injury. This point is further emphasized in a recent study on the distribution of glomerular type VI collagen in diabetes.' The qualitative results of HSPG distribution obtained here in human diabetic glomeruli reflect recent quantitative studies on the glomeruli of animal models. In streptozotocin-induced diabetic rats no differences were found in total proteoglycan per glomerulus extracted from the GBM between rats after 12 weeks treatment and untreated controls. There was, however, a higher type IV collagen content in the rat diabetic membrane, so that when the proteoglycan content was expressed per unit mass of protein, relative decreases were observed in diabetic rats2' A similar observation has also been made in human diabetes.' In addition, unaltered HSPG production in renal cortex has been found after examination of steady-state levels of mRNA
H S P G A N D T Y P E IV C O L L A G E N I N DIABETIC G B M
encoding type IV collagen and HSPG in a mouse model (KKAy) of non-insulin-dependent diabetes. The amount of mRNA encoding for the proteoglycan was unchanged in kidneys from diabetic mice with demonstrable GBM thickening, while in contrast mRNA levels for type IV collagen were significantly elevated.22 Whilst HSPG may not be diminished in the GBM, a change in the structure of the heparan sulphate could account for the increased glomerular permeability. Groggel et al. reported that in established puromycin nephrosis in rats, while the heparan sulphate levels were normal the structure was altered and synthesis was diminished in the early stages of the disease.23 These authors also found that in a rabbit model of membranous glomerulonephritis, while the total amount of glycosaminoglycans did not change, the heparan sulphate structure was altered with a correspondIn the report of ing loss of charge ~electivity.~~ Lelongt et al. of established puromycin nephrosis with proteinuria in rats, although the core protein of HSPG was unchanged there was an increase in the charge density characteristics of the de E O V O synthesized proteoglycans.” These recent findings and the results reported here are of interest in relation to the postulated role of HSPG as a major constituent of the glomerular anionic filtration barrier.” While it is tempting to attribute the increased protein permeability of the diabetic glomerulus to an alteration in relative HSPG content and possible changes in heparan sulphate structure, the overproduction or delayed degradation of basement membrane collagens or other constituents may also be important. Although this ultrastructural study involves only three cases of diabetic glomerulopathy, it includes both diffuse and nodular diabetic glomerular lesions and the results show a consistent increase in type IV collagen in the expanded mesangial regions with no alterations in distribution or amount of HSPG. The increase of type IV collagen appears to be a central event in the development of the diabetic glomerular matrix abnormality, and it is still not known whether the high glucose level or secondary local or systemic factors promote the synthesis of this component.*‘
ACKNOWLEDGEMENTS
We would like to thank Dr Peter Wise, Consultant Endocrinologist at the Charing Cross Hospital, for allowing us to study biopsies from his patients.
57
We would also like to thank Ruth Rahman for printing the electron micrographs. REFERENCES I . Kimmelstiel P, Wilson C. Intercapillary lesions in the glomeruli of the kidney. A m JPuthol1936; 12: 83-98.
2. Falk RJ, Scheinmann JI, Mauer SM, Michael AF. Polyantigenic 3.
4.
5.
6.
7.
8. 9.
expansion of basement membrane constituents in diabetic nephropathy. Diohetrs 1983; 32: 34-39. Bruneval P. Foidart JM, Nochy D, Camilleri JP, Bariety J Glomerular matrix proteins in nodular glomerulosclerosis in association with light chain deposition diseaae and diabetes mellitus. H u m Puthol 1985; 1 6 477-484. Gubler M-C, Noel LH, Mounier F, Gros F. Wieslander J. Immunohistochemical study of extracellular matrix components in diabetic glomerulosclerosis. In: Gubler MC, Sternberg M. eds. Progress in Basement Membrane Research. Renal and Related Aspects in Health and Disease. London; Paris: John Libbey, Eurotext Ltd., 1988; 201-204. Beisswenger PJ, Spiro RG. Studies o n the human glomerular basement membrane. Composition, nature of the carbohydrate units, and chemical changes in diabetes mellitus. Diabetes 1975; 22: 180-193. Shimomura H. Spiro RG. Studies o n macromolecular components of human gloinerular basement membrane and alterations in diabetes. Decreased levels of heparan sulphate proteoglycan and laminin. Dinheres 1987; 36: 374-38 I. Mohan PS, Carter WG, Spiro RG. Occurrence of type VI collagen in extracellular matrix of renal glomeruli and its increase in diabetes. Dinhere5 1990: 39: 31-38. Kim Y , Keppel M. Rutowski R, Mauer MS, Wieslander J, Michael A F . Differential expression of basement membrane collagen chains in diabetic nephropathy. Am J P n t h o l 1 9 9 1 ; 138 413-420. Woodrow D F , Shore I , Moss J, Gower P, Phillips M. Immunoelectron microscopic studies of immune complex deposits and basement membrane components in mesangial IgA nephropathy. J Parhol 1989; 157:47-57.
10. Smith PS, Fanning JC. Aarons 1 The structure of the normal human glomerular basement membrane. Ultrastructural localisation of type IV collagen and laminin. Puthology 1989: 21: 254-258. 11. Moss J, Woodrow D F . Shore I, Gower P. Phillips M , Spiro RG. Ultrastructural immunogold studies on the distribution of heparan sulphate proteoglycan in normal human glomeruli and glomerulonephritis. J Purhd 1990; 161: 137-143. 12 Dixey J, Moss J, Woodrow D F , Shore I , Maini RN. SLE nephritis; an ultrastructural immunogold study to evaluate the relationship between immune complexes and the basement membrane components type I V collagen, fibronectin and heparan sulphate proteoglycans. CIin Nephrol1990; 3495-102. 13. Bendayan M. Alteration in the distribution of type IV collagen in glomerular basal laminae in diabetic- rats as revealed by immunocytochemistry and morphoinetrical approach. Diubetologicu 1985: 28: 373-378. 14. Grimaud J-A, Druguet M, Peyrol S, Chevalier 0, Herbage D, El Badrawy N. Collagen typing in human liver: light and electron microscope study. J Histocheni C ~ r o c h e m1980; 2 8 1145-1 156. 15. Morel-Maroger Striker L, Killen PD, Chi E, Striker GE. The composition of glomerulosclerosis. 1. Studies in focal sclerosis, crescentic glomerulonephritis and membranoproliferative glomerulonephritis. LubInvest 1984;Sl: 181-~192. 16 Mohan PS, Spiro R G . Macromolecular organisation of basement membranes. Characteriration and comparison of glomerular basement membrane and lens capsule components by immunochemical and lectin affinity procedures. J B i d Cheni 1986; 261: 4328-4336. 17. Stow JL, Sawada H, Farquhar MG. Basement membrane heparan sulphate protcoglycans are concentrated in the laminae rarae and in the podocytes of the renal glomerulus. Proc Null Acad Sci USA 1985;
8 2 3296-3300. 18. Rainford D. Woodrow D F , Sloper JC, de Wardener HE, Griffiths 1. Post-meningococcal acute glomerular nephritis. Clin Ncphrol1978; 9 249 253.
58
D. WOODROW ET A L
19. Lelongt B, Makino H, Kanwar YS. Status of glomerular proteoglycans in aminonucleoside nephrosis. Kidney Int 1987;31: 1299-1310. 20. Grant DS, Leblond CP. lmmunogold quantitation oflaminin, type IV collagen, and heparan sulfate proteoglycan in a variety of basement membranes. J Hisrochem Cytochern 1988;3 6 271-283. 21. Templeton DM. Retention of glomerular basement membraneproteoglycans accompanying loss of anionic site staining in experimental diabetes. Lab Invest 1989;61: 202-21 1. 22. Ledbetter S , Copeland JE, Noonan D, Vogeli G, Hassell J. Altered steady state mRNA levels of basement membrane proteins in diabetic mouse kidneys and thromboxane synthase inhibition. Diubetes 1990; 3 9 196203.
23. Groggel GC, Hovingh P, Border WA, Linker A. Changes in glomeruPar hepdrdn sulfate in puromycin aminonucleoside nephrosis. Am J Puthol1987;128: 521-527. 24. Groggel GC, Stevenson J, Hovingh P,Linker A, Border WA. Changes in heparan sulfate correlate with increased glomerular permeability. Kidney Int 1988; 3 3 517-523. 25. Farquhar MG. The glomerular basement membrane: a selective macromolecular filter. In: Hay ED, ed. Cell Biology of Extracellular Matrix. New York: Plenum Press, 1981;335-378. 26. Spiro RG. Pathogenesis of diabetic glomerulopathy: a biochemical view. In: Mogensen CE, ed. The Kidney and Hypertension in Diabetes. Boston: Martinus NijhoK, 1988; 117-130.
167: 49-58 (1992)
DIABETIC GLOMERULOSCLEROSISIMMUNOGOLD ULTRASTRUCTURAL STUDIES ON THE GLOMERULAR DISTRIBUTION OF TYPE IV COLLAGEN AND HEPARAN SULPHATE PROTEOGLYCAN DAVID WOODROW*, JILL MOSS',
IAN SHORE* AND ROBERT G . S P I R O ~
*Department of Histopathology, Charing Cross and Westminster Medical School, London. U.K.; t Elliott P. J o s h Research Laboratory, Harvard Medical School, Boston, U.S.A. Received 25 September 1991 Accepted26 November 1991
SUMMARY We have undertaken an ultrastructural immunogold investigation of the distribution of type IV collagen and heparan sulphate proteoglycan (HSPG) in glomeruli from the kidneys of one normal control and three patients with diabetes mellitus and proteinuria. The sample included both diffuse and nodular diabetic glomcrulosclerosis. In the control and diabetic kidneys, the type IV collagen was present predominantly on the endothelial aspect of the glomerular basement membrane (GBM), and by contrast the HSPG was found mainly on the epithelial side. In the mesangium in both control and diabetic glomeruli, type IV collagen was found predominantly in the central regions, while HSPG was mostly restricted to the region beneath the epithelial cells. Consequently, where there is a marked increase in mesangial matrix with nodule formation in diabetics there is a corresponding increase in the amount of type IV collagen but not of HSPG. Although the three diabetic patients were proteinuric, thc HSPG was not decreased in the thickened GBMs. KEY
WORDS-Immunogold, ultrastructure, glomerular basement membrane, diabetes melhtus, heparan sulphatc proteoglycan, type IV collagen. INTRODUCTION
The glomerular lesions of diabetes mellitus include a n increase in thickness of the G B M together with a diffuse increase in the mesangial matrix which in some cases is accentuated to form conspicuous mesangial nodules in the glomerulithe Kimmelstiel-Wilson lesion.' Immunohistochemical studies indicate that the extracellular matrix of diabetic glomeruli contains the normal G B M constituents type TV collagen, laminin, fibronectin, and HSPG,24 while biochemical analysis of the extracted glomerular matrix has shown a n increase in the amount of type TV and type VI collagen and a decrease in the Addressee for correspondence: Dr D . F. Woodrow, Department of Histopathology. Charing Cross and Westminster Medical School, Fulham Palace Road, London W6 8RF, U.K.
0022-341 7/92/050049-10 $05.00 0 1992 by John Wiley & Sons, Ltd.
amount of laminin and H S P G when expressed as a percentage of the total dry weight of the g I o m e r ~ l i . ~The - ~ more recent studies on diabetic glomeruli have also stressed that the matrix components are not equally distributed throughout the extracellular matrix and that alterations in the mesan ial matrix may occur independently of the G B M .f4.7,* The unequal distribution of G B M components within the human glomerulus has also been demonstrated in immune-complex glomerulonephritis and in control kidneys using the ultrastructural immunogold Type IV collagen is found distributed predominantly in the mesangium and on the endothelial aspect of the G B M , while the H S P G is localized mainly on the epithelial aspect of the G B M with very little within the central regions of mesangial matrix. H u m a n diabetic glomeruli
50
D. WOODROW ET AL.
have not been studied by this method but ultrastructural immunogold investigations of the glomeruli of streptozotocin-induced diabetes in rats have demonstrated an accentuation of type IV collagen on the endothelial aspect of the GBM following induction of the lesion.13 Since alterations in extracellular matrix are an important feature of diabetic glomerulosclerosis, we have used the ultrastructural immunogold technique to study the distribution of type IV collagen and HSPG in early and late diabetes with particular reference to the nodular lesions, and have compared these findings with those from normal glomeruli.
MATERIALS AND METHODS Needle biopsies were obtained from three patients with diabetes mellitus. The first patient was an insulin-dependent 23-year-old female who showed thickened glomerular basement membranes at biopsy. The second patient was a non-insulindependent 54-year-old male with diffuse glomerulosclerosis, and the last patient was an insulindependent female aged 53 years with nodular diabetic glomerulosclerosis. Additional clinical information and routine renal biopsy findings are outlined below. Patient I Female aged 23 years; insulin dependent; diagnosed 2 years before biopsy; normotensive. Renal biopsy was performed because of a transient episode of proteinuria and decreased renal function. The cause was not discovered and renal function returned to normal. There was no further proteinuria. Renal biopsy-Nine glomeruli. There was a slight increase in mesangial matrix in otherwise normal glomeruli (Fig. la). Some tubules had thickened basement membranes. The interstitium was normal. The arterioles were within normal limits and there was focal fibroelastic intimal hyperplasia of an artery. Immune studies were negative. Electron microscop~-Capillary loops were patent and lined by normal endothelium. Many GBMs were evenly thickened (up to 650nm) and there was segmental foot process effacement of epithelial cells. Mesangial regions were normal.
Patient 2 Male aged 54 years; non-insulin-dependent diabetic for 27 years. Proteinuria was consistently about 1 .O g/24 h. Hypertensive. Serum creatinine: 90pmol/l. No retinopathy. Treated with Glibenclamide and Metformin. Renal biopsy-Six glomeruli. One glomerulus was hyalinized while the remainder showed a diffuse increase in mesangial matrix (Fig. lb). Occasional tubules were occluded by intraluminal casts and the interstitium was normal. There was focal hyaline change in the arterioles and fibroelastic intimal hyperplasia of the arteries. Electron microscopy-Capillary loops were patent and lined by normal endothelium. GBMs were evenly thickened (up to 900 nm) and there was segmental foot process effacement of the epithelial cells. The mesangial matrix was focally increased. Patient 3 Female aged 53 years; insulin dependent for 27 years. Hypertensive with proteinuria consistently about 1.0 g/24 h. Serum creatinine: 150-190 ,umol/l. Renal biopsy-Twenty-five glomeruli. Three glomeruli were sclerosed while in the remainder there was diffuse thickening of the capillary loops, together with a diffuse and prominent increase in the mesangial matrix. In some glomeruli the matrix formed nodular aggregates (Fig. Ic). Fibrin cap lesions and capsular drops were also present. There were foci of tubule atrophy and many tubules had thickened basement membranes. The interstitium was fibrosed. The arterioles showed severe hyaline degenerative change which in some glomeruli was seen in both the afferent and the efferent vessels. The arteries were within normal limits. Electron microscopy-Part of the glomerulus showed patent capillary loops lined by normal endothelium. GBMs were evenly thickened (up to 860 nm) and there was a large increase in mesangial matrix. The remainder of the glomerulus was sclerosed. Control patient A biopsy of morphologically normal kidney was obtained from a female, aged 45 years, who had a nephrectomy for a renal adenocarcinoma. By
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
51
Fig. I-Light micrographs showing glomeruli from (a) patient 1, where there is a slight increase in mesangial matrix; (b) patient 2: diffuse diabetic glomerulosclerosis, where a diffuse increase in mesangial matrix is seen (PAS); and (c) patient 3: nodular diabetic glomerulosclerosis, where thickened capillary loops, increased mesangial matrix, and nodules are all present
52
D. WOODROW ET A L .
electron microscopy, GBMs measured up to 290 nm in thickness. Immunological reagents Rabbit anti-human type IV collagen was obtained from the Institut Pasteur de Lyon, France (1041 1). The antigen was extracted from human placenta and was purified before immunization as previously d e ~ c r i b e d .The ' ~ specificity of this antibody had been determined by radioimmunoassay as indicated by the manufacturer, and cross-reactivity with types I, 11, and I11 collagens, and fibronectin was less than 0.5 per cent. In a normal human control kidney by immunofluorescence, the type IV collagen was localized in a linear pattern along the GBM and within the mesangium. It was also present in the Bowman's capsule basement membrane, the tubule basement membranes, and the matrix around the smooth muscle cells of the arterioles, as reported previ0us1y.l~ Rabbit anti-bovine HSPG specificity was verified by solid-phase radioimmunoassays and by immunoblotting of electrophoretically separated antigens." By indirect immunofluorescence on the normal kidney, the HSPG stained the capillary loops of glomeruli in a linear pattern and the mesangial regions were less intensely stained. All tubule basement membranes stained and also the capillary membranes and the matrix around the smooth muscle cells of the arteries, as has been previously reported. l 7 For indirect immunofluoresence, FITC-labelled sheep anti-rabbit immunoglobulin was obtained from Wellcome Research Laboratories, U.K. Goat anti-rabbit IgG conjugated to lOnm gold was obtained from Cambio Ltd.. U.K.
ethanol; infiltrated with Lowicryl K4M resin; and embedded in Lowicryl resin under ultraviolet light at low temperature.' The immunogold technique was carried out at room temperature as previously described.' Ultrathin sections were preincubated in an ovalbumin solution, followed by primary antiserum, washes in buffer, and then gold-labelled secondary antibody. The sections were stained with uranyl acetate and Reynold's lead citrate. Controls consisted of omitting the primary antiserum, or replacing the primary antiserum with normal rabbit serum or immunoglobulin fraction. RESULTS Control kidney Type IV collagen was localized throughout the GBM but predominantly on the endothelial aspect (Fig. 2a). In the mesangium, the gold particles were distributed diffusely in the matrix but were less abundant in the peripheral subepithelial zones. In contrast, the HSPG was found throughout the GBM but more prominently on the epithelial aspect (Fig. 3a). In the mesangium, HSPG was most conspicuous in the subepithelial regions with few particles seen within the central regions of the matrix (Fig. 3b). Patient I The distribution of type IV collagen and HSPG in the thickened GBM and mesangium was essentially the same as the control kidney and there was no obvious decrease in the level of gold staining for HSPG on the epithelial aspect of the GBM (Figs 4a and 4b).
Light microscopy The renal biopsies were fixed in Duboscq-Brazil solution and were routinely processed and stained as previously described.Ix Immunoelectron microscopy Fresh tissue was placed in 4 per cent paraformaldehyde in 0.1 M phosphate buffer, pH 7.2; dissected into 1 mm'; left for 3 h at 4°C; and stored in 1 per cent paraformaldehyde at 4°C for 1 week. The samples were processed to Lowicryl K4M resin (Taab Laboratories Ltd., U.K.). Briefly, the tissue was dehydrated through ascending grades of
Patient 2 Type IV collagen was seen across the entire thickened GBM although there was still more present on the endothelial aspect (Fig. 2b). The expanded areas of mesangial matrix of the diabetic glomerulosclerosis were diffusely and prominently stained. This pattern of staining was in sharp contrast to the HSPG, where, as in the control kidney, only very infrequent gold particles were found in the central regions of the mesangium. (Fig. 5a). The HSPG was also found in all of the capillary loops, where it was distributed particularly on the epithelial aspect of the GBM (Fig. 5b) and was not
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
53
Fig. 2-Electron micrographs ofglomeruli to show thedistribution of type IVcollagen (a) predominantly on the endothelial aspect of the GBM of the control kidney, (b) across the thickened GBM of patient 2, and (c) predominantly in the non-peripheral region of a mesangial nodule of patient 3. (a) x 29 000; (b) x 29 000; (c) x 23 000. (U: urinary space)
54
D. WOODROW ET A L .
Fig. 3-Electron micrographs of thecontrol kidney showing the distribution of HSPG (a) predominantly on the epithelial aspect of the G B M and (b) mainly in subepithelial regions o f t h e mesangium. (a) x 29000; (b) x 29 000. (U: urinary space)
diminished in amount compared with the control kidney.
Patient 3 Type IV collagen was abundant throughout the increased mesangial regions and the nodules, being less evident at the periphery (Fig. 2c). It was again present predominantly on the endothelial aspect of the thickened GBMs but often also seen distributed throughout the entire thickness. Nodules and central regions of increased mesangium contained minimal amounts of HSPG, while all peripheral capillary loops contained HSPG, particularly on the epithelial aspect (Figs 6a and 6b). In none of the loops was the level of staining less conspicuous than in the control glomerulus. In all of the four biopsies examined, no localization of type IV collagen or HSPG was seen in the epithelial, endothelial, or mesangial cells.
DISCUSSION This ultrastructural immunogold study shows that both type 1V collagen and HSPG can be localized in paraformaldehyde-fixed, Lowicrylembedded kidneys, and that there is a marked difference in the distribution of the two components. In the GBM, type IV collagen is present predominantly on the endothelial aspect, although in some of the diabetic capillary loops it extends across the full thickness of the membrane. This distribution is similar to that reported in immunofluorescence studies on diabetic kidneys'-4 and previous ultrastructural immunogold investigations on normal GBM9." and patients with immune-complex g l o m e r u l ~ n e p h r i t i s .It ~ ~is' ~also the pattern found in rats with streptozotocin-induced diabetes meIIitus. I' Type IV collagen is also found in the mesangial matrix and is diffusely and evenly distributed in
HSPG AND TYPE IV COLLAGEN IN DIABETIC GBM
55
Fig. 4-Electron micrographs of a glomerulus from patient 1 to show the predominantly subepithelial distribution of HSPG in (a) the GBM and (b) the mesangium. (a) x 25 000; (b) x 23 000. (U: urinary space)
Fig. 5-Electron micrographs of a glomerulus from patient 2 to show the Predominantly subepithelial distribution of HSPG in (a) the mesangium and (b) the GBM. (a) x 20 000; (b) x 20 000. (U: urinary space)
56
D. WOODROW ET A L .
Fig. &Electron micrographs of a glomerulus from patient 3 to show the predominantly subepithelial distribution ofHSPG in (a) the mesangium and (b) the GBM. (a) x 23 000; (b) x 27 000. (U: urinary space)
central regions both in normal glomeruli and in the expanded diabetic niesangial matrix and nodules, indicating that there is an increase in the amount of type IV collagen present. This reflects the increased amount of type IV collagen found from biochemical analyses and immunofluorescence studies of diabetic g l o r n e r ~ l i . ' ~ ~ ~ ~ In all of the glomeruli examined, the HSPG is predominantly located on the epithelial aspect of the GBM and mesangium with very little found in the central areas of the normal mesangium or the grossly expanded diabetic mesangial matrix and nodules. This distribution is similar to that previously observed in normal kidney and in cases of human glomerulonephritis,"." and contrasts markedly with the type IV collagen findings. There is also no diminution of staining for HSPG around the periphery of the capillary loops in any of the diabetic patients with proteinuria, a finding already reported in human glomerulonephritis.' The only other ultrastructural studies on the distribution of HSPG using antibodies have been carried out in rats, where in aminonucleoside nephrosis there was no alteration in the distribution or amount of glomerular HSPG following the development of the nephrotic syndrome." In this study the HSPG was found to be regularly spaced along
the lamina rarae, whereas other workers showed a different arrangement with the HSPG diffusely distributed throughout the GBM.'" The fact that HSPG is not diminished while type IV collagen is increased in the diabetic glomerulus together with their different distributions demonstrates the need to consider variations in the composition of the matrix in different regions of the glomerulus when investigating the response of the glomerulus to injury. This point is further emphasized in a recent study on the distribution of glomerular type VI collagen in diabetes.' The qualitative results of HSPG distribution obtained here in human diabetic glomeruli reflect recent quantitative studies on the glomeruli of animal models. In streptozotocin-induced diabetic rats no differences were found in total proteoglycan per glomerulus extracted from the GBM between rats after 12 weeks treatment and untreated controls. There was, however, a higher type IV collagen content in the rat diabetic membrane, so that when the proteoglycan content was expressed per unit mass of protein, relative decreases were observed in diabetic rats2' A similar observation has also been made in human diabetes.' In addition, unaltered HSPG production in renal cortex has been found after examination of steady-state levels of mRNA
H S P G A N D T Y P E IV C O L L A G E N I N DIABETIC G B M
encoding type IV collagen and HSPG in a mouse model (KKAy) of non-insulin-dependent diabetes. The amount of mRNA encoding for the proteoglycan was unchanged in kidneys from diabetic mice with demonstrable GBM thickening, while in contrast mRNA levels for type IV collagen were significantly elevated.22 Whilst HSPG may not be diminished in the GBM, a change in the structure of the heparan sulphate could account for the increased glomerular permeability. Groggel et al. reported that in established puromycin nephrosis in rats, while the heparan sulphate levels were normal the structure was altered and synthesis was diminished in the early stages of the disease.23 These authors also found that in a rabbit model of membranous glomerulonephritis, while the total amount of glycosaminoglycans did not change, the heparan sulphate structure was altered with a correspondIn the report of ing loss of charge ~electivity.~~ Lelongt et al. of established puromycin nephrosis with proteinuria in rats, although the core protein of HSPG was unchanged there was an increase in the charge density characteristics of the de E O V O synthesized proteoglycans.” These recent findings and the results reported here are of interest in relation to the postulated role of HSPG as a major constituent of the glomerular anionic filtration barrier.” While it is tempting to attribute the increased protein permeability of the diabetic glomerulus to an alteration in relative HSPG content and possible changes in heparan sulphate structure, the overproduction or delayed degradation of basement membrane collagens or other constituents may also be important. Although this ultrastructural study involves only three cases of diabetic glomerulopathy, it includes both diffuse and nodular diabetic glomerular lesions and the results show a consistent increase in type IV collagen in the expanded mesangial regions with no alterations in distribution or amount of HSPG. The increase of type IV collagen appears to be a central event in the development of the diabetic glomerular matrix abnormality, and it is still not known whether the high glucose level or secondary local or systemic factors promote the synthesis of this component.*‘
ACKNOWLEDGEMENTS
We would like to thank Dr Peter Wise, Consultant Endocrinologist at the Charing Cross Hospital, for allowing us to study biopsies from his patients.
57
We would also like to thank Ruth Rahman for printing the electron micrographs. REFERENCES I . Kimmelstiel P, Wilson C. Intercapillary lesions in the glomeruli of the kidney. A m JPuthol1936; 12: 83-98.
2. Falk RJ, Scheinmann JI, Mauer SM, Michael AF. Polyantigenic 3.
4.
5.
6.
7.
8. 9.
expansion of basement membrane constituents in diabetic nephropathy. Diohetrs 1983; 32: 34-39. Bruneval P. Foidart JM, Nochy D, Camilleri JP, Bariety J Glomerular matrix proteins in nodular glomerulosclerosis in association with light chain deposition diseaae and diabetes mellitus. H u m Puthol 1985; 1 6 477-484. Gubler M-C, Noel LH, Mounier F, Gros F. Wieslander J. Immunohistochemical study of extracellular matrix components in diabetic glomerulosclerosis. In: Gubler MC, Sternberg M. eds. Progress in Basement Membrane Research. Renal and Related Aspects in Health and Disease. London; Paris: John Libbey, Eurotext Ltd., 1988; 201-204. Beisswenger PJ, Spiro RG. Studies o n the human glomerular basement membrane. Composition, nature of the carbohydrate units, and chemical changes in diabetes mellitus. Diabetes 1975; 22: 180-193. Shimomura H. Spiro RG. Studies o n macromolecular components of human gloinerular basement membrane and alterations in diabetes. Decreased levels of heparan sulphate proteoglycan and laminin. Dinheres 1987; 36: 374-38 I. Mohan PS, Carter WG, Spiro RG. Occurrence of type VI collagen in extracellular matrix of renal glomeruli and its increase in diabetes. Dinhere5 1990: 39: 31-38. Kim Y , Keppel M. Rutowski R, Mauer MS, Wieslander J, Michael A F . Differential expression of basement membrane collagen chains in diabetic nephropathy. Am J P n t h o l 1 9 9 1 ; 138 413-420. Woodrow D F , Shore I , Moss J, Gower P, Phillips M. Immunoelectron microscopic studies of immune complex deposits and basement membrane components in mesangial IgA nephropathy. J Parhol 1989; 157:47-57.
10. Smith PS, Fanning JC. Aarons 1 The structure of the normal human glomerular basement membrane. Ultrastructural localisation of type IV collagen and laminin. Puthology 1989: 21: 254-258. 11. Moss J, Woodrow D F . Shore I, Gower P. Phillips M , Spiro RG. Ultrastructural immunogold studies on the distribution of heparan sulphate proteoglycan in normal human glomeruli and glomerulonephritis. J Purhd 1990; 161: 137-143. 12 Dixey J, Moss J, Woodrow D F , Shore I , Maini RN. SLE nephritis; an ultrastructural immunogold study to evaluate the relationship between immune complexes and the basement membrane components type I V collagen, fibronectin and heparan sulphate proteoglycans. CIin Nephrol1990; 3495-102. 13. Bendayan M. Alteration in the distribution of type IV collagen in glomerular basal laminae in diabetic- rats as revealed by immunocytochemistry and morphoinetrical approach. Diubetologicu 1985: 28: 373-378. 14. Grimaud J-A, Druguet M, Peyrol S, Chevalier 0, Herbage D, El Badrawy N. Collagen typing in human liver: light and electron microscope study. J Histocheni C ~ r o c h e m1980; 2 8 1145-1 156. 15. Morel-Maroger Striker L, Killen PD, Chi E, Striker GE. The composition of glomerulosclerosis. 1. Studies in focal sclerosis, crescentic glomerulonephritis and membranoproliferative glomerulonephritis. LubInvest 1984;Sl: 181-~192. 16 Mohan PS, Spiro R G . Macromolecular organisation of basement membranes. Characteriration and comparison of glomerular basement membrane and lens capsule components by immunochemical and lectin affinity procedures. J B i d Cheni 1986; 261: 4328-4336. 17. Stow JL, Sawada H, Farquhar MG. Basement membrane heparan sulphate protcoglycans are concentrated in the laminae rarae and in the podocytes of the renal glomerulus. Proc Null Acad Sci USA 1985;
8 2 3296-3300. 18. Rainford D. Woodrow D F , Sloper JC, de Wardener HE, Griffiths 1. Post-meningococcal acute glomerular nephritis. Clin Ncphrol1978; 9 249 253.
58
D. WOODROW ET A L
19. Lelongt B, Makino H, Kanwar YS. Status of glomerular proteoglycans in aminonucleoside nephrosis. Kidney Int 1987;31: 1299-1310. 20. Grant DS, Leblond CP. lmmunogold quantitation oflaminin, type IV collagen, and heparan sulfate proteoglycan in a variety of basement membranes. J Hisrochem Cytochern 1988;3 6 271-283. 21. Templeton DM. Retention of glomerular basement membraneproteoglycans accompanying loss of anionic site staining in experimental diabetes. Lab Invest 1989;61: 202-21 1. 22. Ledbetter S , Copeland JE, Noonan D, Vogeli G, Hassell J. Altered steady state mRNA levels of basement membrane proteins in diabetic mouse kidneys and thromboxane synthase inhibition. Diubetes 1990; 3 9 196203.
23. Groggel GC, Hovingh P, Border WA, Linker A. Changes in glomeruPar hepdrdn sulfate in puromycin aminonucleoside nephrosis. Am J Puthol1987;128: 521-527. 24. Groggel GC, Stevenson J, Hovingh P,Linker A, Border WA. Changes in heparan sulfate correlate with increased glomerular permeability. Kidney Int 1988; 3 3 517-523. 25. Farquhar MG. The glomerular basement membrane: a selective macromolecular filter. In: Hay ED, ed. Cell Biology of Extracellular Matrix. New York: Plenum Press, 1981;335-378. 26. Spiro RG. Pathogenesis of diabetic glomerulopathy: a biochemical view. In: Mogensen CE, ed. The Kidney and Hypertension in Diabetes. Boston: Martinus NijhoK, 1988; 117-130.
