Ultrastructure of nonenzymatically glycated mesangial matrix in diabetic nephropathy

Kidney Jntemationa4 VoL 48 (1995), pp. 517—526

Ultrastructure of nonenzymatica!ly glycated mesangial matrix in diabetic nephropathy HIR0HJMI MAKINO, KENICHI SHIKATA, KAzuE HIRONAKA, MASAHIKO KusHIRo, YASUSHI Y&r&sA1a,

Hiiaiu SuGIM0T0, ZENSUKE OTA, NoRm A1t\xI, and SEIKOH HoRIucifi Third Department of Internal Medicine, Okayama University Medical School, Okayama, Japan and Department of Biochemistzy, Kumamoto University Medical Schoo4 Kumamoto, Japan

Ultrastructure of noncnzymatically glycated mesangial matrix in diabetic nephropathy. Advanced protein glycation has been proposed as a major factor in the development of diabetic nephropathy. Advanced glycation end products (AGEs) have altered the structure of extracellular matrix component and impaired self association in vitro. To elucidate the role of AGEs in the progression of diabetic nephropathy, the present study was undertaken to localize glomerular AGEs immunohistochemically. Ultrastructural changes of the mesangial matrix were analyzed with high resolution scanning electron microscopy. No glomerular AGEs staining was noted in normal control kidney specimens, or in tissue from glomerulonephritis patients without diabetes mellitus. The mesangium showed a positive AGEs staining in advanced stages of diabetic nephropathy, and the most characteristic finding was the strong AGEs staining in nodular lesions. By high resolution scanning electron microscopy, control and diabetic mesangial matrices revealed a meshwork structure composed of fine fibrils (10 nm in width) and numerous pores (12 to 13 nni in diameter). In the nodular lesions, however, loosening of the meshwork was significant, and the diameter of the pores was enlarged (approximately 24 nm). This study provides the first immunohistochemical evidence that AGEs are localized in diabetic glomeruli, most notably to nodular lesions.

Advanced glycation might play a role in the progression of diabetic nephropathy through impairment of the assembly of matrix proteins in vivo.

Although it remains still controversial, available data indicate that the development of diabetic nephropathy is linked to hyperglycemia [1, 2]. Glucose reacts nonenzymatically with proteins to form Schiff base and Amadori products. Further incubation of these early products leads to the formation of advanced glycation end products (AGEs) that are characterized by fluorescence, a brown color and cross linking [3, 4]. Protein modification by AGE is expected to occur in proteins with relatively long half lives, such as collagen [5], lens crystallins, and myelin [4]. Such a modification

occurs with normal aging and is accelerated in patients with diabetes [2]. Recent immunohistochemical studies with anti-AGE antibod-

ies have demonstrated their presence in serum and in several types of human tissues including arterial wall [6], brain [7] and renal cortex [8]. However, AGEs have not been demonstrated

10] reacts not only with AGE proteins such as AGE-bovine serum albumin, AGE-human serum albumin and AGE-hemoglobin, but also with AGE preparations obtained from such lysine derivatives

as a-tosyl-L-lysine and a-tosyl-L-lysine methyl ester, or from monoaminocarboxylic acids such as e-aminocapronic acid, y-ami-

no-n-butyric acid and (3-alanine. Unmodified counterparts of these preparations do not serve as antigens. Fructosyl-e-capronic acid, a synthetic Amadori compound, and proposed AGE structures such as 2-(2-froyl)-4(5)-(2-furanyl)-1H-imidazole (FF1) [11], pyrraline [12], and pentosidine [13] are not recognized as antigens

by our antibody [9]. Available data, therefore, suggest that our antibody may react with a primary AGE structure common to AGE preparations, although determination of its structure is currently under way.

In the first stage of this study, we inununohistochemically

evaluated the localization of AGEs in the glomeruli from patients with diabetic nephropathy using our anti-AGE antibody. Glyca-

tion seems to play a vital role in the association of matrix interactions and causes changes in the extracellular matrix of a diabetic patient [1, 2, 14]. Type IV collagen is the major component of the matrix meshwork, and laminin, proteoglycans, fibronectin, and entactin/nidogen together form the meshed struc-

ture of the extracellular matrix [15]. The development and maintenance of a normal extracellular matrix depends on a geometrically ordered self-assembly process involving site-specific, end-to-end, and lateral interactions of these macromolecules [16—18]. Advanced protein glycation has been proposed as a major factor in the development of complications via the formation of AGEs [1, 2, 14]. AGE modification of extracellular matrix components may induce an alteration of structure and impairment of self association in vitro [19, 20]. Since the first set of our experiments revealed AGE accumulation in the mesangial matrix of the sclerosing glomeruli, in the second set of experiments the ultrastructural changes of mesangial matrix were examined using high resolution scanning electron microscopy (HRSEM) [21, 22].

directly in diabetic glomerular lesions. Our anti-AGE antibody [9,

Methods Patients

Received for publication January 4, 1995 and in revised form March 10, 1995 Accepted for publication March 13, 1995

© 1995 by the International Society of Nephrology

For immunohistochemical studies, we evaluated kidney specimens from 10 patients with noninsulin-dependent diabetes mdlitus (NIDDM) and diabetic nephropathy with glomerulosclerosis

517

518

£ I t —i VI

Makino et a!: Glycated diabetic mesangial matrix changes

—

A''

•"

-

$_•4• n4 'U V. — ;eaa.- -

•

Fig. 1. Immunohistoclwmical localization of advanced glycation end products (AGEs) by immunoperoxidase. A. No glomerular, tubular, or interstitial staining for AGEs in a specimen from a 62-year-old patients with membranous nephropathy without diabetes mellitus. B. AGEs immunoreactivity can be seen in the mesangial area of a kidney specimen from a 62-year-old diabetic patient. C. Intense AGEs immunoreactivity in a nodular lesion (arrow). AGEs immunoreactivity also can be observed in the tubulus and interstitium in a kidney specimen from a 85-year-old diabetic patient. D. AGEs immunoreactivity in the fibrous crescent of the Bowman's capsule of a scierosed glomerulus in a kidney specimen from a 85-year-old diabetic patient.

of valying degrees of severity. Eight specimens were obtained at renal biopsy and two were obtained at autopsy from six men and four women aged 43 to 85 years (mean 62 11 years). Proteinuria varied in severity ranging from 0.9 to 13.8 g/day (mean 3.9 3.5

were examined in biopsy specimens. Specimens were stained with

hematoxylin and eosin, periodic acid-Schiff and periodic acidsilver methenamine for light microscopy. The controls for electron microscopic studies included kidneys from two age-matched au-

glday). Impairment of renal function was reflected by a serum topsy cases with no renal disease, and the normal part of three creatinine concentration ranging from 0.84 to 5.4 mg/dl (mean 1.8 kidneys removed from patients with kidney tumor at surgery. 1.4 mg/dl). As controls for immunohistochemical studies, kidneys from three patients with no renal disease, that is, hemaPreparation of anti-AGE-antibody tuna, proteinuria or compromised renal functions and specimens from five patients with glomerulonephritis but without diabetes A mouse anti-AGE monoclonal antibody was prepared and its mellitus, were used. Kidneys from five patients with NIDDM were used for electron

microscopy. Two of these patients were autopsy cases with nephrotic syndrome due to advanced diabetic nephropathy, but who had died of non-renal complications. The other three pa-

immunoreactivity was described previously [9]. Briefly the splenic

lymphocytes from BALB/c mice were immunized with AGEsbovine serum albumin and were fused to myeloma P3U1 cells.

These hybrid cells were screened and cultured, and a cell line with positive reactions to AGEs-human serum albumin, but negative tients were biopsy cases with proteinuria, and frozen blocks used reactions to BSA, was selected through successive subclonings. for immunofluorescence studies were utilized. A total of 30 Antibody was produced in the ascitic fluid of BALB!c mice, and glomeruli were examined in autopsy specimens and 10 glomeruli was purified by DEAE-cellulose chromatography.

519

a

r

flu

a

Makino et al: Glycated diabetic mesangial matrix changes

Fig. 2. Transmission electron micrograph of acellular glomendus from controls (A, B) and patients with diabetic nephropathy (C, D). A. All cellular

components have been completely removed, retaining the in situ architecture. x2,000. B. Partial enlargement of A (large arrow) demonstrating a fibriltar mesangial matrix (MM; X6O,000). C. Note the increase in the mesangial matrix (*; X2,000). D, Higher magnification of the fibrillar mesangial matrix. Note many bundles of collagen fibers (small arrows) which are embedded within the mesangial matrix. (X60,000). Abbreviations are: CL, capillary lumen; US, urinary space; GBM, glomerular basement membrane.

Immunohistochemical procedures

The distribution of AGEs was studied by immunoperoxidase techniques using the avidin-biotin method [23]. Formalin-fixed and paraffin-embedded tissue was cut into 4-i.tm thick sections and deparaffinized by treatment with xylene (5 mm, 3 times), followed by graded ethanol. Nonspecific protein binding sites were blocked

by incubating the tissue sections with 1% fetal calf serum in Tris-buffered snline for 20 minutes. Nonspecific staining was blocked by 15 minutes of incubation with avidin and biotin solutions (Vector Lab., Buarlingame, CA, USA), respectively. These sections were then washed with phosphate-buffered saline (PBS) and dipped in 0.3% H202-100% methanol for 20 minutes

520

Makino et al: Glycated diabetic mesangial matrix changes

Fig. 3. High resolution scanning electron micrograph of a control acellulargiomerulus. A. The glomerular basement membrane appears (GBM) as a thin, single-layered sheet with a smooth surface except for in the wrinkling mesangial area. The epithelial surface of the GBM is smooth. Endothelial surfaces are irregular in the centrolobular zones, where the mesangial matrix (*) appears as a fenestrated septum (X4,800). B. Partial enlargement of A (large arrow), demonstrating the meshwork structure of the mesangial matrix composed of pores (arrowheads) and strands (small arrows). Note the fine fibrils over the stomata (large arrows; x81,000). Abbreviations are: CL, capillary lumen; US, urinary space.

to block endogenous peroxidase. The slides were first stained with extensively between each step with distilled water. Acellular renal a biotin-labeled mouse anti-AGE monoclonal antibody in a cortical specimens were separated for transmission and scanning humidity chamber for two hours at 37°C. Slides were washed with electron microscopy preparation.

0.5% Tween/PBS followed by PBS. The slides were then incubated with avidin biotinylated horseradish peroxidase (Vector Lab.). Next, the slides were washed in PBS and then a color reaction was performed by incubating the slides with freshly prepared 3,3-diaminobenzidine (DAB) reagent until the staining was complete. The reaction was stopped by placing the slides into water. After washing in PBS, Meyer hematoxylin was added as a counterstain. As an additional control, positively stained tissues were subjected to a blocking experiment in which they were preincubated with unlabeled anti-AGE serum. Sections preincubated with anti-AGE serum remained uniformly unstained. Preparation of acellular renal cortex Acellular cortical tissues were prepared as has been described

Tissue preparation for transmission electron microscopy

Acellular renal cortical specimens were immersed into 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2, 2 hr, 4°C),

post-fixed in 0.1 M osmium tetroxide (2 hr 4°C), and then embedded in Epon following dehydration in a graded series of ethanol preparations. Thin sections were stained with uranyl acetate and lead citrate and examined with a transmission electron microscope (Hitachi H7100, Japan) at 75 kV. Tissue preparation for scanning electron microscopy

Acellular renal cortical specimens were prepared for high

previously [24, 25] from five patients with NIDDM and two

resolution scanning electron microscopy by tannic acid conductive controls, and studied by the following transmission and scanning staining as has been described previously [22, 26]. Briefly, specielectron microscopy. Briefly, renal cortical tissue blocks were mens were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate minced to 2 mm3. Minced tissue specimens were sequentially buffer (pH 7.2, 2 hr, 4°C). After rinsing with the same buffer, the exposed to 4 mrvi ethylenediamine tetraacetic acid (24 hr, 4°C), tannin-osmium conductive staining [27] was performed as follows: 3% Triton X-100 (24 hr, 4°C), 0.05% deoxyribonuclease (Type I, specimens were immersed three times in 1% osmium tetroxide (2 Sigma, St. Louis, MO, USA) in 1 M NaCI (24 hr, 4°C), and 4% hr, 4°C), and interposed by 2% tannic acid solution (pH 4.2, 2 hr, sodium deoxycholate (2 to 4 hr, room temperature). All solutions room temperature). Between each step, rinses were performed at

contained 0.1% sodium azide. Tissue specimens were rinsed

least three times with distilled water. Following staining, the

Makino et al: Glycated diabetic mesangial matrix changes

521

Fig. 4. High resolution scanning electron micrograph of an acellular glomerulus from a patient with diabetic nephropathy. A. Note the narrowing of the

capillary lumen (CL) due to the expansion of the mesangial matrix (MM; X3,200). B. Higher magnification of the mesangial matrix demonstrating the meshwork structure of the mesangial matrix with numerous pores of varying sizes composed of strands (small arrows). Note small pores (arrowheads) and larger pores (triangle arrowheads), and globular clusters (*; X 198,000). US is urinary space.

Results specimens were dehydrated in a graded series of ethanol preparations and cryofractured in liquid nitrogen. Cryofractured specImmunohistochemical localization of AGEs imens were transferred to t-butyl alcohol, and freeze-dried with a vacuum evaporator (ID-2, Eiko, Japan). The dried specimens AGEs were hardly stained in the renal cortex specimens from were then mounted on aluminum stubs, and examined with an normal kidneys and from patients with glomerulonephritis withHRSEM (S-900, Hitachi) without metal coating, at original out diabetes mellitus (Fig. 1A). However, slight AGE staining was magnifications ranging from X 250 to 200,000, using an accelerat- observed in the glomerular mesangium and tubulus of patients ing voltage of 20 kV. with the early stages of diabetic nephropathy. AGEs were stained

in the mesangium and the capillary and the staining of the Motphometric and statistical analyses

Morphometric analyses of high-resolution scanning electron micrographs were performed as has been described previously [26]. Micrographs were selected for measurements when the meshwork structure could be observed via the en face view at higher magnifications in order to "flatten" the basement mem-

brane surface and to reduce secondary errors caused by its waviness. Measurements of the meshwork's strands and pores were made using these micrographs. The long diameter of the pores and the perpendicular width of the strands were measured with an ocular micrometer equipped with a 0.1 mm scale. Finally,

mesangium in particular increased with the progression of gbmerulosclerosis (Fig. 1B). The most characteristic finding in diabetic patients was strong AGE staining in the nodular lesions (Fig. 1C). Bowman's capsule and periglomerular fibrosis in sclerosing gbomeruli became positive for AGE staining (Fig. 1D). AGE staining was also observed in the renal artery. Also noted were AGE staining in tubular, peritubular, and extracellular matrix, and the extents of the staining increased in the more advanced cases. Transmission electron microscopic findings

Following treatment with a series of detergents, all cellular the mean value and standard deviation for each parameter were determined after the diameter of 100 pores and the width of 50 components were completely removed, while the GBM and strands had been measured. Statistical significance was analyzed mesangial matrix remained intact (Fig. 2). The mesangial matrix using the Student's t-test and the Cocrane Welch test. A P value was less electron-dense than the GBM, and was clearly distinof < 0.05 was considered significant. guishable from the GBM. In controls, the mesangial matrix was

522

Makino et a!: Glycated diabetic mesangial matrix changes

Fig. 5. High resolution scanning electron micrograph of an acellular nodular lesion (NOD) from a patient with diabetic nephropathy. Note the rough surfac of the nodular lesion (X 16,800). GBM is glomerular basement membrane.

confined to the centro-lobular regions and appeared discontinu- mesangial regions were occupied with the mesangial matrix and ous when observed with a transmission electron microscope (Fig. expanded to the peripheral capillaries, and there also was narrow2A). At higher magnifications, the mesangial matrix appeared to ing of the capillary lumens (Fig. 4A). At higher magnifications, the be less compact than the GBM, and revealed a fine fibrillar meshwork structure, which was composed of strands, became structure (Fig. 2B). Globules were observed on these fibrils. Very evident (Fig. 4B). The widths of these strands were relatively few bundles of banded collagen fibers were observed and mesh- uniform (mean SD 10.5 2.7 nm). These strands generally work formation was not clear. branched and anastomosed with neighboring strands at short

In specimens from diabetic patients, an expansion of the mesangial area and an increase in the mesangial matrix were both observed (Fig. 2C). Wrinkling of the GBM was observed in the mesangial portion. At higher magnifications, the fibrous structure of the mesangial matrix could be observed (Fig. 2D). Globules

intervals, but some straight and curvilinear strands also were observed. The diameters of the pores were variable (mean SD 14.0 6.1 nm), however, statistically they did not differ from those of the controls. Globular clusters also were noted on the matrix surface, which might be adhesion plaques.

were noted mainly at the intersection of the fibrils. Abundant An acellular nodular lesion also was observed (Fig. 5). At collagen fibers among the mesangial matrix were observed in higher magnifications the loose meshwork structure appeared to some specimens. A few collagen fibers were located in parallel be composed of various sized strands ranging from 6 nm to 24 nm with one another (Fig. 2D). (mean SD 11.4 3.8 nm; Fig. 6). Occasionally some thick fibers were noted to be running parallel among the strands, corresponding to the transmission electron micrograph image (Fig. 2D). The On HRSEM examination, the mesangial matrix extended onto pore sizes were variable, ranging from 4 nm to 70 nm (mean SD the endothelial surface of the GBM and appeared as a series of 23.6 12.3 nm), and were statistically larger than those of the anastomosing ridges (Fig. 3A). At higher magnification, the controls. Scanning electron microscopic findings

mesangial matrix appeared to consist of an irregular, loosely

Figure 7 illustrates histograms of the distribution of strand

meshed structures composed of relatively uniform strands ranging

widths and pore diameters measured in the mesangial matrix of a

from 3 to 18 nm wide (mean SD 10.4

normal control, a diabetic patient and a patient with nodular

3.3 nm). The pores demonstrated fairly uniform diameters ranging from 4 to 28 nm (mean SD 13.0 5.3 nm; Fig. 3B). In portions, thin bridging fibrils were observed over the stomata. In specimens from diabetic patients, increases in the mesangial matrix and thickening of the GBM could be observed clearly. The

lesions.

Discussion In this investigation, we were able to demonstrate immunohis-

tochemically the preferential localization of AGEs to nodular

Makino et a!: Glycated diabetic mesangial mat,vc changes

tn

523

;

Fig. 6. Higher magnification of the nodular lesion (Fig. 5, large arrow). Note many thick collagen fibers (large arrows) among the loose meshwork

composed of small pores (arrowheads), larger pores (triangle arrowheads), and strands (small arrows) (X 168,000).

lesions of patients with diabetic glomerulosclerosis. Furthermore, the looser meshwork structures of these lesions were visualized using HRSEM. Concerning the role of glycation in the diabetic kidney, Mitsuhashi et al [281 have reported the increased content of AGE in the

studies from our laboratory [22, 26, 29—31] and others [32, 33] clearly demonstrate the meshwork and pores of the basement membranes. In our previous investigation, we observed the polygonal meshwork structures not only in the GBM [22, 26, 29—31], but also in the tubular basement membrane [26, 31], Bowman's capsule basement membrane, and the peritubular capillary base-

renal cortex of a streptozotocin-induced diabetic rat in their immunochemical study. Miyata and Monnier [8] have demon- ment membrane [26]. The strands averaged 6 to 7 nm wide, strated pyrraline only in the fibrous crescent areas of the glomer- whereas the pore sizes within the meshwork were variable and uli or obliterated glomeruli. On the other hand, our antibody, differed according to the basement membrane type. The strands in although it did not react with pyrraline [9], detected AGEs in the mesangial matrix were a little wider than those of other non-obliterated glomeruli and an expanded mesangial matrix, basement membranes in the kidney. The pores in the mesangial especially in nodular lesions. High resolution scanning electron microscopy, equipped with a field emission source, a very short objective lens, and a resolving power of 0.5 nm at 30 kV [21], made it possible to observe the

three-dimensional ultrastructure of the GBM. These HRSEM

matrices also were larger compared with those of the GBM. By transmission electron microscopy, the mesangial matrix appears more coarsely fibrillar and less electron dense [34]. These differences might be due to the dissimilarity of components of the extracellular matrix. Both the GBM and the mesangial matrix

524

Makino et a!: Glycated diabetic mesangial matrix changes

A

C

B

40

20

4

12 20

4

28

12

20 28

5 10

30

50

70

Diameter of pores, nm

D

F

E

40

80

20

40

0 3

9

18

3

9

18

3

9

15

24

Width of strands, nm Fig. 7. Hthtograms of pore diameter and strand width selected in the mesangial matrices representing one control (A, D), one patient with diabetes (B, E), and a nodular patient (C, F).

contain type IV collagen, laminin, entactin, nidogen and heparan late stages, minor components or components not normally sulfate proteoglycan [15]. The following additional components, present in the mesangial matrix (such as type III, type V, and type including fibronectin [35, 36], amyloid-P component [37], actomy- VI collagen) may increase [39, 40, 44]. osin [35], and chondroitin sulfate proteoglycans [38], are localized Following the removal of cellular components, many collagen

exclusively in the mesangial matrix. Mundel et al [34] have fibers were observed in the mesangial matrix by transmission observed abundant microfibrils with a mean thickness of 15 nm in electron microscopy. Thick fibers observed in the mesangial matrix by HRSEM, which corresponded to transmission electron micrograph findings appeared to be type III or type V collagen. Because type V and type VI collagens are the major components of nodular lesions [39, 44], increases in these interstitial and fibril or microfibril collagens may contribute to the formation of wider strands in the mesangial matrix of a nodular lesion. As mesangial cells are known to produce these extracellular matrix components in vitro, phenotypic changes of the mesangial cell may be respon40], and a decrease [41] or no change [42] in heparan sulfate sible for change in these components [45]. These structural changes are accompanied by an impairment of proteoglycans in cases of early diabetes. However, in cases of progressive glomeruloscierosis, decreases in type IV collagen [39, the associative properties of the basement membrane components 40] and heparan sulfate proteoglycans [42,43] have been observed that reduces their ability to interact with each other to form an in immunohistochemical studies. Immunohistochemically, in the ordered polymeric complex. Tarsio, Regfer and Furcht [46] have

the mesangial matrix by transmission electron microscopy using tannic acid. They have suggested that the ubiquitous distribution of fibronectin within the mesangial matrix is due to the corresponding distribution of microfibrils. In diabetics, various compositional changes of matrix proteins have been studied extensively. Biochemical and immunohistochemical studies have indicated an increase in the normal constituents of the matrix, such as type IV collagen, and fibronectin [39,

Makino et al: Glycated diabetic mesangial matrix changes

found that nonen2ymatic glycation of laminin and/or type IV collagen significantly reduces the cooperative binding with hepa-

525

11. Ciss'o JCF, ULRICH PC, Buc R, CEltatI A: Detection of an advanced glycation product bound to protein in situ. J Biol Chem

260:7970—7974, 1985 nfl. They have speculated that this is one of the causes of the 12. HAYASE F, NAGAtW RH, MIYATA S, NI0R0GE FG, MONNIER VM: decrease in heparan sulfate proteoglycan in relation to the Aging of proteins; Immunological detection of a glucose-derived proteinuria found in patients with diabetic nephropathy. pyrrole formed during Maillard reaction in vivo. J Biol Chem 263: 3758—3764, 1989 More recently, Anderson, Tsilibary and Charonis [33] have examined structural changes in the bovine kidney tubular base- 13. SELL DR, MONNIER VM: Structure elucidation of a senescense

ment membrane following in vitro nonenzymatic glycation. Using high-resolution scanning electron microscopy, they have identified an increase in the number of large openings in the meshwork, and

have concluded that nonenzymatic glycation and cross link formation among basement membrane components might lead to modifications of the basement membrane. The preferential localization of AGEs in a nodular lesion suggests that AGEs might cause porosity of the extracellular matrix in vivo. As no metalloprotease that is specific for type VI collagen has been identified thus far [47], AGEs formation might occur preferentially in type VI collagen-rich nodular lesions, which are sites of slow turnover. To our knowledge, we are the first to demonstrate the immunohistochemical localization of AGEs in diabetic nodular lesions and to provide morphologic evidence of the loose meshwork structure of the mesangial matrix in diabetic patients in vivo. Acknowledgments A part of this study was supported by a Grant-in-Aid for Scientific Research to Dr. H. Makino (C-04671481, C-0761252) and Dr. Z. Ota (A-04404039). We are grateful to Mrs. Toshiyo Hashimoto, Mrs. Kumiko

Nakata and Mr. Noboru Kishimoto for their assistance in electron

and the pathogenesis of diabetic complications. Ann Intern Med 101:527—537, 1984

15. K.wAK YS: Biophysiology of glomerular filtration and proteinuria. Lab Invest 51:7—21, 1984

16. WEBER S, ENGEL J, WIEDEMANN H, G.ki'ivILJa RW, TIMPL R:

Subunit structure and self assembly of the globular domain of basement-membrane collagen type IV. Eur J Biochem 139:401—410, 1984 17. YURCHENCO PD, RUBEN GC: Basement membrane structure in situ:

Evidence for lateral association in type IV collagen network. I Cell Biol 105:2559—2568, 1987 18. YURCHENCO PD, CHENG YS, COLOGNAT H: Laminin forms an independent network in basement membranes. J Cell Biol 117:1119—1133, 1992

19. C1-IoNIs AS, REGER LA, DEGE JE, Kouzt-KoLlAKos K, FURCHT LT, WOHLI-JUETER RM, T5IUBAKY EC: Laminin alterations after in vitro nonenzymatic glycosylation. Diabetes 39:807—8 14, 1990 20. CI-IA1oNIs AS, TsILIBARY EC: Structural and functional changes of laminin and type IV collagen after nonenzymatic glycation. Diabetes 41(Suppl 2):49—51, 1992

21. TAI'tKA K, MATU! I, KURODA K, Mrrsusnmt A: A new ultra high resolution scanning electron microscope (HUST-1). Biomedica SEM 14:23—25, 1985 (in Japanese) 22. MAKINO H, YIAsAKi Y, HARAMOTO T, SHIKATA K, HIRONAKA K, OTA Z, KANWAR YS: Ultrastructural changes of extracellular matrices

microscopy.

Reprint requests to Dr. Hirofumi Makino, Third Department of Internal

Medicine, Okayama University Medical School 2—5-1 Shikata-cho, Okayama 700, Japan.

in diabetic nephropathy revealed by high resolution scanning and immunoelectron microscopy. Lab Invest 68:45—55, 1993 23. GUESDON J-L, TERNYNIK T, AviwKAs S: The use of avidin-biotin

interaction in immunoenzyme techniques. J Histochem Cytochem

References 1. BROWNLEE

crosslink from human extracellular matrix: Implication of pentoses in the aging process. I Clin Invest 85:380—384, 1990 14. BROWNLEE M, VIssAKA H, CERAMI A: Nonenzymatic glycosylation

M, CErwu A, VissAIt H: Advanced glycosylation end

products in tissue and the biochemical basis of diabetic complications. NEng1J Med 318:1315—1321, 1988 2. VIAssA1 H, BUcAIA R, STRIKER L: Pathogenic effects of advanced

glycosylation: Biochemical, biologic, and clinical implications for diabetes and aging. Lab Invest 70:138—151, 1994 3. MAILLARD LC: Actions des acides amines sur les sucres: Formation des melanoidines par voie methodique. CRAcad Sci 154:66—68, 1912

4. Doicz.x MH: The significance of the products of the Maillard (Browning) reaction in diabetes. Diab Med 8:505—516, 1991 5. COHEN MP, URDIvIA E, SURMA M, Wu VY: Increased glycosylation of glomerular basement membrane collagen in diabetes. Biochem Biophys Res Comm 95:765—769, 1980

27:1131—1139, 1979

24. MAKINO H: Three-dimensional ultrastructure of rat acellular glomerulus by scanning electron microscopy. I Electron Microsc 37:294—304, 1988 25. MAKINO H, YAMASAKI Y, HIRONAKA K, OTA Z: Glomerular extracel-

lular matrices in rat diabetic glomerulopathy by scanning electron microscopy. Virclwws Arch B Cell Pathol 62:19—24, 1992 26. HIRONAKA K, MAKINO H, YAMASAKI Y, OTA Z: Renal basement

membranes by ultrahigh resolution scanning electron microscopy. Kidney mt 43:334—345, 1993

27, MuxARn T: A metal impregnating method of biological specimens for scanning electron microscopy. Arch Histol Jpn 35:323—326, 1973 28. MITSUHASHI T, NAKAYAMA H, ITOH T, KuwA.nMA 5, Aoiu S, ATSUMI

T, KOIKE T: Immunohistochemical detection of advanced glycation end products in renal cortex from STZ-induced diabetic rat. Diabetes

6. Nic.ui Y, HoRn Y, NISHIN0 T, SHIHU H, SAKAGUCHI Y, KA-

42:826—832, 1993 29. YAMASAKI Y, MAKINO H, HIRONAKA K, OTA Z: Three-dimensional

G05HIMA T, D0HI K, MAKITA Z, VLASSARA H, BucAIA R: Immuno-

architecture of rat glomerular basement membrane by ultra-high

histochemical localization of advanced glycosylation end products in coronary atheroma and cardiac tissue in diabetes mellitus.Am JPathol

resolution scanning electron microscopy. Acta Med Okayama 44:333— 335, 1990 30. HIRONAKA K, MAKINO H, YAM.ksAiu Y, OTA Z: Pores in the glomer-

143:1649—1656, 1993

7. HARRINGTON CR, CoIcAco CALS: Alzheimer's disease. A glycation connection. Nature 370:247—248, 1994 8. MIYATA 5, MONNIER V: Immunohistochemical detection of advanced glycosylation end products in diabetic tissue using monoclonal antibody to pyrraline. J Cliii Invest 89:1102—1112, 1992

9. Hoiuucrn 5, Artiu N, M0RIN0 Y: Immunochemical approach to characterize advanced glycation end products of the Maillard reaction. J Biol Chem 266:7329—7332, 1991

10. Aiwu N, UENO N, CHAKRABARTI B, Moiur'io Y, H0RIUcHI S: Immunochemical presence of advanced glycation end products in human lens proteins and its positive correlation with aging. J Biol Chem 267:10211—10214, 1992

ular basement membrane revealed by ultrahigh resolution scanning electron microscopy. Nephmn 64:647—649, 1993

31. YAMASAKI Y, MA3uto H, OTA Z: Meshwork structure in bovine glomerular and tubular basement membrane by ultra-high-resolution scanning electron microscopy. Nephron 66:189—199, 1994 32. SHIRATO I, ToMINo Y, KOIDE H, SAKAI T: Fine structure of the glomerular basement membrane of rat kidney visualized by highresolution scanning electron microscopy. Cell Tissue Res 266:1—10, 1991 33. ANDERSON SS, TSILIBARY EC, CHARONIS AS: Nonenzymatic glycosy-

lation-induced modifications of intact bovine kidney tubular basement membrane. I Cliii Invest 92:3045—3052, 1993

526

Makino et a!: Glycated diabetic mesangial ,natth changes

34. MUNDEL P, ELGER M, S&ii T, Kiuz W: Microfibrils are a major component of the mesangial matrix in the rat kidney. Cell Tissue Res 254:183—187, 1988 35. HOUSER MT, SCHEINMAN JI, BASGEN J, STEFFES MW, MICHAEL AF:

Preservation of mesangium and immunohistochemically defined antigen in glomerular basement membrane isolated by detergent extraction. J Clin Invest 69:1169—1175, 1982 36. LINDER E, MIETrINEN A, TONROTH T: Fibronectin as a marker for the

glomerular mesangium in immunohistology of kidney biopsies. Lab Invest 42:70—75, 1980 37. DYan RF, LocKwooD CM, KERSHAW M, DUANCE VC, BALTZ ML,

PEPYs MD: Amyloid P component of normal human glomerular basement membrane. J Ep Med 152:1162—1174, 1980

38. Kwoa YS, RosENzwEIo LI, LIrnceR A, JAKuB0wsKI ML: Decreased de novo synthesized sulfated glycosaminoglycans in the gbmerular basement membrane and mesangial matrix. Lab Invest 49: 216—225, 1983

39. F.u RI, SCHEINMMAN JI, MAUER SM, MIchAEL AF: Polyantigenic expression of basement membrane constituents of diabetic nephropathy. Diabetes 32(Suppl 2):34—39, 1983 40. NERLIcIl A, SCHLEICHER E: Inununohistochemical localization of

extracellular matrix components in human diabetic gbomerular lesions. Am J Pathol 139:889—899, 1991

41. KANwAi YS, ROSENZWEIG L, LIR A, JAEunowsiu M: Decreased

de novo synthesis of glomerular proteoglycans in diabetes: Biochemical and autoradiographic evidence. Proc Nat! Acad Sci USA 80:2272— 2275, 1983 42. VEIu.aER RL, STEFFES MW, SIssoN-Ross S, MAUER SM: Heparan sulfate proteoglycan in the glomerular basement membrane in type 1 diabetes mellitus. Kidney nt 41:1070—1080, 1992 43. MAKINO H, IKEDA S, HARAMOTO T, OrA Z: Heparan sulfate proteo-

glycans are lost in patients with diabetic nephropathy. Nephron 61:415—421, 1992 44. MAKINO H, SHIKATA K, WIESLANDER J, WADA J, KASHIHARA K,

YosHlolcA K, OTA Z: Localization of fibril/microfibril and basement membrane collagen in diabetic glomerulosclerosis in type 2 diabetes. Diab Med 11:304—311, 1994

45. ABRASS CL PETERSON CV, RAUGI G: Phenotypic expression of collagen types in mesangial matrix of diabetic and nondiabetic rats. Diabetes 37:1695—1702, 1988

46. TARsI0 JF, REGFER LA, FtJRCHT LT: Molecular mechanisms in basement membrane complications of diabetes. Alterations in heparin, laminin, and type IV collagen association. Diabetes 37:532—539, 1988 47. BIRKEDAL-HANSEN H, Mooan WGI, BODDEN MK, WINDSOR U, BrnsmDa-HANsEN B, DECAaro A, ENGLER JA: Matrix metalloproteinases: A review. Crit Rev Oral Biol Med 4:197—250, 1993