- Email: [email protected]
Vascular cell adhesion molecule-l expression in the renal interstitium of diabetic KKAy mice
Diabetes Research and Clinical Practice 44 (1999) 1 – 8
Vascular cell adhesion molecule-l expression in the renal interstitium of diabetic KKAy mice Keisuke Ina a,*, Hirokazu Kitamura a, Toshimitsu Okeda b, Kaoruko Nagai a, Zhong Y. Liu a, Midori Matsuda a, Yoshihisa Fujikura a b
a Department of Anatomy, Oita Medical Uni6ersity, Hasama-machi, Oita 879 -5593, Japan First Department of Medicine, Oita Medical Uni6ersity, Hasama-machi, Oita 879 -5593, Japan
Received 27 July 1998; received in revised form 8 October 1998; accepted 14 January 1999
Abstract To investigate the mechanism of interstitial inflammation in diabetic nephropathy, we used spontaneously diabetic KKAy mice. Twelve KKAy mice were divided into two groups; six mice were fed standard mouse chow ad libitum and six mice were placed on a diet (i.e. they received the same amount of chow as six control C57BL mice). Diabetic KKAy mice developed hypercholesterolemia and albuminuria. Animals were killed at 16 weeks of age and renal tissues were immunostained for vascular cell adhesion molecule-l (VCAM-1). In diabetic KKAy mice, the renal interstitium was infiltrated by monocytes, lymphocytes, plasma cells, and other cells. The walls of venules near the infiltrating cells were more intensely stained for VCAM-1 when compared with other sites. In contrast, the VCAM-1 staining of arterioles and peritubular capillaries was not significantly increased. There was weak VCAM-1 staining of the infiltrating cells, including lymphocytes, monocytes, and other cells. Electron microscopy demonstrated immunolabeling for VCAM-1 on the cell surface and in the cytoplasm of both infiltrating cells and vascular endothelial cells. In KKAy mice placed on a diet, there was less staining for VCAM-1 and cellular infiltration was also decreased. Thus, increased expression of VCAM-1 by the endothelial cells of venules and VCAM-1 expression by infiltrating cells were demonstrated in the interstitium of kidneys from diabetic mice. These results suggest that increased expression of VCAM-1 by endothelial cells and infiltrating cells contributes to interstitial inflammation in diabetic nephropathy. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Diabetic nephropathy; Interstitial inflammation; Vascular cell adhesion molecule-l
1. Introduction
* Corresponding author.
It has been proposed that diabetic nephropathy is predominantly a glomerular disease. However,
0168-8227/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 9 9 ) 0 0 0 1 1 - X
2
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
a number of observations [1 – 6] have suggested a potentially important role for alterations in the structure and function of the renal interstitium in diabetes. It has been shown that pathologic changes of the interstitium are characterized by inflammation and fibrosis in the early stage of diabetic nephropathy [1–6]. Although it has been suggested that interstitial inflammatory cells play an active role in interstitial fibrogenesis, primarily due to their ability to synthesize fibrogenetic cytokines such as transforming growth factor beta-l (TGF-b1) [7–11], the mechanism initiating interstitial inflammation remains unknown. Recently, the role of adhesion molecules in atherosclerosis [12] and various types of inflammation [8,11,13] has attracted attention. Integrin receptors mediate the adhesion of monocytes to endothelial cells. Very late antigen-4 (VLA-4) binds monocytes to vascular cell adhesion molecule-1 (VCAM-1), which is expressed by the endothelial cells of the peritubular capillaries in the renal interstitium [8,11]. A role of VCAM-1 has been reported in renal interstitial inflammation induced by various causes [7–11,14] other than diabetes mellitus. Those reports indicated that VCAM-1 was expressed by mononuclear cells in the inflammatory lesions and that its expression might be associated with cell adhesion in the lesions. However, the role of VCAM-1 in interstitial inflammation in diabetic kidneys has not been investigated. In the present study, we employed KKAy mice as an animal model of diabetes mellitus. The KKAy mouse is a congenic strain established by transfer of the yellow obese gene (Ay) into KK mice with moderate hyperglycemia through repeated crossing of yellow obese mice and KK mice [15]. KKAy mice show both obesity and a high plasma glucose level, which are features recognized in type II diabetes. To determine the molecular basis of interstitial inflammation in diabetic nephropathy, expression of VCAM-1 was investigated at the light and electron microscopic levels using the immunogold-silver method. Other parameters, such as body weight (wt), plasma glucose, serum lipids, serum creatinine, and the
urinary albumin assessed.
excretion
rate,
were
also
2. Methods
2.1. Experimental design Twelve spontaneously diabetic male KKAy mice and six control male C57BL mice (Japan CREA, Tokyo, Japan), 4 weeks old, were housed in individual cages in a temperature- and lightcontrolled environment, and were fed standard mouse chow ad libitum with free access to water. Six of the KKAy mice were placed on a diet from 5 weeks old. These mice were given the same amount of chow as control animals ate each day, once a day at 5:00 p.m. until the end of the study. Daily food intake, body wt, and the plasma glucose level (PG) (glucose-oxidase method) at 10:00 a.m. were monitored in the mice at 5 and 16 weeks of age. Serum triglyceride, serum cholesterol, serum creatinine, and the urinary albumin excretion rate (urinary albumin adjusted for the urinary creatinine level) were examined at 16 weeks of age. Animals were anesthetized with Nembutal at 16 weeks of age, and both kidneys were collected for histologic evaluation.
2.2. Morphological examinations At the time of sacrifice, the kidneys were sliced and frozen in a compound for tissue specimens and stored at − 80°C until histological examination. Cryostat sections (10 mm thick) of the kidneys were stained with H&E. Serial sections were used for immunohistochemical examination.
2.3. Immunohistochemical studies 2.3.1. Light microscopy Ten-micrometer-thick cryosections of kidneys from each group were evaluated by indirect immunogold-silver staining [16] to determine the pattern of VCAM-1 expression. Sections were hydrated and incubated for 2 h at room temperature
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
with a rat anti-mouse VCAM-1 antibody (Phar Mingen, San Diego, CA, USA). After rinsing in phosphate buffer (pH 7.4), the sections were further incubated for 1 h at room temperature with a gold-labeled goat anti-rat immunoglobulin G (IgG) antibody (Amersham Pharmacia Biotech, Buckinghamshire, UK). Then the sections were developed for 40 min at 20°C in our physical developer [16], which contained 0.8% citric acid, 0.3% bromohydroquinone, 15% gum Arabic, and 0.2% silver nitrate. Finally, the sections were counterstained with Kernechtrot. The intensity of immunostaining for VCAM-1 was scored semiquantitatively in interstitial fields including an arteriole and a venule with or without cell infiltration and in the peritubular capillaries. Forty renal interstitial fields from each group of animals were examined at a 400 × magnification. The following semiquantitative score was used: 0 =negative staining; 1 + = minimal staining, equivalent to the staining intensity in peritubular capillaries from control mice; 2 + = mild staining; 3 + = moderate staining; 4 + = intense staining, equivalent to the staining intensity of arterioles in the renal interstitium from all groups.
2.3.2. Electron microscopy Cryosections of kidneys from each group were fixed for 10 min at room temperature in 3% paraformaldehyde plus 0.5% glutaraldehyde. After being rinsed, the sections were processed in the same manner as for indirect immunohistochemistry at the light microscopic level. After development, the sections were dehydrated and embedded in epon–epoxy resin. Ultrathin sections were cut, mounted on neoprenecoated copper grids, and stained with methanolic uranyl acetate [17] and aqueous lead citrate [18]. The sections were observed and photographed under a transmisson electron microscope (TEM-1200EXII; JEOL, Tokyo, Japan) at 80 kV. Identification of a venule or a capillary was also done electron microscopically.
3
2.4. Statistical analysis All values are expressed as the mean 9 S.D. Statistical significance (PB 0.05) was evaluated using Student’s t-test.
3. Results
3.1. Daily food intake, body wt and PG le6el The chow intake was 1.4- to 2.1-fold in diabetic KKAy mice, compared with C57BL mice at the same age throughout the study. All mice principally ate chow at night. The body wt and PG of KKAy mice tended to be greater than those of control mice at 5 weeks of age [wt, 27.0 9 2.3 vs. 20.891.0 g (PB 0.001); PG, 225976 vs. 1729 14 mg/dl (NS)], and were significantly increased compared with dieting KKAy mice and control mice at 16 weeks of age [wt, 47.59 3.1 (PB 0.001) vs. 35.19 1.5 (PB 0.05) vs. 32.99 l.6 g; PG, 518943 (PB 0.001) vs. 214 9 24 (PB 0.01) vs 173914 mg/dl, respectively] (Table 1).
3.2. Renal function Diabetic KKAy mice showed a significantly elevated urinary albumin excretion rate [3.439 0.46 (PB 0.001) vs. 2.01 9 0.76 (NS) vs. 1.57 + 0.20 mg/g creatinine, respectively; logarithmic scale], but serum creatinine levels were not elevated at 16 weeks of age compared with the other groups (0.149 0.05 vs. 0.169 0.05 vs. 0.189 0.04 mg/dl, respectively) (Table 1).
3.3. Lipid abnormalities Diabetic KKAy mice showed a significant increase of serum cholesterol and triglyceride levels at 16 weeks of age (Table 1). Dieting KKAy mice had markedly lower serum triglyceride levels as well as lower serum cholesterol levels compared with diabetic KKAy mice at the same age [serum cholesterol, 185920 (PB 0.001) vs. 1359 16 (P B0.001) vs. 1059 9 mg/dl; serum triglyceride, 2319 55 (PB 0.001) vs. 859 38 (PB 0.05) vs. 469 8 mg/dl, respectively] (Table 1).
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
4
3.4. H &E staining Cellular infiltration was present close to venules in the renal interstitium of kidneys harvested from KKAy mice with or without dieting (Figs. 1 and 2). However, diabetic KKAy mice had five times more lesions per kidney section when compared with dieting KKAy mice. In diabetic KKAy mice, infiltrating cell numbers tended to be higher than in dieting KKAy mice [54927 vs. 25 96 (NS)]. The infiltrating cells consisted of monocytes, lymphocytes, plasma cells, and other cells (7.7 92.6% vs. 72.1 9 10.4% vs. 16.9 9 7.4% vs. 3.3 91.3% for both KKAy groups). Enlarged glomeruli with mild to moderate mesangial expansion were occasionally seen in cross-sections of kidneys from diabetic KKAy mice, and to a lesser extent from dieting KKAy mice (Fig. 1).
3.5. Light microscopic immunostaining for VCAM-1 Weak staining for VCAM-1 was seen in infiltrating cells, lymphocytes, monocytes, and other cells in the kidneys of KKAy mice with or
without dieting. There was no difference of staining intensity among the infiltrating cells (Figs. 2 and 3). In KKAy mice, the walls of venules close to the infiltrating cells showed significantly more intense staining for VCAM-1 compared with other parts of the venules (Figs. 2 and 3). However, moderate staining was also seen in parts of the venule walls away from infiltrating cells (data not shown). Thus, the vascular wall of a venule was divided into two parts by the relationship with cell infiltration in a cross-section (Fig. 3). Marked staining for VCAM-1 was found in the endothelial cells of arterioles in the renal interstitium from all three groups (Figs. 2 and 3). Peritubular capillaries revealed minimal staining and there was no significant difference among the three groups (Figs. 2 and 3).
3.6. Electron microscopic immunolabeling for VCAM-1 Endothelial cells were the only cells stained for VCAM-1 in the vascular walls (Fig. 4). Immunolabeling was seen on the cell surface and in the cytoplasm of both endothelial cells and infiltrating cells (Fig. 4).
Table 1 Systemic parametersa
Body weight (g) Plasma glucose (mg/dl) Serum creatinine (mg/dl) Serum cholesterol (mg/dl) Serum triglycerides (mg/dl) Urinary albumin/urinary creatinine (mg/g creatinine) (logarithmic scale) a
5 weeks of age
16 weeks of age
Control
Diabetic KKAy
Control
Dieting KKAy
Diabetic KKAy
20.891.0 172914
27.092.3d 225 9 76
32.9 9 1.6 173 914 0.18 90.04 1059 9 46 9 8 1.57 90.20
35.1 9 1.5b 214 9 24c 0.169 0.05 135 916d 85 9 38b 2.01 90.76
47.5 9 3.1f,g 518 943d,g 0.14 9 0.05 185 920d,g 231 955d,g 3.43 9 0.46d,e
Results are the mean9 SD (N= 6 per group). PB0.05 vs control group. c PB0.01 vs control group. d PB0.001 vs control group. e PB0.05 vs dieting KKAy group. f PB0.01 vs dieting KKAy group. g PB0.001 vs dieting KKAy group. b
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
5
the immunoreactions for VCAM-1 were virtually missing.
4. Discussion
Fig. 1. H&E staining from control mice (A), dieting KKAy mice (B), and diabetic KKAy mice (C). In control mice, infiltrated lesions were not seen. In diabetic KKAy mice, cellular infiltrations (arrows) near venules in the interstitium and glomerular lesions (arrowhead) with mesangial expansion were occasionally observed, and to a lesser extent in dieting KKAy mice. v, venule ( ×100).
3.7. Staining with preabsorbed antibody In the light and electron microscopic sections which were reacted with preabsorbed antibody,
Several studies [8,11,19–24] have supported a crucial role for VCAM-1/VLA-4 interactions in mononuclear cell recruitment. They have demonstrated the expression of VCAM-1 on the surface of monocytes, but not an increase of VCAM-1 expression in the vascular walls [8,11]. In the present study, most of the cells infiltrating the renal interstitium in KKAy mice consisted of mononuclear cells (monocytes, lymphocytes, and plasma cells) that were immunostained for VCAM-1. This result suggested that mononuclear cells might be more likely to express VCAM-1 than other circulating cells in diabetics. It was also revealed that the expression of VCAM-1 was increased in the walls of venules close to the infiltrating cells when compared with the venules at other sites. Moderate staining was also found in parts of the venule walls without adjacent infilrating cells. Taken together, these findings suggested that increased expression of VCAM-1 by the endothelial cells of venules might be an initial event in cell infiltration. Interstitial inflammation of the kidneys has various causes, including cyclosporine nephritis [9], unilateral ureteral obstruction [7], protein overload [8], 5/6 nephrectomy [10], dietary hypercholesterolemia [11], and irradiation [14], and the mechanism of underlying inflammation might be different in each case. In the present model, increased expression of VCAM-1 by endothelial cells of the venules may have led to interaction with VLA-4 on circulating cells, which then infiltrated interstitium, allowing VCAM-1 on the surface of infiltrating cells to interact with VLA-4 on adjacent infiltrating cells in the inflammatory lesion. The effects of VCAM-1 promoting cell infiltration may include chemoattraction as well as facilitating cell adhesion. The cause of increased VCAM-1 expression remains unknown. In the diabetic milieu, hemodynamic and/or metabolic alterations might be responsible. However, genetic factors seem to be
6
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
less important because diet decreased both VCAM1 expression and cell infiltration. KKAy mice showed hypercholesterolemia in the present study and hypercholesterolemia appears to promote the accumulation of oxidized low density lipoprotein (ox LDL) in renal tissue [11,25 – 28]. Eddy reported
that ox LDL binds to the scavenger receptor of macrophages in inflammatory lesions and that expression of VCAM-1 might be increased as a result [11]. Also, ox LDL might act as a cytokine that induces the expression of VCAM-1 by endothelial cells and infiltrating cells [11].
Fig. 2. H&E staining (A, C, E) and immunostaining (B, D, F) for VCAM-1 in serial sections from control mice (A, B), dieting KKAy mice (C, D), and diabetic KKAy mice (E, F). In diabetic KKAy mice and dieting KKAy mice, the walls (arrowheads) of the venules close to the sites of cellular infiltration were more strongly immunostained compared with other parts of the venules (B, D, F). Infiltrating cells (arrows) were weakly immunostained and there was almost no difference in staining among the various types of cells (F). Immunostained cells in the infiltrative lesion were demonstrated at higher magnification ( ×790) in the inset of (F). a, arteriole; v, venule ( ×380).
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
7
Fig. 3. Semiquantitative analysis of immunostaining for VCAM-1 in the three groups. ( ) control group; ( ) dieting KKAy group; ( ) diabetic KKAy group. Scoring wasdone as described in Section 2. The staining score of venule walls close to infiltrating cells was significantly increased compared with that for other parts of the venules. * P B0.001 compared with the corresponding control group. C, control group; D, dieting KKAy group; K, diabetic KKAy group; VI, venule wall close to infiltrating cells; V, venule wall away from infiltrating cells; Ca, capillaries; I, infiltrating cells.
In the present study, it was also revealed that enlarged glomeruli with mild to moderate mesangial expansion were occasionally seen in diabetic KKAy mice. It was most likely that albuminuria occurred due to the glomerular lesion rather than the interstitial one. In summary, interstitial inflammation was present at the early stage of diabetic nephropathy. The mechanism of such inflammation was examined in relation to expression of the adhesion molecule VCAM-1. Increased expression of VCAM-1 in the venules and infiltrating cells was demonstrated in the interstitium of kidneys fiom spontaneously diabetic KKAy mice. It was concluded that this increased expression of VCAM-1 might be partly responsible for interstitial inflammation in diabetic nephropathy.
Acknowledgements We are grateful for the technical assistance of Shuji Tatsukawa and the able secretarial assistance of Yukari Goto.
Fig. 4. Electron immunolabeling for VCAM-1 in the renal interstitium of diabetic KKAy mice. The venular wall close to the infiltrating cells (A) and the infiltrating cells (B) are positive. Endothelial cells are exclusively labeled in the venular walls (A). There is immunolabeling of the cytoplasm as well as the cell surface (arrowheads) of both endothelial cells and infiltrating cells (A, B). e, endothelial cell (A, × 9500; B, ×6700).
References [1] R. Bader, H. Bader, K.E. Grund, S. Mackensen-Haen, H. Christ, A. Bohle, Structure and function of the kidney in diabetic glomerulosclerosis: correlations between morphological and functional parameters, Pathol. Res. Pract. 1G7 (1980) 204 – 216. [2] S.M. Mauer, M.W. Steffes, E.N. Ellis, D.E.R. Sutherland, D.M. Brown, F.C. Goetz, Structural – functional relationships in diabetic nephropathy, J. Clin. Invest. 74 (1984) 1143 – 1155.
8
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
[3] A. Bohle, M. Wehrmann, O. Bogenschutz, C. Batz, G.A. Muller, The pathogenesis of chronic renal failure in diabetic nephropathy, Pathol. Res. Pract. 187 (1991) 251– 259. [4] P.H. Lane, M.W. Steffes, P. Fioretto, S.M. Mauer, Renal interstitial expansion in insulin-dependent diabetes mellitus, Kidney Int. 43 (1993) 661–667. [5] F.N. Ziyadeh, S. Goldfarb, The renal tubulointerstitium in diabetes mellitus, Kidney Int. 39 (1991) 464–475. [6] D. Suzuki, M. Miyazaki, K. Jinde, et al., In situ hybridization studies of matrix metalloproteinase-3, tissue inhibitor of metalloproteinase-1 and type IV collagen in diabetic nephropathy, Kidney Int. 52 (1997) 111–119. [7] S. Ishidoya, J. Morrissey, R. McCracken, A. Reyes, S. Klahr, Angiotensin II receptor antagonist ameliorates renal tubulointerstitial fibrosis caused by unilateral ureteral obstruction, Kidney Int. 47 (1995) 1285–1294. [8] A.A. Eddy, C.M. Giachelli, Renal expression of genes that promote interstitial inflammation and fibrosis in rats with protein-overload proteinuria, Kidney Int. 47 (1995) 1546 – 1557. [9] B.A. Young, E.A. Burdmann, R. Johnson, et al., Cellular proliferation and macrophage influx precede interstitial fibrosis in cyclosporine nephrotoxicity, Kidney Int. 48 (1995) 439 – 448. [10] V. Kliem, R.J. Johnson, C.E. Alpers, et al., Mechanisms involved in the pathogenesis of tubulointerstitial fibrosis in 5/6-nephrectomized rats, Kidney Int. 49 (1996) 666– 678. [11] A.A. Eddy, Interstitial inflammation and fibrosis in rats with diet-induced hypercholesterolemia, Kidney Int. 50 (1996) 1139 – 1149. [12] M.I. Cybulsky, M.A.J.r. Gimbrone, Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis, Science 251 (1991) 788–791. [13] N. Yagi, K. Yokono, K. Amano, et al., Expression of intercellular adhesion molecule 1 on pancreatic 13-cells accelerates 13-cell destruction by cytotoxic T-cells in murine autoimmune diabetes, Diabetes 44 (1995) 744– 752. [14] T. Oikawa, M. Freeman, W. Lo, D.E. Vaughan, A. Fogo, Modulation of plasminogen activator inhibitor-1 in vivo:A new mechanism for the anti-fibrotic effect of renin-angiotensin inhibition, Kidney Int. 51 (1997) 164– 172.
.
[15] H. Iwatsuka, A. Shino, Z. Suzuoki, General survey of diabetic features of yellow KK mice, Endocrinol. Jpn. 17 (1970) 23 – 35. [16] K. Takita, T. Shimada, H. Kitamura, S. Fujii, M. Nakamura, Studies on the physical developer for use in immunohistochemistry, Acta Histochem. Cytochem. 23 (1990) 647 – 662. [17] J.H. Stempak, R.T. Ward, An improved staining method for electron microscopy, J. Cell Biol. 22 (1964) 697 – 701. [18] E.S. Reynolds, The use of lead citrate at high pH as an electron opaque stain inelectronmicroscopy, J.Cell Biol. 17 (1963) 208 – 212. [19] C.E. Alpers, K.L. Hudkins, C.L. Davis, et al., Expression of vascular cell adhesion molecule-1 in kidney allograft rejection, Kidney Int. 44 (1993) 805 – 816. [20] D.M. Briscoe, J.S. Pober, W.E. Harmon, R.S. Cotran, Expression of vascular cell adhesion molecule-1 in human renal allografts, J. Am. Soc. Nephrol. 3 (1993) 1180 – 1185. [21] C. Brockmeyer, M. Ulbrecht, D.J. Schendel, et al., Distribution of cell adhesion molecules (ICAM-1, VCAM-1, ELAM-1) in renal tissue during allograft rejection, Transplantation 55 (1993) 610 – 615. [22] Y. Lin, J.A. Kirby, D.A. Browell, et al., Renalallograft rejection expression and function of VCAM-1 on tubular epithelial cells, Clin. Exp. Immunol. 92 (1993) 145 – 151. [23] R.P. Wuthrich, Vascular cell adhesion molecule-l (VCAM-1) expression in murinelupus nephritis, Kidney Int. 42 (1992) 903 – 914. [24] D. Seron, J.S. Cameron, D.O. Haskard, Expression of VCAM-1 in the normal and diseased kidney, Nephrol. Dial. Transplant. 6 (1991) 917 – 922. [25] W.F. Keane, M.P. O’Donnell, B.L. Kasiske, Y. Kim, Oxidative modification of low-density lipoproteins by mesangial cells, J. Am. Soc. Nephrol. 4 (1993) 187 – 194. [26] B.H. Rovin, L.C. Tan, LDL stimulates mesangial fibronectin production and chemoattractant expression, Kidney Int. 43 (1993) 218 – 225. [27] D.C. Wheeler, R.S. Chana, N. Topley, M.M. Petersen, M. Davies, J.D. Williams, Oxidation of low density lipoprotein by mesangial cells may promote glomerular injury, Kidney Int. 45 (1994) 1628 – 1636. [28] J. Rie, S. Silbiger, J. Neugarten, Glomerular macrophages in nephrotoxic serum nephritis are activated to oxidize low-density lipoprotein, Am. J. Kidney Dis. 26 (1995) 362 – 367.
Vascular cell adhesion molecule-l expression in the renal interstitium of diabetic KKAy mice Keisuke Ina a,*, Hirokazu Kitamura a, Toshimitsu Okeda b, Kaoruko Nagai a, Zhong Y. Liu a, Midori Matsuda a, Yoshihisa Fujikura a b
a Department of Anatomy, Oita Medical Uni6ersity, Hasama-machi, Oita 879 -5593, Japan First Department of Medicine, Oita Medical Uni6ersity, Hasama-machi, Oita 879 -5593, Japan
Received 27 July 1998; received in revised form 8 October 1998; accepted 14 January 1999
Abstract To investigate the mechanism of interstitial inflammation in diabetic nephropathy, we used spontaneously diabetic KKAy mice. Twelve KKAy mice were divided into two groups; six mice were fed standard mouse chow ad libitum and six mice were placed on a diet (i.e. they received the same amount of chow as six control C57BL mice). Diabetic KKAy mice developed hypercholesterolemia and albuminuria. Animals were killed at 16 weeks of age and renal tissues were immunostained for vascular cell adhesion molecule-l (VCAM-1). In diabetic KKAy mice, the renal interstitium was infiltrated by monocytes, lymphocytes, plasma cells, and other cells. The walls of venules near the infiltrating cells were more intensely stained for VCAM-1 when compared with other sites. In contrast, the VCAM-1 staining of arterioles and peritubular capillaries was not significantly increased. There was weak VCAM-1 staining of the infiltrating cells, including lymphocytes, monocytes, and other cells. Electron microscopy demonstrated immunolabeling for VCAM-1 on the cell surface and in the cytoplasm of both infiltrating cells and vascular endothelial cells. In KKAy mice placed on a diet, there was less staining for VCAM-1 and cellular infiltration was also decreased. Thus, increased expression of VCAM-1 by the endothelial cells of venules and VCAM-1 expression by infiltrating cells were demonstrated in the interstitium of kidneys from diabetic mice. These results suggest that increased expression of VCAM-1 by endothelial cells and infiltrating cells contributes to interstitial inflammation in diabetic nephropathy. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Diabetic nephropathy; Interstitial inflammation; Vascular cell adhesion molecule-l
1. Introduction
* Corresponding author.
It has been proposed that diabetic nephropathy is predominantly a glomerular disease. However,
0168-8227/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 9 9 ) 0 0 0 1 1 - X
2
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
a number of observations [1 – 6] have suggested a potentially important role for alterations in the structure and function of the renal interstitium in diabetes. It has been shown that pathologic changes of the interstitium are characterized by inflammation and fibrosis in the early stage of diabetic nephropathy [1–6]. Although it has been suggested that interstitial inflammatory cells play an active role in interstitial fibrogenesis, primarily due to their ability to synthesize fibrogenetic cytokines such as transforming growth factor beta-l (TGF-b1) [7–11], the mechanism initiating interstitial inflammation remains unknown. Recently, the role of adhesion molecules in atherosclerosis [12] and various types of inflammation [8,11,13] has attracted attention. Integrin receptors mediate the adhesion of monocytes to endothelial cells. Very late antigen-4 (VLA-4) binds monocytes to vascular cell adhesion molecule-1 (VCAM-1), which is expressed by the endothelial cells of the peritubular capillaries in the renal interstitium [8,11]. A role of VCAM-1 has been reported in renal interstitial inflammation induced by various causes [7–11,14] other than diabetes mellitus. Those reports indicated that VCAM-1 was expressed by mononuclear cells in the inflammatory lesions and that its expression might be associated with cell adhesion in the lesions. However, the role of VCAM-1 in interstitial inflammation in diabetic kidneys has not been investigated. In the present study, we employed KKAy mice as an animal model of diabetes mellitus. The KKAy mouse is a congenic strain established by transfer of the yellow obese gene (Ay) into KK mice with moderate hyperglycemia through repeated crossing of yellow obese mice and KK mice [15]. KKAy mice show both obesity and a high plasma glucose level, which are features recognized in type II diabetes. To determine the molecular basis of interstitial inflammation in diabetic nephropathy, expression of VCAM-1 was investigated at the light and electron microscopic levels using the immunogold-silver method. Other parameters, such as body weight (wt), plasma glucose, serum lipids, serum creatinine, and the
urinary albumin assessed.
excretion
rate,
were
also
2. Methods
2.1. Experimental design Twelve spontaneously diabetic male KKAy mice and six control male C57BL mice (Japan CREA, Tokyo, Japan), 4 weeks old, were housed in individual cages in a temperature- and lightcontrolled environment, and were fed standard mouse chow ad libitum with free access to water. Six of the KKAy mice were placed on a diet from 5 weeks old. These mice were given the same amount of chow as control animals ate each day, once a day at 5:00 p.m. until the end of the study. Daily food intake, body wt, and the plasma glucose level (PG) (glucose-oxidase method) at 10:00 a.m. were monitored in the mice at 5 and 16 weeks of age. Serum triglyceride, serum cholesterol, serum creatinine, and the urinary albumin excretion rate (urinary albumin adjusted for the urinary creatinine level) were examined at 16 weeks of age. Animals were anesthetized with Nembutal at 16 weeks of age, and both kidneys were collected for histologic evaluation.
2.2. Morphological examinations At the time of sacrifice, the kidneys were sliced and frozen in a compound for tissue specimens and stored at − 80°C until histological examination. Cryostat sections (10 mm thick) of the kidneys were stained with H&E. Serial sections were used for immunohistochemical examination.
2.3. Immunohistochemical studies 2.3.1. Light microscopy Ten-micrometer-thick cryosections of kidneys from each group were evaluated by indirect immunogold-silver staining [16] to determine the pattern of VCAM-1 expression. Sections were hydrated and incubated for 2 h at room temperature
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
with a rat anti-mouse VCAM-1 antibody (Phar Mingen, San Diego, CA, USA). After rinsing in phosphate buffer (pH 7.4), the sections were further incubated for 1 h at room temperature with a gold-labeled goat anti-rat immunoglobulin G (IgG) antibody (Amersham Pharmacia Biotech, Buckinghamshire, UK). Then the sections were developed for 40 min at 20°C in our physical developer [16], which contained 0.8% citric acid, 0.3% bromohydroquinone, 15% gum Arabic, and 0.2% silver nitrate. Finally, the sections were counterstained with Kernechtrot. The intensity of immunostaining for VCAM-1 was scored semiquantitatively in interstitial fields including an arteriole and a venule with or without cell infiltration and in the peritubular capillaries. Forty renal interstitial fields from each group of animals were examined at a 400 × magnification. The following semiquantitative score was used: 0 =negative staining; 1 + = minimal staining, equivalent to the staining intensity in peritubular capillaries from control mice; 2 + = mild staining; 3 + = moderate staining; 4 + = intense staining, equivalent to the staining intensity of arterioles in the renal interstitium from all groups.
2.3.2. Electron microscopy Cryosections of kidneys from each group were fixed for 10 min at room temperature in 3% paraformaldehyde plus 0.5% glutaraldehyde. After being rinsed, the sections were processed in the same manner as for indirect immunohistochemistry at the light microscopic level. After development, the sections were dehydrated and embedded in epon–epoxy resin. Ultrathin sections were cut, mounted on neoprenecoated copper grids, and stained with methanolic uranyl acetate [17] and aqueous lead citrate [18]. The sections were observed and photographed under a transmisson electron microscope (TEM-1200EXII; JEOL, Tokyo, Japan) at 80 kV. Identification of a venule or a capillary was also done electron microscopically.
3
2.4. Statistical analysis All values are expressed as the mean 9 S.D. Statistical significance (PB 0.05) was evaluated using Student’s t-test.
3. Results
3.1. Daily food intake, body wt and PG le6el The chow intake was 1.4- to 2.1-fold in diabetic KKAy mice, compared with C57BL mice at the same age throughout the study. All mice principally ate chow at night. The body wt and PG of KKAy mice tended to be greater than those of control mice at 5 weeks of age [wt, 27.0 9 2.3 vs. 20.891.0 g (PB 0.001); PG, 225976 vs. 1729 14 mg/dl (NS)], and were significantly increased compared with dieting KKAy mice and control mice at 16 weeks of age [wt, 47.59 3.1 (PB 0.001) vs. 35.19 1.5 (PB 0.05) vs. 32.99 l.6 g; PG, 518943 (PB 0.001) vs. 214 9 24 (PB 0.01) vs 173914 mg/dl, respectively] (Table 1).
3.2. Renal function Diabetic KKAy mice showed a significantly elevated urinary albumin excretion rate [3.439 0.46 (PB 0.001) vs. 2.01 9 0.76 (NS) vs. 1.57 + 0.20 mg/g creatinine, respectively; logarithmic scale], but serum creatinine levels were not elevated at 16 weeks of age compared with the other groups (0.149 0.05 vs. 0.169 0.05 vs. 0.189 0.04 mg/dl, respectively) (Table 1).
3.3. Lipid abnormalities Diabetic KKAy mice showed a significant increase of serum cholesterol and triglyceride levels at 16 weeks of age (Table 1). Dieting KKAy mice had markedly lower serum triglyceride levels as well as lower serum cholesterol levels compared with diabetic KKAy mice at the same age [serum cholesterol, 185920 (PB 0.001) vs. 1359 16 (P B0.001) vs. 1059 9 mg/dl; serum triglyceride, 2319 55 (PB 0.001) vs. 859 38 (PB 0.05) vs. 469 8 mg/dl, respectively] (Table 1).
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
4
3.4. H &E staining Cellular infiltration was present close to venules in the renal interstitium of kidneys harvested from KKAy mice with or without dieting (Figs. 1 and 2). However, diabetic KKAy mice had five times more lesions per kidney section when compared with dieting KKAy mice. In diabetic KKAy mice, infiltrating cell numbers tended to be higher than in dieting KKAy mice [54927 vs. 25 96 (NS)]. The infiltrating cells consisted of monocytes, lymphocytes, plasma cells, and other cells (7.7 92.6% vs. 72.1 9 10.4% vs. 16.9 9 7.4% vs. 3.3 91.3% for both KKAy groups). Enlarged glomeruli with mild to moderate mesangial expansion were occasionally seen in cross-sections of kidneys from diabetic KKAy mice, and to a lesser extent from dieting KKAy mice (Fig. 1).
3.5. Light microscopic immunostaining for VCAM-1 Weak staining for VCAM-1 was seen in infiltrating cells, lymphocytes, monocytes, and other cells in the kidneys of KKAy mice with or
without dieting. There was no difference of staining intensity among the infiltrating cells (Figs. 2 and 3). In KKAy mice, the walls of venules close to the infiltrating cells showed significantly more intense staining for VCAM-1 compared with other parts of the venules (Figs. 2 and 3). However, moderate staining was also seen in parts of the venule walls away from infiltrating cells (data not shown). Thus, the vascular wall of a venule was divided into two parts by the relationship with cell infiltration in a cross-section (Fig. 3). Marked staining for VCAM-1 was found in the endothelial cells of arterioles in the renal interstitium from all three groups (Figs. 2 and 3). Peritubular capillaries revealed minimal staining and there was no significant difference among the three groups (Figs. 2 and 3).
3.6. Electron microscopic immunolabeling for VCAM-1 Endothelial cells were the only cells stained for VCAM-1 in the vascular walls (Fig. 4). Immunolabeling was seen on the cell surface and in the cytoplasm of both endothelial cells and infiltrating cells (Fig. 4).
Table 1 Systemic parametersa
Body weight (g) Plasma glucose (mg/dl) Serum creatinine (mg/dl) Serum cholesterol (mg/dl) Serum triglycerides (mg/dl) Urinary albumin/urinary creatinine (mg/g creatinine) (logarithmic scale) a
5 weeks of age
16 weeks of age
Control
Diabetic KKAy
Control
Dieting KKAy
Diabetic KKAy
20.891.0 172914
27.092.3d 225 9 76
32.9 9 1.6 173 914 0.18 90.04 1059 9 46 9 8 1.57 90.20
35.1 9 1.5b 214 9 24c 0.169 0.05 135 916d 85 9 38b 2.01 90.76
47.5 9 3.1f,g 518 943d,g 0.14 9 0.05 185 920d,g 231 955d,g 3.43 9 0.46d,e
Results are the mean9 SD (N= 6 per group). PB0.05 vs control group. c PB0.01 vs control group. d PB0.001 vs control group. e PB0.05 vs dieting KKAy group. f PB0.01 vs dieting KKAy group. g PB0.001 vs dieting KKAy group. b
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
5
the immunoreactions for VCAM-1 were virtually missing.
4. Discussion
Fig. 1. H&E staining from control mice (A), dieting KKAy mice (B), and diabetic KKAy mice (C). In control mice, infiltrated lesions were not seen. In diabetic KKAy mice, cellular infiltrations (arrows) near venules in the interstitium and glomerular lesions (arrowhead) with mesangial expansion were occasionally observed, and to a lesser extent in dieting KKAy mice. v, venule ( ×100).
3.7. Staining with preabsorbed antibody In the light and electron microscopic sections which were reacted with preabsorbed antibody,
Several studies [8,11,19–24] have supported a crucial role for VCAM-1/VLA-4 interactions in mononuclear cell recruitment. They have demonstrated the expression of VCAM-1 on the surface of monocytes, but not an increase of VCAM-1 expression in the vascular walls [8,11]. In the present study, most of the cells infiltrating the renal interstitium in KKAy mice consisted of mononuclear cells (monocytes, lymphocytes, and plasma cells) that were immunostained for VCAM-1. This result suggested that mononuclear cells might be more likely to express VCAM-1 than other circulating cells in diabetics. It was also revealed that the expression of VCAM-1 was increased in the walls of venules close to the infiltrating cells when compared with the venules at other sites. Moderate staining was also found in parts of the venule walls without adjacent infilrating cells. Taken together, these findings suggested that increased expression of VCAM-1 by the endothelial cells of venules might be an initial event in cell infiltration. Interstitial inflammation of the kidneys has various causes, including cyclosporine nephritis [9], unilateral ureteral obstruction [7], protein overload [8], 5/6 nephrectomy [10], dietary hypercholesterolemia [11], and irradiation [14], and the mechanism of underlying inflammation might be different in each case. In the present model, increased expression of VCAM-1 by endothelial cells of the venules may have led to interaction with VLA-4 on circulating cells, which then infiltrated interstitium, allowing VCAM-1 on the surface of infiltrating cells to interact with VLA-4 on adjacent infiltrating cells in the inflammatory lesion. The effects of VCAM-1 promoting cell infiltration may include chemoattraction as well as facilitating cell adhesion. The cause of increased VCAM-1 expression remains unknown. In the diabetic milieu, hemodynamic and/or metabolic alterations might be responsible. However, genetic factors seem to be
6
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
less important because diet decreased both VCAM1 expression and cell infiltration. KKAy mice showed hypercholesterolemia in the present study and hypercholesterolemia appears to promote the accumulation of oxidized low density lipoprotein (ox LDL) in renal tissue [11,25 – 28]. Eddy reported
that ox LDL binds to the scavenger receptor of macrophages in inflammatory lesions and that expression of VCAM-1 might be increased as a result [11]. Also, ox LDL might act as a cytokine that induces the expression of VCAM-1 by endothelial cells and infiltrating cells [11].
Fig. 2. H&E staining (A, C, E) and immunostaining (B, D, F) for VCAM-1 in serial sections from control mice (A, B), dieting KKAy mice (C, D), and diabetic KKAy mice (E, F). In diabetic KKAy mice and dieting KKAy mice, the walls (arrowheads) of the venules close to the sites of cellular infiltration were more strongly immunostained compared with other parts of the venules (B, D, F). Infiltrating cells (arrows) were weakly immunostained and there was almost no difference in staining among the various types of cells (F). Immunostained cells in the infiltrative lesion were demonstrated at higher magnification ( ×790) in the inset of (F). a, arteriole; v, venule ( ×380).
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
7
Fig. 3. Semiquantitative analysis of immunostaining for VCAM-1 in the three groups. ( ) control group; ( ) dieting KKAy group; ( ) diabetic KKAy group. Scoring wasdone as described in Section 2. The staining score of venule walls close to infiltrating cells was significantly increased compared with that for other parts of the venules. * P B0.001 compared with the corresponding control group. C, control group; D, dieting KKAy group; K, diabetic KKAy group; VI, venule wall close to infiltrating cells; V, venule wall away from infiltrating cells; Ca, capillaries; I, infiltrating cells.
In the present study, it was also revealed that enlarged glomeruli with mild to moderate mesangial expansion were occasionally seen in diabetic KKAy mice. It was most likely that albuminuria occurred due to the glomerular lesion rather than the interstitial one. In summary, interstitial inflammation was present at the early stage of diabetic nephropathy. The mechanism of such inflammation was examined in relation to expression of the adhesion molecule VCAM-1. Increased expression of VCAM-1 in the venules and infiltrating cells was demonstrated in the interstitium of kidneys fiom spontaneously diabetic KKAy mice. It was concluded that this increased expression of VCAM-1 might be partly responsible for interstitial inflammation in diabetic nephropathy.
Acknowledgements We are grateful for the technical assistance of Shuji Tatsukawa and the able secretarial assistance of Yukari Goto.
Fig. 4. Electron immunolabeling for VCAM-1 in the renal interstitium of diabetic KKAy mice. The venular wall close to the infiltrating cells (A) and the infiltrating cells (B) are positive. Endothelial cells are exclusively labeled in the venular walls (A). There is immunolabeling of the cytoplasm as well as the cell surface (arrowheads) of both endothelial cells and infiltrating cells (A, B). e, endothelial cell (A, × 9500; B, ×6700).
References [1] R. Bader, H. Bader, K.E. Grund, S. Mackensen-Haen, H. Christ, A. Bohle, Structure and function of the kidney in diabetic glomerulosclerosis: correlations between morphological and functional parameters, Pathol. Res. Pract. 1G7 (1980) 204 – 216. [2] S.M. Mauer, M.W. Steffes, E.N. Ellis, D.E.R. Sutherland, D.M. Brown, F.C. Goetz, Structural – functional relationships in diabetic nephropathy, J. Clin. Invest. 74 (1984) 1143 – 1155.
8
K. Ina et al. / Diabetes Research and Clinical Practice 44 (1999) 1–8
[3] A. Bohle, M. Wehrmann, O. Bogenschutz, C. Batz, G.A. Muller, The pathogenesis of chronic renal failure in diabetic nephropathy, Pathol. Res. Pract. 187 (1991) 251– 259. [4] P.H. Lane, M.W. Steffes, P. Fioretto, S.M. Mauer, Renal interstitial expansion in insulin-dependent diabetes mellitus, Kidney Int. 43 (1993) 661–667. [5] F.N. Ziyadeh, S. Goldfarb, The renal tubulointerstitium in diabetes mellitus, Kidney Int. 39 (1991) 464–475. [6] D. Suzuki, M. Miyazaki, K. Jinde, et al., In situ hybridization studies of matrix metalloproteinase-3, tissue inhibitor of metalloproteinase-1 and type IV collagen in diabetic nephropathy, Kidney Int. 52 (1997) 111–119. [7] S. Ishidoya, J. Morrissey, R. McCracken, A. Reyes, S. Klahr, Angiotensin II receptor antagonist ameliorates renal tubulointerstitial fibrosis caused by unilateral ureteral obstruction, Kidney Int. 47 (1995) 1285–1294. [8] A.A. Eddy, C.M. Giachelli, Renal expression of genes that promote interstitial inflammation and fibrosis in rats with protein-overload proteinuria, Kidney Int. 47 (1995) 1546 – 1557. [9] B.A. Young, E.A. Burdmann, R. Johnson, et al., Cellular proliferation and macrophage influx precede interstitial fibrosis in cyclosporine nephrotoxicity, Kidney Int. 48 (1995) 439 – 448. [10] V. Kliem, R.J. Johnson, C.E. Alpers, et al., Mechanisms involved in the pathogenesis of tubulointerstitial fibrosis in 5/6-nephrectomized rats, Kidney Int. 49 (1996) 666– 678. [11] A.A. Eddy, Interstitial inflammation and fibrosis in rats with diet-induced hypercholesterolemia, Kidney Int. 50 (1996) 1139 – 1149. [12] M.I. Cybulsky, M.A.J.r. Gimbrone, Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis, Science 251 (1991) 788–791. [13] N. Yagi, K. Yokono, K. Amano, et al., Expression of intercellular adhesion molecule 1 on pancreatic 13-cells accelerates 13-cell destruction by cytotoxic T-cells in murine autoimmune diabetes, Diabetes 44 (1995) 744– 752. [14] T. Oikawa, M. Freeman, W. Lo, D.E. Vaughan, A. Fogo, Modulation of plasminogen activator inhibitor-1 in vivo:A new mechanism for the anti-fibrotic effect of renin-angiotensin inhibition, Kidney Int. 51 (1997) 164– 172.
.
[15] H. Iwatsuka, A. Shino, Z. Suzuoki, General survey of diabetic features of yellow KK mice, Endocrinol. Jpn. 17 (1970) 23 – 35. [16] K. Takita, T. Shimada, H. Kitamura, S. Fujii, M. Nakamura, Studies on the physical developer for use in immunohistochemistry, Acta Histochem. Cytochem. 23 (1990) 647 – 662. [17] J.H. Stempak, R.T. Ward, An improved staining method for electron microscopy, J. Cell Biol. 22 (1964) 697 – 701. [18] E.S. Reynolds, The use of lead citrate at high pH as an electron opaque stain inelectronmicroscopy, J.Cell Biol. 17 (1963) 208 – 212. [19] C.E. Alpers, K.L. Hudkins, C.L. Davis, et al., Expression of vascular cell adhesion molecule-1 in kidney allograft rejection, Kidney Int. 44 (1993) 805 – 816. [20] D.M. Briscoe, J.S. Pober, W.E. Harmon, R.S. Cotran, Expression of vascular cell adhesion molecule-1 in human renal allografts, J. Am. Soc. Nephrol. 3 (1993) 1180 – 1185. [21] C. Brockmeyer, M. Ulbrecht, D.J. Schendel, et al., Distribution of cell adhesion molecules (ICAM-1, VCAM-1, ELAM-1) in renal tissue during allograft rejection, Transplantation 55 (1993) 610 – 615. [22] Y. Lin, J.A. Kirby, D.A. Browell, et al., Renalallograft rejection expression and function of VCAM-1 on tubular epithelial cells, Clin. Exp. Immunol. 92 (1993) 145 – 151. [23] R.P. Wuthrich, Vascular cell adhesion molecule-l (VCAM-1) expression in murinelupus nephritis, Kidney Int. 42 (1992) 903 – 914. [24] D. Seron, J.S. Cameron, D.O. Haskard, Expression of VCAM-1 in the normal and diseased kidney, Nephrol. Dial. Transplant. 6 (1991) 917 – 922. [25] W.F. Keane, M.P. O’Donnell, B.L. Kasiske, Y. Kim, Oxidative modification of low-density lipoproteins by mesangial cells, J. Am. Soc. Nephrol. 4 (1993) 187 – 194. [26] B.H. Rovin, L.C. Tan, LDL stimulates mesangial fibronectin production and chemoattractant expression, Kidney Int. 43 (1993) 218 – 225. [27] D.C. Wheeler, R.S. Chana, N. Topley, M.M. Petersen, M. Davies, J.D. Williams, Oxidation of low density lipoprotein by mesangial cells may promote glomerular injury, Kidney Int. 45 (1994) 1628 – 1636. [28] J. Rie, S. Silbiger, J. Neugarten, Glomerular macrophages in nephrotoxic serum nephritis are activated to oxidize low-density lipoprotein, Am. J. Kidney Dis. 26 (1995) 362 – 367.