Vascular endothelial growth factor accelerates renal recovery in experimental thrombotic microangiopathy1

Kidney International, Vol. 58 (2000), pp. 2390–2399

Vascular endothelial growth factor accelerates renal recovery in experimental thrombotic microangiopathy1 YOON-GOO KIM, SHIN-ICHI SUGA, DUK-HEE KANG, J. ASHLEY JEFFERSON, MARILDA MAZZALI, KATHERINE L. GORDON, KATSUYUKI MATSUI, SILVANA BREITENEDER-GELEFF, STUART J. SHANKLAND, JEREMY HUGHES, DONTSCHO KERJASCHKI, GEORGE F. SCHREINER, and RICHARD J. JOHNSON Division of Nephrology, University of Washington Medical Center, Seattle, Washington, USA; Division of Nephrology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea; Department of Clinical Pathology, University of Vienna-AKH, Vienna, Austria; and SCIOS Inc., Sunnyvale, California, USA

Vascular endothelial growth factor accelerates renal recovery in experimental thrombotic microangiopathy. Background. Renal microvascular injury characterizes thrombotic microangiopathy (TMA). The possibility that angiogenic growth factors may accelerate recovery in TMA has not been studied. Methods. TMA was induced in rats by the selective right renal artery perfusion of antiglomerular endothelial cell IgG (30 mg/kg). Twenty-four hours later, rats received vascular endothelial growth factor (VEGF121, 100 ␮g/kg/day) or vehicle (control) daily until day 14. To evaluate renal function, the unperfused left kidney was removed at day 14, and rats were sacrificed at day 17. Results. The induction of TMA was associated with loss of glomerular and peritubular capillary endothelial cells and decreased arteriolar density at day 1. Some spontaneous capillary recovery was present by day 17; however, repair was incomplete, and severe tubulointerstitial damage occurred. The lack of complete microvascular recovery was associated with reduced VEGF immunostaining in the outer medulla. VEGFtreated rats had more glomeruli with intact endothelium, less glomerular ischemia (collapsed glomeruli), and greater peritubular capillary density with less peritubular capillary loss. This was associated with less tubulointerstitial fibrosis, less cortical atrophy, and improved renal function. Conclusions. VEGF accelerates renal recovery in this experimental model of TMA. These studies suggest that angiogenic growth factors may provide a new therapeutic strategy for diseases associated with endothelial cell injury.

1

See Editorial by Remuzzi, p. 2594

Key words: angiogenesis, ischemia, endothelium, hemolytic uremic syndrome. Received for publication February 25, 2000 and in revised form June 20, 2000 Accepted for publication June 29, 2000

 2000 by the International Society of Nephrology

The hemolytic uremic syndrome (HUS) and the related thrombotic microangiopathies (TMAs) are clinical syndromes characterized by thrombocytopenia, nonimmune hemolytic anemia, and variable degrees of renal insufficiency [1, 2]. Most cases of HUS are secondary to enteric infection with certain verotoxin-expressing strains of Escherichia coli (particularly O157:H7) [3]. Pathologically, the hallmark of HUS is injury to the glomerular and other renal microvascular endothelium and is characterized by arteriolar and capillary swelling, apoptosis, and/or detachment of endothelial cells with intracapillary thrombi formation [2, 4]. While most patients with E. coli-associated HUS recover from the acute illness, as many as 20 to 40% will show evidence of persistent renal injury, as manifested by proteinuria, hypertension, or chronic renal disease [5–8]. The long-term prognosis correlates with the extent of the initial histologic damage [8]. Recently, we developed a model of TMA in the rat by the selective renal artery perfusion with an antiendothelial cell antibody. This model is characterized by severe glomerular and peritubular capillary endothelial injury resulting in all of the histologic features of TMA. Similar to severe forms of HUS, the renal injury does not resolve, and progressive glomerular and tubulointerstitial damage results [9]. We postulated that the incomplete resolution of renal injury in our TMA model may be due to inadequate recovery of the glomerular and peritubular capillaries. Recently, the endothelial cell mitogen vascular endothelial growth factor (VEGF) [10] has been shown to be stimulate angiogenesis in animal models of peripheral and myocardial ischemia and in selected patients with critical limb ischemia [11–16]. We therefore examined whether VEGF could accelerate the recovery of the mi-

2390

Kim et al: VEGF stimulates renal recovery in TMA

crovasculature and improve histology and renal function in the TMA model. METHODS Experimental thrombotic microangiopathy model Thrombotic microangiopathy was induced in male Sprague-Dawley rats (280 to 330 g; Simonsen Laboratory, Gilroy, CA, USA) by selective perfusion of the right kidney with antiglomerular endothelial cell (GEN) IgG (30 mg/kg body weight). The characteristics of antiGEN antibody [9] and the technique of selective right renal artery perfusion [17] have been described previously. Although the antibody was raised against glomerular endothelial cells, it also cross-reacts to a lesser degree with the endothelium of the peritubular capillaries [9]. Validation of the effectiveness of the perfusion was established in all animals by performing a survival cortical biopsy 10 minutes after perfusion and verifying the presence of goat IgG by immunofluorescence. All rats perfused with anti-GEN IgG exhibited diffuse glomerular staining for goat IgG along all capillary walls in a coarse, uninterrupted pattern consistent with staining of GEN. Peritubular capillaries and arterioles in the kidney were also positive. Quantitative immunofluorescence densitometry of the goat IgG deposition in glomeruli was also performed using a Leitz immunofluorescence microscope with visual output connected to a photomultiplier as previously described. The mean quantitated fluorescence was considered adequate for the study [18]. Experimental protocol The studies were designed to determine whether VEGF hastens capillary repair and renal recovery in the TMA model. Rats with TMA were therefore randomized to receive VEGF or vehicle beginning one day after perfusion of the anti-GEN antibody. VEGF or vehicle injections were continued daily for 14 days. At 14 days, the normal nonperfused kidney was removed so that subsequent renal function measurements would reflect only that of the diseased kidney. Rats underwent sacrifice three days later, at which time blood and renal tissues were collected. VEGF121 (50 ␮g/kg/injection) was administered subcutaneously twice a day from days 1 through 14. Control animals received vehicle [phosphate-buffered saline (PBS), pH 7.2, control group]. The VEGF121 isoform was used, as it is the only isoform of VEGF that has no heparin binding and therefore can result in therapeutically effective plasma levels when administered subcutaneously [19]. The dose of VEGF121 used has been found to result in peak VEGF levels of 50 ng/mL at 100 minutes and 5 ng/mL at 6 hours, which are effective levels to stimulate angiogenesis in a rat ischemic hindlimb model (data not shown).

2391

To obtain baseline data on the severity of the capillary injury at day 1, a separate set of rats with TMA (N ⫽ 4) was sacrificed at day 1, and sagittal sections of the kidney were obtained for analysis. Control animals for this part of the study consisted of rats perfused with nonimmunized goat IgG in PBS, pH 7.2 (N ⫽ 4), and sacrificed at day 1. Renal histology Kidneys obtained at the time of sacrifice were divided longitudinally, and one half of the kidney was fixed in methyl Carnoy’s solution, processed, paraffin embedded, sectioned (4 ␮m), and stained with the periodic acidSchiff (PAS) reagent with hematoxylin counterstain. Additional 4 ␮m sections were immunostained using an indirect avidin-biotin immunoperoxidase method with the following primary antibodies: ␣SM-1 (1A4), a monoclonal IgG2a to ␣-smooth muscle actin (␣-SMA) (Sigma Chemical Co., St. Louis, MO, USA); RECA-1 (a monoclonal IgG1 antibody specific for a 60 kD rat endothelial cell membrane antigen) [20], and JG-12, a monoclonal antibody to rat endothelial cells (characterized later in this article). VEGF was stained with affinity-purified rabbit anti-VEGF polyclonal antibody (Santa Cruz Biotech, Inc., Santa Cruz, CA, USA). Controls included omission of the primary antibody or substitution with an irrelevant antibody of the same species and isotype. Capillary endothelial proliferation was detected by double immunostaining tissue sections with the antiendothelial cell antibody (JG-12) and an antibody to the proliferating nuclear cell antigen (PCNA; 19A2; Coulter Immunology, Hialeah, FL, USA). Controls consisted of substituting the primary antibody with an irrelevant monoclonal antibody of the same isotype as described previously [21]. Preparation and characterization of JG-12 antibody to rat microvascular endothelium Rat glomeruli were isolated by graded sieving, and a membrane protein fraction was prepared by incubation in 200 mmol/L Na2CO3 (pH ⫽ 11) followed by phase separation in Triton X-114, as described [22]. Monoclonal antibodies were raised [23] and screened for selective endothelial staining pattern on unfixed cryostat sections of normal rat kidneys and clone JG-12 was selected. Monoclonality was verified by three rounds of limiting dilution, and the antibody was further shown to be of the IgG1 class. To characterize the antigen of the JG-12 antibody, rat glomeruli were isolated, dissolved in sodium dodecyl sulfate (SDS) sample buffer, the protein lysate separated by SDS-polyacrylamide gel electrophoresis (SDSPAGE), and transferred to nitrocellulose [23]. Immunoblotting was performed using alkaline phosphatase coupled with sheep anti-mouse IgG (Promega, Madison,

2392

Kim et al: VEGF stimulates renal recovery in TMA

Fig. 1. Characteristics of JG-12 antibody. (a) Immunoblotting using alkaline phosphatase coupled with sheep anti-mouse IgG. JG-12 bound to a 70 kD protein in glomerular lysates. Lane A, Coomassie-Blue stained pattern of proteins in glomerular lysate. Lane B, Control incubation with irrelevant monoclonal antibody of the same subtype as JG-12 (that is, IgG1). Lane C, Immunoblot with monoclonal antibody JG-12, that specifically labels a broad 70 kD band. (b) Immunoelectron microscopy of renal glomeruli (CL, capillary lumen; GBM, glomerular basement membrane; P, podocyte; US, urinary space). JG-12 labeling was strictly confined to the luminal aspect of the endothelial cells. (c and d). Renal endothelial cell staining of normal rat kidneys with RECA-1 (c) and JG-12 (d) show a similar staining pattern (⫻160).

WI, USA) and developed by 5⬘-bromo-4-chloro-3-indolyl phosphate toluidine salt. JG-12 was found to specifically bind a 70 kD protein in the glomerular lysates (Fig. 1a). Similar results were obtained in the lung (data not shown). By immunohistochemistry, intense labeling of glomerular and peritubular capillaries was observed, and the pattern was similar to that seen with the RECA-1 antibody (Fig. 1 c, d). Labeling intensity was highest in the capillaries and decreased with the caliber of the blood vessel, and unlike RECA-1, staining was insignificant in arterial vessels. By immunoelectron microscopy on renal glomeruli, the specific labeling of the endothelial cells was apparent, especially in oblique sections that exposed the endothelial cell bodies (Fig. 1b). Labeling was strictly confined to the luminal aspect of the endothelial cells, and no other component of the glomerulus was labeled.

Quantitation of morphology and immunohistochemical staining Using computer-assisted image analysis software (Optimas, version 6.2; Media Cybernetics, Silver Springs, MD, USA), we measured the total area of the cortex, the outer medulla, and cortical thickness on the longitudinal section of each biopsy. Cortical thickness was defined as the mean length between the renal surface and the most inner (juxtamedullary) glomeruli utilizing 10 measurements for each biopsy. Glomerular sclerosis was defined as segmental or global capillary collapse with increased matrix deposition in over 25% of glomerular surface area. Glomerular collapse (which is generally considered a measurement of glomerular ischemia) was defined as shrinkage of the glomerular tuft to one half or less of the diameter of Bowman’s capsule. For each biopsy, all

Kim et al: VEGF stimulates renal recovery in TMA

glomeruli (N ⬎ 140) on the whole longitudinal section of a kidney were examined [24]. Tubulointerstitial injury was defined as increased tubular cellularity, basement membrane thickening, tubular dilation, atrophy, sloughing, or interstitial widening and was graded two different ways. The first method used a blinded semiquantitative scoring system (0 through 5) as follows: grade 0 ⫽ no tubulointerstitial injury present; grade 1 ⫽ ⬍10%; grade 2 ⫽ 10 to 25%; grade 3 ⫽ 26 to 50%; grade 4 ⫽ 51 to 75%; and grade 5 ⫽ ⬎75% of tubulointerstitium injured [24]. For each biopsy, the entire cortical and outer medullar regions on longitudinal section were evaluated under low power (⫻100) through 10 ⫻ 10 eyepiece grid (1 mm2), and a mean score per biopsy was calculated [25]. The second method used computer-assisted image analysis software (Optimas), in which the percentage area of tubulointerstitial injury was measured utilizing the entire cortex and outer medulla on longitudinal section at ⫻50 magnification. All analyses were performed blinded. Assessment of microvasculature Changes in endothelial cell morphology and density were confirmed by staining tissue sections with two antibodies (RECA-1 and JG-12) to different endothelial cell antigens. Although changes in capillary density appeared qualitatively similar using both antibodies, JG-12 gave a lower background and more intense staining than RECA-1. Therefore, only data obtained with JG-12 are shown (Fig. 1). Capillary loss was assessed using a rarefaction index [26]. JG-12–immunostained sections were examined through 10 ⫻ 10 eyepiece grid under a ⫻10 objective. At this magnification, the grid covered an area of 1 mm2. Each square (0.1 mm2) within the grid that did not contained JG-12–positive cells was scored. At least 50 fields in the cortex and outer medulla were examined on the longitudinal section of each kidney, and a mean score per biopsy was calculated. The minimum possible capillary score is 0, and the maximum score is 100, where 100 would indicate a complete absence of JG-12–positive cells [26]. Glomeruli with intact endothelium were defined when strong staining for JG-12 was present in over 75% of the glomerular tuft. Renal arterioles and interlobular arteries were identified by ␣smooth muscle actin staining, and the arteriolar density was reported as the number of arterioles and interlobular arteries per mm2 cortex. Additional measurements Blood urea nitrogen (BUN) was determined colorimetrically with a commercial kit (Sigma Diagnostics, St. Louis, MO, USA). Urinary nitrate (NO3⫺) plus nitrite (NO2⫺) excretion was measured by conversion of nitrate to nitrite using nitrate reductase followed by the addition of the Griess reagent (Clontech, Palo Alto, CA, USA).

2393

Statistical methods Values are expressed as mean ⫾ SE. Differences between the VEGF and vehicle-treated groups were evaluated with the unpaired Student’s t-test. When more than two groups were compared, analysis was by analysis of variance (ANOVA) with the Bonferonni correction for multiple comparisons. The correlation between capillary rarefaction and tubulointerstitial injury was evaluated by linear regression. RESULTS Perfusion with anti-GEN IgG results in microvascular endothelial injury with tissue damage at day 1 Control rats perfused with normal goat IgG had normal capillary morphology, no capillary rarefaction (score 0), and normal histology (PAS staining) at day 1. Arterioles and interlobular arteries in the cortex were also intact with a density of 4.95 ⫾ 0.31 vessels/mm2 (Table 1). In contrast, the infusion of anti-GEN IgG resulted in injury to both glomerular and peritubular capillary endothelium, as reported previously [9]. At day 1, there was extensive loss of peritubular capillaries (capillary rarefaction score, 41.5 ⫾ 2.8; Fig. 2). Capillary loss was particularly severe in the outer medullary area (rarefaction score 54.0 ⫾ 2.6) compared with that in the cortex (score 29.0 ⫾ 3.1). In glomeruli, a decrease in JG-12 staining (⬍75% of glomerular tuft) was present in 37.2 ⫾ 3.8% of glomeruli consistent with acute GEN loss. Arterioles and interlobular arteries in the cortex also frequently demonstrated denuded or swollen endothelium. There was a decrease in ␣-smooth muscle actin–positive arterioles and interlobular arteries in the cortex (2.36 ⫾ 0.26 vessels/mm2) compared with controls (4.95 ⫾ 0.31 vessels/mm2, P ⬍ 0.0001; Fig. 2). The renal microvascular injury was accompanied by thrombi in the glomerular capillary loops and tubular damage with tubular necrosis, sloughing of tubular epithelial cells, and intratubular casts (PAS staining; Fig. 2). In control rats perfused with normal IgG, VEGF expression was prominent in proximal tubules, the medullary thick ascending limb, and collecting ducts in the outer medulla and medullary rays. Glomeruli showed weak staining in glomerular epithelial cells. In rats with TMA, VEGF expression was increased focally in tubules in the cortex and in glomeruli at day 1; the latter is localized to visceral glomerular epithelial cells as reported previously [9]. However, in the medulla, VEGF expression was markedly decreased compared with control rats (Fig. 3 a, b). Effect of VEGF on renal recovery in TMA Rats with TMA were randomized to receive VEGF or vehicle beginning at day 1 after induction of disease,

2394

Kim et al: VEGF stimulates renal recovery in TMA Table 1. Effect of vascular endothelial growth factor (VEGF) treatment in experimental thrombotic microangiopathy (TMA) Day 1

Arteriole and interlobular artery density /mm2 Capillary rarefaction scorea Glomeruli with intact endothelium % Glomerulosclerosis % Collapsed glomeruli % Cortical thickness mm Tubulointerstitial injury % Area Score

Day 17

Control

TMA

Vehicle

VEGF group

P valuea

4.95 ⫾ 0.31 0

2.36 ⫾ 0.26 41.5 ⫾ 2.8

3.25 ⫾ 0.38 4.28 ⫾ 1.1

4.58 ⫾ 0.23 1.67 ⫾ 0.4

0.014 0.001

100 0 0 1.52 ⫾ 0.03

62.8 ⫾ 3.8 ND ND ND

80.0 ⫾ 3.0 14.7 ⫾ 2.9 5.30 ⫾ 1.20 1.11 ⫾ 0.02

90.6 ⫾ 2.7 7.8 ⫾ 2.7 1.54 ⫾ 0.45 1.32 ⫾ 0.05

0.025 0.116 0.015 0.003

ND ND

25.7 ⫾ 4.6 3.12 ⫾ 0.43

10.2 ⫾ 3.7 1.72 ⫾ 0.30

0.027 0.023

0 0

ND is not done. Data are mean ⫾ SE. a Comparison of vehicle-treated to VEGF-treated groups at day 17

Fig. 2. Renal histologic injury at day 1 in experimental thrombotic microangiopathy (TMA). (a) At day 1, rats with TMA display thrombi in glomerular capillary loops, tubular necrosis, sloughing of tubular epithelial cells, and tubular cast formation (PAS, ⫻200). (b) A loss of glomerular and peritubular capillary endothelial cell staining is present at day 1 (JG-12 immunostaining, ⫻200). (c and d) Compared with control animals (c), arterioles and interlobular arteries, identified by ␣-smooth muscle actin (␣-SMA) staining, are decreased in rats with TMA (d, ⫻100).

and the injections were continued until day 14. At this time, the nonperfused kidney was removed so that any observed changes in renal function would reflect the diseased kidney, and the rat was followed for three more days until sacrifice. Examination of the diseased kidney at day 17 in the vehicle-treated rats documented marked atrophy or dilation of tubules and interstitial fibrosis.

The tubulointerstitial injury was particularly severe in the outer medulla and medullary rays. VEGF expression remained decreased in the outer medulla, especially at sites of interstitial fibrosis (Fig. 3c). The focal areas of increased VEGF immunostaining in glomeruli and outer cortex observed at day 1 were no longer present. The tubulointerstitial injury was also associated with signifi-

Kim et al: VEGF stimulates renal recovery in TMA

2395

Fig. 3. Vascular endothelial growth factor (VEGF) expression in experimental thrombotic microangiopathy (TMA). (a) In control rats administered nonimmune IgG, VEGF immunostaining is prominent in tubules in the outer medulla and medullary rays. (b) In rats with TMA at day 1, there is a loss of VEGF in the outer medulla and focal increases in the cortex. (c) At day 17, vehicle-injected rats with TMA continued to show decreased VEGF expression, particularly at sites of tubulointerstitial injury. (⫻50).

cantly reduced renal function, with a mean BUN of 172 mg/dL (compared with normal values of 24 ⫾ 0.4 mg/dL, P ⬍ 0.01). In contrast, VEGF-treated rats had significantly less tubulointerstitial injury (PAS staining) in diseased kidneys at day 17 (Fig. 4 a, b). Both the percentage area of tubulointerstitial injury (10.2 ⫾ 3.7%) and the mean score (1.72 ⫾ 0.30) were significantly less in the VEGF group compared with the vehicle group (25.7 ⫾ 4.6%, 3.12 ⫾ 0.43, P ⬍ 0.05, respectively). VEGF-treated animals also had less cortical atrophy, as reflected by preservation of the cortical thickness (P ⬍ 0.01; Table 1). The BUN was lower by a factor of 2.5 in the VEGF group compared with the vehicle group (BUN 68 ⫾ 8 vs. 172 ⫾ 40 mg/dL, P ⫽ 0.03; Fig. 5). Five of the six control rats had BUN levels greater than 100 mg/dL versus one of six of VEGF-treated rats. The differences between VEGF- and vehicle-treated TMA rats are summarized in Table 1. Effect of VEGF on endothelial cell recovery Vehicle-treated rats had a substantial recovery of both the glomerular and peritubular capillary endothelium at day 17 (Table 1). However, VEGF-treated rats showed a significantly greater recovery. The number of glomeruli with positive staining for endothelial cells was greater in VEGF-treated rats (1.98 ⫾ 0.10 glomeruli/mm2) compared with vehicle-treated rats (1.62 ⫾ 0.10 glomeruli/ mm2, P ⬍ 0.02). The percentage of glomeruli with endothelial cell staining was also greater in the VEGF group compared with the vehicle group (P ⬍ 0.02; Table 1). VEGF treatment was also associated with a threefold

reduction in the number of glomeruli showing collapse of the tuft, consistent with less glomerular ischemia. Glomerular sclerosis tended to be lower in the VEGFtreated group, but the difference between the VEGF (7.8 ⫾ 2.7%) and vehicle groups (14.7 ⫾ 2.9%) did not reach statistical difference (Table 1). Peritubular capillary density was also greater in the VEGF group (Fig. 4 c, d and Table 1). In particular, VEGF-treated rats showed less capillary rarefaction than vehicle-treated controls (Table 1). Cortical arteriole and interlobular artery density (␣-SMA staining) was also significantly greater in the VEGF group (4.58 ⫾ 0.23 vessels/mm2) than in the vehicle group (3.25 ⫾ 0.38 vessels/mm2, P ⫽ 0.01) and was similar to that observed in the day 1 control rats (4.95 ⫾ 0.31 vessels/mm2, P ⫽ NS by analysis of variance with correction for multiple comparisons; Fig. 4 e, f). Peritubular capillary proliferation (identified by JG-12 and PCNA double staining) was present in both vehicle- and VEGF-treated groups. However, proliferating cells were rare in both VEGF (0.04 ⫾ 0.01 cells/mm2) and vehicle groups (0.02 ⫾ 0.003 cells/mm2), and there was no statistical difference between two groups (P ⬎ 0.05) at day 17 (which was 3 days after the VEGF infusion was stopped). Endothelial cell proliferation was most commonly observed in or near the sites of tubulointerstitial injury. Glomerular endothelial cell proliferation was minimal in both groups. Since VEGF-mediated angiogenesis acts through the nitric oxide system, we examined the effect of VEGF on urinary nitrite excretion. VEGF-treated rats had significantly greater excretion of urinary nitrates/nitrites than vehicle-treated controls (1913 ⫾ 180 vs. 786 ⫾ 400 nmol/

2396

Kim et al: VEGF stimulates renal recovery in TMA

Fig. 4. Effect of VEGF on experimental TMA. VEGF-treated rats had less tubulointerstitial injury (a, PAS) than vehicle-injected rats (b) at day 17. Glomerular collapse (b, arrows) was more prominent in vehicle-injected rats. Both capillary density (JG-12 staining, c and d) and arteriolar density (␣-SMA staining, e and f ) were greater in the VEGF-treated group (c and e) compared to vehicle-injected rats (d and f, ⫻100).

day, respectively, P ⫽ 0.02) Indeed, the urinary nitrate/ nitrite excretion in VEGF-treated rats approached that observed in the rats that had received nonimmune IgG (2401 ⫾ 243 nmol/day). Correlation of capillary rarefaction with tubulointerstitial injury and renal function Capillary rarefaction was most pronounced in areas of tubulointerstitial injury in both vehicle- and VEGFtreated rats. Within individual animals, the degree of capillary rarefaction correlated with the degree of tubu-

lointerstitial injury (TI) injury (r ⫽ 0.60, P ⫽ 0.038, Fig. 6). The degree of capillary loss (capillary rarefaction score) also correlated with BUN levels (r ⫽ 0.63, P ⫽ 0.026). DISCUSSION In this study, we tested the hypothesis that the administration of an angiogenic factor may accelerate renal recovery in a model of TMA. TMA was induced by the selective renal artery perfusion of an antibody to glomer-

Kim et al: VEGF stimulates renal recovery in TMA

Fig. 5. VEGF treatment improves tubulointerstitial injury and renal function in experimental TMA. The percentage of (A) tubulointerstitial injury was lower in VEGF treated rats at day 17, and (B) BUN levels were also reduced 2.5-fold.

Fig. 6. Capillary rarefaction correlates significantly with tubulointerstitial fibrosis in experimental TMA. Individual values for VEGFtreated rats (䊊) and vehicle-injected rats (䊏) are shown.

ular endothelial cells that also cross-reacts with peritubular capillary and renal arterial endothelium. We have previously shown that this antibody leads to an acute complement-dependent apoptosis of endothelial cells [27]. Consistent with this finding was the observation in the current study of marked loss of glomerular and peritubular capillary staining at day 1 in association with significant glomerular and tubulointerstitial injury. Rats were then randomized to receive VEGF or vehicle for 14 days, and 3 days later, the rats were sacrificed. While clearly some spontaneous capillary repair had occurred in the control kidneys, recovery of the glomerular and peritubular capillary network was incomplete, and there was significant chronic tubulointerstitial fibrosis and renal failure. In contrast, VEGF-treated rats had significantly greater recovery of the microvasculature, and this was associated with better renal function and less fibrosis. To our knowledge, this is the first time that an angiogenic factor has been used successfully to treat kidney disease. Since the study design was to determine whether VEGF could accelerate capillary repair, treatment was not started until 24 hours after induction of disease. At this time, renal injury is well established, as evidenced

2397

by the presence of glomerular thrombi and tubulointerstitial damage in the diseased rats as compared with control rats that were perfused with nonimmune IgG. Marked glomerular and peritubular capillary loss was also documented in the diseased rats by staining for two different endothelial cell antigens (identified by the RECA-1 and JG-12 antibodies). This strongly suggests that the loss of endothelial cell staining was not simply due to the loss of a particular endothelial cell antigen secondary to the disease process. The most severe tubular injury at day 1 was in the outer medulla and medullary rays, which is consistent with studies suggesting that tubular cells in these region are the most susceptible to ischemic injury [28, 29]. These tubules in the outer medulla and medullary rays also constitutively express VEGF, and it was therefore not surprising to find a reduction of VEGF staining in diseased animals at this time point. Since VEGF has trophic, prosurvival, and angiogenic properties for endothelial cells [30, 31], the acute loss of VEGF in the tubules at day 1 provided a rationale for investigating whether the infusion of VEGF could hasten endothelial recovery and improve overall outcome. We therefore administered VEGF121 to rats with this model of TMA from day 1 to day 14 in order to determine whether reconstitution of this angiogenic growth factor could improve renal recovery. To determine whether VEGF treatment could improve renal function in the diseased kidney, we removed the normal, nonperfused kidney at day 14 and determined the renal function three days later. Vehicle-treated rats showed a remarkable recovery of the glomeruli, which is consistent with a prior study documenting an early glomerular endothelial cell proliferative response in conjunction with up-regulation of podocyte VEGF expression [9]. There was also a substantial recovery of the peritubular capillaries in vehicletreated rats. Nevertheless, the recovery of the peritubular capillaries was incomplete, and at day 17, substantial tubulointerstitial fibrosis and renal failure were present. The observation that VEGF expression remained low in the outer medulla in conjunction with the inadequate capillary repair in vehicle-treated rats suggested that the continued lack of this angiogenic factor might have contributed to the poor recovery. Vascular endothelial growth factor-treated rats had a more complete recovery of the glomerular endothelium as well as more marked recovery of the peritubular capillaries compared with vehicle-infused rats. Capillary rarefaction was threefold less in VEGF-infused rats, and capillary density was visibly improved (Fig. 4 a, b). VEGF-infused rats also had markedly less tubulointerstitial fibrosis, preservation of cortical volume (as reflected by cortical thickness), and BUN levels that were 2.5-fold lower than control rats. Further evidence for a

2398

Kim et al: VEGF stimulates renal recovery in TMA

relationship between the microvascular recovery and the improvement in renal structure and function was the clear correlation in individual rats between capillary rarefaction and tubulointerstitial fibrosis (r ⫽ 0.6) and BUN (r ⫽ 0.63). Presumably, the improvement in blood supply would have resulted in less hypoxic damage to the tubules [32]. We do not believe that the improvement in capillary density observed in VEGF-infused rats was an artifact related to less tubulointerstitial expansion. Although vehicle-treated rats had more tubulointerstitial disease, they also had more cortical atrophy (less cortical thickness), which would have a tendency to result in crowding of glomeruli and the microvasculature and would therefore be expected to increase and not decrease capillary densities as was observed. Another question relates to how the relatively subtle differences in capillary rarefaction at day 17 (1.7 vs. 4.2%, respectively) could explain the marked differences in renal fibrosis and function. However, since the capillary loss was primarily restricted to the areas of tubulointerstitial injury (10 and 25% of the kidneys, respectively), the differences in capillary density are more marked than that reflected by the rarefaction scores, which are measurements of the entire kidney. In addition, the capillary rarefaction score was probably relatively insensitive compared with the dramatic changes in capillary density observed (Fig. 4). Further evidence for an angiogenic effect of the VEGF was the observation that urinary nitrites (metabolites of nitric oxide) were increased in the urine of VEGF-treated rats, as it is known that the angiogenic effects of VEGF act through stimulation of nitric oxide production [33, 34]. It remains possible, however, that some of the benefits of VEGF therapy may have been via direct actions on other cell populations within the kidney. In summary, we have demonstrated that VEGF treatment initiated after injury to the renal microvascular endothelium in a model of TMA-stimulated capillary repair reduced tubulointerstitial injury and improved renal function. These findings suggest that renal microvascular injury may mediate progressive tubulointerstitial injury, and that infusion of VEGF or other proangiogenic factors may benefit patients with HUS or TMA. ACKNOWLEDGMENTS Y-G. Kim is a recipient of the International Fellowship Award of the International Society of Nephrology. Additional support was from U.S. Public Health grants DK-52121, DK-43422, and DK-47695 and a small grant from SCIOS (Sunnyvale, CA, USA), Samsung Biomedical Research Institute (C-A0-018-1), and Samsung Medical Center. Reprint requests to Richard J. Johnson, M.D., Chief, Renal Section, Baylor College of Medicine, 6650 Fannin, Suite 1273, Houston, Texas 77030, USA. E-mail: [email protected]

REFERENCES 1. Remuzzi G, Ruggenenti P: The hemolytic uremic syndrome. Kidney Int 47:2–19, 1995 2. Brandt JR, Avner ED: Hemolytic uremic syndrome and thrombotic thrombocytopenic purpura, in Immunologic Renal Diseases, edited by Neilson E, Couser W, New York, Raven Press, 1996 3. Boyce T, Swerdlow D, Griffin P: Escherichia coli O157:H7 and the hemolytic uremic syndrome. N Engl J Med 333:364–368, 1995 4. Remuzzi G, Ruggenenti P: The hemolytic uremic syndrome. Kidney Int 53(Suppl 66):S54–S57, 1998 5. Siegler RL, Milligan MK, Burningham TH, Christofferson RD, Chang SY, Jorde LB: Long-term outcome and prognostic indicators in the hemolytic-uremic syndrome. J Pediatr 118:195– 200, 1991 6. Siegler RL, Pavia AT, Christofferson RD, Milligan MK: A 20-year population-based study of postdiarrheal hemolytic uremic syndrome in Utah. Pediatrics 94:35, 1994 7. Fitzpatrick M, Shah V, Trompeter RS, Dillon MJ, Barratt TM: Long-term renal outcome of childhood haemolytic uraemic syndrome. BMJ 303:489–492, 1991 8. Gagnadoux MF, Habib R, Gubler MC, Bacri JL, Broyer M: Long-term (15-25 years) outcome of childhood hemolytic-uremic syndrome. Clin Nephrol 46:39–41, 1996 9. Nangaku M, Alpers CE, Pippin J, Shankland SJ, Adler S, Kurokawa K, Couser WG, Johnson RJ: A new model of renal microvascular endothelial injury. Kidney Int 52:182–194, 1997 10. Ferrara N: Role of vascular endothelial growth factor in the regulation of angiogenesis. Kidney Int 56:794–814, 1999 11. Takeshita S, Zheng LP, Brogi E, Kearney M, Pu LQ, Bunting S, Ferrara N, Symes JF, Isner JM: Therapeutic angiogenesis: A single intraarterial bolus of vascular endothelial growth factor augments revascularization in a rabbit ischemic hind limb model. J Clin Invest 93:662–670, 1994 12. Harada K, Friedman M, Lopez JJ, Wang SY, Li J, Prasad PV, Pearlman JD, Edelman ER, Sellke FW, Simons M: Vascular endothelial growth factor administration in chronic myocardial ischemia. Am J Physiol 270:H1791–H1802, 1996 13. Isner JM, Walsh K, Symes J, Pieczek A, Takeshita S, Lowry J, Rosenfield K, Weir L, Brogi E, Jurayj D: Arterial gene transfer for therapeutic angiogenesis in patients with peripheral artery disease. Hum Gene Ther 7:959–988, 1996 14. Baumgartner I, Pieczek A, Manor O, Blair R, Kearney M, Walsh K, Isner JM: Constitutive expression of phVEGF165 after intramuscular gene transfer promotes collateral vessel development in patients with critical limb ischemia. Circulation 97:1114– 1123, 1998 15. Isner JM, Baumgartner I, Rauh G, Schainfeld R, Blair R, Manor O, Razvi S, Symes JF: Treatment of thromboangiitis obliterans (Buerger’s disease) by intramuscular gene transfer of vascular endothelial growth factor: Preliminary clinical results. J Vasc Surg 28:964–973, 1998 16. Banai S, Jaklitsch MT, Shou M, Lazarous DF, Scheinowitz M, Biro S, Epstein SE, Unger EF: Angiogenic-induced enhancement of collateral blood flow to ischemic myocardium by vascular endothelial growth factor in dogs. Circulation 89:2183–2189, 1994 17. Alpers CE, Hudkins KL, Pritzl P, Johnson RJ: Mechanisms of clearance of immune complexes from peritubular capillaries in the rat. Am J Pathol 139:855–867, 1991 18. Nangaku M, Alpers CE, Pippin J, Shankland SJ, Kurokawa K, Adler S, Morgan BP, Johnson RJ, Couser WG: CD59 protects glomerular endothelial cells from immune-mediated thrombotic microangiopathy in rats. J Am Soc Nephrol 9:590–597, 1998 19. Ferrara N, Davis-Smyth T: The biology of vascular endothelial growth factor. Endocr Rev 18:4–25, 1997 20. Duijvestijn AM, van Goor H, Klatter F, Majoor GD, van Bussel E, van Breda Vriesman PJ: Antibodies defining rat endothelial cells: RECA-1, a pan-endothelial cell-specific monoclonal antibody. Lab Invest 66:459–466, 1992 21. Iruela-Arispe L, Gordon K, Hugo C, Duijvestijn AM, Claffey KP, Reilly M, Couser WG, Alpers CE, Johnson RJ: Participation of glomerular endothelial cells in the capillary repair of glomerulonephritis. Am J Pathol 147:1715–1727, 1995

Kim et al: VEGF stimulates renal recovery in TMA 22. Kerjaschki D, Ojha PP, Susani M, Horvat R, Binder S, Hovorka A, Hillemanns P, Pytela R: A beta 1-integrin receptor for fibronectin in human kidney glomeruli. Am J Pathol 134:481–489, 1989 23. Breiteneder-Geleff S, Matsui K, Soleiman A, Meraner P, Poczewski H, Kalt R, Schaffner G, Kerjaschki D: Podoplanin, novel 43-kd membrane protein of glomerular epithelial cells, is down-regulated in puromycin nephrosis. Am J Pathol 151:1141– 1152, 1997 24. Lombardi D, Gordon KL, Polinsky P, Suga S, Schwartz SM, Johnson RJ: Salt-sensitive hypertension develops after short-term exposure to angiotensin II. Hypertension 33:1013–1019, 1999 25. Thomas SE, Anderson S, Gordon KL, Oyama TT, Shankland SJ, Johnson RJ: Tubulointerstitial disease in aging: Evidence for underlying peritubular capillary damage, a potential role for renal ischemia. J Am Soc Nephrol 9:231–242, 1998 26. Gerber H-P, Hillan KJ, Ryan AM, Kowalski J, Keller GA, Rangell L, Wright BD, Radtke F, Aguet M, Ferrara N: VEGF is required for growth and survival of neonatal mice. Development 126:1149–1159, 1999 27. Hughes J, Nangaku M, Alpers CE, Shankland SJ, Couser WG, Johnson RJ: The C5b-9 membrane attack complex mediates endothelial cell apoptosis in experimental glomerulonephritis. Am J Physiol 278:F747–F757, 2000

2399

28. Epstein FH, Agmon Y, Brezis M: Physiology of renal hypoxia. Ann NY Acad Sci 718:72–81, 1994 29. Brezis M, Rosen S: Hypoxia of the renal medulla: Its implications for disease. N Engl J Med 332:647–655, 1995 30. Alon T, Hemo I, Itin A, Pe’er J, Stone J, Keshet E: Vascular endothelial growth factor acts as a survival factor for newly formed retinal vessels and has implications for retinopathy of prematurity. Nat Med 1:1024–1028, 1995 31. Yuan F, Chen Y, Dellian M, Safabakhsh N, Ferrara N, Jain RK: Time-dependent vascular regression and permeability changes in established human tumor xenografts induced by an anti-vascular endothelial growth factor/vascular permeability factor antibody. Proc Natl Acad Sci USA 93:14765–14770, 1996 32. Fine LG, Orphanides C, Norman JT: Progressive renal disease: The chronic hypoxia hypothesis. Kidney Int 53(Suppl 65):S74–S78, 1998 33. Papapetropoulos A, Garcia-Cardena G, Madri JA, Sessa WC: Nitric oxide production contributes to the angiogenic properties of vascular endothelial growth factor in human endothelial cells. J Clin Invest 100:3131–3139, 1997 34. Ziche M, Morbidelli L, Choudhuri R, Zhang HT, Donnini S, Granger HJ, Bicknell R: Nitric oxide synthase lies downstream from vascular endothelial growth factor-induced but not basic fibroblast growth factor-induced angiogenesis. J Clin Invest 99:2625– 2634, 1997