J Am Soc Nephrol 14: 1254–1271, 2003 Urokinase Receptor Deficiency Accelerates Renal Fibrosis in Obstructive Nephropathy

GUOQIANG ZHANG,* HEUNGSOO KIM,* XIAOHE CAI,* JESU´ SM.LO´ PEZ-GUISA,* CHARLES E. ALPERS,† YOUHUA LIU,† PETER CARMELIET,§ and ALLISON A. EDDY* *University of Washington and Children’s Hospital and Regional Medical Center, Division of Nephrology, Seattle, Washington; †Department of Pathology, University of Washington, Seattle, Washington; ‡Department of Pathology, University of Pittsburgh, Pittsburgh, Pennsylvania; and §The Center For Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, Leuven, Belgium.

Abstract. The urokinase cellular receptor (uPAR) recognizes wild-type animals after UUO (62 Ϯ 20 versus 135 Ϯ 13 units the N-terminal growth factor domain of urokinase-type plas- at day 3 UUO; 74 Ϯ 17 versus 141 Ϯ 16 at day 7 UUO; 98 Ϯ minogen activator (uPA) and is expressed by several cell types. 20 versus 165 Ϯ 10 at day 14 UUO; Ϫ/Ϫ versus ϩ/ϩ). In The present study was designed to test the hypothesis that contrast, renal expression of several genes that regulate plas- uPAR regulates the renal fibrogenic response to chronic injury. min activity were similar in both genotypes, including uPA, Groups of uPAR wild-type (ϩ/ϩ) and deficient (Ϫ/Ϫ) mice tPA, PAI-1, protease nexin-1, and ␣2-antiplasmin. Worse renal were investigated between 3 and 14 d after unilateral ureteral fibrosis in the uPARϪ/Ϫ mice appears to be TGF-␤-indepen- obstruction (UUO) or sham surgery. Not detected in normal dent, as TGF-␤ activity was actually reduced by 65% in the kidneys, uPAR mRNA was expressed in response to UUO in Ϫ/Ϫ mice despite similar renal TGF-␤1 mRNA levels. Signif- the ϩ/ϩ mice. By in situ hybridization, uPAR mRNA tran- icantly lower levels of the major 2.3-kb transcript and the scripts were detected in renal tubules and interstitial cells of the 69-kd active protein of hepatocyte growth factor (HGF), a obstructed uPARϩ/ϩ kidneys. The severity of renal fibrosis, known anti-fibrotic growth factor, in the uPARϪ/Ϫ mice based on the measurement of total collagen (13.5 Ϯ 1.5 versus suggests a potential link between HGF and the renoprotective 9.8 Ϯ 1.0 ␮g/mg kidney on day 14; Ϫ/Ϫ versus ϩ/ϩ) and effects of uPAR. These data suggest that renal uPAR attenuates interstitial area stained by Masson trichrome (22 Ϯ 4% versus the fibrogenic response to renal injury, an outcome that is 14 Ϯ 3% on day 14; Ϫ/Ϫ versus ϩ/ϩ) was significantly mediated in part by urokinase-dependent but plasminogen- greater in the uPARϪ/Ϫ mice. In the absence of uPAR, renal independent functions. uPA activity was significantly decreased compared with the

Progressive renal disease, characterized histologically by tubu- ogen activator/plasmin enzymes also have direct and indirect lar atrophy and the accumulation of extracellular matrix pro- extracellular matrix-degrading actions (5,6). Somewhat para- teins in the renal interstitium, is associated with declining renal doxically, plasmin also activates the latent fibrogenic cytokine function (1,2). In addition to increased matrix protein synthe- TGF-␤ in vitro, but whether this is a significant in vivo action sis, impaired degradation of interstitial matrix proteins also of plasmin is unclear (7). The rate-limiting step for plasmin appears to play an important role in renal fibrogenesis (3). generation is the activity of tissue-type (tPA) and/or urokinase- Although the mechanisms regulating matrix turnover within type (uPA) plasminogen activators. PA activity is tightly reg- the renal interstitium are poorly understood, recent studies ulated at several levels, including pro-enzyme synthesis, ex- suggest that the plasmin cascade plays a significant role (4). In tracellular activation of the latent enzymes, and their inhibition addition to their well-known fibrinolytic activity, the plasmin- by specific inhibitors such as plasminogen activator inhibitor-1 (PAI-1), PAI-2, and possibly protease nexin-1. Within normal kidneys, high levels of uPA are synthesized by renal tubules. Received August 12, 2002. Accepted January 31, 2003. Generally undetected in normal kidneys, PAI-1 is expressed de Dr. Eric Rondeau served as Guest Editor and supervised the review and final novo during the active phase of renal fibrosis (reviewed in disposition of this manuscript. reference 8). Although mice with the PAI-1 null mutation Correspondance to Dr. Allison A. Eddy, Children’s Hospital and Regional develop less severe renal (9) and pulmonary fibrosis (10) than Medical Center, Division of Nephrology, Mail Stop 5G-1, 4800 Sand Point Way NE, Seattle, WA 98105. Phone: 206-987-2524; Fax: 206-987-2636; do wild-type mice, it remains unclear whether the blunted E-mail: [email protected] fibrogenic response is entirely due to inhibition of plasminogen 1046-6673/1405-1254 activator/plasmin proteolytic activity or also related to the Journal of the American Society of Nephrology pro-inflammatory and pro-angiogenic effects of PAI-1. Copyright © 2003 by the American Society of Nephrology A cellular receptor has been identified for the N-terminal DOI: 10.1097/01.ASN.0000064292.37793.FB growth factor domain of uPA (uPAR), also known as CD87 J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1255

(11,12). It is expressed by cells of several lineages, including half kidney were frozen at Ϫ80°C for total collagen measurement lymphohematopoietic cells (monocytes, neutrophils, and acti- (approximately one fourth) and for the total RNA and protein extrac- vated T cells), resident kidney cells (glomerular and tubular tion (approximately three fourths). epithelial cells and mesangial cells), endothelial cells, fibro- blasts, and myofibroblasts (5,13–18). This highly glycosylated Analysis of Tubulointerstitial Fibrosis 50-kD to 65-kD protein is linked to the plasma membrane by Total renal collagen was measured biochemically as described glycosylphosphatidylinositol (GPI). The uPAR binds to both previously (26). In brief, an accurately weighed portion of the kidney the latent and active form of uPA. Once receptor-bound, the was homogenized in distilled water, hydrolyzed in 10N HCl and latent enzyme can be activated while the active enzyme retains incubated at 110°C for 18 h. The hydrolysate was dried by speed vacuum centrifugation and redissolved in buffer (25 g of citric acid, 6 its enzymatic activity. The inhibitor PAI-1 may also bind to the ml of glacial acetic acid, 60 g of sodium acetate, and sodium hydrox- receptor-bound enzyme, an interaction that promotes internal- ide [17 g in 500 ml], pH 6.0). Total hydroxyproline in the hydrolysate ization and degradation of uPA and PAI-1 (5,19). Soluble was determined according to the chemical method of Kivirikko et al. forms of uPAR also exist, generated by proteolytic cleavage of (27). Total collagen in the tissue was calculated on the assumption that the transmembrane domain (20). collagen contains 12.7% hydroxyproline by weight. Final results were It is now evident that uPAR is a multifunctional receptor that expressed as ␮g/mg kidney wet weight. is involved not only in cell-surface uPA activity and plasmin Kidney sections (day 14 UUO and sham) were stained with Masson generation, but also in mediating protease-independent effects, trichrome (Sigma, St. Louis, MO), and the percent aniline blue- including cell adhesion and migration and outside-in signaling stained tubulointerstitial area was measured using a point-counting (12,21). Less clear is the role of receptor-bound uPA and method (28). subsequent plasmin generation in the context of tissue remod- eling. In particular, little is known about the expression and Immunohistology function of uPAR during renal fibrosis. Immunohistochemical Immunohistochemical studies were performed on 4-␮m paraffin- studies have reported uPAR expression on normal human embedded renal sections. Primary antibodies used were rabbit anti- tubules in one study (18) but not in another (16). Increased human LDL receptor-related protein (LRP) provided by Dr. D.K. Strickland, American Red Cross, Rockville, MD (29), rabbit anti- kidney uPAR expression has been described in several renal mouse urokinase (American Diagnostica Inc., Greenwich, CT), rabbit disease states such as endotoxemia (15,22), acute tubular ne- anti-human TGF-␤ (Santa Cruz Biotechnology, Santa Cruz, CA), and crosis (16), thrombotic microangiopathy (16), nephrotoxic se- mouse anti-human HGF (30). Immunoperoxidase staining was per- rum nephritis (23), pyelonephritis (22), and chronic allograft formed using the ABC ELITE kit (Vector Laboratories Inc., Burlin- rejection (24). The present study was designed to investigate game, CA). Sections stained with the secondary antibodies alone were the functional role of uPAR in the renal fibrogenic response to negative. For HGF staining, the primary and secondary antibodies sustained injury by investigating the response to ureteral ob- were pre-complexed before incubation with the tissue sections to struction in uPAR-deficient mice compared with wild-type minimize crossreactivity (31). mice of the same genetic background. To evaluate the extent of tubular injury, sections were stained with biotinylated Phaseolus vulgaris agglutinin-E (PHA-E) (Vector Labo- ratories Inc.), a lectin that binds to proximal tubular brush border (32) Materials and Methods or with a mouse anti-kidney-specific cadherin (Ksp) monoclonal Animals and Experimental Design antibody (Zymed Laboratories Inc, San Francisco, CA), which is UPARϪ/Ϫ and ϩ/ϩ mice on a C57Bl6 background were bred in expressed on the basolateral membrane of collecting ducts (33). our animal facility and allowed to grow to a minimum weight of 20 g PHA-E staining was used to grade tubular injury on a scale of 1 to 4 before the initiation of the study (25). The genotypes of the mice were as described by Mizuno et al. (34). Loss of Ksp-cadherin expression confirmed by Southern blot analysis of DNA extracted from tails. was determined as a measure of early tubular injury and expressed as Groups of weight-matched and gender-matched uPAR-deficient and % positive tubules. wild-type mice were studied 3, 7, and 14 d after unilateral ureteral ϭ obstruction (UUO) and 7 d after sham surgery (n 8 per group). In Situ Hybridization UUO surgery was performed under general anesthesia. The left ureter uPAR in situ hybridization was performed as described previously was ligated with 4.0 silk at two separate points in the UUO groups. All using a 35S-labeled riboprobe prepared using the Riboprobe Combi- mice were sacrificed by exsanguination under general anesthesia. All nation System-T3/T7 RNA Polymerase kit (Promega Corp., Madison, procedures were performed in compliance with the guidelines estab- WI) (35). In brief, formalin-fixed, paraffin-embedded 4-␮m kidney lished by National Research Council Guide for the Care and Use of tissue sections were deparaffinized and rehydrated using standard Laboratory Animals. procedures. Sections were washed with 0.5ϫ SCC and digested with 10 ␮g/ml proteinase K (Sigma). Prehybridization was performed for Kidney Tissue Preparation 2 h by adding 50 ␮l of prehybridization buffer (0.3 M NaCl, 20 mM After exsanguination, the left kidney was procured, the capsule Tris, pH 8.0, 5 mM ethylenediaminetetraacetic acid, 1ϫ Denhardt removed, and the kidney weighed. The kidney was divided longitu- solution, 10% dextran sulfate, 10 mM dithiothreitol, and 50 ␮g/ml dinally and subdivided. One half was prepared for histologic studies: yeast tRNA). Hybridization was initiated by adding 500,000 cpm of one piece (approximately one third) was fixed in 10% buffered the 35S-labeled uPAR riboprobe (sense or anti-sense) in 50 ␮lof formalin and paraffin-embedded; the remaining piece (approximately prehybridization buffer and incubated overnight at 50°C. Sections two thirds) was imbedded in Tissue-Tek OCT compound (Sakura were treated with RNase A (20 ␮g/ml; Sigma) followed by three Finetek, Torrence, CA) and snap-frozen. Sections from the second high-stringency washes in 0.1ϫ SSC/0.5% Tween 20 (Sigma) for 40 1256 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

min at 50°C and several 2ϫ SSC washes. After the tissue was (Sigma) as substrate. Molecular markers and human MMP-2 and dehydrated and air-dried, it was dipped in NTB2 emulsion, (Kodak, MMP-9 standards (Chemicon International Inc., Temecula, CA) were Rochester, NY) and developed in the dark for 10 wk. After develop- also loaded into the outer wells. After protein separation by electro- ment, the sections were counterstained with hematoxylin and eosin, phoresis, the gel was rinsed in 2.5% Triton X-100 at room temperature dehydrated, and cover-slipped. with gentle shaking for 30 min. After incubation for 17 to 20 h at 37°C

in a solution containing 50 mM Tris and 10 mM CaCl2, pH 7.8, the Northern Blot Analysis gel was stained with 0.002% Coomassie blue and photographed. The size of each lytic band was measured using the NIH image analysis Total kidney RNA was isolated by the phenol/guanidine isothio- program. cyanate extraction method using TRIzol-BRL reagent (Life Technol- Casein plasminogen zymography was performed to evaluate renal ␮ ogies). Total kidney RNA (18 g) from each experimental animal was uPA and tPA activity using the methods of Roche et al. (48) with loaded into individual wells and separated by 1.0% agarose formal- minor modifications as described previously (9). The procedure used dehyde gel electrophoresis. A photomicrograph of the ethidium bro- was similar to gelatin zymography, except that the zymography gel mide-stained gel was obtained to determine RNA loading equality. was made of 10% SDS-polyacrylamide containing 2 mg/ml ␣-casein The RNA was transferred to a hybridization membrane (GeneScreen and 10 ␮g/ml plasminogen (Sigma). Molecular weight markers and Plus; New England Nuclear Life Science Products, Boston, MA) and human urokinase standards (Calbiochem Co., San Diego, CA) were ultraviolet light crosslinked (UV Crosslinker; Hoeffer Scientific In- loaded into the outer wells. PA-specific bands were verified by their struments, San Francisco, CA). Complementary DNA probes were disappearance when re-run in an identical gel that lacked radiolabeled with 32P dCTP (3000 Ci/mmol) by random priming with 7 plasminogen. T Quick Prime kit (Pharmacia Biotech, Piscataway, NJ). The blots Total kidney plasmin activity was measured using a plasmin- were hybridized with the radiolabeled cDNA probes using the Quick- specific chromogenic substrate, Chromozym PL (Boehringer Mann- Hyb hybridization buffer (Stratagene, La Jolla, CA). Autoradiographs heim, Indianapolis, IN) as described previously (9). were obtained and the density of each band quantified using the NIH Image program. The 18-s ribosomal bands in the ethidium bromide- TGF-␤ Bioactivity stained gels were used to adjust for RNA loading equality as described TGF-␤ bioactivity was measured in kidney protein extracts using a previously (36). TGF-␤–responsive mink lung epithelial cell (MLEC) line (a generous The cDNA probes used were murine uPAR1 (from Dr. Niels gift from Dr. Daniel B. Rifkin, New York University Medical Center, Behrendt, Finsen Laboratory, Copenhagen, Denmark) (37), rat fi- New York, NY) (49). This cell line was generated by fusing a bronectin lambda-rlf-1 (from Dr. R. Hynes, Center for Cancer Re- truncated TGF-␤-inducible PAI-1 promoter to a firefly luciferase search, Massachusetts Institute of Technology, Cambridge, MA) (38), reporter gene and transfected into MLEC. This bioassay is specific for ␣ mouse 1(I) procollagen (from Dr. S. Thorgeirsson, National Cancer active TGF-␤ with a detection limit of approximately 5 to 10 pg/ml. ␤ Institute, Bethesda, MD) (39), rat TGF- 1 (from Dr. S.W. Qian, Kidney protein samples (15 ␮g) diluted in serum-free media (tripli- National Cancer Institute) (40), mouse hepatocyte growth factor (41), cate samples from three individual animals per group) were added rat PAI-1 (from Dr. T.D. Gelehrter, University of Michigan, Ann directly to monolayers of confluent MLEC cultures. A standard curve Arbor, MI) (42), rat uPA (from Dr. J. Degen, Children’s Hospital of TGF-␤1 activity was generated using serial dilutions of recombi- Research Foundation, University of Cincinnati, Cincinnati, OH), nant human active TGF-␤1 (ED ϭ 0.05 to 0.1 ng/ml; Amersham mouse tPA (from Dr. D.S. Strickland, The Rockefeller University, 50 Pharmacia). After overnight incubation at 37°Cin5%CO2, cells were New York, NY) (43), mouse plasminogen (American Tissue Culture harvested and luciferase activity measured using the Enhanced Lucif- Collection, Rockville, MD), mouse protease nexin-1 (from Dr. Ve- erase Assay Kit (Pharmingen, San Diego, CA). jsada, University of Geneva, Geneva, Switzerland) (44), and mouse ␣2-antiplasmin (from Dr. A. Sappino, University of Geneva, Geneva, HGF Western Blotting Switzerland) (45). Pro-HGF and the active form of the HGF protein (␣ chain) were detected by Western blot analysis as described by Grenier et al. (50). Protease Activity Protein samples (40 ␮g) were separated by 10% SDS-PAGE under Protein was isolated from kidney tissue that had been stored at reducing conditions. The proteins were transferred to a nitrocellulose Ϫ80°C. Pieces were individually ground into a fine powder under membrane, and the immunoreactive protein was visualized using ECL liquid nitrogen conditions, using a mortar and pestle that had been enhanced chemiluminescence (Amersham Pharmacia Biotech Inc., prechilled with dry ice. For gelatin zymography, the powder was Piscataway, NJ). The primary antibody was monoclonal anti-human HGF (clone 8), known to crossreact with rat HGF (30); the secondary homogenized in extraction buffer (0.05 M Tris, 0.01 M CaCl2, 2.0 M guanidine HCl, 0.2% Triton X-100, pH 7.5) and dialyzed using antibody was HRP-conjugated goat anti-mouse IgG (Sigma). Protein dialysis membrane Spectra/PorR 1 (Spectrum Medical Industries, Inc., loading equality was determined by amido black staining. Houston, TX) against 0.05 M Tris, 0.2% Triton X-100, pH 7.5, for 48hat4°C. For casein plasminogen zymography, the powder was Statistical Analyses Ϯ mixed with homogenizing buffer (50 mmol/L Tris, pH 7.6, 1% SDS). All results were expressed as mean 1 SD. Results were analyzed Ͻ Individual samples were then centrifuged for 5 min (14,000 ϫ g), and by the Mann Whitney U test. A P value 0.05 was considered the protein concentration was measured in the supernatants using the statistically significant. Bradford protein assay kit (Bio-Rad, Hercules, CA). Samples were aliquoted and stored at Ϫ80°C for zymographic studies. Results MMP-9 and MMP-2 activity were measured by gelatin zymogra- Renal Expression of uPAR and LDL-Receptor phy according to the method reported by Kenagy et al. (46,47). In Related Protein brief, protein samples (10 ␮g/well) were loaded without heating onto Renal uPAR mRNA was undetectable by Northern blotting a 7% polyacrylamide gel containing 1 mg/ml porcine skin gelatin in sham-operated kidneys and remained undetectable in all J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1257

UUO kidneys from uPARϪ/Ϫ mice. UPAR mRNA was ex- (uPARϩ/ϩ and Ϫ/Ϫ, respectively; Figure 4). The increase in pressed in UUO kidneys from the uPARϩ/ϩ mice animals kidney collagen was significantly higher in the uPARϪ/Ϫ after3dofobstruction (Figure 1). In situ hybridization studies mice: 7.6 Ϯ 0.9 compared with 6.3 Ϯ 0.6 ␮g/mg kidney at day on kidneys after 7 and 14 d of obstruction identified uPAR 7 UUO and 13.5 Ϯ 1.5 versus 9.8 Ϯ 1.0 ␮g/mg kidney at day transcripts in uPARϩ/ϩ kidneys, especially in tubular and 14 UUO (uPARϪ/Ϫ versus uPARϩ/ϩ; P Ͻ 0.05). Evaluation interstitial cells (Figure 2). Expression of LRP by interstitial of the interstitial area stained blue with Masson trichrome also cells, an endocytosis receptor that functions as an uPAR co- demonstrated that interstitial fibrosis was significantly more receptor, was increased to a greater extent in the uPARϩ/ϩ extensive in the uPARϪ/Ϫ kidneys (Figure 4). Tubular injury mice than the uPARϪ/Ϫ mice after ureteral obstruction (Fig- was more severe in the uPARϪ/Ϫ mice. The number of ure 3). Ksp-cadherin–positive collecting ducts was significantly re- duced in the uPARϪ/Ϫ group within 3 d after UUO (Figure 5), Severity of Tubulointerstitial Fibrosis and they were undetectable in both genotypes by 14 d. Using In response to UUO, total kidney collagen increased with PHA-E lectin staining to evaluate late proximal tubular dam- time in the mice of both genotypes, reaching a 2.6-fold to age, even after 14 d of obstruction the uPARϩ/ϩ mice had less 3.6-fold increase relative to the sham kidneys by day 14 UUO extensive tubular destruction (Figure 5).

Figure 1. Renal urokinase cellular receptor (uPAR) gene expression. Kidney total RNA from individual mice (18 ␮g each) was separated and transferred to a nylon blot. The membrane was probed with 32P-dCTP–labeled mouse uPAR1 cDNA and exposed for 120 h. A 1.5-kb uPAR mRNA band was detected in kidneys of uPARϩ/ϩ mice after unilateral ureteral obstruction (UUO). 1258 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 2. uPAR in situ hybridization. Using 35S-labeled uPAR anti-sense and sense probes, the site of uPAR gene expression was determined after UUO surgery. Anti-sense probes failed to produce positive signals with sham uPARϩ/ϩ (A) and uPARϪ/Ϫ (B) kidneys and uPARϪ/Ϫ obstructed kidneys (D). After7dofUUO, several tubular cells (examples indicated with arrows) and interstitial cells (highlighted with figure insert) were positive, as indicated by the black silver grains (C). Low levels of background activity are indicated by hybridization of obstructed uPARϩ/ϩ (E) and uPARϪ/Ϫ (F) kidneys with uPAR sense probes. Magnification: ϫ400.

Renal Expression of PA and PA Inhibitor Genes in a similar pattern in both genotypes, including enhanced Renal expression of plasminogen activator genes was increased expression in dilated atrophic tubules found within fibrotic loci to a similar extent in the mice of both genotypes in response to (Figure 8). Renal plasmin activity did not differ between UUO (twofold and threefold increase in uPA and tPA, respec- uPARϩ/ϩ and Ϫ/Ϫ mice: sham: 2.7 Ϯ 0.8 versus 2.4 Ϯ 0.5; tively, at day 7 UUO; Figures 6 and 7). Renal mRNA levels of the day 3 UUO: 3.2 Ϯ 2.4 versus 5.3 Ϯ 1.0; day 14: 2.0 Ϯ 1.5 two renal serine protease inhibitors, PAI-1 and protease nexin-1, versus 3.1 Ϯ 2.1 ϫ 10Ϫ4 units/ml, respectively. were also strongly induced in response to UUO and to a similar Gelatin gel zymography demonstrated that MMP-9 was the level in uPARϩ/ϩ and Ϫ/Ϫ mice (Figures 6 and 7). In contrast, most abundant gelatinase in the kidney. After UUO surgery, renal mRNA levels for the plasmin inhibitor ␣2-antiplasmin were MMP-9 activity decreased compared with the sham kidneys; significantly downregulated in mice of both genotypes in response MMP-9 activity was significantly lower in the uPARϪ/Ϫ mice to UUO (Figures 6 and 7). Renal plasminogen mRNA was un- compared with uPARϩ/ϩ mice on days 3 and 14 (Figure 9). In detectable in all kidneys. contrast, renal MMP-2 activity increased in response to UUO without a consistent difference between the two genotypes: Renal Protease Activity MMP-2 levels slightly lower (nonsignificant) on days 3 and 7 but As illustrated in casein plasminogen gel zymogram (Figure higher (significant) by day 14 in the uPARϪ/Ϫ mice (Figure 9). 8), uPA is the predominant plasminogen activator detected in the kidney after UUO. The activity of both uPA and tPA TGF-␤1 increased with time after UUO, but both were significantly Renal TGF-␤1 mRNA was strongly induced in UUO kid- lower in the uPARϪ/Ϫ mice compared with uPARϩ/ϩ mice. neys from mice of both genotypes after 3, 7, 14 d of obstruc- By immunostaining, uPA protein was detected in renal tubules tion: eightfold, fivefold, and fourfold in the uPARϩ/ϩ mice J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1259

Figure 3. Renal expression of LDL receptor-related protein (LRP) by interstitial cells detected by immunohistochemical staining. In comparison with sham-operated uPARϩ/ϩ (A) and uPARϪ/Ϫ (B) kidneys, the number of LRP-positive interstitial cells increased after7dofureteral obstruction, although to a greater extent in the uPARϩ/ϩ kidneys (C) compared with the uPARϪ/Ϫ kidneys (D). Magnification: ϫ400. and sixfold, sevenfold, and fivefold in the uPARϪ/Ϫ mice at Extracellular Matrix Genes 3, 7, and 14 d, respectively (nonsignificant differences between Renal procollagen ␣1(I) and fibronectin mRNA level was the genotypes; Figure 10). Immunohistochemical staining of significantly elevated in mice of both genotypes after UUO the obstructed kidneys identified TGF-␤1 protein primarily in compared with sham-operated groups (Figure 12). As shown, renal tubules and occasional interstitial cells (Figure 10). In no differences were noted between the uPARϪ/Ϫ and contrast to the mRNA levels, TGF-␤1 bioactivity was actually uPARϩ/ϩ mice on day 14 after UUO. Although not shown, significantly lower in the uPARϪ/Ϫ kidneys compared with matrix gene expression was also similar on days 3 and 7. the uPARϩ/ϩ kidneys after7dofobstruction (Figure 10). Discussion Hepatocyte Growth Factor The uPAR is a multifunctional cellular receptor that is In vitro studies have reported that urokinase activates the involved in such diverse biologic processes as angiogenesis, anti-fibrotic growth factor HGF (51). Seven days after UUO inflammation, wound repair, and tumor metastasis (12,21). The surgery, all four HGF transcripts (6.0, 3.1, 2.3, and 1.5 kb) results of the present study identify an important modulating were significantly increased; levels of the major 2.3-kb tran- role for uPAR during the renal fibrogenic response that devel- script were significantly lower in the uPARϪ/Ϫ mice (Figure ops in response to chronic injury caused by ureteral obstruc- 11). By immunohistochemistry, the HGF protein was detected tion. Although there are now several reports of increased uPAR within the interstitium. Western blot analysis (day 7) showed expression in damaged kidneys, the functional significance of significantly lower levels of active HGF (␣ chain) in the these changes remains unclear. In the present study, renal uPARϪ/Ϫ mice despite similar amounts of pro-HGF protein in uPAR was significantly increased after ureteral obstruction in both genotypes (Figure 11). By day 14, HGF protein could not wild-type mice, while the absence of uPAR in the uPAR null be detected by Western blotting. mice was associated with a more aggressive fibrogenic re- 1260 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 4. Renal fibrosis severity. The mean total kidney collagen content expressed as ␮g/mg wet kidney weight was significantly higher in uPARϪ/Ϫ mice (dotted line) than in uPARϩ/ϩ mice (solid line) by day 7 and day 14 of UUO (A). The interstitial area stained blue by Masson trichrome (day 14) was significantly greater in the uPARϪ/Ϫ mice (B). Light photomicrographs of Masson trichrome-stained uPARϩ/ϩ (C) and uPARϪ/Ϫ (D) sham kidneys and uPARϩ/ϩ (E) and uPARϪ/Ϫ (F) obstructed kidneys illustrate the difference in the extent of interstitial fibrosis between the two genotypes in response to ureteral obstruction. Magnification: ϫ400. Values in the graphs (A and B) represent the mean Ϯ 1SD(n ϭ 6 per group; *P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice). sponse. The ability of uPAR to attenuate renal fibrosis appears after receptor binding, pro-urokinase can be activated while the to be related, at least in part, to enhanced activity of proteases active enzyme is stabilized. The specific mechanism(s) of that degrade extracellular matrix proteins, including uPA, tPA, pro-urokinase activation within the kidney is unknown; several and possibly MMP-9. enzymes can achieve this function in vitro, including plasmin, The kidney is the major site of production of the uPAR kallikrein, cathepsins, matrix metalloproteinase-3, and Pump-1 ligand, uPA. Other known ligands are kininogen and vitronec- metalloproteinase (53). In the present study, the renal expres- tin (12,52). In response to UUO, renal pro-urokinase mRNA sion of uPAR in the wild-type mice after UUO was associated levels were significantly increased in mice of both genotypes with significantly higher levels of uPA activity than was ob- but uPA activity was significantly higher in the uPARϩ/ϩ served in the kidneys lacking uPAR. animals, consistent with the known ability of uPAR to stabilize Urokinase initiates several extracellular events that could uPA activity at cellular surfaces. uPAR binds both the single accelerate the rate of degradation of matrix proteins. It has chain pro-urokinase and the two-chain active enzyme. Even limited ability to directly degrade some matrix proteins such as J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1261

Figure 5. Tubular damage severity. Immunohistochemical staining with the biotinylated lectin Phaseolus vulgaris agglutinin-E (PHA-E) was used to grade the severity of proximal tubular damage (A through F). Binding of PHA-E was limited to the proximal tubular brush border of normal sham kidneys (C and F; uPARϩ/ϩ). After 14 d of ureteral obstruction, proximal tubular damage was evident as characterized by brush border loss, epithelial cell simplification, dilation of tubular lumina often with cast formation, and tubular atrophy (A, B, D, and E). Compared with uPARϩ/ϩ mice (A and D), proximal tubular damage was more extensive in the uPARϪ/Ϫ mice (B and E). The results are shown graphically, expressed as the mean score Ϯ 1 SD; * P Ͻ 0.05 (M). The Ksp-cadherin is detected on collecting ducts of normal sham kidneys by immunohistochemical staining (G and J; uPARϩ/ϩ and uPARϪ/Ϫ, respectively). Three days after UUO, collecting ducts were ectatic and epithelial injury was evident by loss of Ksp-cadherin expression (H and K; uPARϩ/ϩ and uPARϪ/Ϫ, respectively). Expressed as percent positive tubules Ϯ 1 SD, the loss of Ksp-cadherin was more extensive in the uPARϪ/Ϫ mice (open bars) compared with the UPARϩ/ϩ mice (solid bars). After7dofUUO, very few Ksp-1–positive tubule were detected in mice of both genotypes (I and L; uPARϩ/ϩ and uPARϪ/Ϫ, respectively); * P Ͻ 0.05, UUO versus sham of the same genotype; ϩ P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice (N). Magnification: ϫ100 in A through C; ϫ400 in D through L. 1262 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 6. Renal expression of PA and PA inhibitor genes. Representative Northern blot from day 7 studies illustrates specific bands for plasminogen activators uPA (A) and tPA (B), and the inhibitors PAI-1 (C), protease nexin-1 (D), ␣2-antiplasmin (E). The changes in response to UUO were similar between the uPARϩ/ϩ and Ϫ/Ϫ genotypes except for ␣2-antiplasmin levels that were downregulated to a lesser extent in the uPARϪ/Ϫ mice (P Ͻ 0.05). The bands of 18S rRNA in the ethidium bromide stained gel showed the equality of RNA loading.

fibronectin (54). It also activates certain latent matrix metallo- extent of renal fibrosis. Whether alterations in MMP-1 or proteinases, including membrane type (MT)-MMP-1 and MT- MMP-3 activity are involved remains speculative, as we have MMP-2 (55). However, it is the ability of uPA to catalyze the not been able to detect MMP-1 or MMP-3 activity using conversion of plasminogen to plasmin and the subsequent kidney protein extracts and casein and gelatin zymography. plasmin-dependent activation of metalloproteinases, especially Recent fibrosis studies in plasminogen-deficient mice have interstitial collagenase (MMP-1) and stromelysin-1 (MMP-3), yielded conflicting data and suggest that the plasmin-depen- that has been considered to be the primary pathway whereby dent effects could be tissue-specific. Accelerated fibrosis has uPA promotes the degradation of extracellular matrix (56). been reported in the model of bleomycin-induced pulmonary Significantly less MMP-9 activity, the predominant renal ge- fibrosis in mice genetically deficient in plasminogen (57), latinase, was detected in the uPARϪ/Ϫ kidneys 3 and 14 d while renal fibrosis induced by UUO was not accentuated in after UUO, which may be one factor contributing to the greater the absence of plasminogen (58). By contrast, in acute glomer- J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1263

Figure 7. Renal expression of PA and PA inhibitor genes. Bar graphs show the results of the quantitative densitometric analysis of the Northern blot studies, as illustrated in Figure 5 for day 7. Shaded bars represent the uPARϩ/ϩ groups; open bars are the uPARϪ/Ϫ groups. Results are means Ϯ 1 SD expressed in arbitrary densitometric units, with one unit representing the mean value obtained for the sham uPARϩ/ϩ kidneys, except for PAI-1, where gene expression was not detected in the sham kidneys. * P Ͻ 0.05, UUO versus sham of the same genotype; ϩ P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. ular injury associated with fibrin deposition, plasminogen de- higher in the wild-type mice, yet the net effect of upregulated ficiency exacerbates the disease severity (59). In the present uPAR expression was enhanced PA activity and less interstitial study, plasminogen mRNA was not detected in any of the collagen deposition. kidneys, and measured renal plasmin activity was unaltered by Urokinase has been reported to induce the expression and genotype. We cannot discount the possibility that technical activation of HGF, a growth factor with important anti-fibrotic limitations of the plasmin assay accounted for the failure to activities (51). HGF has been reported to attenuate renal fibro- detect genotype-dependent differences in renal plasmin activ- sis in several experimental models (34,60–63). In this study, ity. Alternatively, other regulators of plasmin activity may the renal response to UUO was characterized by an increase in overshadow plasminogen activator activity in this model. Par- the expression of all four HGF mRNA transcripts. The pre- adoxically, plasmin may also activate latent TGF-␤, at least in dominant 2.3-kb transcript was significantly higher in the vitro (7); whether this is a significant in vivo response is uPARϩ/ϩ than the uPARϪ/Ϫ mice. HGF protein deposited in unclear. In the present study, TGF-␤ bioactivity was actually the interstitium of the obstructed kidneys and may have con- 1264 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 8. Zymogram of renal plasminogen activator activity. Lytic bands representing renal tPA (65 kD) and uPA (40 kD) activity, as measured by casein plasminogen zymography in uPARϩ/ϩ and Ϫ/Ϫ mice (A). Semiquantitative evaluation of the uPA and tPA (B) was performed using the NIH Image program, and the results are expressed as the mean Ϯ 1SD(n ϭ 3 per group). * P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. Photomicrograph of uPA immunoperoxidase staining, illustrating numerous positive tubules in an uPARϪ/Ϫ kidney 7 days after UUO. Magnification: ϫ400.

tributed to the dampened fibrogenic response observed in the The other known uPAR ligands, high–molecular weight uPAR wild-type mice given that levels of the active HGF alpha kininogen (64) and vitronectin, are also increased in the kidney chain were higher in the obstructed wild-type kidneys. in response to ureteral obstruction. While not investigated in J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1265

Figure 9. Zymogram of renal gelatinase activity. Lytic bands represent renal MMP-9 and MMP-2 activity, as measured by gelatin zymography in uPARϩ/ϩ and Ϫ/Ϫ mice (A). Semiquantitative measurement of the size of the lytic bands was performed using the NIH Image program and expressed as arbitrary units relative to the lytic activity in the sham kidneys of the uPARϩ/ϩ mice. The results are expressed as the mean standardized activity units Ϯ 1SD(n ϭ 3 per group). Shaded bars are uPARϩ/ϩ mice; open bars are uPARϪ/Ϫ mice. * P Ͻ 0.05, uPARϩ/ϩ versus uPARϪ/Ϫ mice. this study, it is possible that uPAR-kininogen interactions also kidney model, a therapy that was associated with increased mediate renoprotective effects, as kallikrein gene therapy has urinary kinin excretion (65). High–molecular weight kininogen been reported to attenuate glomerulosclerosis in the remnant may indirectly promote the activation of uPAR-bound 1266 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 10. Renal TGF-␤1 gene expression and bioactivity. The Northern blot autoradiograph illustrates renal TGF-␤ mRNA levels 3, 7, and 14 d after UUO surgery (A through C). The lower panel below each Northern blot illustrates the ethidium bromide–stained 18S ribosomal bands that were used to adjust for any RNA loading inequality. By densitometric analysis, the mean values expressed in arbitrary units (uPARϩ/ϩ versus uPARϪ/Ϫ) were: day 3 UUO: 7.5 Ϯ 1.0 versus 8.5 Ϯ 1.4; day 7 UUO: 5.2 Ϯ 0.5 versus 5.1 Ϯ 0.4; and day 14 UUO: 3.7 Ϯ 0.2 versus 3.6 Ϯ 0.2 (P is nonsignificant at all time points). Renal TGF-␤ bioactivity, measured after 7 d of obstruction using the MLEC cell line transfected with PAI-1 promoter/luciferase gene, showed significantly greater TGF-␤ bioactivity in the uPARϩ/ϩ mice (shaded bars) than in the uPARϪ/Ϫ mice (open bars) (D). The TGF-␤ bioactivity is expressed as mean Ϯ 1 SD pg/ml of triplicate samples from three mice per group. By immunohistochemical staining, TGF-␤ protein was present in several renal tubules and occasional interstitial cells as illustrated in a uPARϪ/Ϫ kidney 7 d after UUO (E). Magnification: ϫ250. ϩ P Ͻ 0.05 compared with sham group of the same genotype; * P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1267

Figure 11. Renal hepatocyte growth factor (HGF) expression. Northern blot autoradiograph illustrates the expression of the four HGF mRNA transcripts 7 d after UUO surgery (A). Quantitative densitometric analysis (corrected for RNA loading using the 18S ethidium bromide-stained bands) found significantly higher levels of the major 2.3-kb band in the uPARϩ/ϩ kidneys (shaded bars) compared with the uPARϪ/Ϫ kidneys (open bars) (B). Immunohistochemical staining of an uPARϪ/Ϫ kidney 7 d after UUO demonstrating HGF-positive interstitial cells (C). Magnification: ϫ400. Western blot illustrating kidney levels of pro-HGF and mature HGF ␣-chain 3 and 7 d after UUO surgery (D). Quantitative analysis demonstrated lower levels of active HGF in the uPARϪ/Ϫ mice (open bars) compared with the uPARϩ/ϩ mice (shaded bars) at both time points (E). *P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. 1268 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

Figure 12. Renal matrix gene expression. Fibronectin and procollagen ␣1(I) mRNA levels measured by Northern blotting 14 d after UUO surgery and expressed as mean densitometric score Ϯ 1 SD for uPARϩ/ϩ (shaded bars) and uPARϪ/Ϫ (open bars) kidneys. Differences were not statistically significant between the genotypes. Although not shown, changes in matrix gene expression were also similar in uPARϩ/ϩ and uPARϪ/Ϫ kidneys 3 and 7 d after UUO. * P Ͻ 0.05 compared with sham group of the same genotype. prourokinase, as it serves as a prekallikrein receptor, thereby of matrix synthesis do not explain the differences. In contrast juxtaposing latent uPA and one of its activators at cellular with the findings in this study, deficiency of uPAR failed to surfaces (66). alter the degree of pulmonary interstitial fibrosis in mice In addition to its ability to amplify uPA activity, uPAR may treated with bleomycin (57). However, in that model tPA may modulate cellular behavior (12,67,68). Although uPAR itself is be a more important plasminogen activator as fibrin (the pri- attached to the plasma membrane via a GPI anchor that lacks mary tPA substrate) accumulation is significant and may play an intracellular domain, uPAR can dimerize with other cellular an important pathogenetic role, although data on the latter receptors to initiate numerous cell surface and intracellular point are conflicting (10,72–75). It may also turn out that events. Most relevant to the present study, uPAR interacts with plasmin activity is more important in the lung than it is in the the scavenger receptor LDL receptor-related protein (LRP) kidney on the basis of the initial studies using plasminogen- (69). This interaction appears to be the primary pathway for deficient mice. Nonetheless, these findings suggest that uPA- clearance of extracellular PAI-1 after it binds to uPAR-uPA based therapies may represent a novel therapeutic approach for (5). In the present study, renal interstitial cells were shown to progressive renal disease. In fact, uPA therapy administered to express LRP protein by immunostaining and the area stained animals with bleomycin-induced lung disease using adenoviral was reduced in the uPARϪ/Ϫ mice. The uPAR is also ex- gene vectors or recombinant protein, has been reported to pressed by several interstitial cells in the wild-type mice, reduce fibrosis (76,77). Although the results of uPA therapy in indicating that the uPAR-LRP scavenger receptor complex is chronic tubulointerstitial disease have not yet been reported, co-expressed by renal interstitial cells and is theoretically recombinant tPA treatment has been reported to reduce the available for PAI-1 clearance. Increased PAI-1 accumulation extent of glomerular matrix accumulation in rats with Thy-1 could explain why tPA activity, as well as uPA activity, was nephritis (78). significantly reduced in the uPARϪ/Ϫ mice. In summary, the results of the present study suggest that It is also noteworthy that LRP is an endocytosis receptor for within the kidney uPAR helps to regulate the intensity of the multiple ligands, including fibronectin (70), and has recently fibrogenic response to chronic damage, at least in part by been identified as the receptor for connective tissue growth increasing the activity of plasminogen-activating proteases. factor (71), both important components of the fibrogenic re- Whether polymorphisms in the uPAR gene (79) correlate with sponse. uPAR is also known to dimerize with members of the the risk of renal disease progression and whether therapeutic integrin superfamilies (␤1, ␤3, and ␤5), interactions that may interventions designed to enhance uPAR expression and activ- trigger intracellular signaling responses and promote cell ad- ity can be protective are questions deserving of future hesion and migration. Given that uPA and its receptor are also investigation. strongly expressed by tubular cells during UUO, it is possible that uPA-uPAR-dependent cellular signaling responses may promote tubular survival within a fibrosing environment. Al- Acknowledgments though the present study does not prove that the observed This work was funded by grant support from the National Institutes increase in plasminogen activator activity is the major reason of Health DK54500 (AAE), DK58925 (JL-G), DK47659 (CEA), DK61408 (YL), and the Northwest Kidney Foundation (JL-G). for the less aggressive fibrogenic response observed in the uPAR wild-type mice, it is noteworthy that the levels of increased expression of the genes encoding the interstitial References matrix proteins fibronectin and procollagen I were similar in 1. Eddy AA: Molecular insights into renal interstitial fibrosis. JAm the wild-type and knockout mice, suggesting that altered rates Soc Nephrol 7: 2495–2508, 1996 J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1269

2. Eddy AA: Molecular basis of renal fibrosis. Pediatr Nephrol 15: 21. Dear AE, Medcalf RL: The urokinase-type-plasminogen-activa- 290–301, 2000 tor receptor (CD87) is a pleiotropic molecule. Eur J Biochem 3. Gonza´lez-Avila G, Vadillo-Ortega F, Pe´rez-Tamayo R: Experi- 252: 185–193, 1998 mental diffuse interstitial renal fibrosis. A biochemical approach. 22. Florquin S, van den Berg JG, Olszyna DP, Claessen N, Opal SM, Lab Invest 59: 245–252, 1988 Weening JJ, van der Poll T: Release of urokinase plasminogen 4. Rerolle J-P, Hertig A, Nguyen G, Sraer J-D, Rondeau E: Plas- activator receptor during urosepsis and endotoxemia. Kidney Int minogen activator inhibitor type 1 is a potential target in renal 59: 2054–2061, 2001 fibrogenesis. Kidney Int 58: 1841–1850, 2000 23. Xu Y, Berrou J, Chen X, Fouqueray B, Callard P, Sraer JD, 5. Vassalli JD, Sappino AP, Belin D: The plasminogen activator/ Rondeau E: Induction of urokinase receptor expression in neph- plasmin system. J Clin Invest 88: 1067–1072, 1991 rotoxic nephritis. Exp Nephrol 9: 397–404, 2001 6. Mignatti P: Extracellular matrix remodeling by metalloprotein- 24. Tang WH, Friess H, di Mola FF, Schilling M, Maurer C, Graber ase and plasminogen activators. Kidney Int 47:S-12–S-14, 1995 HU, Dervenis C, Zimmermann A, Buchler MW: Activation of 7. Lyons RM, Gentry LE, Purchio AF, Moses HL: Mechanism of the serine proteinase system in chronic kidney rejection. Trans- activation of latent recombinant transforming growth factor beta plantation 65: 1628–1634, 1998 1 by plasmin. J Cell Biol 110: 1361–1367, 1990 25. Dewerchin M, Van Nuffelen A, Wallays G, Bouche´ A, Moons L, 8. Eddy AA: Plasminogen activator inhibitor-1 and the kidney. Carmeliet P, Mulligan RC, Collen D: Generation and character- Am J Physiol Renal Physiol 283: F209–220, 2002 ization of urokinase receptor-deficient mice. J Clin Invest 97: 9. Oda T, Jung YO, Kim H, Cai x, Lopez-Guisa J, Ikeda Y, Eddy 870–878, 1996 AA: PAI-1 deficiency attenuates the fibrogenic response to ure- 26. Eddy AA, Giachelli CM: Renal expression of genes that promote teral obstruction. Kidney Int 30: 587–596, 2001 interstitial inflammation and fibrosis in rats with protein-over- 10. Eitzman DT, McCoy RD, Zheng X, Fay WP, Shen T, Ginsburg load proteinuria. Kidney Int 47: 1546–1557, 1995 D, Simon RH: Bleomycin-induced pulmonary fibrosis in trans- 27. Kivirikko KI, Laitinen O, Prockop DJ: Modifications of a spe- genic mice that either lack or overexpress the murine plasmino- cific assay for hydroxyproline in urine. Analyt Biochem 19: gen activator inhibitor-1 gene. J Clin Invest 97: 232–237, 1996 249–255, 1967 11. Vassalli JD, Baccino D, Belin D: A cellular binding site for the 28. Zhang G, Oldroyd SD, Huang LH, Yang B, Li Y, Ye R, El Nahas Mr 55,000 form of the human plasminogen activator, urokinase. AM: Role of apoptosis and Bcl-2/Bax in the development of J Cell Biol 100: 86–92, 1985 tubulointerstitial fibrosis during experimental obstructive ne- 12. Preissner KT, Kanse SM, May AE: Urokinase receptor: a mo- phropathy. Exp Nephrol 9: 71–80, 2001 lecular organizer in cellular communication. Curr Opin Cell Biol 29. Ulery PG, Beers J, Mikhailenko I, Tanzi RE, Rebeck GW, 12: 621–628, 2000 Hyman BT, Strickland DK: Modulation of beta-amyloid precur- 13. Rondeau E, Ochi S, Lacave R, He CJ, Medcalf R, Delarue F, sor protein processing by the low density lipoprotein receptor- Sraer JD: Urokinase synthesis and binding by glomerular epithe- related protein (LRP). Evidence that LRP contributes to the lial cells in culture. Kidney Int 36: 593–600, 1989 pathogenesis of Alzheimer’s disease. J Biol Chem 275: 7410– 14. Nguyen G, Li X-M, Peraldi M-N, Zacharias U, Hage`ge J, Ron- 7415, 2000 deau E, Sraer J-D: Receptor binding and degradation of uroki- 30. Yang J, Chen S, Huang L, Michalopoulos GK, Liu Y: Sustained nase-type plasminogen activator by human mesangial cells. Kid- expression of naked plasmid DNA encoding hepatocyte growth ney Int 46: 208–215, 1994 factor in mice promotes liver and overall body growth. Hepatol- 15. Almus-Jacobs F, Varki N, Sawdey MS, Loskutoff DJ: Endotoxin ogy 33: 848–859, 2001 stimulates expression of the murine urokinase receptor gene in 31. Hierck BP, Iperen LV, Gittenberger De Groot AC, Poelmann vivo. Am J Pathol 147: 688–698, 1995 RE: Modified indirect immunodetection allows study of murine 16. Xu Y, Hagege J, Mougenot B, Sraer JD, Rønne E, Rondeau E: tissue with mouse monoclonal antibodies. J Histochem Cytochem Different expression of the plasminogen activation system in 42: 1499–1502, 1994 renal thrombotic microangiopathy and the normal human kidney. 32. Tamura K, Manabe N, Uchio K, Miyamoto M, Yamaguchi M, Kidney Int 50: 2011–2019, 1996 Ogura A, Yamamoto Y, Nagano N, Furuya Y, Miyamoto H: 17. Shetty S, Kumar A, Johnson AR, Pueblitz S, Holiday D, Raghu Characteristic changes in carbohydrate profile in the kidneys of G, Idell S: Differential expression of the urokinase receptor in hereditary nephrotic mice (ICGN strain). J Vet Med Sci 62: fibroblasts from normal and fibrotic human lungs. Am J Respir 379–390, 2000 Cell Mol Biol 15: 78–87, 1996 33. Igarashi P, Shashikant CS, Thomson RB, Whyte DA, Liu-Chen 18. Wagner SN, Atkinson MJ, Wagner C, Hofler H, Schmitt M, S, Ruddle FH, Aronson PS: Ksp-cadherin gene promoter. II. Wilhelm O: Sites of urokinase-type plasminogen activator ex- Kidney-specific activity in transgenic mice. Am J Physiol 277: pression and distribution of its receptor in the normal human F599–F610, 1999 kidney. Histochem Cell Biol 105: 53–60, 1996 34. Mizuno S, Kurosawa T, Matsumoto K, Mizuno-Horikawa Y, 19. Vilhardt F, Nielsen M, Sandvig K, van Deurs B: Urokinase-type Okamoto M, Nakamura T: Hepatocyte growth factor prevents plasminogen activator receptor is internalized by different mech- renal fibrosis and dysfunction in a mouse model of chronic renal anisms in polarized and nonpolarized Madin-Darby canine kid- disease. J Clin Invest 101: 1827–1834, 1998 ney epithelial cells. Mol Biol Cell 10: 179–195, 1999 35. Eitner F, Cui Y, Hudkins KL, Stokes MB, Segerer S, Mack M, 20. Hoyer-Hansen G, Ronne E, Solberg H, Behrendt N, Ploug M, Lewis PL, Abraham AA, Schlondorff D, Gallo G, Kimmel PL, Lund LR, Ellis V, Dano K: Urokinase plasminogen activator Alpers CE: Chemokine receptor CCR5 and CXCR4 expression cleaves its cell surface receptor releasing the ligand-binding in HIV-associated kidney disease. J Am Soc Nephrol 11: 856– domain. J Biol Chem 267: 18224–18229, 1992 867, 2000 1270 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1254–1271, 2003

36. Jones CL, Buch S, Post M, McCulloch L, Liu E, Eddy AA: Renal 53. Bachmann F: Plasminogen-plasmin enzyme system. In: Hemo- extracellular matrix accumulation in acute puromycin amino- stasis and Thrombosis. Basic Principles and Clinical Practice. nucleoside nephrosis in rats. Am J Pathol 141: 1381–1396, 1992 Edited by Colman RW, et al., Philadelphia, Lippincott Williams 37. Kristensen P, Eriksen J, Blasi F, Dano K: Two alternatively & Wilkins, 2001, pp 275–320 spliced mouse urokinase receptor mRNAs with different histo- 54. Gold LI, Schwimmer R, Quigley JP: Human plasma fibronectin logical localization in the gastrointestinal tract. J Cell Biol 115: as a substrate for human urokinase. Biochem J 262: 529–534, 1763–1771, 1991 1989 38. Schwarzbauer JE, Tamkun JW, Lemischka IR, Hynes RO: Three 55. Kazes I, Delarue F, Hagege J, Bouzhir-Sima L, Rondeau E, Sraer different fibronectin mRNAs arise by alternative splicing within JD, Nguyen G: Soluble latent membrane-type 1 matrix metallo- the coding region. Cell 35: 421–431, 1983 protease secreted by human mesangial cells is activated by 39. Nakatsukasa H, Nagy P, Evarts RP, Hsia C-C, Marsden E, urokinase. Kidney Int 54: 1976–1984, 1998 Thorgeirsson SS: Cellular distribution of transforming growth 56. Loskutoff DJ, Quigley JP: PAI-1, fibrosis, and the elusive pro- factor-␤1 and procollagen types I III, and IV transcripts in carbon visional fibrin matrix. J Clin Invest 106: 1441–1443, 2000 tetrachloride-induced rat liver fibrosis. J Clin Invest 85: 1833– 57. Swaisgood CM, French EL, Noga C, Simon RH, Ploplis VA: 1843, 1990 The development of bleomycin-induced pulmonary fibrosis in 40. Qian SW, Kondaiah P, Roberts AB, Sporn MB: cDNA cloning mice deficient for components of the fibrinolytic system. Am J by PCR of rat transforming growth factor ␤-1. Nucl Acids Res Pathol 157: 177–187, 2000 18: 3059, 1990 58. Edgtton KL, Carmeliet P, Kitching AK: Endogenous plasmin is 41. Liu Y, Michalopoulos GK, Zarnegar R: Molecular cloning and not protective in unilateral ureteric ligation induced renal inter- characterization of cDNA encoding mouse hepatocyte growth stitial fibrosis. J Am Soc Nephrol 13: 539A, 2002 factor. Biochim Biophys Acta 1216: 299–303, 1993 59. Kitching AR, Holdsworth SR, Ploplis VA, Plow EF, Collen D, 42. Zehab R, Gelehrter TD: Cloning and sequencing of cDNA for the Carmeliet P, Tipping PG: Plasminogen and plasminogen activa- rat plasminogen activator inhibitor-1. Gene 73: 459–468, 1988 tors protect against renal injury in crescentic glomerulonephritis. 43. Rickles RJ, Darrow AL, Strickland S: Molecular cloning of J Exp Med 185: 963–968, 1997 complementary DNA to mouse tissue plasminogen activator 60. Liu Y, Rajur K, Tolbert E, Dworkin LD: Endogenous hepatocyte mRNA and its expression during F9 teratocarcinoma cell differ- growth factor ameliorates chronic renal injury by activating entiation. J Biol Chem 263: 1563–1569, 1988 matrix degradation pathways. Kidney Int 58: 2028–2043, 2000 44. Niclou SP, Suidan HS, Pavlik A, Vejsada R, Monard D: Changes 61. Mizuno S, Matsumoto K, Nakamura T: Hepatocyte growth factor in the expression of protease-activated receptor 1 and protease suppresses interstitial fibrosis in a mouse model of obstructive nexin-1 mRNA during rat nervous system development and after nephropathy. Kidney Int 59: 1304–1314, 2001 nerve lesion. Eur J Neurosci 10: 1590–1607, 1998 62. Yang J, Dai C, Liu Y: Systemic administration of naked plasmid 45. Menoud P-A, Sappino N, Boudal-Khoshbeen M, Vassalli J-D, encoding hepatocyte growth factor ameliorates chronic renal Sappino A-P: The kidney is a major site of ␣2-antiplasmin fibrosis in mice. Gene Ther 8: 1470–1479, 2001 production. J Clin Invest 97: 2478–2484, 1996 63. Yang J, Liu Y: Blockage of tubular epithelial to myofibroblast 46. Kenagy RD, Nikkari ST, Welgus HG, Clowes AW: Heparin transition by hepatocyte growth factor prevents renal interstitial inhibits the induction of three matrix metalloproteinases (strome- fibrosis. J Am Soc Nephrol 13: 96–107, 2002 lysin, 92-kD gelatinase, and collagenase) in primate arterial 64. el-Dahr SS, Dipp S: Differential effects of ureteral obstruction on smooth muscle cells. J Clin Invest 93: 1987–1993, 1994 rat kininogen gene family. J Am Soc Nephrol 5: 102–109, 1994 47. Kim H, Oda T, Lopez-Guisa J, Wing D, Edwards DR, Soloway 65. Wolf WC, Yoshida H, Agata J, Chao L, Chao J: Human tissue PD, Eddy AA: TIMP-1 deficiency does not attenuate interstitial kallikrein gene delivery attenuates hypertension, renal injury, fibrosis in obstructive nephropathy. J Am Soc Nephrol 12: 736– and cardiac remodeling in chronic renal failure. Kidney Int 58: 748, 2001 730–739, 2000 48. Roche PC, Campeau JD, Shaw STJ: Comparative electrophoretic 66. Colman RW, Pixley RA, Najamunnisa S, Yan W, Wang J, Mazar analysis of human and porcine plasminogen activators in SDS- A, McCrae KR: Binding of high molecular weight kininogen to polyacrylamide gels containing plasminogen and casin. Biochem human endothelial cells is mediated via a site within domains 2 Biophys Acta 745: 82–89, 1983 and 3 of the urokinase receptor. J Clin Invest 100: 1481–1487, 49. Abe M, Harpel JG, Metz CN, Nunes I, Loskutoff DJ, Rifkin DB: 1997 An assay for transforming growth factor-beta using cells trans- 67. Blasi F: uPA, uPAR, PAI-1: key intersection of proteolytic, fected with a plasminogen activator inhibitor-1 promoter-lucif- adhesive and chemotactic highways? Immunol Today 18: 415– erase construct. Anal Biochem 216: 276–284, 1994 417, 1997 50. Grenier A, Chollet-Martin S, Crestani B, Delarche C, El Benna 68. Ossowski L, Aguirre-Ghiso JA: Urokinase receptor and integrin J, Boutten A, Andrieu V, Durand G, Gougerot-Pocidalo MA, partnership: coordination of signaling for cell adhesion, migra- Aubier M, Dehoux M: Presence of a mobilizable intracellular tion and growth. Curr Opin Cell Biol 12: 613–620, 2000 pool of hepatocyte growth factor in human polymorphonuclear 69. Herz J, Strickland DK: LRP: a multifunctional scavenger and neutrophils. Blood 99: 2997–3004, 2002 signaling receptor. J Clin Invest 108: 779–784, 2001 51. Naldini L, Vigna E, Bardelli A, Follenzi A, Galimi F, Comoglio 70. Salicioni AM, Mizelle KS, Loukinova E, Mikhailenko I, Strick- PM: Biological activation of pro-HGF (hepatocyte growth fac- land DK, Gonias SL: The low density lipoprotein receptor- tor) by urokinase is controlled by a stoichiometric reaction. J Biol related protein mediates fibronectin catabolism and inhibits fi- Chem 270: 603–611, 1995 bronectin accumulation on cell surfaces. J Biol Chem 277: 52. Schmaier AH: The plasma kallikrein-kinin system counterbal- 16160–16166, 2002 ances the renin-angiotensin system. J Clin Invest 109: 1007– 71. Segarini PR, Nesbitt JE, Li D, Hays LG, Yates JR, 3rd, Car- 1009, 2002 michael DF: The low density lipoprotein receptor-related pro- J Am Soc Nephrol 14: 1254–1271, 2003 uPAR in Renal Fibrosis 1271

tein/alpha2- macroglobulin receptor is a receptor for connective fibrosis in fibrinogen-null mice. J Clin Invest 106: 1341– tissue growth factor. J Biol Chem 276: 40659–40667, 2001 1350, 2000 72. Olman MA, Mackman N, Gladson CL, Moser KM, Loskutoff 76. Sisson TH, Hattori N, Xu Y, Simon RH: Treatment of bleomy- DJ: Changes in procoagulant and fibrinolytic gene expression cin-induced pulmonary fibrosis by transfer of urokinase-type during bleomycin-induced lung injury in the mouse. J Clin Invest plasminogen activator genes. Hum Gene Ther 10: 2315–2323, 96: 1621–1630, 1995 1999 73. Idell S, James KK, Gillies C, Fair DS, Thrall RS: Abnormalities 77. Hart DA, Whidden P, Green F, Henkin J, Woods DE: Partial of pathways of fibrin turnover in lung lavage of rats with oleic reversal of established bleomycin-induced pulmonary fibrosis acid and bleomycin-induced lung injury support alveolar fibrin by rh-urokinase in a rat model. Clin Invest Med 17: 69–76, deposition. Am J Pathol 135: 387–399, 1989 1994 74. Wilberding JA, Ploplis VA, McLennan L, Liang Z, Cornelissen 78. Haraguchi M, Border WA, Huang Y, Noble NA: t-PA promotes I, Feldman M, Deford ME, Rosen ED, Castellino FJ: Develop- glomerular plasmin generation and matrix degradation in exper- ment of pulmonary fibrosis in fibrinogen-deficient mice. Ann N imental glomerulonephritis. Kidney Int 59: 2146–2155, 2001 Y Acad Sci 936: 542–548, 2001 79. Kohonen-Corish MR, Wang Y, Doe WF: A highly polymorphic 75. Hattori N, Degen JL, Sisson TH, Liu H, Moore BB, Pandrangi CA/GT repeat in intron 3 of the human urokinase receptor gene RG, Simon RH, Drew AF: Bleomycin-induced pulmonary (PLAUR). Hum Genet 97: 124–125, 1996