Urokinase Receptor Modulates Cellular and Angiogenic Responses in Obstructive Nephropathy

Home , Urokinase receptor

J Am Soc Nephrol 14: 1234–1253, 2003 Urokinase Receptor Modulates Cellular and Angiogenic Responses in Obstructive Nephropathy

GUOQIANG ZHANG,* HEUNGSOO KIM,* XIAOHE CAI,* JESUS M. LOPEZ-GUISA,* PETER CARMELIET,† and ALLISON A. EDDY* *University of Washington, Children’s Hospital and Regional Medical Center, Division of Nephrology, Seattle, Washington; and †The Center For Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, Leuven, Belgium.

Abstract. Interstitial cells have been implicated in the patho- fibrosis in the Ϫ/Ϫ mice based on Sirius red staining (4.6 Ϯ genesis of renal fibrosis. Given that the urokinase receptor 0.9 versus 2.3 Ϯ 0.9% area at 14 d UUO). Absence of the (uPAR) is known to play a role in cell adhesion, migration, and uPAR scavenger receptor was associated with significantly angiogenesis, the present study was designed to evaluate the greater accumulation of plasminogen activator inhibitor-1 pro- role of uPAR in the regulation of the phenotypic composition tein (PAI-1) (20.5 Ϯ 3.5 versus 9.1 Ϯ 2.9% area, day 14 UUO) of interstitial cells (macrophages, myofibroblasts, capillaries) and vitronectin protein (2.4 Ϯ 1.1 versus 0.9 Ϯ 0.4% area, day in response to chronic renal injury. Groups of uPAR wild-type 14 UUO). By immunostaining ␣SMAϩ cells, CD34ϩ cells, (ϩ/ϩ) and knockout (Ϫ/Ϫ) mice were investigated between 3 vitronectin and PAI-1 co-localized to the same tubulointersti- and 14 d after unilateral ureteral obstruction (UUO) or sham tial area. The number of apoptotic cells increased in response surgery (n ϭ 8 mice per group). The density of F4/80ϩ to UUO but was significantly higher in the Ϫ/Ϫ mice (2.0 Ϯ interstitial macrophages (M␾) was significantly lower in the 0.2 versus 1.2 Ϯ 0.2 per 100 tubulointerstitial cells, day 14 Ϫ/Ϫ mice (3.3 Ϯ 0.4 versus 6.9 Ϯ 1.7% area at day 3 UUO; UUO) while the number of proliferating cells was significantly 10.8 Ϯ 1.6 versus 15.7 Ϯ 1.0% at day 14 UUO; Ϫ/Ϫ versus lower in the uPARϪ/Ϫ mice. These data suggest that uPAR ϩ/ϩ). In contrast, in the Ϫ/Ϫ mice there were significantly deficiency suppresses renal M␾ recruitment, but the absence of more ␣ smooth muscle actin (␣SMA)–positive cells (12.9 Ϯ this scavenger receptor actually accentuates the fibrogenic 3.2 versus 7.8 Ϯ 1.5% area at day 3 UUO; 21.0 Ϯ 4.7 versus response, likely due in part to the delayed clearance of angio- 9.7 Ϯ 1.9% at day 14 UUO) and CD34-positive endothelial genic/profibrotic molecules such as PAI-1 and decreased re- cells (8.4 Ϯ 1.9 versus 4.0 Ϯ 1.1% area at day 14 UUO). These ceptor-associated uPA activity. differences were associated with significantly more interstitial

Interstitial fibrosis and subsequent tubular atrophy are pivotal RANTES, integrins, and the urokinase-plasmin cascade, have pathologic changes that lead to end-stage kidney disease (1). been implicated in the processes of macrophage adhesion and Renal fibrosis is the end result of a series of events often migration (3). Myofibroblasts usually become the predominant initiated by the recruitment of monocytes from the circulation, interstitial cell type with chronic inflammation. These cells are their activation, and differentiation into macrophages. Macro- considered to be the major source of the extracellular matrix phage activities are numerous and may include trophic effects components that accumulate during renal fibrosis. Myofibro- (growth promoting, cellular differentiation, tissue repair, an- blasts appear de novo in areas of future fibrosis in response to giogenesis), cytotoxic tissue injury, and scavenging activities. stimuli such as basic fibroblast growth factor, platelet-derived Tubulointerstitial infiltration by macrophages is thought to growth factor, TGF-␤1, metalloproteinase-2 (MMP-2), and play an active role in tissue fibrogenic reactions (1,2). Several plasminogen activator inhibitor-1 (PAI-1). Despite significant molecules, including macrophage chemoattractant protein-1 advances, knowledge about the regulation and function of (MCP-1), osteopontin, transforming growth factor-␤ (TGF-␤), macrophages and myofibroblasts during renal fibrosis remains incomplete, a fact that has hampered the development of ef- Received January 9, 2002. Accepted February 10, 2003. fective therapy for patients with progressive renal disease. Dr. Eric Rondeau served as Guest Editor and supervised the review and final A high-affinity cellular receptor for urokinase-type plasmin- disposition of this manuscript. ogen activator (uPAR or CD87) has been identified on the Correspondence to Dr. Allison A. Eddy, Children’s Hospital and Regional plasma membrane of a variety of cell types, including mono- Medical Center, Division of Nephrology, Mail Stop CH 5G-1, 4800 Sand Point Way NE, Seattle, WA 98105. Phone: 206-987-2524; Fax: 206-987-2636; cytes, neutrophils, activated T cells, endothelial cells, glomer- E-mail: [email protected] ular epithelial and mesangial cells, tubular epithelial cells, 1046-6673/1405-1234 fibroblasts, and myofibroblasts (4–10). First identified in 1985 Journal of the American Society of Nephrology (11), uPAR is a highly glycosylated 50-kD to 60-kD protein. Copyright © 2003 by the American Society of Nephrology Its currently known ligands are uPA, vitronectin, and kinino- DOI: 10.1097/01.ASN.0000064701.70231.3F gen (12). uPAR itself lacks a transmembrane domain. It is J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1235 anchored to the plasma membrane by a glycosyl phosphatidyl- of functional units with these cellular proteins allows uPAR to inositol (GPI) moiety. Soluble forms also exist. Tubular uPAR generate cell surface concentrated proteolysis required for cell immunoreactivity has been reported in normal human kidneys migration and also to contribute to non-proteolytic cellular (10). Changes in the pattern of uPAR expression have not yet adhesion by interacting with ␤2-leukocyte integrins or ␣v␤3or been extensively investigated in renal disease. Increased renal ␣v␤5 integrins and vitronectin (18–20). Due to these cellular production of uPAR has been reported in humans and mice functions, uPAR is thought to regulate cellular responses dur- with endotoxemia and in human kidneys with chronic pyelo- ing angiogenesis, inflammation, wound repair, and tumor me- nephritis, acute tubular necrosis, and chronic allograft rejection tastasis (12,19,21,22). In addition to its ability to modulate cell (7,8,13,14). migration, uPAR may mediate molecular crosstalk at cellular It is now clear that uPAR is a multifunctional protein surfaces, cytoskeletal reorganization, endocytosis-dependent (15,16,17). The glycolipid-anchored uPAR co-localizes peri- scavenging, and cellular apoptosis (12,17,23). cellularly with components of the urokinase-plasmin activation Given these important functions for uPAR together with the system, including uPA and PAI-1, and endocytosis receptors observation that PAI-1 and uPA may be upregulated during such as members of the LDL receptor related protein (LRP) renal fibrosis, the present study was designed to investigate the family. uPAR is frequently co-expressed with caveolin and role of uPAR in the renal cellular response that follows ureteral members of the beta integrin superfamily (12). The formation obstruction.

Figure 1. Urokinase receptor (uPAR) genotype and its renal expression. (A) Southern blot analysis demonstrating the 5-kb band of the wild-type uPAR gene in the uPARϩ/ϩ mice and the 3.5-kb mutant band in the uPARϪ/Ϫ mice. (B) Northern blot analysis shows specific uPAR1 mRNA bands in the kidneys of uPARϩ/ϩ mice 7 d after unilateral ureteral obstruction (UUO).

Figure 2. uPAR expression. After7dofUUO, uPAR protein was not detected in uPARϪ/Ϫ kidneys (A). At this time, uPAR was expressed by interstitial cells (B) and tubular epithelial cells (C). Magnification: ϫ400. 1236 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

were processed for cryostat and microtome sectioning. The day 7 UUO kidneys and their sham controls were processed for protein and RNA extraction. Pieces to be embedded in paraffin were fixed in 10% buffered formalin, and those for cryostat sectioning were embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrence, CA) and snap-frozen in pre-chilled 2-methylbutane. Tissues for protein and mRNA extraction were snap-frozen in liquid nitrogen and stored at Ϫ80°C for subsequent use. For protein isolation, frozen kidney tissue from each animal was individually ground into a fine powder using a pre-chilled mortar and pestle, 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 dialysis membrane Spectra/PorR 1 (Spectrum Medical Industries, Inc., Houston, Texas) against 0.05 M Tris, 0.2% Triton X-100, pH 7.5, for 48 h at 4°C. The samples were centrifuged for 5 min (14,000 ϫ g). The supernatant was aliquoted on the basis of the protein concentration measured using the Bradford protein assay (Bio-Rad, Hercules, CA). The aliquoted samples were stored at Ϫ80°C.

Northern Blot Analysis Total kidney RNA was isolated by a modified phenol and guanidine isothiocyanate method using TRIzol reagent (Life Technologies BRL Life Technologies, Grand Island, NY) according to the manu- facturer’s instructions. Total kidney RNA (15 ␮g) from each animal was separated by a 1% agarose formaldehyde gel electrophoresis. After a photomicrograph of the ethidium bromide-stained gel was obtained to evaluate RNA loading equality, the RNA was transferred to a hybridization membrane (GeneScreen Plus, New England Nu- clear Life Science Products, Boston, MA) and fixed by ultraviolet Figure 3. F4/80-positive interstitial macrophages. (A through D) cross-linking (UV Crosslinker, Hoeffer Scientific Instruments, San Photomicrographs of F4/80 immunohistochemical staining in 3 d Francisco, CA). Complementary DNA probes for mouse uPAR1 (a UUO (A and B) and 14 d UUO (C and D) kidneys. At both times, 1.5-kb fragment), provided by Dr. Niels Behrendt, Finsen Laboratory, there are significantly fewer macrophages in the uPARϪ/Ϫ mice (B Copenhagen, Denmark (25), rat osteopontin, provided by Dr. C. and D) compared with the uPARϩ/ϩ mice (A and C). Magnification: Giachelli, University of Washington, Seattle, WA (26), and mouse ϫ400. The F4/80-positive tubulointerstitial area expressed as mean Ϯ 1 MCP-1 provided by Dr. B. Rollins, Dana Farber Cancer Institute, 32 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.001, Boston, MA (27), were labeled with P dCTP (3,000 Ci/mmol) by uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.01, UUO compared with sham random priming with the T7 Quick Prime kit (Pharmacia Biotech, group of same genotype and sex. Piscataway, NJ). The membranes were hybridized with the radiola- beled cDNA probe using the QuickHyb hybridization solution (Strat- agene, La Jolla, CA). Autoradiographs were developed and the density of each band quantified using the NIH Image program. The Materials and Methods density of the 18 s ribosomal bands in the formaldehyde gels were Animals and Experimental Protocol used for RNA loading control. UPAR-deficient (uPARϪ/Ϫ) and wild-type (uPARϩ/ϩ) mice on a C57BL/6 genetic background used for this study have been previously Histologic Studies described (24). Mice were bred in our animal facility and allowed to Immunohistochemical studies were performed on frozen or paraf- grow to a minimum weight of 20 g before the study began. The fin-embedded renal tissue sections (4 ␮m). Immunoperoxidase stain- genotype of the mice was confirmed by Southern blot analysis of ing using ABC ELITE kits (Vector Laboratories Inc, Burlingame, DNA extracted from tails. Five groups of gender-matched, age- CA) was performed on paraffin sections. Primary antibodies (Ab1) Ϫ Ϫ ϩ ϩ matched, and weight-matched uPAR / and / mice were stud- included anti-mouse uPAR (R&D Systems, Minneapolis, MN), rat ϭ ϭ ied: 3 d after UUO (female, n 8 each), 7 d after UUO (male, n anti-mouse F4/80, rat anti-mouse CD11b monoclonal antibodies (Se- ϭ 8 each), 14 d after UUO (female, n 8 each), and 7 d after sham rotec Ltd., Oxford, UK), goat anti-mouse MCP-1 (Santa Cruz Bio- ϭ surgery (n 8 males and 8 females). UUO surgery was performed technology, CA), anti-mouse osteopontin (Santa Cruz Biotechnolo- under general anesthesia. The left ureter was ligated with 4.0 silk at gy), horseradish peroxidase (HRP)-conjugated mouse anti-human two separate locations in the UUO groups. Mice were killed by smooth muscle actin (SMA) monoclonal antibody (DAKO Corp., exsanguination under general anesthesia. All procedures were per- Carpinteria, CA), HRP-conjugated mouse-anti-proliferating cell nu- formed in compliance with the guidelines established by National clear antigen (PCNA; DAKO), murine anti-mouse PAI-1 monoclonal Research Council Guide for the Care and Use of Laboratory Animals. antibody (MA-33H1F7, a generous gift from Dr. Declerck, Katholieke University, Belgium) (28), and rat anti-mouse CD34 antibody Kidney Tissue Preparation (Pharmingen, San Diego, CA). For PAI-1 staining, Ab1 and Ab2 were Following exsanguination, the obstructed left kidney was harvested pre-complexed before incubation with the tissue sections to minimize and the capsule removed. Day 3 and day 14 UUO and sham kidneys cross-reactivity between Ab2 and murine tissue IgG (29). Cryosec- J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1237

Figure 4. CD11b-positive interstitial macrophages. (A through D) Photomicrographs of CD11b immunohistochemical staining in 3 d UUO (A and B) and 14 day UUO (C and D) kidneys. At both times, there are significantly fewer macrophages in the uPARϪ/Ϫ mice (B and D) compared with the uPARϩ/ϩ mice (A and C). Magnification: ϫ750. (E) The CD11b-positive cells per 100 tubulointerstitial cells, expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.01, UUO compared with sham group. tions were stained with antibodies to vitronectin (rabbit anti-mouse were FITC-conjugated goat anti-rabbit IgG (Organon Teknika Corp., polyclonal antibody; a generous gift from Dr. David Loskutoff, The West Chester, PA) or FITC-conjugated rabbit anti-goat IgG (Southern Scripps Research Institute, CA) (30) and ␣v integrin (goat anti-mouse Biotechnology Associates). For co-localization studies, a few addi- ␣v integrin antiserum; Santa Cruz Biotechnology). The Ab2 used tional serial UUO paraffin sections were stained for ␣SMA and CD34, 1238 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 5. Renal ␣v integrin expression. (A through D) Photomicrographs of ␣v integrin immunofluorescence staining in sham (A and B) and 14 day UUO (C and D) kidneys. Weak tubular staining is present in a few tubules of sham uPARϩ/ϩ kidneys (A) and uPARϪ/Ϫ kidneys (B). Tubular staining is increased on day 14 UUO but to a greater degree in the ϩ/ϩ kidneys (C) than the Ϫ/Ϫ kidneys (D). Magnification: ϫ400. (E) The ␣v integrin-positive tubulointerstitial area expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ. Closed bars are uPARϩ/ϩ. * P Ͻ 0.001, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.01, UUO compared with sham group.

vitronectin, or PAI-1. ␣SMA and CD34 double-staining were per- software (Optimas version 6.5, Optimas Corp., Bothell, WA) as formed with HRP–conjugated and alkaline phosphatase-conjugated previously reported (31). A point-counting method was used to quan- secondary antibodies. Sections stained with Ab2 only were run in tify F4/80ϩ macrophage staining. Results were expressed as percent- parallel as a negative control. The stained tubulointerstitial area was age of total measured tubulointerstitial area. CD11bϩ interstitial cells measured using a computerized image analysis system and Optimas and PCNA-positive tubular and interstitial cells were counted manu- J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1239

Figure 6. Monocyte chemoattractant protein-1 (MCP-1) expression. A Northern blot autoradiograph 7 d after UUO demonstrates significantly increased renal MCP-1 mRNA compared with sham kidneys. MCP-1 mRNA was significantly more abundant in the uPARϪ/Ϫ compared with the uPARϩ/ϩ kidneys (A). Western blotting confirmed higher MCP-1 protein concentrations in the uPARϪ/Ϫ mice after 7 d of UUO (B). Immunostaining detected MCP-1 primarily in renal tubules and occasional interstitial cells (C). Magnification: ϫ400. The histograms illustrate the results expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. ϩ P Ͻ 0.05 UUO compared with sham mice of the same genotype. ally using an eyepiece grid and expressed as the percent positive sections were deparaffinized and stripped of proteins by incubation tubulointerstitial cells as described previously (32). with 20 ␮g/ml proteinase K in PBS at 37°C for 15 min. Endogenous

peroxidase was inactivated by immersing the sections in 3% H2O2 for Western Blot Analysis 5 min. After pre-incubation with equilibration buffer, the samples Protein samples (20 or 80 ␮g) were separated by 10% SDS-PAGE). were incubated with TdT in reaction buffer (containing bio-14-dUTP) Proteins were transferred to a nitrocellulose membrane and the im- at 37°C for 60 min. Rinsed with PBS, the tissue sections were stained munoreactive protein visualized using ECL-enhanced chemilumines- with the ABC ELITE kit (Vector Lab. Inc.). Color was developed with cence (Amersham Pharmacia Biotech Inc., Piscataway, NJ). Ab1 used AEC Substrates Chromogen (DAKO Corp.) and counterstained with were rabbit anti-mouse CD14 antiserum (Santa Cruz Biotechnology), hematoxylin. Negative controls were obtained by eliminating TdT in mouse anti-murine PAI-1 monoclonal antibody, HRP-conjugated reaction buffer while other steps were run in parallel. Apoptotic mouse anti-PCNA monoclonal antibody, and HRP-conjugated mouse tubulointerstitial cell nuclei were counted in 10 random cortical fields anti-human ␣-SMA monoclonal antibody. Ab2 were HRP-conjugated (ϫ400 magnification). Results were expressed as the number positive goat anti-rabbit IgG antiserum (Chemicon International Inc.) and per 100 nuclei. For day 3 UUO kidneys, the number of apoptotic HRP-conjugated goat anti-mouse IgG antiserum (Sigma Chemical nuclei within tubular cross-sections and the interstitium were evalu- Co.). For PAI-1 probing, Ab1 and Ab2 were premixed before incu- ated separately. bation with the blots to block the crossreaction of Ab2 to tissue mouse IgG (29). Ponceau S red or amido black staining of the blots was performed to determine loading equality Interstitial Fibrosis Evaluated by Sirius Red Staining Picrosirius red staining was performed to evaluate histologically In Situ End Labeling of Apoptosis the interstitial area occupied by collagen fibrils as described previ- Apoptotic cell nuclei were detected by in situ end labeling of ously (31). Sections were examined by polarized light microscopy. endonuclease-cleaved DNA as described previously (33). Briefly, Photographs of six random cortical fields (ϫ400) from each animal 1240 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 7. Osteopontin (OPN) expression. A Northern blot autoradiograph 7 d after UUO demonstrates significantly increased renal osteopontin (OPN) mRNA compared with sham kidneys. OPN mRNA was significantly less abundant in the uPARϪ/Ϫ compared with the uPARϩ/ϩ kidneys (A). Immunostaining detected OPN protein in renal tubules; the lower expression levels in the uPARϪ/Ϫ animals appeared to be attributed to more extensive tubular destruction (B: uPARϩ/ϩ day 7 UUO; C: uPARϪ/Ϫ day 7 UUO; D: uPARϩ/ϩ day 14 UUO; E: uPARϪ/Ϫ day 14 UUO; F: uPARϩ/ϩ sham). Magnification: ϫ400. The histograms illustrate the day 7 results expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ0.05, uPARϪ/Ϫ versus uPARϩ/ϩ mice. ϩ P Ͻ 0.05 UUO compared with sham mice of the same genotype. J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1241

Figure 8. Alpha smooth muscle actin (␣-SMA) expression. (A through D) Photomicrographs of ␣-SMA immunohistochemical staining in 3 d (A and B) and 14 d UUO (C and D) kidneys. Significantly less staining is present in the interstitium of uPARϩ/ϩ kidneys (A and C) than in the uPARϪ/Ϫ kidneys (B and D). Magnification: ϫ400. (E) The ␣-SMA-positive tubulointerstitial area expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.05, ** P Ͻ 0.01, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.01, UUO compared with correspondent sham. 1242 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 9. CD34-positive microvasculature. (A through D) Photomicrographs of CD34 immunohistochemical staining in sham (A and B) and 14 d UUO (C and D) kidneys. In the sham uPARϩ/ϩ (A) and Ϫ/Ϫ (B) kidneys, CD34 antigen is expressed by peritubular capillaries and glomerular endothelial cells (arrow). Following ureteral obstruction, the CD34-positive interstitial area is reorganized with expansion in fibrotic areas (stars). These changes are more marked in the uPARϪ/Ϫ mice (D) compared with the uPARϩ/ϩ (C) mice. Magnification: ϫ400. (E) The CD34 tubulointerstitial area expressed as mean Ϯ 1 SD. Open bars: uPARϪ/Ϫ; closed bars: uPARϩ/ϩ.*P Ͻ 0.01, uPARϪ/Ϫ versus uPARϩ/ϩ. J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1243

demonstrated undetectable uPAR message in the kidneys of uPARϪ/Ϫ mice and the sham uPARϩ/ϩ mice. In the kidneys of the uPARϩ/ϩ mice, uPAR mRNA was present 7 d after UUO (Figure 1B). uPAR immunohistochemical staining was negative on sham and all uPARϪ/Ϫ kidneys. In response to UUO, uPAR expression was detected on interstitial and tubular cells in the uPARϩ/ϩ kidneys (Figure 2).

Interstitial Macrophages and ␣v Integrin Expression In response to UUO and compared with the sham-operated kidneys, the number of F4/80ϩ interstitial macrophages was significantly increased at 3 d only in the uPARϩ/ϩ mice; by day 14, the number of F4/80ϩ interstitial macrophages was significantly increased in the mice of both genotypes (Figure 3). However, at both time points, there were significantly fewer F4/80ϩ cells in the uPARϪ/Ϫ mice compared with the ϩ/ϩ mice. This difference in the number of renal macrophages was reconfirmed when CD11bϩ interstitial cells were counted (Figure 4). Western blot analysis for CD14, another murine monocyte/macrophage antigen, on day 7, also showed that CD14 protein levels were 1.5-fold higher in uPARϩ/ϩ mice (1.00 Ϯ 0.11 versus 0.63 Ϯ 0.02 arbitrary units). Given that previous studies have suggested that uPAR may interact with ␣v␤3 and ␣v␤5 integrin receptors to facilitate leukocyte adhesion and migration, expression of the ␣v integrin chain was examined (18,19,22). In the ϩ/ϩ mice ␣v integrin was expressed at low levels on a few cortical tubules in the sham control kidneys. ␣v protein was increased tenfold in the day 14 UUO group, but its expression was primarily restricted to tubules (Figure 5). Although tubular expression of ␣v integrin also increased in uPARϪ/Ϫ mice during UUO, this adhesive molecule was expressed at a significantly lower level in uPARϪ/Ϫ mice on days 14 UUO compared with ϩ/ϩ mice. Several chemoattractant molecules are also known to participate in renal monocyte recruitment triggered by ureteral obstruction such as monocyte chemoattractant protein-1 Figure 10. Co-localization of interstitial myofibroblasts and neovas- (MCP-1) and osteopontin. In response to7dofUUO, renal ␣ cularization. Double-staining showing co-localization of -SMA MCP-1and osteopontin mRNA levels were significantly in- (DAB with nickel reaction product in dark brown) and CD34 (Fast Ϫ Ϫ creased, but only osteopontin expression was attenuated in the Red TR/naphthol phosphate reaction product in red) in a uPAR / Ϫ Ϫ ϩ ϩ ϩ uPAR / mice compared with uPAR / mice (Figures 6 kidney on day 3 UUO (A). Higher magnification shows CD34 Ϫ Ϫ endothelium (solid black arrow) and perivascular smooth muscle cells and 7). In fact, MCP-1 levels were higher in the uPAR / and interstitial myofibroblasts (open block arrow) (B). Magnifica- mice on day 7. Western blot analysis reconfirmed the differ- tions: ϫ400 in A; ϫ750 in B. ence in MCP-1 expression between the uPARϩ/ϩ and Ϫ/Ϫ mice (Figure 6). Immunostaining of the obstructed kidneys were taken using a SPOT camera and the percent positive tubuloin- detected MCP-1 protein mainly in renal tubules but some terstitial area measured using the Optimas program. interstitial reactivity was also present. Osteopontin expression in the UUO kidneys was limited to a subpopulation of renal Statistical Analyses tubules, and the difference between the genotypes appeared to Ϫ Ϫ All data were expressed as mean Ϯ 1 SD unless otherwise stated. be due to greater tubular destruction in the uPAR / mice, Results were analyzed by the Mann-Whitney U test or t test using the especially on day 14 when several of the osteopontin-positive SPSS or Excel software. A P value Ͻ 0.05 was considered statistically cells appeared as tubular remnants within the interstitium (Fig- significant. ure 7).

Results Interstitial Myofibroblasts and Capillaries uPAR Genotype and Renal Expression Smooth muscle cell-specific ␣ actin immunostaining was The genotypes of the experimental mice were confirmed by confined to the perivascular cells of the arterioles in sham Southern blot analysis (Figure 1A). Northern blot analysis groups of both genotypes. In response to UUO, the number 1244 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 11. Tubulointerstitial cell apoptosis. Photomicrographs illustrating in situ end labeling (ISEL) of apoptotic nuclei after 14 day UUO in uPARϩ/ϩ (A) and Ϫ/Ϫ (B) kidneys. Arrows indicate apoptotic cells with typical condensed nuclei within the tubules and interstitial area. Sections were counterstained with hematoxylin. Magnification: ϫ400. The graphs show the total number of apoptotic tubulointerstitial cells (C) and the number of apoptotic tubular cells (TC) and interstitial cells (IC) counted separately at 3 day UUO (D) expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ0.05, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.05, UUO versus sham (C). of positive interstitial cells significantly increased with acterized by increased angiogenesis, as defined by the den- time, but the response was more intense in the uPARϪ/Ϫ sity of CD34ϩ microvascular endothelial cells (34). The mice (Figure 8). The mean number of interstitial myofibro- degree of neovascularization was significantly greater in the blasts was 1.8-fold and 2.1-fold higher in uPARϪ/Ϫ com- uPARϪ/Ϫ group compared with the ϩ/ϩ group (Figure 9). pared with the ϩ/ϩ mice on days 3 and 14 UUO, respec- These CD34ϩ microvessels co-localized with ␣-SMA-pos- tively. This difference between the two genotypes was itive interstitial myofibroblasts (Figure 10). There was a reconfirmed on day 7 UUO by Western blot analysis that significant positive correlation between the numerical detected 2.4-fold more ␣-SMA protein in the kidneys of the values for the area of the tubulointerstitium stained for uPARϪ/Ϫ mice (data not shown). CD34 and ␣SMA (r ϭ 0.81; P Ͻ 0.05; Spearman rank The interstitial cellular response to UUO was also char- correlation). J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1245

blot studies indicated that proliferation was an early response to UUO (Figure 12). However, at both 3 and 14 d, PCNA protein levels were significantly higher in the uPARϩ/ϩ kidneys compared with the uPARϪ/Ϫ kidneys. By semiquantita- tive immunohistochemistry it was determined that most PCNA-positive cells in the UUO kidneys were tubular epithelial cells although rare positive interstitial cells were also detected (Figure 12).

Accumulation of Vitronectin and PAI-1 Vitronectin, a PAI-1-binding extracellular matrix protein that is a ligand of ␣v␤3 and ␣v␤5 integrins (18,19), was restricted to the vasculature and glomeruli in the sham-operated kidneys. After UUO, vitronectin was deposited within atrophic tubules and the interstitium to a greater extent in the uPARϪ/Ϫ mice compared with ϩ/ϩ mice (Figure 13). By immunostaining PAI-1 protein, not detected in the sham kidney, accumulated in interstitial areas, often co-localizing with ␣-SMA-positive cells (Figure 14). Western blot analysis demonstrated significantly more PAI-1 protein (1.5-fold increased) in uPARϪ/Ϫ compared with ϩ/ϩ kidneys on days 7 after UUO (Figure 15).

Interstitial Fibrosis Sirius red staining showed an impressive increase in interstitial collagen fibrils after UUO, reaching a tenfold increase in the uPAR-deficient mice by 14 d (Figure 16). The Sirius red-positive interstitial area was significantly less in the uPARϩ/ϩ mice compared with the Ϫ/Ϫ mice after 14 d of ureteral obstruction.

Discussion The results of this study indicate that the urokinase receptor Figure 12. Tubulointerstitial cell proliferation. Western blot analysis serves to dampen the severity of the renal fibrogenic response detected a significant increase in PCNA expression 3 d (A) and 14 d that is initiated by ureteral obstruction. In this experimental (B) after UUO between uPARϩ/ϩ and Ϫ/Ϫ mice. Densitometric model, we have previously reported that renal uPA gene ex- analysis of the blots showed significantly less PCNA protein, reflect- pression and enzyme activity are significantly increased (35). ing lesser mitotic activity, in the uPARϪ/Ϫ mice (open bars) com- From the present study, it is evident that uPAR plays a pivotal pared with the uPARϩ/ϩ mice (closed bars) (C). PCNA immunohis- role in the regulation of the cellular responses to ureteral tochemical staining identified most PCNA-positive (black) nuclei as obstruction but that its role is highly cell specific. Cells of tubular epithelial cells (TC) with lesser numbers of positive interstitial multiple lineages may express uPAR including resident kidney cells (IC). Fewer proliferating TC and IC were detected in the cells (epithelial, mesangial, and endothelial), inflammatory uPARϪ/Ϫ kidneys (D; day 3 UUO). Photomicrographs are represen- cells (monocytes, activated T cells, and neutrophils), and fi- tative fields 3 d (E: uPARϩ/ϩ; G: uPARϪ/Ϫ) and 14 d after UUO (F: uPARϩ/ϩ; H: uPARϪ/Ϫ). Magnification: ϫ400. Results are broblasts/myofibroblasts. In response to ureteral obstruction, shown graphically as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed renal uPAR gene expression was significantly upregulated. bars are uPARϩ/ϩ.*P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ The uPAR protein was identified on both interstitial cells and 0.05, UUO versus sham mice of the same genotype. renal tubules in wild-type kidneys after UUO. The expression of uPAR in these regions of the kidney coupled with significant differences in the number of interstitial monocytes, myofibro- Tubulointerstitial Cell Apoptosis and Proliferation blasts, and interstitial endothelial cells in uPAR wild-type mice Apoptotic tubulointerstitial cells were rarely detected in compared with uPAR null mice suggests that recruitment and sham kidneys of both genotypes, but the number increased perhaps function of these cells are modulated by uPAR. with time following UUO (Figure 11). uPAR deficiency re- Monocytes and macrophages that pervade the interstitium of sulted in the appearance of significantly more apoptotic tubular chronically damaged kidneys are thought to be one of the and interstitial cells on day 3 and day 14 UUO. Genotype also mediators of fibrosis due their ability to synthesize several affected the proliferative response to injury. Using PCNA pro-fibrotic molecules (36). The present study is one of the first expression levels as an estimate of mitotic activity, Western to demonstrate that the phenotype of the inflammatory renal 1246 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 13. Vitronectin accumulation. (A through D) are photomicrographs of vitronectin immunofluorescence staining in sham (A and B) and 14 d UUO (C and D) kidneys. Vitronectin is present in glomeruli (block arrow) and vessels (single arrow) of sham uPARϩ/ϩ kidneys (A) and uPARϪ/Ϫ kidneys (B). Vitronectin accumulated in the tubulointerstitium, especially within dilated or atrophic tubules, by day 14 UUO. Compared with the Ϫ/Ϫ kidneys (D), the extent of vitronectin deposition was less in the ϩ/ϩ kidneys (C) and was limited to a few fibrotic loci. Magnifications: ϫ250 in A and B; ϫ400 in C and D. (E) The vitronectin-positive tubulointerstitial area expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.05, uPARϪ/Ϫ versus uPARϩ/ϩ. ϩ P Ͻ 0.05 UUO versus sham of the same genotype.

interstitial cells is a critical factor that determines whether the delay the clearance of molecules that promote fibrosis. Al- macrophages function primarily as scavengers to minimize though many molecules may be involved, our data suggest that injury or as villains that perpetrate damage. In the absence of delayed clearance of PAI-1 and perhaps apoptotic cells by uPAR, not only is interstitial macrophage recruitment im- uPAR-bearing cells may contribute to a more aggressive fi- paired, but the absence of this “scavenging” receptor appears to brotic response. In addition to several in vitro studies docu- J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1247

Figure 14. Renal plasminogen activator inhibitor-1 protein (PAI-1) accumulation. Immunohistochemical staining on day 14 UUO shows greater PAI-1 accumulation in the interstitium of the uPARϪ/Ϫ mice (B) compared with the uPARϩ/ϩ mice (A). Staining of serial sections illustrates co-localization of PAI-1 protein (C) to areas of interstitial ␣SMAϩ myofibroblasts (D). Stars highlight regions stained for both PAI-1 and ␣-SMA. PAI-1 protein is not detected in uPARϪ/Ϫ sham kidneys (E). Magnification: ϫ400. The graph illustrates the PAI-1-positive tubulointerstitial area on day 14. * P Ͻ 0.01.

menting a role for uPAR in cellular movement, impaired macrophages may also occur (40). In the present study, most of migration of tumor cells and neutrophils has also been reported the PCNAϩ cells were tubular cells, and a difference in the in uPAR-deficient mice (22,37). In the mouse model of bleo- number of positive interstitial cells was relatively small but mycin-induced lung fibrosis, delayed macrophage recruitment significant. Whether the proliferating cells were macrophages has also been observed in uPAR-deficient mice (38). The or myofibroblasts was not determined. Urokinase is known to findings in the present study of significantly fewer renal mac- play a role in cell migration due to its ability to facilitate rophages in uPARϪ/Ϫ mice contrasts with the results of an cell-cell and cell-matrix interactions. UPAR is known to asso- earlier study of acute crescentic glomerulonephritis that dem- ciate with the leukocyte integrin CD11b/CD18 and L-selectin onstrated that neither uPAR nor uPA-deficiency modified the to regulate leukocyte migration and cellular signaling severity of glomerular inflammation or renal dysfunction, pos- (20,41,42). In addition, the uPA-uPAR complex physically sibly because increased uPA is not a significant feature in that cooperates with certain members of the integrin superfamily model of acute glomerular injury (39). including ␣v␤5 and ␣v␤3 to direct cell adhesion to and mi- The increased number of interstitial mononuclear cells in gration along vitronectin (18,19,43). Co-clustering and reso- chronically damaged kidneys is thought to be the consequence nance energy transfer between uPAR and ␣v␤5or␣v␤3 inte- of the migration of circulating monocytes into the interstitium grins has been observed to transduce migratory signals to cells although limited in situ proliferation of resident interstitial adherent to vitronectin (44). In addition to serving as an inte- 1248 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 15. Day 7 renal PAI-1 protein. Western blot of proteins isolated form kidneys after7dofUUOandstained with anti-PAI-1 antiserum. Results of densitometric analysis after correcting for protein loading based on ponceau S staining showed significantly more PAI-1 in the knockout (KO) compared with the wild-type (WT) kidneys.

grin-associated protein, the cell surface glycolipid-anchored served in the uPARϪ/Ϫ mice also contributed to the blunted uPAR may also serve as an integrin ligand, thereby mediating inflammatory response. In fact, osteopontin-induced cell mi- direct contact with adjacent cells (45). In the present study, gration may be dependent on uPA-uPAR activity (51). expression of the ␣v integrin chain was enhanced in response Despite the fact that uPAR deficiency dampened the inten- to UUO but it was most abundant on tubular cells. Expression sity of the interstitial inflammatory response to obstruction, the of the ␣v integrin chain by tubular cells was attenuated in the severity of fibrosis was worse in the uPAR-deficient mice. This uPAR-deficient mice at 14 d, perhaps an indication that, like outcome is likely due to the more aggressive myofibroblastic the leukocyte integrin CD11b/CD18, clustering of uPAR may response that developed in the uPAR-deficient mice. The de- induce ␣v integrin expression (46) or that the two receptors are velopment of a myofibroblastic phenotype in the renal inter- coordinately expressed (19). stitium has been highly predictive of renal functional deterio- Furthermore, uPA may cleave its receptor to release soluble ration due to fibrosis (52–54). Myofibroblasts are currently uPAR, a molecule with monocyte chemoattractant properties considered to be a major source of the matrix proteins that (47). It has recently been reported that this chemotactic re- accumulate in the kidney during fibrosis. The specific cellular sponse is triggered by interactions of soluble uPAR with the origin of interstitial myofibroblasts remains controversial but FPRL1/LXA4 receptor (formyl-methionyl-leucyl-leucyl-pro- possibilities included transformed resident interstitial fibro- line[fMLP]-like receptor-1/lipoxin A4 receptor) (48). Addi- blasts, transdifferentiated tubular epithelial cells, migratory tional studies will be necessary to determine if uPAR-integrin vascular cells and pericytes cells, circulating mesenchymal and/or soluble uPAR-FPRL1/LXA4 receptor interactions fully precursor cells, and perhaps even transformed monocytic cells explain why the interstitial recruitment of uPAR-deficient (3,55). Given the unknown origin of the interstitial myofibro- monocytes is impaired in obstructive uropathy. Several che- blasts, it is impossible to determine if uPAR plays a direct role mokines and adhesion molecules have been implicated in the in promoting or impairing their migration. Our in vivo obser- genesis of the interstitial inflammatory response to obstruction, vations argue against differences in myofibroblast proliferation including MCP-1 (49) and osteopontin (50). We cannot elim- and/or apoptosis as an explanation but this possibility should inate the possibility that the lower levels of osteopontin ob- be investigated more carefully in vitro. Of note is the in vivo J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1249

Figure 16. Interstitial collagen accumulation. (A through D) Photomicrographs illustrating picrosirius red staining of sham uPARϩ/ϩ and Ϫ/Ϫ kidneys (A and B, respectively) and day 14 UUO uPARϩ/ϩ and Ϫ/Ϫ kidneys (C and D, respectively). Magnification: ϫ400. The graph shows the Sirius red-positive tubulointerstitial area expressed as mean Ϯ 1 SD. Open bars are uPARϪ/Ϫ; closed bars are uPARϩ/ϩ.*P Ͻ 0.01, Ϫ/Ϫ versus ϩ/ϩ. ϩ P Ͻ 0.01, UUO versus sham. observation that uPAR deficiency does not affect the migration null mice may have resulted in the appearance of significantly of smooth muscle cells (56). Our results also suggest the more myofibroblasts as a secondary consequence. Parallel gen- possibility that impaired “scavenging” activities in the uPAR otype-dependent differences in the extent of interstitial angio- 1250 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

Figure 17. Schematic summary of the potential anti-fibrotic effects of uPAR. The uPAR binds both single-chain pro-urokinase and active two-chain uPA resulting in the generation of high pericellular proteolytic activity. Plasmin has multiple effects, including the activation of certain latent matrix-degrading metalloproteinases. Both uPA and uPAR exhibit direct monocyte chemoattractant properties. Although uPAR itself is anchored to the cell membrane via a glycosyl-phosphatidylinositol (GPI) moiety and lacks an intracellular domain, it frequently partners with a variety of integrin receptors to promote cellular adhesion to vitronectin matrices. In addition, the uPAR-integrin complexes may collaborate in the initiation of intracellular signaling reactions although the relevance of these effects to fibrosis remains unexplored. Through interactions with scavenger receptors such the LDL-receptor related protein (LRP) and the uPAR-associated protein (uPARAP), uPAR appears to facilitate endocytosis and degradation of pro-fibrotic molecules such as PAI-1. genesis, as defined by the number of cells expressing the Over-expression of PAI-1 is a feature of most progressive endothelial antigen CD34, and the observed co-localization of renal diseases (65). We have recently reported that genetic regions of neovascularization with interstitial myofibroblasts PAI-1 deficiency resulted in significantly fewer interstitial suggest a significant relationship between these two processes. myofibroblasts and decreased renal fibrosis in mice with ob- UPAR has been characterized as a scavenger receptor by structive nephropathy suggesting that the increased PAI-1 ac- virtue of its ability to work in collaboration with other scav- cumulation may be relevant to pathogenesis of the enhanced enger receptors, especially the LDL receptor-related protein fibrosis that was observed in the uPAR null mice (35). While (LRP) to delete “unneeded” molecules by endocytosis (57– decreased plasmin-dependent proteolysis may partially explain 60). LRP-independent internalization of uPAR ligands has also the pro-fibrotic effects of PAI-1 deficiency, PAI-1 may also been reported (61). This endocytotic pathway is the primary regulate the migration of fibroblasts along vitronectin matrices. route of elimination of extracellular PAI-1 (4,62). UPA and Vitronectin (or protein-S) is an adhesive protein that accumu- PAI-1 are subsequently degraded within lysosomes while lates within extracellular matrices during the course of injury uPAR is recycled to the cell surface. Significantly more PAI-1 and repair (66,67). In addition to certain integrin receptors, accumulated in the kidneys of the uPARϪ/Ϫ mice after UUO. including ␣v␤3, the classic vitronectin receptor, uPAR also has This finding, coupled with the fact that there was no difference a vitronectin-binding site (30,68). Vitronectin, the primary in renal PAI-1 mRNA levels between uPAR-deficient and PAI-1 binding protein, binds both uPAR-bound and matrix- wild-type mice (63), suggests a key role for uPAR in PAI-1 bound PAI-1. It has been suggested that vitronectin may func- protein turnover in the kidney. The physiologic internalization tion as a shuttle to facilitate PAI-1 transport to and phagocy- of the urokinase-PAI-1 complex is triggered by the interaction tosis by uPAR (66). In the present study the extent of of PAI-1 with a receptor belonging to the LRP family, and vitronectin accumulation in response to obstruction was greater involves the formation of a macro-quaternary structure of in the kidneys of the uPAR null mice, suggesting that the uPAR, uPA, LRP, and PAI-1 (4). In addition, an alternative delayed clearance of PAI-1 protein in these mice may be internalization process has also been described whereby uPAR related to the absence of a functional vitronectin-uPAR path- acts as the anchoring structure on the plasma membrane and way. While the mechanism that accounts for greater vitronectin LRP subsequently works as the endocytic trigger (64). accumulation cannot be addressed by this in vivo study, there J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1251 are reasons to speculate that it may be an indirect consequence 5. Rondeau E, Ochi S, Lacave R, He CJ, Medcalf R, Delarue F, of uPAR deficiency. Vitronectin is endocytosed and degraded Sraer JD: Urokinase synthesis and binding by glomerular epithe- by the ␣v␤5 integrin receptor (69). Not only does uPAR lial cells in culture. Kidney Int 36: 593–600, 1989 interact with ␣v␤5 integrin (18), it is also possible that, like 6. Nguyen G, Li X-M, Peraldi M-N, Zacharias U, Hage`ge J, Ron- ␣v␤3, it may be coordinately expressed with uPAR (70). deau E, Sraer J-D: Receptor binding and degradation of uroki- Increased PAI-1 accumulation may also be relevant to the nase-type plasminogen activator by human mesangial cells. Kid- ney Int 46: 208–215, 1994 differences in neovacularization observed between the uPAR 7. Almus-Jacobs F, Varki N, Sawdey MS, Loskutoff DJ: Endotoxin wild-type and deficient mice. In a study of transplanted ma- stimulates expression of the murine urokinase receptor gene in lignant keratinocytes, genetic PAI-1 deficiency was associated vivo. Am J Pathol 147: 688–698, 1995 with a less robust angiogenic response, resulting in less exten- 8. Xu Y, Hagege J, Mougenot B, Sraer JD, Rønne E, Rondeau E: sive local tumor invasion (71). Recent data suggest that the Different expression of the plasminogen activation system in angiogenic effects of PAI-1 are dependent on its ability to renal thrombotic microangiopathy and the normal human kidney. inhibit proteolytic activity rather than due to its interactions Kidney Int 50: 2011–2019, 1996 with vitronectin and integrins (72). 9. Shetty S, Kumar A, Johnson AR, Pueblitz S, Holiday D, Raghu The uPAR may also modify the rate of renal cell death by G, Idell S: Differntial expression of the urokinase receptor in apoptosis. Tubular cell apoptosis is currently considered to be fibroblasts from normal and fibrotic human lungs. Am J Respir a major pathway leading to tubular atrophy in progressive renal Cell Mol Biol 15: 78–87, 1996 disease. In the present study, uPAR deficiency resulted in the 10. Wagner SN, Atkinson MJ, Wagner C, Hofler H, Schmitt M, appearance of significantly more apoptotic tubular cells; at the Wilhelm O: Sites of urokinase-type plasminogen activator ex- same time proliferation-dependent tubular cell regeneration pression and distribution of its receptor in the normal human kidney. Histochem Cell Biol 105: 53–60, 1996 was blunted. Cultured human glioma cells exposed to uPAR 11. Vassalli JD, Baccino D, Belin D: A cellular binding site for the anti-sense have been reported to undergo more apoptotic cell Mr 55,000 form of the human plasminogen activator, urokinase. death, an observation that was associated with upregulated J Cell Biol 100: 86–92, 1985 expression of the pro-apoptotic gene BAX (23). Interactions 12. Preissner KT, Kanse SM, May AE: Urokinase receptor: A mo- ␣ ␤ between uPAR and the v 3 integrin may also enhance cell lecular organizer in cellular communication. Curr Opin Cell Biol survival via anti-apoptotic mechanisms (73). 12: 621–628, 2000 Changes in the cellular responses to ureteral obstruction 13. Florquin S, van den Berg JG, Olszyna DP, Claessen N, Opal SM, were not the only differences observed in mice lacking uPAR. Weening JJ, van der Poll T: Release of urokinase plasminogen As predicted, renal plasminogen activator activity was signif- activator receptor during urosepsis and endotoxemia. Kidney Int icantly decreased despite similar renal mRNA levels of the 59: 2054–2061, 2001 plasminogen activators and their known inhibitors (63). 14. Tang WH, Friess H, di Mola FF, Schilling M, Maurer C, Graber In summary, uPAR plays an important role in directing HU, Dervenis C, Zimmermann A, Buchler MW: Activation of changes in the cellular phenotype of tubulointerstitial cells that the serine proteinase system in chronic kidney rejection. Trans- is associated with the fibrogenic response to ureteral obstruc- plantation 65: 1628–1634, 1998 tion. Our data suggest that uPAR deficiency impairs monocyte/ 15. Ossowski L, Aguirre-Ghiso JA: Urokinase receptor and integrin partnership: Coordination of signaling for cell adhesion, migra- macrophage recruitment and diminishes scavenger receptor tion and growth. Curr Opin Cell Biol 12: 613–620, 2000 function resulting in delayed clearance of PAI-1 and vitronec- 16. Dear AE, Medcalf RL: The urokinase-type-plasminogen-activa- tin (Figure 17). As a consequence, ureteral obstruction induced tor receptor (CD87) is a pleiotropic molecule. Eur J Biochem in the absence of uPAR is characterized by a more intense 252: 185–193, 1998 myofibroblastic response, neovascularization, and tubular cell 17. Mondino A, Resnati M, Blasi F: Structure and function of the death resulting in more extensive renal destruction by fibrosis. urokinase receptor. Thromb Haemost 82[Suppl 1]: 19–22, 1999 18. Yebra M, Parry GCN, Stro¨mblad S, Mackman N, Rosenberg S, Acknowledgments Mueller BM, Cheresh DA: Requirement of receptor-bound This work was funded by grant support from the National Institutes urokinase-type plasminogen activator for integrin ␣v␤5-directed of Health DK-54500 (A.A.E.). Part of this work was previously cell migration. J Biol Chem 271: 29393–29399, 1996 published in abstract form (J Am Soc Nephrology 12:723A, 2001). 19. Khatib AM, Nip J, Fallavollita L, Lehmann M, Jensen G, Brodt P: Regulation of urokinase plasminogen activator/plasmin-medi- References ated invasion of melanoma cells by the integrin vitronectin 1. Eddy AA: Experimental insights into the tubulointerstitial dis- receptor alphaVbeta3. Int J Cancer 91: 300–308, 2001 ease accompanying primary glomerular lesions. J Am Soc Neph- 20. Chapman HA, Wei Y: Protease crosstalk with integrins: The rol 5: 1273–1287, 1994 urokinase receptor paradigm. Thromb Haemost 86: 124–129, 2. van Goor H, van der Horst ML, Fidler V, Grond J: Glomerular 2001 macrophage modulation affects mesangial expansion in the rat 21. Preissner KT, Kanse SM, Chavakis T, May AE: The dual role of after renal ablation. Lab Invest 66: 564–571, 1992 the urokinase receptor system in pericellular proteolysis and cell 3. Eddy AA: Molecular basis of renal fibrosis. Pediatr Nephrol 15: adhesion: Implications for cardiovascular function. Basic Res 290–301, 2000 Cardiol 94: 315–321, 1999 4. Vassalli JD, Sappino AP, Belin D: The plasminogen activator/ 22. Gyetko MR, Sud S, Kendall T, Fuller JA, Newstead MW, Stan- plasmin system. J Clin Invest 88: 1067–1072, 1991 diford TJ: Urokinase receptor-deficient mice have impaired neu- 1252 Journal of the American Society of Nephrology J Am Soc Nephrol 14: 1234–1253, 2003

trophil recruitment in response to pulmonary Pseudomonas mice deficient for components of the fibrinolytic system. Am J aeruginosa infection. J Immunol 165: 1513–1519, 2000 Pathol 157: 177–187, 2000 23. Kin Y, Chintala SK, Go Y, Sawaya R, Mohanam S, Kyritsis AP, 39. Kitching AR, Holdsworth SR, Ploplis VA, Plow EF, Collen D, Rao JS: A novel role for the urokinase-type plasminogen acti- Carmeliet P, Tipping PG: Plasminogen and plasminogen activa- vator receptor in apoptosis of malignant gliomas. Int J Oncol 17: tors protect against renal injury in crescentic glomerulonephritis. 61–65, 2000 J Exp Med 185: 963–968, 1997 24. Dewerchin M, Van Nuffelen A, Wallays G, Bouche´ A, Moons L, 40. Lan HY, Nikokic-Paterson DJ, Mu W, Atkins RC: Local mac- Carmeliet P, Mulligan RC, Collen D: Generation and character- rophage proliferation in the progression of glomerular and tubu- ization of urokinase receptor-deficient mice. J Clin Invest 97: lointerstitial injury in rat anti-GBM glomerulonephritis. Kidney 870–878, 1996 Int 48: 753–760, 1995 25. Kristensen P, Eriksen J, Blasi F, Dano K: Two alternatively 41. Simon DI, Wei Y, Zhang L, Rao NK, Xu H, Chen Z, Liu Q, spliced mouse urokinase receptor mRNAs with different histo- Rosenberg S, Chapman HA: Identification of a urokinase recep- logical localization in the gastrointestinal tract. J Cell Biol 115: tor-integrin interaction site. Promiscuous regulator of integrin 1763–1771, 1991 function. J Biol Chem 275: 10228–10234, 2000 26. Giachelli C, Bae N, Lombardi D, Majesky M, Schwartz S: 42. Sitrin RG, Pan PM, Blackwood RA, Huang J, Petty HR: Cutting Molecular cloning and characterization of 2B7, a rat mRNA edge: evidence for a signaling partnership between urokinase which distinguishes smooth muscle cell phenotypes in vitro and receptors (CD87) and L-selectin (CD62L) in human polymor- is identified to osteopontin (secreted phosphoprotein I, 2aR). phonuclear neutrophils. J Immunol 166: 4822–4825, 2001 Biochem Biophys Res Com 177: 867–873, 1991 27. Rollins BJ, Morrison ED, Stiles CD: Cloning and expression of 43. Carriero MV, Del Vecchio S, Capozzoli M, Franco P, Fontana L, JE, a gene inducible by platelet-derived growth factor and whose Zannetti A, Botti G, D’Aiuto G, Salvatore M, Stoppelli MP: product has cytokine-like properties. Proc Natl Acad Sci USA 85: Urokinase receptor interacts with alpha(v)beta5 vitronectin re- 3738–3742, 1988 ceptor, promoting urokinase-dependent cell migration in breast 28. Bijnens AP, Gils A, Knockaert I, Stassen JM, Declerck PJ: cancer. Cancer Res 59: 5307–5314, 1999 Importance of the hinge region between alpha-helix F and the 44. Xue W, Mizukami I, Todd RF 3rd, Petty HR: Urokinase-type main part of serpins, based upon identification of the epitope of plasminogen activator receptors associate with beta1 and beta3 plasminogen activator inhibitor type 1 neutralizing antibodies. integrins of fibrosarcoma cells: dependence on extracellular ma- J Biol Chem 275: 6375–6380, 2000 trix components. Cancer Res 57: 1682–1689, 1997 29. Hierck BP, Iperen LV, Gittenberger De Groot AC, Poelmann 45. Tarui T, Mazar AP, Cines DB, Takada Y: Urokinase-type plas- RE: Modified indirect immunodetection allows study of murine minogen activator receptor (CD87) is a ligand for integrins and tissue with mouse monoclonal antibodies. J Histochem Cytochem mediates cell-cell interaction. J Biol Chem 276: 3983–3990, 42: 1499–1502, 1994 2001 30. Deng G, Curriden SA, Wang S, Rosenberg S, Loskutoff DJ: Is 46. Sitrin RG, Pan PM, Harper HA, Todd RF, 3rd, Harsh DM, plasminogen activator inhibitor-1 the molecular switch that gov- Blackwood RA: Clustering of urokinase receptors (uPAR; erns urokinase receptor-mediated cell adhesion and release? CD87) induces proinflammatory signaling in human polymor- J Cell Biol 134: 1563–1571, 1996 phonuclear neutrophils. J Immunol 165: 3341–3349, 2000 31. Kim H, Oda T, Lopez-Guisa J, Wing D, Edwards DR, Soloway 47. Hoyer-Hansen G, Ronne E, Solberg H, Behrendt N, Ploug M, PD, Eddy AA: TIMP-1 deficiency does not attenuate interstitial Lund LR, Ellis V, Dano K: Urokinase plasminogen activator fibrosis in obstructive nephropathy. J Am Soc Nephrol 12: 736– cleaves its cell surface receptor releasing the ligand-binding 748, 2001 domain. J Biol Chem 267: 18224–18229, 1992 32. Eddy AA, Michael AF: Acute tubulointerstitial nephritis associ- 48. Resnati M, Pallavicini I, Wang JM, Oppenheim J, Serhan CN, ated with aminonucleoside nephrosis. Kidney Int 33: 14–23, Romano M, Blasi F: The fibrinolytic receptor for urokinase 1988 activates the G protein-coupled chemotactic receptor FPRL1/ 33. Zhang G, Oldroyd SD, Huang LH, Yang B, Li Y, Ye R, El Nahas LXA4R. Proc Natl Acad Sci USA 99: 1359–1364, 2002 AM: Role of apoptosis and Bcl-2/Bax in the development of 49. Diamond JR, Kees-Folts D, Ding G, Frye JE, Restrepo NC: tubulointerstitial fibrosis during experimental obstructive ne- Macrophages, monocyte chemoattractant peptide-1, and TGF-␤1 phropathy. Exp Nephrol 9: 71–80, 2001 in experimental hydronephrosis. Am J Physiol 226: F926–F933, 34. Konda R, Sato H, Sakai K, Sato M, Orikasa S, Kimura N: 1994 Expression of platelet-derived endothelial cell growth factor and its potential role in up-regulation of angiogenesis in scarred 50. Ophascharoensuk V, Giachelli CM, Gordon K, Hughes J, Pichler kidneys secondary to urinary tract diseases. Am J Pathol 155: R, Brown P, Liaw L, Schmidt R, Shankland SJ, Alpers CE, 1587–1597, 1999 Couser WG, Johnson RJ: Obstructive uropathy in the mouse: role 35. Oda T, Jung YO, Kim H, Cai x, Lopez-Guisa J, Ikeda Y, Eddy of osteopontin in interstitial fibrosis and apoptosis. Kidney Int 56: AA: PAI-1 deficiency attenuates the fibrogenic response to ure- 571–580, 1999 teral obstruction. Kidney Int 30: 587–596, 2001 51. Tuck AB, Hota C, Chambers AF: Osteopontin(OPN)-induced 36. Eddy AA: Role of cellular infiltrates in response to proteinuria. increase in human mammary epithelial cell invasiveness is uroki- Am J Kidney Dis 37[Suppl 2]: 525–529, 2001 nase (uPA)-dependent. Breast Cancer Res Treat 70: 197–204, 37. May AE, Kanse SM, Lund LR, Gisler RH, Imhof BA, Preissner 2001 KT: Urokinase receptor (CD87) regulates leukocyte recruitment 52. Alpers CE, Hudkins KL, Floege J, Johnson RJ: Human renal via beta 2 integrins in vivo. J Exp Med 188: 1029–1037, 1998 cortical interstitial cells with some features of smooth muscle 38. Swaisgood CM, French EL, Noga C, Simon RH, Ploplis VA: cells participate in tubulointerstitial and crescentic glomerular The development of bleomycin-induced pulmonary fibrosis in injury. J Am Soc Nephrol 5: 201–210, 1994 J Am Soc Nephrol 14: 1234–1253, 2003 uPAR Modulates Cellular Responses in Renal Fibrosis 1253

53. Goumenos DS, Brown CB, Shortland J, El Nahas AM: Myofi- 64. Ippoliti R, Lendaro E, Benedetti PA, Torrisi MR, Belleudi F, broblasts, predictors of progression of mesangial IgA nephropa- Carpani D, Soria MR, Fabbrini MS: Endocytosis of a chimera thy? Nephrol Dial Transplant 9: 1418–1425, 1994 between human pro-urokinase and the plant toxin saporin: an 54. Roberts ISD, Burrows C, Shanks JH, Venning M, McWilliam unusual internalization mechanism. FASEB J 14: 1335–1344, LJ: Intersitial myofibroblasts: Predictors of progression in mem- 2000 branous nephropathy. J Clin Pathol 50: 123–127, 1997 65. Jernigan SM, Eddy AA, eds: Experimental insights into the 55. Grimm PC, Nickerson P, Jeffery J, Savani RC, Gough J, McK- mechanisms of tubulo-interstitial scarring. In: Mechanisms and enna RM, Stern E, Rush DN: Neointimal and tubulointerstitial Clinical Management of Chronic Renal Failure, edited by El infiltration by recipient mesenchymal cells in chronic renal- Nahas M, Harris K, Anderson S, Oxford, Oxford University allograft rejection. N Engl J Med 345: 93–97, 2001 Press, 2000, pp 104–145 56. Carmeliet P, Moons L, Dewerchin M, Rosenberg S, Herbert JM, 66. Wei Y, Waltz DA, Rao N, Drummond RJ, Rosenberg S, Chap- Lupu F, Collen D: Receptor-independent role of urokinase-type man HA: Identification of the urokinase receptor as an adhesion plasminogen activator in pericellular plasmin and matrix metal- receptor for vitronectin. J Biol Chem 269: 32380–32388, 1994 loproteinase proteolysis during vascular wound healing in mice. 67. Reilly JT, Nash JR: Vitronectin (serum spreading factor): Its J Cell Biol 140: 233–245, 1998 localisation in normal and fibrotic tissue. J Clin Pathol 41: 57. Kounnas MZ, Henkin J, Argraves WS, Strickland DK: Low 1269–1272, 1988 density lipoprotein receptor-related protein/alpha 2-macroglobu- 68. Waltz DA, Natkin LR, Fujita RM, Wei Y, Chapman HA: Plas- lin receptor mediates cellular uptake of pro-urokinase. J Biol min and plasminogen activator inhibitor type 1 promote cellular Chem 268: 21862–21867, 1993 motility by regulating the interaction between the urokinase 58. Conese M, Nykjaer A, Petersen CM, Cremona O, Pardi R, receptor and vitronectin. J Clin Invest 100: 58–67, 1997 Andreasen PA, Gliemann J, Christensen EI, Blasi F: alpha-2 69. Panetti TS, McKeown-Longo PJ: The alpha v beta 5 integrin Macroglobulin receptor/Ldl receptor-related protein (Lrp)- de- receptor regulates receptor-mediated endocytosis of vitronectin. pendent internalization of the urokinase receptor. J Cell Biol 131: J Biol Chem 268: 11492–11495, 1993 1609–1622, 1995 70. Nip J, Rabbani SA, Shibata HR, Brodt P: Coordinated expression 59. Czekay RP, Kuemmel TA, Orlando RA, Farquhar MG: Direct of the vitronectin receptor and the urokinase- type plasminogen Binding of Occupied Urokinase Receptor (uPAR) to LDL Re- ceptor- related Protein Is Required for Endocytosis of uPAR and activator receptor in metastatic melanoma cells. J Clin Invest 95: Regulation of Cell Surface Urokinase Activity. Mol Biol Cell 12: 2096–2103, 1995 1467–1479, 2001 71. Bajou K, Noel A, Gerard RD, Masson V, Brunner N, Holst- 60. Herz J, Strickland D: LRP: a multifunctional scavenger and Hansen C, Skobe M, Fusenig NE, Carmeliet P, Collen D, signaling receptor. J Clin Invest 108: 779–784, 2001 Foidart JM: Absence of host plasminogen activator inhibitor 1 61. Rajagopal V, Kreitman RJ: Recombinant toxins that bind to the prevents cancer invasion and vascularization. Nat Med 4: urokinase receptor are cytotoxic without requiring binding to the 923–928, 1998 alpha(2)-macroglobulin receptor. J Biol Chem 275: 7566–7573, 72. Bajou K, Masson V, Gerard RD, Schmitt PM, Albert V, Praus M, 2000 Lund LR, Frandsen TL, Brunner N, Dano K, Fusenig NE, Weidle 62. Vilhardt F, Nielsen M, Sandvig K, van Deurs B: Urokinase-type U, Carmeliet G, Loskutoff D, Collen D, Carmeliet P, Foidart JM, plasminogen activator receptor is internalized by different mech- Noel A: The plasminogen activator inhibitor PAI-1 controls in anisms in polarized and nonpolarized Madin-Darby canine kid- vivo tumor vascularization by interaction with proteases, not ney epithelial cells. Mol Biol Cell 10: 179–195, 1999 vitronectin. Implications for antiangiogenic strategies. J Cell Biol 63. Zhang G, Kim H, Cai X, Lo´pez-Guisa JM, Alpers CE, Liu Y, 152: 777–784, 2001 Carmeliet P, Eddy AA: Urokinase receptor deficiency acceler- 73. Brassard DL, Maxwell E, Malkowski M, Nagabhushan TL, ates renal fibrosis in obstructive nephropathy. J Am Soc Nephrol Kumar CC, Armstrong L: Integrin alpha(v)beta(3)-mediated ac- 14: 1254–1271, 2003 tivation of apoptosis. Exp Cell Res 251: 33–45, 1999