J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin-Mediated Collagen Matrix Remodeling by Rat Mesangial Cells Is Differentially Regulated by Transforming Growth Factor- and Platelet-Derived Growth Factor-BB
SHOJI KAGAMI,* SHUJI KONDO,* KLEMENS LOSTER,¨ † WERNER REUTTER,† TAKASHI KUHARA,* KOJI YASUTOMO,* and YASUHIRO KURODA* *Department of Pediatrics, School of Medicine, University of Tokushima, Tokushima, Japan; and †Institut fu¨r Molekularbiologie und Biochemie, Freie Universitat Berlin, Berlin-Dahlem, Germany.
Abstract. Pathologic remodeling of mesangial matrix after glo- mRNA analysis experiments demonstrated that TGF-, but not merular injury is the central biologic feature of glomerular PDGF-BB, increases the expression of ␣11 integrin in mes- scarring (sclerosis). Transforming growth factor- (TGF-) angial cells cultured on plastic surface and in collagen gels. and platelet-derived growth factor (PDGF)-BB have been im- The upregulation of ␣11 integrin expression by TGF- cor- plicated in the development of glomerular scarring in rat and related with increases in gel contraction and collagen-depen- human glomerulonephritis. To clarify molecular and cellular dent adhesion but not migration of mesangial cells. On the mechanisms involved in abnormal mesangial remodeling, this other hand, PDGF-BB enhanced MC-mediated gel contraction study focused on the role of ␣11 integrin, a collagen/laminin and migration without affecting cell adhesion to collagen I. receptor, in rat mesangial cells, using collagen gel contraction Growth factor-induced collagen-dependent adhesion, migra- as an experimental model of in vivo collagen matrix remodel- tion, and gel contraction were significantly attenuated by in- ing and scar formation. In addition, the influence of TGF- and cubation with anti-␣1, 1 subunit antibodies. Thus, these data PDGF-BB on mesangial cell (MC)-mediated collagen gel con- indicate that ␣11 integrin-mediated collagen matrix remod- traction in association with the ␣11 integrin expression was eling can be modulated by TGF- and PDGF-BB via different evaluated. Integrin function blocking studies using anti-␣1, 1 mechanisms. ␣11 integrin-mediated mesangial matrix remod- subunit antibodies indicated that MC-␣11 integrin is essen- eling induced by TGF- or PDGF-BB may be a pathogenic tially required not only for collagen-dependent adhesion/mi- mechanism leading to glomerular scarring. gration, but also for gel contraction. Protein synthesis and
Abnormal mesangial extracellular matrix (ECM) remodeling associated ␣ and  subunits, mainly mediate cell/ECM inter- after glomerular injury is characterized by increased expression actions. 1 integrins may play a critical role in cell prolifera- of ECM components such as fibronectin, laminin, and collagen tion, migration, differentiation, and matrix assembly, and thus type IV (collagen IV) and/or neoexpression of collagen I and participate in tissue development and remodeling after injury (7). III. In addition, it is the central biologic feature of progressive Recently, transforming growth factor- (TGF-) and plate- glomerular diseases leading to glomerular dysfunction and let-derived growth factor (PDGF) have been recognized to be scarring (sclerosis) (1–4). One current concept emerging from key mediators in the development of glomerular scarring in molecular cell biologic studies in progressive kidney diseases experimental and human kidney diseases (1–4). In this context, is that this pathologic mesangial ECM remodeling is caused by 1 integrins are effective biologic targets to modulate various uncontrolled interaction between mesangial cells, ECM, and growth factor actions (8,9). In fact, our previous experimental growth factors (5,6). studies using a rat model of acute glomerulonephritis (GN) and  A large body of studies demonstrated that 1 integrins, a mesangial cell (MC) culture system suggested that enhanced family of cell surface receptors consisting of noncovalently expression of glomerular TGF- concomitantly increases the expression of 1 integrins (␣11, ␣51 integrins) and ECM components (fibronectin, collagen I, IV) and thereby contrib- Received March 16, 1998. Accepted September 24, 1998. utes to pathologic mesangial ECM accumulation (2,10). In This work was presented in part at the 29th annual meeting of the American Society of Nephrology, November 3 to 6, 1996, and has appeared in abstract addition, our immunohistologic study of human GN found form (J Am Soc Nephrol 7: A2550, 1996). significant changes of glomerular 1 integrins in IgA nephrop- Correspondence to Dr. Shoji Kagami, Department of Pediatrics, School of athy and lupus GN that paralleled both the severity of the Medicine, University of Tokushima, Kuramoto-cho-3-chome, Tokushima 770- glomerular lesions such as ECM deposition and cellularity, and 8503, Japan. Phone: 011 81 886 31 3111; Fax: 011 81 886 31 8697; E-mail:   [email protected] glomerular synthesis of TGF- (11). However, 1 integrin- mediated MC action in abnormal mesangial ECM build-up 1046-6673/1004-0779$03.00/0 Journal of the American Society of Nephrology after injury remains obscure. Copyright © 1999 by the American Society of Nephrology Many types of cells including MC are able to reorganize 780 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 779–789, 1999 collagen fibers and to perform gel contraction when cultured in Japan). A rabbit polyclonal antiserum (AB1936) reactive with the three-dimensional collagen I gels (12–17). This system has cytoplasmic domains of ␣2 subunit was purchased from Chemicon been extensively used as a model of ECM remodeling similar International (Temecula, CA) (26). Function-blocking monoclonal ␣ to wound healing and pathologic scarring (12–14). Gel con- mouse anti-rat 1 subunit antibody (mAb 33.4) (IgG1) and polyclonal   traction involves multiple cellular properties including forma- rabbit anti-rat 1 subunit IgG (pAb anti- 1) were produced and characterized as described previously (27). Each antibody is directed tion of contract sites to collagen fibers, reorganization of the to the extracellular domain of 1or␣1 integrin subunit, respectively, network formed by collagen fibrils, and cell migration within and can disrupt MC adhesion to ECM proteins with ligand specificity collagen lattice (14–17). Several studies have clearly docu- (10). A mouse myeloma IgG1(Zymed Laboratories, San Francisco, mented that the collagen and laminin receptors—the ␣11 and CA), normal rabbit serum, and rabbit IgG (R&D Systems, Minneap- ␣21 integrins—are necessary components in the process of olis, MN) were used as control antibodies. collagen matrix reorganization and collagen gel contraction (18–24). Of the two, the ␣11 integrin was identified as the Immunohistochemistry major collagen binding receptor of rat glomerular mesangial Immunofluorescence (IF) analysis was performed to examine the cells in vivo (2,24). expression of ␣1, ␣2, 1 integrin subunits in mesangial cells cultured To investigate the potential role of ␣11 integrin in abnor- on plastic or in gels, respectively. Briefly, mesangial cells cultured on mal mesangial matrix remodeling observed in progressive glo- multichamber slides (Nunc, Naperville, IL) were washed with cold merular diseases, we employed an in vitro collagen gel system phosphate-buffered saline (PBS), then fixed and permeabilized with using isolated rat mesangial cells and focused on how this 2% paraformaldehyde, 0.1% Triton X-100 for 10 min. Cells were  ␣ ␣ integrin contributes to the contraction process. Additional ex- incubated with a rabbit antiserum against 1, 1, or 2 subunit for 1 h periments have been performed to study the effect of two at 37°C, followed by FITC-coupled donkey anti-rabbit IgG antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) for1hat representative fibrogenic growth factors, TGF- and PDGF- ␣  ␣  37°C. After washing, the slides were coverslipped and examined BB, on MC- 1 1 integrin expression and on 1 1 integrin- under a fluorescence microscope. For IF examination of collagen I gel mediated collagen gel contraction. cultures, the gels were removed from the 24-well plates after culture This report shows that ␣11 integrin is required for collag- for 24 h and snap-frozen in OCT embedding compound (Miles, Inc., en-dependent MC adhesion, migration, and gel contraction. Elkhart, IN). Cryostat sections (5 m) were placed on albumin-coated The ability of TGF- to stimulate gel contraction was found to slides, then fixed and stained as described above. Some sections were be dependent on increased ␣11 integrin synthesis, whereas double-stained with a combination of monoclonal mouse anti-rat ␣1 PDGF-BB enhancement of gel contraction and cell migration subunit antibody (mAb) 33.4 and polyclonal rabbit anti-rat 1 subunit was revealed to be dependent on ␣11 integrin activity. IgG (pAb anti-1)for1hat37°C, followed by a combination of FITC-coupled donkey anti-mouse IgG antibody and tetramethyl-rho- damine isothiocyanate-coupled donkey anti-rabbit IgG antibody Materials and Methods (Jackson ImmunoResearch Laboratories) for1hat37°C. Cell Culture Rat mesangial cells were obtained from intact glomeruli of 4- to 35S-Methionine Labeling of Mesangial 6-wk-old Sprague Dawley rats and characterized according to pub- Cells and Immunoprecipitation lished methods (25). Mesangial cells were used between passages 5 Mesangial cells on monolayer or in gel culture were metabolically and 9 and were maintained in RPMI 1640 (Sigma Chemical Co., St. 35 labeled with S-methionine for 18 h in methionine-free medium. Louis, MO) supplemented with 18% fetal bovine serum (Life Tech- After metabolic labeling, collagen gels were immersed in 25 mM nologies, Grand Island, NY), 100 U/ml penicillin, 100 g/ml strep- Tris-HCl, pH 7.4, containing 100 mM N-octyl--D-glycopyranoside tomycin, 0.1 U/ml insulin, and 25 mM Hepes buffer at 37°C in a 5% (Sigma), 150 mM NaCl, 1 mM CaCl , and 1 mM MgCl (lysis CO incubator. All experiments were performed after a 2-d incubation 2 2 2 buffer), minced into fine pieces, repeatedly aspirated into syringe, and in serum-free RPMI 1640 to reach mesangial cells quiescence (25). forced through needles with decreasing diameter. Monolayer cultures Cell proliferation and protein synthesis assays were performed as were also lysed in lysis buffer. Radiolabeled protein in cell lysates was described previously (25). counted, and a standard amount of radioactivity was used in the immunoprecipitation analysis. The lysate supernatants (5 ϫ 106 cpm Extracellular Matrix Proteins, Growth of 35S-labeled proteins/ml) precleared by incubation with normal Factors, and Antibodies rabbit serum and protein A-Sepharose (Pharmacia LKB Biotechnol-  Fibronectin was purified from human plasma by affinity chroma- ogy Inc.) were immunoprecipitated with rabbit antisera against 1, ␣ ␣ tography on gelatin-Sepharose 4B (Pharmacia LKB Biotechnology, 1, 2 integrin subunits for 16 h at 4°C. Immune complexes were Uppsala, Sweden) (10). Rat tail collagen type I (collagen I) and mouse recovered by binding to protein A-Sepharose. Beads were washed laminin were obtained from Collaborative Biomedical Products (Bed- with lysis buffer, and immunoprecipitates were analyzed by 4 to 12% ford, MA). Human placental collagen IV was purchased from Sigma SDS-polyacrylamide gel electrophoresis under nonreducing condi- Chemical Co. Purity of these matrix proteins was verified by sodium tions as described previously (2). Integrin bands were quantified by dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. Recombi- densitometeric scanning using an LKB UltroScan XL apparatus nant human TGF- was generously supplied from Kirin Co. (Tokyo, (Pharmacia LKB Biotechnology). Japan) (10). Recombinant human PDGF-BB was purchased from Boehringer Mannheim Biochemical (Indianapolis, IN). Rabbit anti- Northern Blot Analysis sera reactive with the cytoplasmic domains of ␣1 and 1 integrin The total cellular RNA was isolated from mesangial cells in mono- subunits were kind gifts from Dr. K. Iwamoto (University of Osaka, layer or gel culture using RNAzol (Biotecx Laboratories, Houston, J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin in Collagen Matrix Remodeling 781
Figure 1. Analysis of ␣11 and ␣21 integrins expressed on rat mesangial cells either on plastic (PL) or in collagen gels (COL). Lysates of 35S-methionine-labeled mesangial cells were immunoprecipitated with rabbit antisera to the 1 subunit (lane 1), the ␣1 subunit (lane 2), and the ␣2 subunit (lane 3). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed under nonreducing conditions. The lines to the right indicate the position of the 180-kD ␣1 subunit, the 135-kD ␣2 subunit, and the 115-kD 1 subunit. Molecular weight markers are shown on the left.
Figure 2. Immunolocalization of ␣11 and ␣21 integrins expressed on rat mesangial cells cultured in collagen gels. Mesangial cells cultured in collagen gels were either double-stained with the combination of monoclonal anti-rat ␣1 subunit antibody (mAb, 33.4) (a) and polyclonal rabbit anti-1 subunit antibody (pAb anti-1) (b), or stained only with polyclonal rabbit anti-␣2 subunit antiserum (AB1936) (c) at 24 h of culture. Double immunostaining showed a widespread colocalization of ␣1 and 1 integrin subunits along the cell membrane of spindle-shaped mesangial cells (a and b). The ␣2 subunit was only faintly seen with same distribution (c). Magnification, ϫ400.
TX). For Northern analysis, RNA was denatured and fractionated by inson Labware, Lincoln Park, NJ) were coated overnight at 4°C electrophoresis through a 1.2% agarose gel (5 g/lane) and transferred with either collagen I, collagen IV, or laminin (10 g/ml, respec- to a Hybond nylon membrane (Amersham Corp., Paisley, United tively) and then blocked with 10 mg/ml bovine serum albumin Kingdom). Nucleic acids were immobilized by ultraviolet irradiation (BSA) for1hat37°C. Quiescent mesangial cells were treated with (Stratagene, La Jolla, CA) and hybridized with 1 ϫ 106 cpm of or without TGF- or PDGF-BB for 24 h, subsequently detached, 32P-labeled cDNA probes/ml in hybridization buffer (50% form- washed, and pelleted. Cells were suspended in RPMI 1640 con- ϫ amide, 19% Denhardt’s solution, 0.1% SDS, 5 SSC, 200 g/ml taining 1 mM MgCl2 and 1 mM CaCl2, plated onto the ECM- denatured salmon sperm DNA). The cDNA probes were radiolabeled coated wells at 104 cells/well and incubated at 37°C for 50 min. with 32P-dCTP by random oligonucleotide priming (Boehringer Cell adhesion was evaluated by staining adherent cells with 0.1% Mannheim Biochemicals, Indianapolis, IN). The cDNA probes used crystal violet, solubilizing cells with 0.2% Triton X-100 in PBS, in this study included a mouse 1 integrin subunit cDNA (28) and a and reading the absorbance at 630 nm with a Corona Microplate rat ␣1 integrin subunit cDNA (29). To normalize for signal intensity Reader (Corona Electric Co., Ibaragi, Japan). This procedure of of the cDNA probes, a rat glyceraldehyde-3-phosphate dehydrogenase quantification was demonstrated to give values proportional to the (GAPDH) cDNA was used (30). The hybridization product was number of adherent cells (10). When the effect of anti-␣1 subunit visualized by autoradiography. Densitometric scans were performed antibody (mAb, 33.4) or anti-1 subunit antibody (pAb anti-1) on from autoradiographs to quantify the intensity of hybridization. MC adhesion was assessed, antibodies were incubated with mes- angial cells for 30 min at 20°C before adding them to ECM Cell Adhesion Assay protein-containing wells. Adhesion studies by mesangial cells were performed to examine whether ␣11 integrin functions as a collagen/laminin receptor for Collagen Gel Contraction Assay cell adhesion to collagen I, collagen IV, or laminin as described Collagen gel assays were performed to examine the ability of previously (10). Briefly, 96-well flat-bottom plates (Becton Dick- mesangial cells to reorganize and contract three-dimensional collagen 782 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 779–789, 1999
only cells that had migrated through the membrane remained. The membrane was fixed with methanol, stained with Diff-Quik (Baxter, McGaw Park, IL), and excised from the plastic supports and allowed to air-dry on a glass slide with Permount (Sigma Chemical Co.). Cells were counted at ϫ400 magnification in 10 microscope fields per filter. Background cell numbers due to migration to wells containing no collagen were subtracted from those values that reflect collagen- dependent migration. Assays were performed in triplicate and values presented are the means Ϯ SD of 30 fields from three filters.
Statistical Analyses Results are presented as mean Ϯ SD. A value of P Ͻ 0.05 was used to determine statistical significance (t test). Triplicate wells were analyzed for each experiment, and each experiment was performed independently a minimum of three times.
Results Characterization of Collagen Binding Integrins of Figure 3. Effects of anti-1 subunit antibody (pAb anti-1) and Mesangial Cells Cultured on Plastic or in anti-␣1 monoclonal antibody (mAb) 33.4 on mesangial cells induced Collagen Gels collagen gel contraction. The degree of gel contraction by mesangial To determine which of the collagen-binding integrins con- cells in the presence of pAb anti-1 (200 g/ml) or mAb 33.4 (50 taining the 1 subunit play an important role for cell adhesion g/ml) was compared to that in the presence of appropriate control to collagens and during collagen gel contraction, we first antibodies (normal rabbit IgG (Rab IgG) or mouse IgG1 (MIgG1)) at examined MC expression of ␣11or␣21 integrin by IF and 48h(P Ͻ 0.01 versus control). immunoprecipitation. As shown in Figure 1, mesangial cells synthesize mainly ␣11 integrin (180/115 kD) when cultured on plastic, whereas synthesis of the ␣21 integrin (135/115 I gels. Quiescent mesangial cells were harvested and suspended in the absence or presence of anti-1 (pAb anti-1) or anti-␣1 integrin kD) is comparatively low. This pattern of integrin subunit subunit antibody (mAb 33.4) at a concentration of 5 ϫ 105 cells/ml in expression is not changed when mesangial cells are grown 0.5 ml of 1.25ϫ RPMI 1640 for 30 min at 20°C. The cell suspension within collagen gels, indicating that the ␣11 integrin is the was mixed on ice with 0.5 ml of collagen I solution (59% 1.25ϫ major collagen receptor of mesangial cells. The quantitative RPMI 1640, 40% rat tail collagen [3.75 mg/ml], 1% 0.2N NaOH). difference in integrin subunit expression is confirmed by IF. Collagen/cell suspensions (500 l each) were incubated in 24-well Consistent with protein synthesis, ␣1 and 1 integrin subunits plates (Costar, Cambridge, MA) at 37°C to polymerize the collagen. were strongly expressed, but ␣2 integrin subunit was weakly The diameter of the hydrated gels was measured by use of an inverted expressed on mesangial cells cultured on plastic, as reported microscope at 24 and 48 h. To examine the effects of growth factors previously (24). At 24 h after gel formation, ␣1 and 1  on gel contraction, TGF- or PDGF-BB was added to collagen/cell subunits were abundantly colocalized on mesangial cells with suspension before gel polymerization. an elongated shape in collagen gels, whereas ␣2 subunit on mesangial cells was only faintly positive and showed a scat- Migration Assays tered distribution (Figure 2). Both findings support our previ- Motility assays were performed using a modification of the assay ous study showing that mesangial cells use predominantly as described previously (31). Transwell chambers (6.5 mm; Costar) ␣  with an 8-m pore polycarbonate membrane were used. The under- 1 1 integrin for cell adhesion to collagen I and laminin, and ␣  side of the polycarbonate membrane was coated overnight at 4°C with that 2 1 integrin plays a relatively minor role in MC adhesion 20 g/ml collagen I. Control wells contained no collagen I. to these matrix components (10). The application of function- The lower chamber was rinsed twice with RPMI 1640–0.1% BSA. blocking antibodies against ␣1, 1 integrin subunits in colla- Growth-arrested mesangial cells were detached, washed, and sus- gen gel contraction assays was further used to elucidate the pended in RPMI 1640–0.1% BSA. To assess the effect of antibodies functional significance of ␣11 integrin in gel contraction. on cell migration, mesangial cells were diluted to 50 ϫ 105 cells/ml This assay system has been proposed to mimic in vivo matrix and incubated with either control rabbit IgG (200 g/ml), rabbit reorganization during wound healing or scarring (12–14).   anti- 1 subunit IgG (pAb anti- 1) (200 g/ml), control mouse IgG1 When collagen gels were supplemented with rabbit polyclonal (50 g/ml) or mAb 33.4 (50 g/ml) for 30 min at 20°C just before anti-1 subunit antibody (pAb anti-1) (200 g/ml), gel con- loading the chamber. Cells were added to the upper chamber of the traction by mesangial cells was completely abolished at 24 h Transwell at 5 ϫ 105 cells/well. A total of 600 l of RPMI 1640– 0.1% BSA containing the same amount of control rabbit IgG, pAb and this effect continued for more than 48 h (Figure 3). A anti-1, control mouse IgG1, or mAb 33.4 was added to the lower weaker but comparable effect was produced by application of chamber. In some assays, the indicated amount of either PDGF-BB or the ␣1 subunit-specific monoclonal antibody (mAb) 33.4 (50 TGF- was included in the lower chamber. After6hofincubation, g/ml), which also reduces the ability of mesangial cells to mesangial cells on the upper surface were removed by scraping so that contract collagen gels (P Ͻ 0.01). Taken together, these find- J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin in Collagen Matrix Remodeling 783 ings indicate that ␣11 integrin acts as a functional collagen time-dependent manner when mesangial cells were grown in receptor in collagen matrix reorganization. monolayer cultures. Here, we investigated in parallel experi- ments the effect of PDGF-BB on ␣11 integrin synthesis Effects of TGF- and PDGF-BB on ␣11 Integrin compared with TGF-. As shown in Figure 4, A and B, Expression by Mesangial Cells Cultured on PDGF-BB did not influence mRNA expression of either inte- Plastic and in Collagen Gels grin subunit at any time point (0 to 48 h) or at any dose (0 to To assess whether MC-␣11 integrin is a target for fibro- 100 ng/ml) examined. Corresponding to these results, there is genic growth factors, we examined the influence of TGF- and no change in protein synthesis of ␣11 integrin with respect to PDGF-BB on the expression of ␣11 integrin at the protein PDGF-BB treatment (Figure 4C). In collagen gels, TGF- and and mRNA levels. As we have reported previously (10), PDGF-BB treatment showed similar changes in ␣11 integrin TGF- enhances the expression of ␣11 integrin in a dose- and expression as observed on plastic. TGF- increased the protein
Figure 4. Influence of platelet-derived growth factor-BB (PDGF-BB) on ␣1 and 1 integrin subunit mRNA expression and protein synthesis in mesangial cells cultured on plastic (A through C). (A) Quiescent mesangial cells were incubated with the indicated doses of PDGF-BB for 24 h. Total RNA (5 g/lane) was isolated and subjected to Northern blot hybridization using 32P-labeled cDNA (1 ϫ 106 cpm/ml in hybridization buffer) specific for 1 and ␣1 integrin subunits. mRNA expression of GAPDH was used for signal normalization. (B) Quiescent mesangial cells were exposed to PDGF-BB (50 ng/ml), and total RNA was extracted at the indicated times. Time point 0 represents the control. Total RNA (5 g/lane) was analyzed by Northern blot as described above. (C) Quiescent mesangial cells were incubated with the indicated doses of PDGF-BB for 24 h, and in the last 18 h, mesangial cells were radiolabeled with 35S-methionine. Cell lysates (5 ϫ 106 cpm of 35S-labeled proteins/ml) were immunoprecipitated with rabbit antiserum to the ␣1 subunit, and immunoprecipitates were resolved by SDS-PAGE under nonreducing conditions. 784 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 779–789, 1999 and mRNA expression of ␣11 integrin on mesangial cells in creases not only cell adhesion to collagen I, collagen IV, or a dose-dependent manner (Figure 5), whereas PDGF-BB did laminin as previously shown (10), but also stimulates gel not have any influence on ␣11 integrin expression by mes- contraction by mesangial cells (P Ͻ 0.01, respectively) (Fig- angial cells in gels (data not shown). ures 6 and 7A). Both findings reflected the levels of MC- ␣11 integrin expression on plastic or in gels as we reported previ- Effects of TGF- and PDGF-BB on MC Adhesion to ously (10), and as shown in Figure 5. The polyclonal anti-1 ECM Substrates and MC-Mediated Collagen antibody (pAb anti-1) almost completely blocked TGF-- Gel Contraction induced collagen gel contraction and significantly inhibited We next investigated the functional significance of MC- TGF--induced MC adhesion to ECM components. The an- ␣11 integrin expression on cell adhesion and gel contraction ti-␣1 integrin subunit antibody (mAb 33.4) also strongly after stimulation with TGF- or PDGF-BB, using function- blocked both of these two growth factor-induced phenomena, blocking anti-integrin antibodies. TGF- significantly in- suggesting that TGF- stimulates cell adhesion to collagen and laminin and gel contraction by increasing the number of ␣11 integrin on the cell surface (Figure 7A and Figure 8, A and B). On the other hand, PDGF-BB had no significant effect on MC adhesion to these ECM substrates but enhanced gel contraction by mesangial cells without the accompanying changes in ␣11 integrin expression observed with TGF- (Figure 7B and Fig- ure 8, C and D). Enhanced gel contraction by PDGF-BB was almost completely blocked with pAb anti-1 incubation and was significantly inhibited using mAb 33.4, suggesting that PDGF-BB-induced gel contraction is also dependent on ␣11 integrin activity (Figure 8, C and D).
Effects of TGF- and PDGF-BB on MC Migration to Collagen I The observation that PDGF-BB-enhanced capacity of mes- angial cells to contract collagen gels does not indicate an increased rate in ␣11 integrin synthesis suggests that PDGF-BB may act via stimulation of ␣11 integrin-mediated migration of mesangial cells in collagen gels. The importance of cell migration for gel contraction and the stimulatory effect
Figure 5. Influence of transforming growth factor- (TGF-)on␣1 and 1 integrin subunit mRNA expression and protein synthesis in collagen gel cultured mesangial cells (A and B). (A) Mesangial cells in collagen gel were incubated with the indicated doses of TGF- for 24 h. Total RNA was isolated and subjected to Northern blot hybrid- ization using 32P-labeled cDNA (1 ϫ 106 cpm/ml in hybridization Figure 6. Effect of TGF- on collagen gel contraction by mesangial buffer) specific for 1, ␣1 integrin subunits. (B) Mesangial cells in cells. Mesangial cells cultured in collagen gel were treated with collagen gel were incubated with the indicated doses of TGF- for indicated doses of TGF- for 24 h. The degree of gel contraction by 24 h, and in the last 18 h, cells were radiolabeled with 35S-methionine. TGF--treated cells was compared with that by untreated cells (con- Immunoprecipitation was performed similarly to Figure 4. trol) (P Ͻ 0.01 versus control). J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin in Collagen Matrix Remodeling 785
These findings substantially support the important role of ␣11 integrin in collagen-dependent migration by mesangial cells stimulated with or without PDGF-BB. Lack of complete inhi- bition of cell migration toward collagen I by mAb 33.4 leaves open the possibility that 1 containing integrins other than ␣11 are involved in migration processes. TGF- (10 ng/ml) did not influence cell migratory activity under these experi- mental conditions (Figure 9).
Discussion In this study, we used a collagen gel contraction assay as a model system for studying whether MC-␣11 integrin could mediate collagen matrix remodeling, a process that has impor- tance in glomerular wound healing and irreversible scarring (sclerosis) after injury (5,6,34–36). Persistent, abnormal mes- angial reorganization of collagen I, III, and IV is a prominent feature in sclerotic lesions of several progressive glomerular diseases. Mesangial cells are key producers of these mesangial matrix proteins, suggesting that MC may play a significant role in pathologic collagen matrix remodeling (1–6,34–36). The ␣11 integrin investigated in this study is a collagen-binding receptor expressed on rat glomerular mesangial cells and has been reported to be increased with enhanced expression of its ligands (collagens I and IV) in rat and human glomerulone- phritis (2,11). Immunofluorescence, metabolic labeling followed by immu- noprecipitation, and functional studies of cultured rat mesan- gial cells with function-blocking anti-␣1,1 subunit antibodies clearly indicated that ␣11 integrin is an important collagen- binding receptor in vitro and works as a functional molecule to mediate collagen-dependent adhesion/migration and collagen matrix contraction by mesangial cells. Gel contraction by mes- Figure 7. Effects of rabbit anti-1 subunit IgG (pAb anti-1) and angial cells was completely inhibited by the incubation with anti-␣1 monoclonal antibody (mAb) 33.4 on TGF-- or PDGF-BB- anti-1 subunit antibody (pAb anti-1), but this complete treated mesangial cell adhesion to collagen I (COL I), collagen IV inhibition could not be achieved by anti-␣1 subunit antibody (COL IV), or laminin (LM) (A and B). Mesangial cells were treated (mAb 33.4) alone even when administered at high concentra- with either TGF- (10 ng/ml) (A), PDGF-BB (50 ng/ml) (B), or tions (200 g/ml). This suggests that collagen-binding recep- vehicle (control) for 24 h, then detached and preincubated with either tors other than the ␣11 integrin may also contribute to this  rabbit IgG (Rab IgG, 200 g/ml), pAb anti- 1(200 g/ml), mouse process. Indeed, several reports have shown that another col- IgG1 (MIgG1, 50 g/ml) or mAb 33.4 (50 g/ml) before adding them lagen binding receptor, the ␣21 integrin, is an essential re- to extracellular matrix-coated wells. Each attachment of TGF-- quirement for gel contraction by many cell types, including treated (A) or PDGF-BB-treated (B) cells preincubated with anti- fibroblasts, melanoma cells, and osteoblasts; thus, it has been integrin antibody or nonimmune IgG was appropriately compared ␣  with nonstimulated, control cell adhesion preincubated with Rab IgG proposed in these cell types that the 2 1 integrin is the or MIgG1. *P Ͻ 0.01 versus control. primary integrin in collagen matrix remodeling (19–21). How- ever, these reports did not adequately address the substantial role of ␣11 integrin in gel contraction because the level of of PDGF on cell migration have been shown in several cell ␣11 expression on these cells is very low, and function- culture studies (15,16,32,33). Migration assays using function- blocking anti-␣1 subunit antibodies were not used (19–21). blocking antibodies were introduced to test this possibility. In Using rat cardiac fibroblasts expressing both ␣11 and ␣21 these assays, mesangial cells were stimulated with growth integrins almost at an equal level, Carver et al. recently re- factors for 6 h after the cells were plated on collagen I-coated ported that ␣11 integrin is, at least in part, responsible for wells. As shown in Figure 9, mesangial cells stimulated with collagen reorganization and gel contraction (22). Furthermore, PDGF-BB (50 ng/ml) showed significant increase of migratory Gotwals et al. demonstrated the substantial role of ␣11 inte- responses toward collagen I compared with control, nonstimu- grin in gel contraction and cell migration to collagen matrix by lated cells (P Ͻ 0.01). Both control migration and migration contrasting two different smooth muscle cell (SMC) lines, one induced by PDGF-BB were almost completely inhibited by of these expressing ␣11 and ␣21 integrins, and the other pAb anti-1 and were significantly blocked by mAb 33.4. expressing only ␣11 integrin. Both studies indicate that ␣11 786 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 779–789, 1999
Figure 8. Effects of rabbit anti-1 subunit IgG (pAb anti-1) and anti-␣1 monoclonal antibody (mAb) 33.4 on TGF- (10 ng/ml)- or PDGF-BB (50 ng/ml)-induced collagen gel contraction by mesangial cells (A through D). The degree of TGF--induced (A) or PDGF-BB-induced (C) gel contraction by mesangial cells in the presence of pAb anti-1 or normal rabbit IgG (Rab IgG) (200 g/ml, respectively) was compared with the degree of control gel contraction by Rab IgG-treated mesangial cells at 24 h. The degree of TGF--induced (B) or PDGF-BB-induced (C) gel contraction by mesangial cells in the presence of mAb 33.4 or mouse IgG1 (MIgG1) (50 g/ml, respectively) was compared with the degree of control gel contraction by MIgG1-treated mesangial cells at 24 h. *P Ͻ 0.01 versus control; **P Ͻ 0.05 versus control.
integrin is a critical collagen receptor involved in matrix re- integrin expression on mesangial cells to the gel contraction modeling after vascular wall injury (23). Finally, Riikonen et process. al. suggested the possibility that ␣11 and ␣21 integrins It is well established that TGF- and PDGF-BB are impor- could replace each other as the functional collagen receptor for tant regulators of ECM expansion and cell proliferation, two gel contraction in some osteogenic cell lines (21). This concept processes involved in glomerular wound repair and/or irrevers- supports our data showing the critical role of ␣11 integrin for ible glomerular scarring (1–4). In addition, both growth factors gel contraction because of the overwhelming quantitative dif- have been shown to stimulate collagen gel contraction, an in ference between ␣11 and ␣21 integrin expression on mes- vitro model of tissue healing and scarring, in several types of angial cells. In addition, ␣31 integrin has recently been pro- cell lines (15,18,21). Increased expression of TGF- or posed to function as a collagen receptor of cultured mesangial PDGF-BB has been identified in various types of human and cells, suggesting that MC-␣31 integrin may play a role in experimental GN that lead to glomerular scarring (1–4). Re- collagen gel contraction (24,37). Additional experiments using cently, Isaka et al. using in vivo gene transfer techniques, function-blocking antibodies to the ␣2- and ␣3-integrin sub- demonstrated that overexpression of either the TGF- or the units are needed to precisely define the contribution of ␣11 PDGF-BB gene induced glomerulosclerosis in rats (38). We J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin in Collagen Matrix Remodeling 787
did not affect the expression of MC-␣11 integrin at the protein and mRNA levels but enhanced MC-mediated gel contraction. The function-blocking antibody experiments indi- cated that PDGF-BB-induced gel contraction is also signifi- cantly mediated by ␣11 integrin activity. Since no mitogenic effects of PDGF-BB on mesangial cells could be demonstrated even at concentrations as high as 50 ng/ml in present collagen gel cultures, mitogenic effects of PDGF-BB are not likely to be responsible for stimulation of gel contraction by mesangial cells. In contrast, our assay system showed that PDGF-BB stimulates MC migration to collagen I, but not adhesion to collagen I, and that this PDGF-BB-induced MC migration is ␣11 integrin-dependent, suggesting that the effect of Figure 9. Migration of TGF-- or PDGF-BB-treated mesangial cells toward collagen I in the presence of anti-1 subunit IgG (pAb anti- PDGF-BB on gel contraction may be linked to its MC migra- ␣  1) or anti-␣1 monoclonal antibody (mAb) 33.4. Mesangial cells tion-stimulatory activity involved in 1 1 integrin function. were preincubated with either purified rabbit IgG (Rab IgG) (200 However, several cell culture studies have reported that PDGF g/ml), pAb anti-1(200 g/ml), mouse IgG1(MIgG1) (50 g/ml), or induces a rapid reorganization of actin filaments in cells mAb 33.4 (50 g/ml) for 30 min at 20°C, and added to the upper (39,40). Therefore, we cannot rule out the possibility that the chamber of a Transwell apparatus containing a membrane coated on stimulation of gel contraction by PDGF-BB is also related to lower side with 20 g/ml collagen I. The lower chamber containing the effects of PDGF on the cytoskeleton, since gel contraction either Rab IgG, pAb anti-1, MIgG1, or mAb 33.4 at same concen- is dependent on cytoskeletal proteins (41,42). On the basis of  tration of upper chamber was supplemented with or without TGF- these findings, it seems likely that PDGF-BB also influences (10 ng/ml) or PDGF-BB (50 ng/ml). Cells penetrating to the lower the process of collagen matrix remodeling/scarring by mesan- chamber were quantified over 6 h. Results represent the mean Ϯ SD gial cells in nephritic glomeruli through the ␣11 integrin of migrated cells in 30 high-power fields. *P Ͻ 0.01 versus control. activity. In conclusion, our present data suggest that MC play a critical role in mesangial matrix remodeling after injury have previously reported that a close relationship exists be- through the expression of ␣11 integrin that mediates cell tween the expression of TGF- and 1 integrins including adhesion/migration to collagen and collagen matrix remodel- ␣11 integrin in abnormal glomerular matrix remodeling in rat ing. In addition, overexpression of either the TGF- or the and human GN. This suggests that TGF- might affect the PDGF-BB gene in the glomerulus often observed in progres- matrix remodeling process through the action of mesangial 1 sive kidney diseases may cause excess mesangial matrix re- integrins (2,11). The present study demonstrates that TGF- modeling by increasing ␣11-integrin-dependent collagen ma- and PDGF-BB stimulate MC-mediated collagen gel contrac- trix reorganization, and thus lead to pathologic healing and tion via different mechanisms and that both effects were sig- glomerular scarring. nificantly dependent on ␣11 integrin function. In general, the collagen gel contraction process is composed Acknowledgments of several phases including cell migration, cell attachment to This work was supported by grants from Japan Ministry and collagen fibers, followed by reorganization of the surrounding Welfare (8670894). We are grateful to Dr. S. Tominaga (Jichi Medical collagen fibers into a more dense and compact arrangement School, Tochigi, Japan) for providing the mouse 1 integrin subunit within the collagen lattice (14–17). In the present study, the cDNA, Dr. L. F. Leichardt (University of California, San Francisco, action of TGF- on gel contraction by mesangial cells showed CA) for providing the rat ␣1 integrin subunit cDNA, and Dr. K. a dose-dependent effect; the enhanced gel contraction corre- Iwamoto (University of Osaka, Osaka, Japan) for the gift of rabbit lates well with the increased expression of MC-␣11 integrin antiserum against ␣1 and 1 integrin subunits. We thank Dr. Markus induced by TGF-. Adhesion assay indicated that the increased Kettler (Free University of Berlin, Germany) for helpful discussions. expression of ␣11 integrin is accompanied by the increased MC adhesion to collagen I. Finally, TGF--induced gel con- References traction or MC adhesion to collagen I was dramatically inhib- 1. Border WA, Noble NA: Transforming growth factor- in tissue ited by antibodies against the ␣1 and 1 integrin subunits. fibrosis. N Engl J Med 331: 1286–1292, 1994 These results suggested that TGF- increases collagen matrix 2. Kagami S, Border WA, Ruoslahti E, Noble NA: Coordinated  reorganization by increasing MC-␣11 integrin expression, expression of integrins and transforming growth factor- -in- duced matrix proteins in glomerulonephritis. Lab Invest 69: which leads to an increased number of adhesion sites for the  68–76, 1993 mesangial cells to collagen I. Thus, overexpression of TGF- 3. Johnson RJ: The glomerular response to injury: Progression or in nephritic glomeruli may induce MC-mediated excess colla- resolution? Kidney Int 45: 1769–1782, 1994 gen matrix remodeling reaction/mesangial scarring with the 4. Johnson RJ: Role of cytokines and growth factors in glomeru- loss of mesangial architecture and function. lonephritis: A chance for future therapeutic intervention. Interestingly, another fibrogenic growth factor, PDGF-BB, Nephron 73: 506–514, 1996 788 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 779–789, 1999
5. Haralson MA: Extracellular matrix and growth factors: An inte- tima formation in rat arteries and mediates collagen matrix grated interplay controlling tissue repair and progression to dis- reorganization. J Clin Invest 97: 2469–2477, 1996 ease. Lab Invest 69: 369–372, 1993 24. Menderick DL, Kelly DM, DuMont SS, Sandstrom DJ: Glomer- 6. Rupprecht HD, Scho¨cklmann HO, Sterzel RB: Cell-matrix inter- ular epithelial and mesangial cells differentially modulate the actions in the glomerular mesangium. Kidney Int 49: 1575–1582, binding specificities of VLA-1 and VLA-2. Lab Invest 72: 367– 1996 375, 1995 7. Ruoslahti E, Noble NA, Kagami S, Border WA: Integrins. Kid- 25. Kagami S, Border WA, Miller DE, Noble NA: Angiotensin II ney Int 45[Suppl 44]: s17–s22, 1994 stimulates extracellular matrix synthesis through induction of 8. Heino J, Ignotz RA, Hemler ME, Crouse C, Massague J: Regu- transforming growth factor- expression in rat glomerular mes- lation of cell adhesion receptors by transforming growth fac- angial cells. J Clin Invest 93: 2431–2437, 1994  tor- . J Biol Chem 264: 380–388, 1989 26. Caniggia I, Han B, Liu J, Wang J, Tanswell AK, Post M: 9. Gailit J, Xu J, Bueller H, Clark RAF: Platelet-derived growth Differential expression of collagen-binding receptors in fetal rat factor and inflammatory cytokines have differential effects on lung cells. Am J Physiol 268: L136–L143, 1995 the expression of integrins ␣11 and ␣51 by human dermal 27. Lo¨ster K, Voigt S, Heidrich C, Hofmann W, Reutter W: Cell- fibroblasts in vitro. J Cell Physiol 169: 281–289, 1996 collagen adhesion is inhibited by monoclonal antibody 33.4 10. Kagami S, Kuhara T, Yasutomo K, Okada K, Lo¨ster K, Reutter against the rat ␣ -integrin subunit. Exp Cell Res 212: 155–160, W, Kuroda Y: Transforming growth factor-(TGF-) stimulates 1 1994 the expression of 1 integrins and adhesion by rat mesangial 28. Tominaga S: Murine mRNA for the -subunit of integrin is cells. Exp Cell Res 229: 1–6, 1996 11. Kuhara T, Kagami S, Kuroda Y: Expression of 1-integrins on increased in BALB/c-3T3 cells entering the G1 phase from the activated mesangial cells in human glomerulonephritis. JAmSoc G0 state. FEBS Lett 238: 315–319, 1988 Nephrol 8: 1679–1687, 1997 29. Ignatius MJ, Large TH, Houde M, Tawil JW, Barton A, Esch F, 12. Kitamura M, Maruyama N, Yoshida H, Nagasawa R, Mitarai T, Carbonetto S, Reichardt LF: Molecular cloning of the rat alpha Sakai O: Extracellular matrix contraction by cultured mesangial 1-subunit: A receptor for laminin and collagen. J Cell Biol 111: cells: An assay system for mesangial cell matrix interaction. Exp 709–720, 1990 Mol Pathol 54: 181–200, 1991 30. Fort PH, Marty L, Piechaczk M, Sabrouty E, Dani C, Jeanteur P, 13. Broberg A, Heino J: Integrin ␣21-dependent contraction of Blanchard JM: Various rat adult tissue express only one major floating collagen gels and induction of collagenase are inhibited mRNA species from the glyceraldehyde-3-phosphate-dehydro- by tyrosine kinase inhibitors. Exp Cell Res 228: 29–35, 1996 genase multigenic family. Nucleic Acids Res 13: 1431–1443, 14. Grinnell F: Fibroblasts, myofibroblasts, and wound contraction. 1985 J Cell Biol 124: 401–404, 1994 31. Keely PJ, Fong AM, Zutter MM, Santoro SA: Alteration of 15. Ehrlich HP, Rajaratnam JBM: Cell locomotion forces versus cell collagen-dependent adhesion, motility, and morphogenesis by contraction forces for collagen lattice contraction: An in vitro the expression of antisense ␣2 integrin mRNA in mammary cells. model of wound contraction. Tissue Cell 22: 407–417, 1990 J Cell Sci 108: 595–607, 1995 16. Eastwood M, Porter R, Khan U, McGrouther G, Brown R: 32. Siegbahn A, Hammacher A, Westermark B, Heldin CH: Differ- Quantitative analysis of collagen gel contractile forces generated ential effects of the various isoforms of platelet-derived growth by dermal fibroblasts and the relationship to cell morphology. factor on chemotaxis of fibroblasts, monocytes, and granulo- J Cell Physiol 166: 33–42, 1996 cytes. J Clin Invest 85: 916–920, 1990 17. Yamato M, Adachi E, Yamamoto K, Hayashi T: Condensation of 33. Ferns GAA, Sprugel KH, Seifert RA, Bowen-Pope DF, Kelly JD, collagen fibrils to the direct vicinity of fibroblasts as a cause of Murray M, Raines EW, Ross R: Relative platelet-derived growth gel contraction. J Biochem 117: 940–946, 1995 factor receptor subunit expression determines cell migration to 18. Gullberg D, Tingstro¨m A, Thuresson AC, Olsson L, Terracio L, different dimeric forms of PDGF. Growth Factors 3: 315–324,  Borg T, Rubin K: 1 integrin-mediated collagen gel contraction 1990 is stimulated by PDGF. Exp Cell Res 186: 264- 272, 1990 34. Glick AD, Jacobson HR, Haralson MA: Mesangial deposition of 19. Schiro JA, Chan BMC, Roswit WT, Kassner PD, Pentland AP, type I collagen in human glomerulosclerosis. Hum Pathol 23: Hemler ME, Eisen AZ, Kupper T: Integrin ␣21(VLA-2) medi- 1373–1379, 1992 ates reorganization and contraction of collagen matrices by hu- 35. Jacot TA, Striker GE, Stetler-Stevenson MA, Striker LJ: Mes- man cells. Cell 67: 403–410, 1991 angial cells from transgenic mice with progressive glomerulo- 20. Klein CE, Dressel D, Steinmayer T, Mauch C, Eckes B, Krieg T, Bankert RB, Weber L: Integrin ␣21 is upregulated in fibro- sclerosis exhibit stable, phenotypic changes including undetect- blasts and highly aggressive melanoma cells in three-dimen- able MMP-9 and increased type IV collagen. Lab Invest 75: sional collagen lattices and mediates the reorganization of col- 791–799, 1996, lagen I fibrils. J Cell Biol 115: 1427–1436, 1991 36. Miralem T, Whiteside CI, Templeton DM: Collagen type I 21. Riikonen T, Koivisto L, Vihinen P, Heino J: Transforming enhances endothelin-mediated contraction and induces nonpro- growth factor- regulates collagen gel contraction by increasing liferating phenotype in mesangial cells. Am J Physiol 39: F960– ␣21 integrin expression in osteogenic cells. J Biol Chem 270: F970, 1996 376–382, 1995 37. Petermann A, Fees H, Grenz H, Goodman SL, Sterzel RB: 22. Carver W, Molano I, Reaves TA, Borg TK, Terracio L: Role of Polymerase chain reaction and focal contact formation indicate ␣11 integrin complex in collagen gel contraction in vitro by integrin expression in mesangial cells. Kidney Int 44: 997–1005, fibroblasts. J Cell Physiol 165: 425–437, 1995 1993 23. Gotwals PJ, Chi-Roso G, Lindner V, Yang J, Ling L, Fawell SE, 38. Isaka Y, Fujiwara Y, Ueda N, Kaneda Y, Kamada T, Imai E: Koteliansky VE: The ␣11 integrin is expressed during neoin- Glomerulosclerosis induced by in vivo transforming growth fac- J Am Soc Nephrol 10: 779–789, 1999 ␣11 Integrin in Collagen Matrix Remodeling 789
tor- or platelet-derived growth factor gene into the rat kidney. kinase and rac in platelet-derived growth factor-induced actin reor- J Clin Invest 92: 2597–2601, 1993 ganization and chemotaxis. Exp Cell Res 234: 434–441, 1997 39. Hammacher A, Mellstro¨m K, Heldin CH, Westermark B: Iso- 41. Guidry C, Grinnell F: Studies on the mechanism of hydrated form-specific induction of actin reorganization by platelet-de- collagen gel reorganization by human skin fibroblasts. J Cell Sci rived growth factor suggests that the functionally active receptor 79: 67–81, 1985 is a dimer. EMBO J 8: 2489–2495, 1989 42. Arora P, McCulloch CAG: Dependence of collagen remodelling 40. Hooshmand-Rad R, Claesson-Welsh L, Wennstro¨m S, Yokote K, on ␣-smooth muscle actin expression by fibroblasts. J Cell Siegbahn A, Heldin CH: Involvement of phosphatidylinositide 3Ј- Physiol 159: 161–175, 1994