Rac regulates integrin-mediated endothelial cell adhesion and migration on laminin-8

Experimental Cell Research 292 (2004) 67 – 77 www.elsevier.com/locate/yexcr

Rac regulates integrin-mediated endothelial cell adhesion and migration on laminin-8 Hironobu Fujiwara, a,b Jianguo Gu, a and Kiyotoshi Sekiguchi a,b,* a

b

Institute for Protein Research, Osaka University, Suita, Osaka 565-0871, Japan Sekiguchi Biomatrix Signaling Project, ERATO, Japanese Science and Technology Corporation, Aichi Medical University, Nagakute-cho, Aichi 480-1195, Japan Received 6 February 2003, revised version received 1 July 2003

Abstract Blood vessel formation requires endothelial cell interactions with the extracellular matrix through cell surface receptors, and signaling events that control endothelial cell adhesion, migration, and lumen formation. Laminin-8 (a4h1g1) is present in all basement membranes of blood vessels in fetal and adult tissues, but despite its importance in vessel formation, its role in endothelial cell adhesion and migration remains undefined. We examined adhesion and migration of HMEC-1 human microvascular endothelial cells on laminin-8 with an emphasis on the integrin-mediated signaling events, as compared with those on laminin-10/11 and fibronectin. We found that laminin-8 was less potent in HMEC-1 cell adhesion than laminin-1, laminin-10/11, and fibronectin, and mediated cell adhesion through a6h1 integrin. Despite its weak cell-adhesive activity, laminin-8 was as potent as laminin-10/11 in promoting cell migration. Cells adhering to laminin-8 displayed streaks of thin actin filaments and formed lamellipodia at the leading edge of the cells, as observed with cells adhering to laminin-10/11, while cells on fibronectin showed thick actin stress fibers and large focal adhesions. Pull-down assays of GTP-loaded Rho, Rac, and Cdc42 demonstrated that Rac, but not Rho or Cdc42, was preferentially activated on laminin-8 and laminin-10/11, when compared with fibronectin. Furthermore, a dominant-negative mutant of Rac suppressed cell spreading, lamellipodial formation, and migration on laminin-8, but not on fibronectin. These results, taken together, indicate that Rac is activated during endothelial cell adhesion to laminin-8, and is pivotal for a6h1 integrinmediated cell spreading and migration on laminin-8. D 2003 Elsevier Inc. All rights reserved. Keywords: Basement membrane; Laminin; Endothelial cell; Integrin; Rac

Introduction Blood vessel formation is a complex process requiring interactions of endothelium with both soluble factors and the extracellular matrix. The extracellular matrix is involved in this process as a scaffold for endothelium and also as bioactive signaling molecules regulating cell adhesion, migration, differentiation, proliferation, and apoptosis through interaction with various cell surface receptors [1]. Abbreviations: FBS, fetal bovine serum; HUVECs, human umbilical vein endothelial cells; mAb, monoclonal antibody; PBS, phosphatebuffered saline; GST-RBD, a fusion protein of glutathione S-transferase to the Rho-binding domain of rhotekin; GST-CRIB, a fusion protein of glutathione S-transferase to the Cdc42/Rac-interactive-binding domain of PAK1; BSA, bovine serum albumin. * Corresponding author. Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan. Fax: +81-6-68798619. E-mail address: [email protected] (K. Sekiguchi). 0014-4827/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2003.08.010

Laminins are heterotrimeric glycoproteins composed of a, h, and g chains, and found in all basement membranes including those of blood vessels [2,3]. The a chains are recognized by cell surface receptors such as integrins and adystroglycan, and therefore, are the primary determinants of the binding specificity of laminin isoforms toward their receptors [4]. To date, five a, three h, and three g chains have been identified, and are known to give rise to at least 15 laminin isoforms [3]. The distribution of these laminin isoforms is tissue-specific and developmentally regulated; suggesting that they are functionally distinct. The a5 chaincontaining laminins, that is, laminin-10 (a5h1g1) and laminin-11 (a5h2g1), are the most abundantly expressed isoforms in mammalian tissues [5,6], whereas the a1 chaincontaining laminins are the most restricted [7,8]. Laminins composed of the a2 chain are deposited predominantly in basement membranes of the cells of mesodermal origin, such as muscles and peripheral nerves [9,10], while the a3 chain-containing laminins are found in most epithelial

68

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

basement membranes, particularly those of epidermal cells [11– 13]. The laminin a4 chain is a component of three known laminin isoforms, laminin-8 (a4h1g1), laminin-9 (a4h2g1), and laminin-14 (a4h2g3) [3,14]. It is mainly expressed in endothelial cells, adipocytes, and muscle cells, and is distributed in all vascular endothelial basement membranes of fetal and adult tissues [10,15 – 17]. In most tissues, two main laminin isoforms, laminin-8 and laminin-10, are present in endothelial basement membranes [5,17 – 19]. Mice lacking the a4 chain have been shown to exhibit neonatal hemorrhages due to unstable microvessels, indicating that laminin-8 seems to be an important component of the newly formed endothelial basement membranes [20]. Recently, we purified laminin-8 and demonstrated that cell adhesion to laminin-8 was mediated through a6h1 and a3h1 integrins [21]. On the other hand, laminin-8 purified from blood platelets as well as recombinant laminin-8 expressed in human 293 cells was shown to mediate cell adhesion through a6h1 integrin exclusively [22 – 24]. The signaling events that result from integrin-mediated cell adhesion to laminin-8 remain unclear. In this study, we examine endothelial cell adhesion and migration on laminin-8 with an emphasis on the integrinmediated signaling events. We found that laminin-8 was more potent than fibronectin in stimulating endothelial cell migration in an a6h1 integrin-dependent manner, although it was less potent in mediating cell adhesion than laminin10/11 and fibronectin. We also showed the importance of Rac activation in endothelial cell adhesion and migration on laminin-8 by using a panel of dominant-negative mutants of Rho family GTPases.

11H against human laminin a4 chain was produced by fusion of Sp2/O mouse myeloma cells with spleen cells from mice immunized with a glutathione S-transferase fusion protein containing the domains I/II of human laminin a4 chain (Thr274 – Tyr444). The mAb 2-11H specifically reacted with laminin-8 among a panel of purified human laminins in ELISA and recognized only the laminin a4 chain on immunoblots under reducing conditions. The mAbs against laminin a4 chain (8B12) and laminin h1 chain (4F5) were produced in our laboratory as described previously [21]. The mAb against laminin h1 chain (DG10) was kindly provided by Dr. Ismo Virtanen (University of Helsinki, Finland) and used in immunoblotting. The mAb against laminin g1 chain (#22) was purchased from Transduction Laboratories (Lexington, KY). The function-blocking mAbs against integrin a2 subunit (P1E6) was obtained from Chemicon (Temecula, CA). The function-blocking mAb against integrin a6 subunit (AMCI 7-4) was kindly provided by Dr. Masahiko Katayama (Eisai, Tsukuba, Japan) [26]. Hybridoma cells secreting the function-blocking mAb against integrin h1 subunit (AIIB2), developed by Dr. Caroline Damsky (University of California, San Francisco, CA), were obtained from the Developmental Studies Hybridoma Bank (University of Iowa, IA). Function-blocking mAbs specific for integrin a3 (3G8) and a5 (8F1) were produced in our laboratory [27,28]. The mAbs against paxillin and Rac were obtained from Transduction Laboratories; the mAb against Rho was from Cytoskeleton (Denver, Co); and the polyclonal antibody against Cdc42 was from Santa Cruz Biotechnology (Santa Cruz, CA). Rhodamine-labeled phalloidin was obtained from Molecular Probes (Eugene, OR). Cell-adhesive proteins

Materials and methods Cell culture Human glioblastoma cell line T98G was obtained from the Japanese Cancer Research Resources Bank (Tokyo, Japan) and grown in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (FBS). Human dermal microvascular endothelial cell line HMEC-1 was obtained from the Centers for Disease Control and Prevention (Atlanta, GA) and maintained in MCDB131 containing 10% FBS, 10 ng/ml epidermal growth factor, and 1 Ag/ml hydrocortisone. Primary cultures of human umbilical vein endothelial cells (HUVECs) were prepared as described previously [25], and grown in Ham’s F12K medium containing 20% FBS, 10 ng/ml basic fibroblast growth factor, and heparin. HUVECs were used between passages 2 and 8. Reagents and antibodies Y27632 was provided by Yoshitomi Pharmaceutical Industries (Osaka, Japan). The monoclonal antibody (mAb) 2-

Mouse laminin-1 was purified from mouse Engelbreth – Holm – Swarm tumor tissues by the method of Paulsson et al. [29]. Human laminin-10/11 was purified from the conditioned medium of A549 cells according to Kikkawa et al. [30], except that the mAb 5D6 was used as an immunoaffinity ligand. Human plasma fibronectin was purified from outdated plasma by gelatin affinity chromatography [31]. Purification of laminin-8 T98G cells were grown in 1700 cm2 roller bottles as described previously [21]. The conditioned medium (f3 l) was clarified by centrifugation, precipitated with 45% saturated ammonium sulfate, and then subjected to gel filtration. Fractions containing laminin-8 were passed through a gelatin-Sepharose column to remove fibronectin. The flow-through fractions from the gelatin-Sepharose were pooled and split into halves, and then applied to immunoaffinity columns containing either the mAb 2-11H against the laminin a4 chain or the mAb 4F5 against the laminin h1 chain. The bound proteins were eluted from each column with 0.1 M triethylamine (pH 11.5), immediately neutralized

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

with 2 M NaH2PO4 (pH 3.6), and dialyzed against phosphate-buffered saline (PBS). SDS-PAGE and immunoblotting SDS-PAGE was carried out according to Laemmli [32] using 5.5% acrylamide gels. Separated proteins were visualized by silver staining or transferred onto PVDF membranes. The membranes were probed with mAbs against individual laminin chains, followed by visualization using an ECL detection kit (Amersham Pharmacia Biotech, UK). Expression vectors The pEF-BOS-HA constructs encoding N19Rho, N17Rac, or N17Cdc42 and the expression vector for GSTCRIB (a fusion protein of glutathione S-transferase to the Cdc42/Rac-interactive- and binding domain of PAK1) were kindly provided by Dr. Kozo Kaibuchi (Nagoya University, Japan). The expression vector for GST-RBD (a fusion protein of glutathione S-transferase to the Rho-binding domain of rhotekin) was kindly provided by Dr. Martin Schwartz (The Scripps Institute, La Jolla, CA).

69

to attach. The dishes were placed in a built-in CO2 incubator on the stage of a Zeiss Axiovert 25 microscope, and subjected to time-lapse video microscopy for 8 h. Cell migration was quantified by tracking the position of the nuclei of migrating cells using Image-Pro Plus. Cell migration inhibition assays were performed as described above except that cells were preincubated with 20 Ag/ml of function-blocking mAbs for 20 min at room temperature before plating in the glass-bottom culture dishes coated with cell-adhesive proteins. To examine the roles of the Rho family GTPases in cell migration, cDNAs encoding dominant-negative mutants of Rho, Rac, or Cdc42 were microinjected together with the cDNA encoding GFP at a 1:1 ratio (the total DNA concentration was 0.2 mg/ml) into the nuclei of adhering cells at 30 min after cell plating. Expressions of the proteins from injected cDNAs were confirmed by immunostaining of HA tags fused to each GTPases at 2 h after injection. Since GFP-positive cells were also HA-positive, we identified the cells expressing dominant-negative mutants by GFP fluorescence in cell migration assays. Two hours after microinjection, cell migration was examined by time-lapse video microscopy over 8 h. Cell migration was quantified by

Cell spreading assay Cell spreading assays ware performed as described previously [27]. Briefly, 96-well microtiter plates were coated with 50 Al of cell-adhesive proteins and then blocked with 2% heat-denatured bovine serum albumin (BSA)/PBS. Cell suspensions (3  104 cells) were added to each well of the coated plates and incubated at 37jC for 30 min. After the plates were washed, cells were fixed, stained with Diff-Quik (International Reagents, Japan), and spread cells (i.e., cells that had become flattened with their long axis more than twice the diameter of the nucleus) counted in three independent fields/well. For cell spreading inhibition assays, trypsinized cells (2  104 cells) were incubated with 20 Ag/ml mAbs against different integrin subunits for 20 min at room temperature, added to wells that had been coated with different cell-adhesive proteins, and then incubated for 30 min at 37jC. Spread cells were counted as above. Cell migration assay and microinjection Cell migration on substrates coated with laminin-8 (20 nM), laminin-10/11 (5 nM), or fibronectin (20 nM) was examined by time-lapse video microscopy using the image processing software Image-Pro Plus (Media Cybernetics, Silver Spring, MD). Glass-bottom culture dishes fitted with / 8-mm coverslips were coated with cell-adhesive proteins and blocked with heat-denatured 2% BSA/PBS. Two milliliters of cell suspensions (4  104 cells/ml in 1% FBS-containing medium) were added to each coated dishes and incubated at 37jC for 30 min to allow cells

Fig. 1. Adhesion of HMEC-1 cells to laminin-1, laminin-8, laminin-10/11, and fibronectin. HMEC-1 cells were incubated at 37jC for 30 min on 96well microtiter plates coated with increasing concentrations of laminin-1 (LN-1), laminin-8 (LN-8), laminin-10/11 (LN-10/11), or fibronectin (FN). Adherent cells were fixed and stained as described in Materials and methods. (A) Cells spreading on the substrates were quantified and expressed as the mean and SD in triplicate assays. Representative images of cells spreading at each protein concentration are shown in (B). Bar, 100 Am.

70

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

tracking the position of the nuclei of GFP-positive cells using Image-Pro Plus. Immunofluorescence staining Glass coverslips were coated with laminin-8 (20 nM), laminin-10/11 (5 nM), or fibronectin (20 nM) overnight at 4jC and blocked with heat-denatured 2% BSA/PBS. HMEC-1 cells were serum-starved overnight and then replated on the coverslips in medium containing 0.5% BSA. Cells were incubated for 2 h at 37jC, and then fixed with 3.7% formaldehyde in PBS for 10 min and permeabilized with 0.1% Triton X-100 for 5 min. Focal adhesions were visualized by incubating cells with mouse anti-paxillin antibody, followed by incubation with Alexa Fluor 488-

conjugated goat antibody against mouse IgG (Molecular Probes). Actin filaments were stained with rhodamineconjugated phalloidin. Coverslips were mounted and observed with a Zeiss LSM PASCAL confocal microscope. For immunofluorescence staining of cells transfected with dominant-negative mutants of Rho family GTPases, HMEC-1 cells were transfected with each expression vector using the LipofectAMINE reagent (Invitrogen, Carlsbad, CA) under manufacturer’s instructions. Fortyeight hours after transfection, cells were replated on the coverslips, incubated for 2 h at 37jC, fixed, permeabilized, and stained as described above. Cells expressing dominant-negative mutants of Rho family GTPases were detected by staining of the HA-tag fused to each mutant small GTPase.

Fig. 2. Effects of anti-integrin mAbs on cell adhesion onto laminin-8, laminin-10/11, and fibronectin. (A) Inhibition of spreading of HMEC-1 cells on laminin8, laminin-10/11, or fibronectin by anti-integrin mAbs. Ninety-six-well microtiter plates were coated with laminin-8 (LN-8, 20 nM), laminin-10/11 (LN-10/11, 5 nM), or fibronectin (FN, 20 nM). HMEC-1 cells were preincubated with 20 Ag/ml of function-blocking mAbs against integrin subunits for 20 min at room temperature and then added to the precoated wells. The mAbs used were anti-a2 subunit (P1E6), anti-a3 subunit (3G8), anti-a5 subunit (8F1), anti-a6 subunit (AMCI 7-4), and anti-h1 subunit (AIIB2). After 30 min incubation at 37jC, cells spreading on the substrates were counted as described in Materials and methods. The number of spreading cells is expressed as a percentage of the spreading cells in the presence of control mouse IgG. Each column and bar represents the mean of triplicate assays and the SD, respectively. (B) SDS-PAGE and immunoblot analysis of purified laminin-8. Laminin-8 was purified from the same conditioned medium of T98G cells on immunoaffinity columns containing either mAb 4F5 (anti-laminin h1 chain) or 2-11H (anti-laminin a4 chain). Purified laminin-8 was analyzed by SDS-PAGE using 5.5% gels under reducing conditions. Separated proteins were visualized by silver staining or transferred onto PVDF membranes followed by staining with mAbs against laminin a4 (8B12), h1 (DG10), or g1 (#22) chains. The positions of molecular size markers are shown in the left margin. (C) Effects of anti-integrin mAbs on spreading of T98G cells on laminin-8 purified on different immunoaffinity columns, that is, the 4F5 anti-laminin h1 mAb column and the 2-11H anti-laminin a4 mAb column.

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

Detection of GTP-loaded Rho, Rac, and Cdc42 HMEC-1 cells were serum-starved overnight, detached, and kept in suspension in the medium containing 0.5% BSA for 60 min. Cells were plated in dishes precoated with laminin-8 (20 nM), laminin-10/11 (5 nM), or fibronectin (20 nM) and incubated for 30 min at 37jC. Cells were washed with PBS and then lysed in the buffer containing 1% Nonidet P-40, 50 mM Tris –HCl (pH 7.4), 100 mM NaCl, 10 mM MgCl2, 1 mM dithiothreitol, 10 Ag/ml leupeptin, 10 Ag/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Cell lysates were clarified by centrifugation at 13 000  g for 10 min at 4jC, and then equal volumes of lysates were incubated with GST-RBD- or GST-CRIBbound glutathione beads (20 Ag of protein per sample) for 60 min at 4jC. The beads were washed three times with lysis buffer, and bound proteins were detected by Western blotting with anti-Rho, anti-Rac, or anti-Cdc42 antibodies. The total amounts of Rho, Rac, and Cdc42 in cell lysates were also detected for normalization.

Results

71

were obtained when HUVECs were used in cell adhesion assays. HMEC-1 adhesion to laminin-10/11 was partially inhibited by the mAb against integrin a3 subunit and completely inhibited by the mAb against integrin h1 subunit. Cell adhesion to fibronectin was inhibited by the mAbs against integrin a5 or h1. These results suggest that HMEC-1 adhesion to laminin-8, laminin-10/11, and fibronectin is mainly mediated by integrin a6h1, a3h1, and a5h1, respectively. Previously, we reported that adhesion of T98G glioma cells to laminin-8 was inhibited by a combination of antibodies against the integrin a3 and a6 subunits, but not by either anti-a3 or anti-a6 mAb alone [21]. However, HMEC-1 adhesion to laminin-8 was strongly inhibited by the mAb against integrin a6 alone. This apparent discrepancy could be due to the difference in either the protocols of laminin-8 purification or the cell types used in the cell adhesion assays. In our previous study, we purified laminin-8 using the mAb 4F5 against the laminin h1 chain as an immunoaffinity ligand [21], although we used the mAb 2-11H against the laminin a4 chain in the present study. To see whether the apparent discrepancy in the integrin binding specificity was due to

HMEC-1 cell adhesion to laminin-8 Although laminin-8 is present in all basement membranes of blood vessels, its cell-adhesive activity toward endothelial cells has not been extensively studied. We examined the cell-adhesive activity of laminin-8 in comparison with that of laminin-1, laminin-10/11, and fibronectin, using HMEC-1 human dermal microvascular endothelial cells. Among the extracellular matrix proteins tested, laminin-8 was found to be least potent in mediating HMEC-1 adhesion to the substrates (Fig. 1A). To attain half-maximal adhesion, laminin-8 required f10 nM of coating concentration, while fibronectin and laminin-1 required 5 nM, and laminin-10/11 required as low as 2.5 nM. Cells adhering to laminin-8 assumed an elongated, often fan-shaped morphology with thin projections (Fig. 1B), resembling the morphology on laminin-1. Cells adhering to laminin-10/11 and fibronectin showed well-spread, cobblestone-like morphology with greater cell – substratum contact area. A comparable level of HMEC-1 adhesion, being close to the maximal level, was observed with laminin-8 at 20 nM, laminin-10/11 at 5 nM, and fibronectin at 20 nM. Unless otherwise specified, we compared cellular responses to laminin-8, laminin-10/11, and fibronectin at these coating concentrations. To elucidate the integrin types mediating adhesion of HMEC-1 cells to laminin-8, we examined the effects of function-blocking antibodies against various integrin subunits on the HMEC-1 adhesion to laminin-8. HMEC-1 adhesion to laminin-8 was strongly inhibited by the mAbs against integrin a6 and h1 (Fig. 2A). No significant inhibition was observed with the anti-a3 antibody. Similar results

Fig. 3. Migration of HMEC-1 cells on substrates coated with laminin-8, laminin-10/11, and fibronectin. (A) HMEC-1 cells were plated on glassbottom culture dishes coated with laminin-8 (LN-8, 20 nM), laminin-10/11 (LN-10/11, 5 nM), or fibronectin (FN, 20 nM), and images of migrating cells were collected by time-lapse video microscopy at 37jC for 8 h. Cell migration paths were tracked and quantified using the Image-Pro Plus image processing software. Paths of 20 cells were tracked in individual assays. Each column and bar represents the mean migration distance and SD in three independent experiments. (B) HMEC-1 cells were preincubated with 20 Ag/ml of function-blocking mAbs against integrin subunits, and then added to the precoated dishes. The mAbs used were anti-a3 subunit (3G8), anti-a6 subunit (AMCI 7-4), and anti-h1 subunit (AIIB2). Cell migration images were collected by time-lapse video microscopy at 37jC for 8 h.

72

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

the differences in the mAbs used in laminin-8 purification, we purified laminin-8 from the same conditioned medium of T98G cells on two separate immunoaffinity columns containing either mAb 4F5 or 2-11H, and performed cell adhesion inhibition assays with a panel of anti-integrin mAbs using T98G. SDS-PAGE profiles after sliver staining or immunoblotting for each subunit chain did not show any significant difference in their subunit size, precluding any selective proteolytic degradation of either one of the laminin-8 preparations (Fig. 2B). Nevertheless, susceptibility of their cell-adhesive activity to anti-integrin mAbs was markedly different (Fig. 2C). Cell adhesion to laminin-8 purified on a 4F5 column was not inhibited by either mAb against integrin a3 or a6 subunit alone, but was completely inhibited by a mixture of these antibodies. In contrast, adhesion of T98G cells to laminin-8 purified on a 2-11H column was completely blocked by the antia6 antibody alone. Thus, the discrepancy in the integrin binding specificity of laminin-8 may result from the difference in the purification protocol of laminin-8, possibly due to an unknown minor contaminant in the laminin8 purified on an anti-laminin h1 chain mAb column (see Discussion).

Laminin-8 stimulates HMEC-1 migration through a6b1 integrin Interaction of endothelial cells with extracellular matrix proteins is essential for cell migration during vessel tube formation. We examined the activity of laminin-8 to promote HMEC-1 cell migration in comparison with that of laminin-10/11 and fibronectin by using time-lapse video microscopy. Migration of HMEC-1 cells on laminin-8 was 2-fold faster than that on fibronectin, whereas it was slightly slower than that on laminin-10/11 (Fig. 3A). To elucidate the integrin types mediating HMEC-1 migration on laminin-8, we examined the effects of function-blocking anti-integrin mAbs against a3, a6, and h1 subunits on HMEC-1 migration on laminin-8. The anti-a6 antibody strongly inhibited laminin-8-mediated HMEC-1 migration, but the effect of anti-a3 antibody was marginal (Fig. 3B). The anti-h1 antibody completely inhibited HMEC-1 migration on laminin-8. These results demonstrate that HMEC-1 migration on laminin-8 is mainly mediated by the a6h1 integrin, but that the a3h1 integrin may also be involved as an auxiliary receptor.

Fig. 4. Cytoskeletal reorganization and focal adhesion formation in cells adhering to laminin-8, laminin-10/11, and fibronectin. Serum-starved HMEC-1 cells were plated on coverslips coated with laminin-8 (LN-8, 20 nM), laminin-10/11 (LN-10/11, 5 nM), or fibronectin (FN, 20 nM), and incubated for 2 h at 37jC. Cells were then fixed and stained with rhodamine-labeled phalloidin to detect F-actin or anti-paxillin antibody to detect focal adhesions. Representative cells on each substrate are shown in (A). Arrowheads point to membrane ruffles induced on laminin-8 and laminin-10/11. (B) Quantification of stress fiber and focal adhesion formation in HMEC-1 cells adhering to laminin-8, laminin-10/11, and fibronectin. Cells stained with rhodamine-labeled phalloidin and anti-paxillin antibody on each substrate were categorized into three groups based on the following criteria; minus ( ), cells with no clear stress fibers or focal adhesions; plus (+), cells with thin stress fibers or small focal adhesions (see the cells on laminin-10/11 in A); double plus (++), cells with thick stress fibers or large focal adhesions (see the cells on fibronectin in A). Relative abundance of each group of cells is expressed as a percentage of the total cells. Bar, 50 Am.

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

Laminin-8 induces the formation of lamellipodia but not stress fibers and focal adhesions The actin cytoskeleton and focal adhesions are involved in a variety of fundamental cellular functions such as cell adhesion, spreading, and migration. Their dynamic properties not only define cell shape and polarity, but also serve as the driving force for cell locomotion. To explore the differences in signals regulating the actin reorganization and focal adhesion assembly, we stained actin filaments and focal adhesions of cells adhering to laminin-8, laminin-10/11, and fibronectin with rhodamine-conjugated phalloidin or the mAb against paxillin. HMEC-1 cells adhering to fibronectin showed thick actin filaments, referred to as stress fibers, all over the cell body, although cells on laminin-8 displayed streaks of thin actin filaments and formed lamellipodia at the leading edge of the cell (Fig. 4A). Cells on laminin-10/11 also showed streaks of thin actin filaments and formed lamellipodia at the leading edge, but contained a small number of thick actin filaments. Focal adhesions were clearly detected in cells adhering to fibronectin but not in cells adhering to laminin-8 or laminin-10/11. Instead, cells on laminin-8 showed small, less prominent focal adhesions, often referred to as focal complexes, in lamellipodia. Cells spread on laminin-10/11 showed some focal adhesions,

73

while they were significantly fewer and smaller than those formed on fibronectin. Cells adhering to each adhesive protein were categorized into three groups based on their degree of stress fiber and focal adhesion formation (Fig. 4B). The criteria of categorization are described in the legend of Fig. 4. The results clearly showed that laminin8 was incapable of inducing stress fibers and focal adhesions, whereas fibronectin was very potent in inducing prominent stress fibers and focal adhesions in most cells. Laminin-10/11 moderately induced stress fibers and focal adhesions. Laminin-8 stimulates Rac activation Actin cytoskeletal reorganization associated with cell spreading and migration is regulated by the Rho family GTPases. It seems likely that Rho activation regulates the assembly of actin filaments, leading to the formation of stress fibers; Rac and Cdc42 regulate the formation of lamellipodia and filopodia, respectively. To examine the matrix-dependent actin reorganization and cell motility with respect to the activation of Rho family GTPases, the levels of GTP-loaded Rho, Rac, and Cdc42 were determined by pull-down assays in the cells adhering to laminin-8, laminin10/11, or fibronectin. Although the difference in the levels

Fig. 5. Adhesion-dependent activation of Rho, Rac, and Cdc42 in HMEC-1 cells. (A) Serum-starved HMEC-1 cells were detached and kept in suspension for 60 min, and then replated on dishes coated with laminin-8 (LN-8, 20 nM), laminin-10/11 (LN-10/11, 5 nM), or fibronectin (FN, 20 nM) for 30 min. Cells were then lysed and incubated with GST-RBD- or GST-CRIB-bound glutathione beads. GTP-loaded Rho bound to GST-RBD and GTP-loaded Rac and Cdc42 bound to GST-CRIB were detected by Western blotting with antibodies against Rho, Rac, or Cdc42 (upper panel). Total cell lysates were also probed with the same antibodies to demonstrate that equal amounts of total protein were used in individual assays (lower panel). Results are representative of three independent experiments. Sus, cells were kept in suspension throughout the assays. (B) The ratios of GTP-loaded GTPases to total GTPases were determined by densitometry of the bands on immunoblots. The data are expressed as relative values after normalization against the ratio in the cells in suspension. Each column and bar represents the mean of three independent assays and the SD, respectively.

74

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

of activated Rho and Cdc42 was marginal, Rac was significantly activated in the cells adhering to laminin-8 and laminin-10/11 than fibronectin (Fig. 5A). Densitometric analysis showed that the ratios of GTP-loaded Rac to total Rac in cells adhering to laminin-8 and laminin-10/11 were more than twice greater than that in cells adhering to fibronectin (Fig. 5B). Overexpression of dominant-negative Rac inhibits cell spreading and migration on laminin-8 To investigate the role of Rho family GTPases in the spreading and migration of HMEC-1 cells on laminin-8, HMEC-1 cells were transfected with dominant-negative mutants of Rho, Rac, or Cdc42. Cell spreading on laminin-8 and laminin-10/11 was significantly reduced by dominant-negative Rac (N17Rac), while cell spreading on fibronectin was not inhibited (Fig. 6A). Membrane ruffles induced by adhesion to laminin-8 or laminin-10/11 were also inhibited by N17Rac. Although dominant-negative Rho (N19Rho) reduced the assembly of actin filaments of HMEC-1 cells irrespective of the adhesive proteins used, it did not inhibit cell spreading. Similarly, Y27632, an inhibitor of ROCK (Rho kinase acting downstream of Rho), inhibited the stress fiber formation, but did not inhibit the spreading of HMEC-1 cells on laminin-8 and laminin10/11. In contrast, the expression of dominant-negative Cdc42 (N17Cdc42) rather potentiated actin filament reorganization, leading to the enhanced stress fiber formation.

Stimulation of stress fiber formation by N17Cdc42 was prominent in cells adhering to fibronectin. N17Cdc42, however, never inhibited cell spreading. To examine the effects of dominant-negative mutants of Rho, Rac, and Cdc42 on HMEC-1 migration on laminin-8, cells adhering to laminin-8 were microinjected with expression vectors encoding GFP, N19Rho, N17Rac, or N17Cdc42. N17Rac inhibited cell migration on laminin-8 by 65%, while N19Rho and N17Cdc42 were only weakly inhibitory (Fig. 6B). These results indicate that Rac activation following cell adhesion to laminin-8 is essential for laminin-8-induced spreading and migration of cells.

Discussion Although laminin-8 and laminin-10 are the major laminin isoforms present in the basement membranes of blood vessels [5,17 –19], there are only a few studies dealing with the interaction of endothelial cells with these laminin isoforms [23,33]. This is partly because these laminins are relatively newly identified isoforms [15,34], and their purification protocols have been only recently established [21 – 23,30]. Furthermore, it is difficult to isolate these laminin isoforms from tissues in an intact form, because of their large size and insolubility due to the covalent cross-linking among them and to other basement membrane components. The interaction of endothelial cells with basement membranes has, therefore, been primarily studied using mouse

Fig. 6. Effects of dominant-negative Rho, Rac, and Cdc42 on spreading and migration of HMEC-1 cells. (A) HMEC-1 cells were transiently transfected with the plasmids encoding GFP, dominant-negative Rho (N19Rho), Rac (N17Rac), or Cdc42 (N17Cdc42), and plated on coverslips coated with laminin-8 (LN-8, 20 nM), laminin-10/11 (LN-10/11, 5 nM), or fibronectin (FN, 20 nM) for 2 h. Cells were then fixed and stained with rhodamine-labeled phalloidin for F-actin and anti-HA antibody for HA-tagged GTPases. Representative images of rhodamine-labeled phalloidin staining are shown. Arrows point to HA-positive cells. Cells treated with Y27632 (25 AM) for 30 min at 37jC were also replated on the precoated coverslips for 2 h and stained with rhodamine-labeled phalloidin. (B) HMEC-1 cells adhering to glass-bottom dishes coated with laminin-8 (20 nM) were microinjected with the plasmids encoding GFP, N19Rho, N17Rac, or N17Cdc42, as described in Materials and methods. Two hours after injection, cell migration was quantified by time-lapse video microscopy for 8 h. The paths of transfected cells were tracked, quantified, and averaged for 20 individual cells in each assay. Each column and bar represents the mean migration distance and SD in three independent experiments. Bar, 50 Am.

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

laminin-1 that can be easily purified from mouse Engelbreth – Holm – Swarm tumor, although laminin-1 has been shown to be absent from basement membranes of blood vessels [7,8,19]. In this study, we examined endothelial cell adhesion and migration on laminin-8 with an emphasis on integrin-mediated signaling events. Although laminin-8 was significantly less potent than laminin-10/11 and fibronectin in mediating endothelial cell adhesion, it was highly active in stimulating cell migration, comparable with the activity of laminin-10/11. We also showed that Rac was activated through endothelial cell adhesion to laminin-8, and played a central role in integrin-mediated cell adhesion and migration on laminin-8. The strength of cell adhesion to extracellular matrix ligands depends on the expression levels of the adhesion receptors and their affinities toward their ligands. Thus, the surface expression level of a6h1 integrin in HMEC-1 cells and its affinity to laminin-8 should be the major determinants of the adhesive strength of HMEC-1 cells to laminin8. Flow cytometric analysis showed that the surface expression of the integrin a6 subunit is lesser than that of the a2, a3, and a5 subunits in HMEC-1 cells, and that the h4 subunit that preferentially dimerizes with the a6 subunit is also expressed in HMEC-1 cells (H.F., unpublished observation). Since the h4 and h1 subunits compete for the a6 subunit [35,36], the expression level of the a6h1 integrin on HMEC-1 cells should be very low, when compared with other integrin heterodimers such as the a2h1, a3h1, and a5h1 integrins. The low expression level of the a6h1 integrin may account for the weak adhesive strength of HMEC-1 cells to laminin-8. Furthermore, low binding affinity of the a6h1 integrin to laminin-8 may also contribute to the weak adhesiveness of HMEC-1 cells to laminin-8. Integrin a6h1 purified from human placenta and reconstituted into liposomes has a very low affinity to laminin-8, as compared with other laminin isoforms such as laminin-1, laminin-5, and laminin-10/11 (Nishiuchi, R., K.S., unpublished observation), consistent with the weak cell-adhesive activity of laminin-8. Adhesive strength to the substratum is also a factor regulating cell migration. High cell – substratum adhesiveness often suppresses cell migration by preventing the release of adhesions at the rear of the cell. It has been postulated that the transmembrane linkage of actin cytoskeleton with substrate-immobilized adhesive ligands through integrins is the major determinant defining the strength that resists cell detachment at the rear [37,38]. Such a transmembrane linkage becomes more stable when integrins are assembled into large clusters, that is, focal adhesions, which are further stabilized by anchorage to actin stress fibers. This is indeed the case with the cells adhering to fibronectin-coated substrates; the pronounced formation of focal adhesions and stress fibers on fibronectin suppresses the breakage of the transmembrane linkage between actin cytoskeleton and substrate-adsorbed fibronectins, thereby retarding cell locomotion. In contrast, cells on

75

laminin-8 and laminin-10/11 did not promote the formation of focal adhesions and stress fibers, thereby facilitating the detachment at the rear of the cell. It seems likely, therefore, that the major determinant defining the adhesive strength resisting cell detachment at the rear is not the potency to mediate initial attachment and spreading on the substrates, but rather the ability to stabilize the transmembrane linkage between cytoskeleton and the substrate-immobilized adhesive ligands, typically through formation of focal adhesions. Actin dynamics within the cell have been shown to be regulated by Rho family GTPases. Rho is responsible for formation of stress fibers and focal adhesions, whereas Rac antagonizes Rho [39,40] and promotes mobilization of actins into the lamellipodia, thereby promoting cell migration through suppression of focal adhesion formation at the rear and generation of the tracking force at the cell front. Our results show that both laminin-8 and laminin-10/11 preferentially activate Rac, not Rho, consistent with their potency in promoting cell migration. Despite the similarities in their potency to preferentially activate Rac and promote cell migration, laminin-8 and laminin-10/11 are clearly different in their ability to mediate cell-substratum adhesion; it requires about four times more laminin-8 than laminin-10/11 to attain comparable levels of cell-adhesive activity. The clear difference in their celladhesive potency is mirrored in their integrin binding specificity; laminin-8 binds specifically to integrin a6h1, whereas laminin-10/11 preferentially binds to integrin a3h1. It should be also noted that morphologies of cells adhering to laminin-8 and laminin-10/11 were significantly different; cells on laminin-8 assume an elongated, often fanshaped morphology, while cells on laminin-10/11 became more spread and cobblestone-like. These differences in morphologies raise the possibility that signals transduced from laminin-8 through integrin a6h1 may not be identical to those transduced from laminin-10/11 through integrin a3h1 with regard to cytoskeletal reorganization, although both laminins up-regulate Rac among Rho family GTPases. Although the mechanism(s) defining such differences in morphology remain to be elucidated, one possible scenario could be that the ability of the integrins a3h1 and a6h1 to associate with tetraspanins, a family of proteins that span the plasma membrane four times, directs the signaling events downstream of these integrins. Thus, CD151, one of the tetraspanins, has been shown to tightly associate with integrin a3h1, but not a6h1, forming detergent-resistant stable complexes [41]. A potential role for CD151 in integrin a3h1-mediated signaling events is implied by its association with conventional-type PKC [42] and phosphatidylinositol 4-kinase [41]. There was an apparent discrepancy in integrin binding specificity between laminin-8 purified on an anti-laminin h1 mAb column and that purified on an anti-laminin a4 mAb column. T98G adhesion to the former was mediated by both a6h1 and a3h1 integrins, while cell adhesion to the latter was mediated by the a6h1 integrin alone. Although the

76

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77

reason for this discrepancy is not clear, addition of a small fraction of laminin-10/11 into laminin-8 (e.g., 1:40 molar ratio), which could not be detected by immunoblot analysis with the laminin a5 chain mAb 15H5, was capable of endowing strong a3h1 integrin-dependent cell-adhesive activity (H.F., data not shown), indicating that a small amount of contamination of laminin-10/11 or other laminin isoforms may account for the discrepancy in integrin binding specificity of laminin-8 purified on different immunoaffinity columns. HMEC-1 migration on laminin-8 was strongly inhibited by the antibody against the integrin a6 subunit and the dominant-negative mutant of Rac, suggesting the involvement of a6h1 integrin and Rac activation in endothelial cell migration on laminin-8. Recently, laminin-8 has been shown to promote migration of monocytes, lymphocytes, and bone marrow progenitor cells in an a6h1 integrin-dependent manner [43 – 46]. Rac activation may thus be involved in the migration of these cells on laminin-8. The interaction of the a6h1 integrin with laminin-8 might also promote endothelial cell migration in vivo, since their expression patterns overlap with those of newly formed capillaries, where endothelial cells are actively migrating [5,17,19,47]. Furthermore, laminins have been shown to be deposited around individual cells at the migrating tips of newly formed vessels [48]. These reports, together with our observations, suggest that laminin-8 is secreted and deposited by endothelial cells at invading tips of blood vessels and promotes endothelial cell migration in a6h1 integrin- and Rac-dependent manners. Activation of Rac and its downstream pathways are believed to be important in blood vessel formation [49]. The capillary tube assembly of endothelial cells induced by the Matrigel overlay has been shown to be inhibited by dominant-negative Rac, but not dominant-negative Cdc42 or the Rho-inhibitor C3 [50]. Capillary lumen formation in collagen or fibrin gels was also inhibited by dominantnegative Rac [51]. In this study, we demonstrated that Rac is preferentially activated upon endothelial cell adhesion to laminin-8 and plays a central role in endothelial cell spreading and migration. Furthermore, endothelial cell adhesion to laminin-10/11 strongly activates Rac, making it likely that these laminin isoforms regulate vessel formation via Rac activation. Further studies on the signals transduced from these laminins, as well as analyses of their functions in vessel formation in vivo, should shed light on the roles these laminins play in blood vessel formation during development, tissue regeneration, and tumor progression.

Acknowledgments We thank Dr. Masahiko Katayama for the mAb against integrin a6 subunit (AMCI 7-4), Dr. Caroline Damsky for the mAb against integrin h1 subunit (AIIB2), Dr. Ismo Virtanen for the mAb against laminin h1 chain (DG10),

and Noriko Sanzen for mAbs against laminin a5 chain (5D6) and laminin h1 chain (4F5). Y27632 was kindly provided by Yoshitomi Pharmaceutical Industries. We also thank Dr. Kozo Kaibuchi for supplying expression vectors for GST-CRIB, N19Rho, N17Rac, and N17Cdc42, and Dr. Martin Schwartz for providing expression vectors for GSTRBD. We are grateful to Ohoshi Murayama and Akemi Fujibayashi for supplying mouse laminin-1 and human laminin-10/11. This work was supported by the Special Coordination Funds and Grants-in-Aid from the Ministry of Education, Culture, Science, Sports, and Technology of Japan. H.F. is supported by the Research Fellowship of the Japan Society for the Promotion of Science for Young Scientists. References [1] B.P. Eliceiri, D.A. Cheresh, Adhesion events in angiogenesis, Curr. Opin. Cell Biol. 13 (2001) 563 – 568. [2] D.S. Grant, H.K. Kleinman, Regulation of capillary formation by laminin and other components of the extracellular matrix, Exs 79 (1997) 317 – 333. [3] H. Colognato, P.D. Yurchenco, Form and function: the laminin family of heterotrimers, Dev. Dyn. 218 (2000) 213 – 234. [4] R. Timpl, D. Tisi, J.F. Talts, Z. Andac, T. Sasaki, E. Hohenester, Structure and function of laminin LG modules, Matrix Biol. 19 (2000) 309 – 317. [5] J.H. Miner, B.L. Patton, S.I. Lentz, D.J. Gilbert, W.D. Snider, N.A. Jenkins, N.G. Copeland, J.R. Sanes, The laminin alpha chains: expression, developmental transitions, and chromosomal locations of a1 – 5, identification of heterotrimeric laminins 8 – 11, and cloning of a novel a3 isoform, J. Cell Biol. 137 (1997) 685 – 701. [6] M. Ekblom, M. Falk, K. Salmivirta, M. Durbeej, P. Ekblom, Laminin isoforms and epithelial development, Ann. N.Y. Acad. Sci. 857 (1998) 194 – 211. [7] M. Falk, M. Ferletta, E. Forsberg, P. Ekblom, Restricted distribution of laminin a1 chain in normal adult mouse tissues, Matrix Biol. 18 (1999) 557 – 568. [8] I. Virtanen, D. Gullberg, J. Rissanen, E. Kivilaakso, T. Kiviluoto, L.A. Laitinen, V.P. Lehto, P. Ekblom, Laminin a1-chain shows a restricted distribution in epithelial basement membranes of fetal and adult human tissues, Exp. Cell Res. 257 (2000) 298 – 309. [9] I. Leivo, E. Engvall, Merosin, a protein specific for basement membranes of Schwann cells, striated muscle, and trophoblast, is expressed late in nerve and muscle development, Proc. Natl. Acad. Sci. U. S. A. 85 (1988) 1544 – 1548. [10] B.L. Patton, J.H. Miner, A.Y. Chiu, J.R. Sanes, Distribution and function of laminins in the neuromuscular system of developing, adult, and mutant mice, J. Cell Biol. 139 (1997) 1507 – 1521. [11] W.G. Carter, M.C. Ryan, P.J. Gahr, Epiligrin, a new cell adhesion ligand for integrin a3h1 in epithelial basement membranes, Cell 65 (1991) 599 – 610. [12] M.P. Marinkovich, G.P. Lunstrum, D.R. Keene, R.E. Burgeson, The dermal – epidermal junction of human skin contains a novel laminin variant, J. Cell Biol. 119 (1992) 695 – 703. [13] J. Lohi, I. Leivo, K. Franssila, I. Virtanen, Changes in the distribution of integrins and their basement membrane ligands during development of human thyroid follicular epithelium, Histochem. J. 29 (1997) 337 – 345. [14] R.T. Libby, M.F. Champliaud, T. Claudepierre, Y. Xu, E.P. Gibbons, M. Koch, R.E. Burgeson, D.D. Hunter, W.J. Brunken, Laminin expression in adult and developing retinae: evidence of two novel CNS laminins, J. Neurosci. 20 (2000) 6517 – 6528.

H. Fujiwara et al. / Experimental Cell Research 292 (2004) 67–77 [15] A. Iivanainen, K. Sainio, H. Sariola, K. Tryggvason, Primary structure and expression of a novel human laminin a4 chain, FEBS Lett. 365 (1995) 183 – 188. [16] T. Niimi, C. Kumagai, M. Okano, Y. Kitagawa, Differentiation-dependent expression of laminin-8 (a4h1g1) mRNAs in mouse 3T3-L1 adipocytes, Matrix Biol. 16 (1997) 223 – 230. [17] N. Petajaniemi, M. Korhonen, J. Kortesmaa, K. Tryggvason, K. Sekiguchi, H. Fujiwara, L. Sorokin, L.E. Thornell, Z. Wondimu, D. Assefa, M. Patarroyo, I. Virtanen, Localization of laminin a4-chain in developing and adult human tissues, J. Histochem. Cytochem. 50 (2002) 1113 – 1130. [18] L.M. Sorokin, F. Pausch, M. Frieser, S. Kroger, E. Ohage, R. Deutzmann, Developmental regulation of the laminin a5 chain suggests a role in epithelial and endothelial cell maturation, Dev. Biol. 189 (1997) 285 – 300. [19] M. Sixt, B. Engelhardt, F. Pausch, R. Hallmann, O. Wendler, L.M. Sorokin, Endothelial cell laminin isoforms, laminins 8 and 10, play decisive roles in T cell recruitment across the blood – brain barrier in experimental autoimmune encephalomyelitis, J. Cell Biol. 153 (2001) 933 – 946. [20] J. Thyboll, J. Kortesmaa, R. Cao, R. Soininen, L. Wang, A. Iivanainen, L. Sorokin, M. Risling, Y. Cao, K. Tryggvason, Deletion of the laminin a4 chain leads to impaired microvessel maturation, Mol. Cell. Biol. 22 (2002) 1194 – 1202. [21] H. Fujiwara, Y. Kikkawa, N. Sanzen, K. Sekiguchi, Purification and characterization of human laminin-8. Laminin-8 stimulates cell adhesion and migration through a3h1 and a6h1 integrins, J. Biol. Chem. 276 (2001) 17550 – 17558. [22] T. Geberhiwot, S. Ingerpuu, C. Pedraza, M. Neira, U. Lehto, I. Virtanen, J. Kortesmaa, K. Tryggvason, E. Engvall, M. Patarroyo, Blood platelets contain and secrete laminin-8 (a4h1g1) and adhere to laminin-8 via a6h1 integrin, Exp. Cell Res. 253 (1999) 723 – 732. [23] J. Kortesmaa, P. Yurchenco, K. Tryggvason, Recombinant laminin-8 (a4h1g1). Production, purification, and interactions with integrins, J. Biol. Chem. 275 (2000) 14853 – 14859. [24] J. Kortesmaa, M. Doi, M. Patarroyo, K. Tryggvason, Chondroitin sulphate modification in the a4 chain of human recombinant laminin-8 (a4h1g1), Matrix Biol. 21 (2002) 483 – 486. [25] E.A. Jaffe, R.L. Nachman, C.G. Becker, C.R. Minick, Culture of human endothelial cells derived from umbilical veins. Identification by morphologic and immunologic criteria, J. Clin. Invest. 52 (1973) 2745 – 2756. [26] M. Katayama, N. Sanzen, A. Funakoshi, K. Sekiguchi, Laminin g2chain fragment in the circulation: a prognostic indicator of epithelial tumor invasion, Cancer Res. 63 (2003) 222 – 229. [27] R. Manabe, N. Ohe, T. Maeda, T. Fukuda, K. Sekiguchi, Modulation of cell-adhesive activity of fibronectin by the alternatively spliced EDA segment, J. Cell Biol. 139 (1997) 295 – 307. [28] Y. Kikkawa, N. Sanzen, H. Fujiwara, A. Sonnenberg, K. Sekiguchi, Integrin binding specificity of laminin-10/11: laminin-10/11 are recognized by a3h1, a6h1 and a6h4 integrins, J. Cell Sci. 113 (Pt 5) (2000) 869 – 876. [29] M. Paulsson, M. Aumailley, R. Deutzmann, R. Timpl, K. Beck, J. Engel, Laminin – nidogen complex. Extraction with chelating agents and structural characterization, Eur. J. Biochem. 166 (1987) 11 – 19. [30] Y. Kikkawa, N. Sanzen, K. Sekiguchi, Isolation and characterization of laminin-10/11 secreted by human lung carcinoma cells. Laminin10/11 mediates cell adhesion through integrin a3h1, J. Biol. Chem. 273 (1998) 15854 – 15859. [31] K. Sekiguchi, S. Hakomori, Domain structure of human plasma fibronectin. Differences and similarities between human and hamster fibronectins, J. Biol. Chem. 258 (1983) 3967 – 3973. [32] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680 – 685. [33] M. Doi, J. Thyboll, J. Kortesmaa, K. Jansson, A. Iivanainen, M. Parvardeh, R. Timpl, U. Hedin, J. Swedenborg, K. Tryggvason, Recombinant human laminin-10 (a5h1g1). Production, purification, and

[34]

[35]

[36]

[37] [38]

[39] [40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

77

migration-promoting activity on vascular endothelial cells, J. Biol. Chem. 277 (2002) 12741 – 12748. J.H. Miner, R.M. Lewis, J.R. Sanes, Molecular cloning of a novel laminin chain, a5, and widespread expression in adult mouse tissues, J. Biol. Chem. 270 (1995) 28523 – 28526. A. Sonnenberg, F. Hogervorst, A. Osterop, F.E. Veltman, Identification and characterization of a novel antigen complex on mouse mammary tumor cells using a monoclonal antibody against platelet glycoprotein Ic, J. Biol. Chem. 263 (1988) 14030 – 14038. M.E. Hemler, C. Crouse, A. Sonnenberg, Association of the VLA a2;6 subunit with a novel protein. A possible alternative to the common VLA h1 subunit on certain cell lines, J. Biol. Chem. 264 (1989) 6529 – 6535. E.A. Cox, A. Huttenlocher, Regulation of intetrin-mediated adhesion during cell migration, Microsc. Res. Tech. 43 (1998) 412 – 419. S.P. Palecek, A. Huttenlocher, A.F. Horwitz, D.A. Lauffenburger, Physical and biochemical regulation of integrin release during rear detachment of migrating cells, J. Cell Sci. 111 (1998) 929 – 940. K. Rottner, A. Hall, J.V. Small, Interplay between Rac and Rho in the control of substrate contact dynamics, Curr. Biol. 9 (1999) 640 – 648. E.E. Sander, J.P. ten Klooster, S. van Delft, R.A. van der Kammen, J.G. Collard, Rac downregulates Rho activity: reciprocal balance between both GTPases determines cellular morphology and migratory behavior, J. Cell Biol. 147 (1999) 1009 – 1022. R.L. Yauchi, F. Berditchevski, M.B. Harler, J. Reichner, M.E. Hemler, Highly stoichiometric, stable, and specific association of integrin a3h1 with CD151 provides a major link to phosphatidylinositol 4-kinase, and may regulate cell migration, Mol. Biol. Cell 9 (1998) 2751 – 2765. X.A. Zhang, A.R. Kazarov, X. Yang, A.L. Bontrager, C.S. Stipp, M.E. Hemler, Function of the tetraspanin CD151-a6h1 integrin complex during cellular morphogenesis, Mol. Biol. Cell 13 (2002) 1 – 11. C. Pedraza, T. Geberhiwot, S. Ingerpuu, D. Assefa, Z. Wondimu, J. Kortesmaa, K. Tryggvason, I. Virtanen, M. Patarroyo, Monocytic cells synthesize, adhere to, and migrate on laminin-8 (a4h1g1), J. Immunol. 165 (2000) 5831 – 5838. T. Geberhiwot, D. Assefa, J. Kortesmaa, S. Ingerpuu, C. Pedraza, Z. Wondimu, J. Charo, R. Kiessling, I. Virtanen, K. Tryggvason, M. Patarroyo, Laminin-8 (a4h1g1) is synthesized by lymphoid cells, promotes lymphocyte migration and costimulates T cell proliferation, J. Cell Sci. 114 (2001) 423 – 433. M. Sixt, R. Hallmann, O. Wendler, K. Scharffetter-Kochanek, L.M. Sorokin, Cell adhesion and migration properties of h2-integrin negative polymorphonuclear granulocytes on defined extracellular matrix molecules. Relevance for leukocyte extravasation, J. Biol. Chem. 276 (2001) 18878 – 18887. Y.C. Gu, J. Kortesmaa, K. Tryggvason, J. Persson, P. Ekblom, S.E. Jacobsen, M. Ekblom, Laminin isoform-specific promotion of adhesion and migration of human bone marrow progenitor cells, Blood 101 (2002) 877 – 885. L.M. Sorokin, M.A. Maley, H. Moch, H. von der Mark, K. von der Mark, L. Cadalbert, S. Karosi, M.J. Davies, J.K. McGeachie, M.D. Grounds, Laminin a4 and integrin a6 are upregulated in regenerating dy/dy skeletal muscle: comparative expression of laminin and integrin isoforms in muscles regenerating after crush injury, Exp. Cell Res. 256 (2000) 500 – 514. D.M. Form, B.M. Pratt, J.A. Madri, Endothelial cell proliferation during angiogenesis. In vitro modulation by basement membrane components, Lab. Invest. 55 (1986) 521 – 530. G.E. Davis, K.J. Bayless, A. Mavila, Molecular basis of endothelial cell morphogenesis in three-dimensional extracellular matrices, Anat. Rec. 268 (2002) 252 – 275. J.O. Connolly, N. Simpson, L. Hewlett, A. Hall, Rac regulates endothelial morphogenesis and capillary assembly, Mol. Biol. Cell 13 (2002) 2474 – 2485. K.J. Bayless, G.E. Davis, The Cdc42 and Rac1 GTPases are required for capillary lumen formation in three-dimensional extracellular matrices, J. Cell Sci. 115 (2002) 1123 – 1136.