Developmental Biology 237, 116–129 (2001) doi:10.1006/dbio.2001.0363, available online at http://www.idealibrary.com on
View metadata, citation and similar papers at core.ac.uk brought to you by CORE ␣111 Integrin Is a Receptor for Interstitial provided by Elsevier - Publisher Connector Collagens Involved in Cell Migration and Collagen Reorganization on Mesenchymal Nonmuscle Cells
Carl-Fredrik Tiger,* Francoise Fougerousse,† Gunilla Grundstro¨m,‡ Teet Velling,‡ and Donald Gullberg‡,1 *Department of Cell and Molecular Biology, Biomedical Center, Box 596, Uppsala University, S-75124 Uppsala, Sweden; †Laboratoire de Histoembryologie et de Cytoge´ne´tique, Faculte´ Cochin Port Royal, Paris 75014, France; and ‡Department of Medical Biochemistry and Microbiology, Biomedical Center, Box 582, Uppsala University, S-75123 Uppsala, Sweden
␣111 integrin constitutes a recent addition to the integrin family. Here, we present the first in vivo analysis of ␣11 protein and mRNA distribution during human embryonic development. ␣11 protein and mRNA were present in various mesenchymal cells around the cartilage anlage in the developing skeleton in a pattern similar to that described for the transcription factor scleraxis. ␣11 was also expressed by mesenchymal cells in intervertebral discs and in keratocytes in cornea, two sites with highly organized collagen networks. Neither ␣11 mRNA nor ␣11 protein could be detected in myogenic cells in human embryos. The described expression pattern is compatible with ␣111 functioning as a receptor for interstitial collagens in vivo. To test this hypothesis in vitro, full-length human ␣11 cDNA was stably transfected into the mouse satellite cell line C2C12, lacking endogenous collagen receptors. ␣111 mediated cell adhesion to collagens I and IV (with a preference for collagen I) and formed focal contacts on collagens. In addition, ␣111 mediated contraction of fibrillar collagen gels in a manner similar to ␣21, and supported migration on collagen I in response to chemotactic stimuli. Our data support a role for ␣111 as a receptor for interstitial collagens on mesenchymally derived cells and suggest a multifunctional role of ␣111 in the recognition and organization of interstitial collagen matrices during development. © 2001 Academic Press Key Words: ␣111 integrin; immunohistochemistry; in situ hybridization; human embryogenesis; collagen binding; collagen gel contraction; cell migration.
INTRODUCTION subunits, which can assemble into 24 different het- erodimers, are known (Velling et al., 1999). Collagens are major constituents of various types of Cellular interactions with the extracellular matrix (ECM) specialized ECM. Until recently, only two members of the are important for fundamental biological processes such as integrin family, ␣11 and ␣21, were known to act as cell proliferation, cell differentiation, cell migration, apo- collagen receptors. More recently, ␣101 (Camper et al., ptosis, morphogenesis, and organogenesis. Integrins are one 1998) and ␣111 (Velling et al., 1999) have been added to family of proteins mediating cell–ECM interactions, and this subfamily of collagen-binding integrins. The ␣31 many members have proven essential for embryonic devel- integrin, initially described as a collagen receptor (Wayner opment (Brakebusch et al., 1997; Hynes, 1996; Sheppard, and Carter, 1987), has later been shown to serve as an 2000). Integrins are heterodimeric transmembrane recep- important laminin-5 receptor (Carter et al., 1991). Genetic tors consisting of noncovalently assembled ␣ and  sub- data, however, have indicated that ␣31 can affect the units (Hynes, 1992). To date, 8 different  and 18 different ␣ activity of the collagen receptor ␣21 through receptor cross-talk (Hodivala-Dilke et al., 1998). 1 To whom correspondence should be addressed. Fax: ϩ46-18- The distribution of the ␣11 and ␣21 in vivo has been 471-4862. E-mail: [email protected]. studied in detail. During embryonic development, ␣11 and
0012-1606/01 $35.00 Copyright © 2001 by Academic Press 116 All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 117
␣21 are widely expressed (Duband et al., 1992; Gardner et heart and skeletal muscle (Velling et al., 1999). However, al., 1996; Wu and Santoro, 1994). Postnatally, the ␣1 chain neither the cellular origin of the ␣11 mRNA in these tissues is strongly expressed on microcapillary endothelium and nor the expression pattern of ␣11 during development is pericytes as well as in smooth muscle cells (Voigt et al., known. 1995). The ␣2 chain is present on platelets, haematopoetic In the current study, we have determined ␣11 protein and cells, and a variety of epithelia (Gardner et al., 1996; Keely mRNA expression in human embryos and have performed et al., 1995; Wu and Santoro, 1994). Both ␣1 and ␣2 are in vitro studies of cell adhesion, cell migration, and colla- present on fibroblasts (Voigt et al., 1995; Wu and Santoro, gen gel contraction. The distribution pattern shows that 1994). ␣11 during embryonic development displays a restricted The ligand specificities of ␣11 and ␣21 differ for expression on mesenchymal nonmuscle cells in areas of collagens I, IV, and XIII (Dickeson and Santoro, 1998; highly organized interstitial collagen networks. The re- Kapyla et al., 2000; Kern et al., 1994; Nykvist et al., 2000; stricted distribution suggests that ␣11 is under strict tran- Tuckwell et al., 1996). Whereas the binding of the intersti- scriptional control and might be one of the sought-after cell tial collagen I by ␣11 and ␣21 involves helical GFOGER adhesion genes acting downstream of scleraxis, involved in and related sequences (Knight et al., 1998, 2000; Xu et al., prefiguring the axial and appendicular skeleton. The in 2000), the ␣11 binding of the basement membrane colla- vitro data show that ␣111 can take part in collagen- gen IV involves specific R and D residues present on mediated events such as cell migration, collagen deposi- different collagen ␣-chains (Eble et al., 1993; Golbik et al., tion, and collagen reorganization, suggesting that ␣111 in 2000). The residues in collagen XIII recognized by ␣11 vivo could fulfill these important functions and play an (Nykvist et al., 2000) have not been identified. Neither the important collagen organizing role during cartilage, inter- exact collagen specificity of ␣101 and ␣111 nor the vertebral disc, and cornea organogenesis. regions within collagens recognized by ␣101 and ␣111 are known. An important in vivo function of cell–collagen interac- tions is to reorganize the collagen matrix within developing EXPERIMENTAL PROCEDURES tendons, ligaments, periosteum, capsules of organs, and in healing wounds (Harris et al., 1981; Stopak and Harris, Cells 1982). The ability of cells to contract three-dimensional collagen matrices in vitro reflects their potential to modu- Normal diploid human fibroblasts (AG1518; Genetic Mutant late a collagen-rich matrix. ␣21 serves as a major 1 Cell Repository, Camden, NJ) were obtained from Dr. Kristofer integrin receptor involved in contraction of collagen gels Rubin (Uppsala University). Skin fibroblasts and keratocytes were (Klein et al., 1991), whereas the ability of ␣11 to contract obtained from Natalie Isnard (Laboratoire de Recherche en collagen gels seems to depend on the cell type (Carver et al., Opthamologie, Paris). Human satellite cells (XXVI) isolated from a 1995; Gotwals et al., 1996; Jenkins et al., 1999; Kagami et 2.5-year-old child were provided by Dr. Helen Blau (Stanford ␣ al., 1999; Racine-Samson et al., 1997). For some cell types, University). Other human cell lines tested for 11 expression an ␣v3-dependent contraction of collagen gels has been included Caco-2 (colon adenocarcinoma, ATCC No. HTB-37), JAR (human choriocarcinoma cells, ATCC No. HTB-144), RD (rhabdo- observed (Agrez et al., 1991; Cooke et al., 2000; Jones et al., myosarcoma, ATCC No. CCL-136), HT-29 cells (human colon 1997; Nunohiro et al., 1999). Induction of collagenases adenocarcinoma, ATCC No. HTB-38), and HT1080 (kindly pro- leading to denaturation of collagen or deposition of RGD- vided by S. Johansson). Murine C2C12 myoblast cells from the containing ligands has been suggested to underlie the ␣v3- American Type Culture Collection were provided by A. Starzinski- dependency of this contraction. Powitz. Cells were cultured at 37°C in Dulbecco’s modified Eagle’s The biological significance of cell–collagen interactions medium, supplemented with 10% fetal bovine serum (FBS) and during development and in adult tissues is still far from antibiotics (Statens veterina¨rmedicinska anstalt, Uppsala). The clear. Mice deficient in ␣1 integrin chain are viable, but cells were grown to subconfluency and passaged every 2–3 days. display a proliferation defect and a disturbed regulation of collagen synthesis in skin fibroblasts (Gardner et al., 1996, 1999; Pozzi et al., 1998). Tumor angiogenesis is also defec- Antibodies tive in mice lacking ␣1 integrin chain (Pozzi et al., 2000). Data from inactivation of ␣2, ␣10, and ␣11 integrin genes in The rabbit antibodies to the cytoplasmic tail of ␣11 integrin have mice are not yet available. been described previously (Velling et al., 1999). To immunoprecipi-   ␣111 was first identified as a major integrin in cultured tate 1 integrins, a polyclonal antibody to rat integrin 1 chain was used (Gullberg et al., 1989). Monoclonal antibodies (MoAb) to skeletal muscle cells (Gullberg et al., 1995b; Velling et al., ␣ ␣ ␣ integrin 1 chain (FB12, MAB 1973) and integrin 2 chain (BHA2.1, 1999). During myogenic differentiation in vitro, the 11 MAB 1998) were obtained from Chemicon International Inc. As a chain is up-regulated at both the protein and the mRNA marker for forming cartilage, two collagen II MoAbs were used: levels. Analysis of mRNA from a panel of adult human mouse MoAb CIIC1 (Klareskog et al., 1986) and rat MoAb 126.7 (P. tissues reveals expression of ␣11 in a wide variety of tissues Wernhoff, unpublished observations), both obtained from Dr. with the highest level of expression in uterus, followed by Patrik Wernhoff (Lund University).
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 118 Tiger et al.
Embryos Immunoprecipitation and Electrophoresis Morphologically normal human embryos (aged from 4 to 8 Cell cultures were washed three times in DMEM medium postovulatory weeks) were obtained from abortions induced by devoid of cysteine and methionine and metabolically labeled Mifepristone (RU486) at Hopital Broussais in Paris. All procedures overnight in the presence of 25 Ci/ml of 35S methionine/cysteine were approved by the Ethical Committee of Saint-Vincent de Paul (pro-Mix [35S] cell labeling mix; Amersham Pharmacia Biotech). Hospital in Paris. Each sample was first examined macroscopically Proteins were extracted from the tissue culture dishes by the during dissection under a stereo-microscope; the developmental addition of 1 ml of solubilization buffer (1% Triton X-100, 0.15 M
stage of the embryos was determined by using established criteria. NaCl, 20 mM Tris–HCl, pH 7.4, 1 mM MgCl2, 1 mM CaCl2) Embryos were collected shortly after delivery and frozen within the containing protease inhibitors (1 mM Pefabloc SC; Roche Diagnos- first 24-h postmortem on dry ice and stored at Ϫ80°C until used. tics), 1% aprotinin, 1 g/ml pepstatin, 1 g/ml leupeptin). Solubi- Seven-micron-thick cryostat sections were mounted on slides lized proteins were centrifuged for 10 min at 15,000g. The centri- previously coated with a 2% 3-aminopropyl-triethoxysilane solu- fuged supernatant was precleared by incubating with preimmune tion in acetone. For in situ hybridization, sections were postfixed IgG 100 g/ml and protein A-Sepharose CL 4B (Pharmacia Biotech for 20 min in 2% paraformaldehyde in 0.1 M phosphate buffer (pH AB) for 2 h. Following centrifugation, immune IgG was incubated 7.4) at room temperature, rinsed (once for 2 min) in phosphate with the extract for 2 h. Specifically bound proteins were recovered buffer, rinsed briefly in water, and dehydrated with a series of with protein A–Sepharose. The precipitate was washed three times washes in ethanol solutions (50, 75, 100%). Sections were then with buffer A (1% Triton X-100, 0.5 M NaCl, 20 mM Tris–HCl, pH
air-dried and finally stored at Ϫ80°C. 7.4, 1 mM MgCl2, 1 mM CaCl2), three times with buffer B (0.1% Triton X-100, 0.15 M NaCl, 20 mM Tris–HCl, pH 7.4, 1 mM
MgCl2, 1 mM CaCl2) prior to solubilization in electrophoresis In Situ Hybridization sample buffer. Proteins were separated on 6% SDS–polyacrylamide gels and processed for fluorography. Precipitated antisense riboprobe derived from 3Ј end of integrin ␣11 cDNA sequence (nt 2130–3893; Velling et al., 1999) was labeled with [␣-33P]UTP (Dupont-Nemours) using T7 RNA poly- Generation of ␣11 Full-Length cDNA merase (Promega). Control sense riboprobe was synthesized on the opposite strand with the T3 RNA polymerase. In situ hybridization A full-length ␣11 cDNA was generated by joining clones 1.1 (nt medium contained 50% formamide, 0.3 M NaCl, 1ϫ Denhardt’s 37–917) and 2.1 (nt 361-3747) (Velling et al., 1999) via an overlap- solution, 0.5 mg/ml tRNA, 10% dextran sulfate, 20 mM Tris (pH ping ApoI site. Briefly, clone 1.1 was excised with ApoI and placed 7.4), 5 mM EDTA (pH 8.0), and [␣-33P]UTP-labeled riboprobe (5 ϫ in a pUC19 vector modified to lack the EcoRI/ApoI site. A fragment 104 cpm/l). After denaturation, 100 l of hybridization solution encompassing nt 740-3747 was excised from clone 2.1 and ligated was deposited on each slide, covered with a coverslip, and incu- into the ApoI site of clone 1.1 creating a full-length cDNA (nt bated in a humidified chamber at 55°C overnight. After hybridiza- 37–3747) followed by an untranslated repeat of nt 740–917. The tion, the sections were washed, once for 30 min at 42°C in 5ϫ SSC, full-length cDNA (including the repeat nucleotides) was inserted once for 20 min at 60°C in 2ϫ SSC containing 50% formamide, into the pBJ-1 expression vector (Takebe et al., 1988) and used for once for 40 min at 37°C in 1ϫ washing solution containing 20 transfection. mg/ml RNAse A, once for 15 min at 37°C in 1ϫ washing solution, ϫ once for 15 min at 37°C in 2 SSC, and finally once for 15 min at Transfection 37°C in 0.1ϫ SSC. Sections were then dehydrated with a series of graded concentrations of ethanol and exposed to Kodak C2C12 cells, cultured in a 25-cm2 culture flask to 80% conflu- Biomax-MR films for 7 days or dipped in Kodak NTB-2 nuclear ency, were transfected with 5 g ␣2-pBJ-1(Kamata et al., 1994) and track emulsion and exposed up to 1 month prior to development. ␣11-pBJ-1 constructs using Superfect (Qiagen) according to the protocol of the manufacturer. The ␣2-pBJ-1 (Kamata et al., 1994) and ␣11-pBJ-1 DNA constructs were co-transfected with the PGK Indirect Immunofluorescence puro vector (Skerjanc et al., 1994) at a ratio of 10:1 and clones were Cells cultured on coverslips were washed in serum-free medium selected in 10 g/ml of puromycin. Individual clones were isolated ␣ ϩ and fixed for 2 min in methanol at Ϫ20°C. Nonspecific binding either by limiting dilution in microtiter plates (C2C12- 11 ), or by sites were blocked by incubating with 10% goat serum (Statens colony picking from petri dishes. Selected clones were further ␣ ␣ veterina¨rmedicinska anstalt, Uppsala) diluted in phosphate buff- assayed for 2 and 11 expression with immunoprecipitation and ␣ ␣ ered saline (PBS). Primary antibodies were incubated with fixed immunohistchemistry using specific antibodies for 2 and 11. cells for 1.5 h at 37°C. Bound antibodies were detected using Cy3-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch). Cell attachment assay Cryosections were left unfixed prior to blocking of nonspecific binding sites with goat serum. For double immunofluorescence, Twenty-four-well cell culture plates (Nunc) were coated with sections were incubated with both primary antibodies simulta- ligands (500 l to a 2-cm2 well) diluted in PBS over night at 4°C, neously. Bound antibodies were detected by using a mixture of followed by blocking with 2% BSA in PBS for1hat37°C and then Cy3-conjugated goat anti-rabbit IgG and FITC-conjugated goat washed in PUCK’s saline (137 mM NaCl, 5 mM KCl, 4 mM
anti-rat IgG (Jackson, multiple labelling grade). Na2CO3, 5.5 mM D-glucose, equilibrated with 5% CO2 at 37°C). Stained cells and tissue sections were mounted in Vectashield Ligands used were: collagen type I (Vitrogen), collagen type IV mounting medium (Vector Laboratories Inc.), visualized, and pho- (Sigma), and human plasma fibronectin (provided by S. Johansson, tographed under a Zeiss Axiophot microscope equipped with optics Uppsala University). Wells were filled with PUCK’s saline and
for observing fluorescence. MgCl2 ϩ CaCl2 was added to obtain a final concentration (after
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 119
addition of cells) of 2 mM MgCl2 and 0.01 mM CaCl2. For cation-dependency assays, only one type of cation was added at various concentrations. Cells were trypsinized and washed four times in PUCK’s saline, and seeded into the wells at a concentra- tion of 250,000 cells/well, and were allowed to attach for 45 min at
37°C and 5% CO2. Wells were washed three times in PUCK’s saline and plates were rapidly frozen at Ϫ20°C for later assay by using the hexoseaminidase test as follows: Attached cells were lysed in 200 l S-solution (3.75 mM p-nitrophenol-N-acetyl--D- glucoseamide, 0.1 M sodium acetate, pH 5, and 0.25% Triton X-100) was added. The plates were incubated at 37°C for a mini- mum of 2 h. Fifty microliters of the cell lysates were transferred to a mictrotiter plate (Nunc) and mixed with 75 l developing buffer (4.5 mM EDTA and 45 mM glycine, pH 10.4). The absorbance at 405 nm was read and used as a measure of cell number. For each cell line used, a cell number standard was made. Each experiment was performed in triplicates. In order to minimize errors from unequal trypsinization stress between cell lines and handling of plates, etc., data were normalized as follows: For each plate, the adhesion to fibronectin (10 g/ml) was used as 100% reference and the background found on BSA-only-coated wells were used as a 0% reference. In all tests, the magnitude of absolute measurements between fibronectin and BSA was always greater than 100:1.
Cell Migration
Cell migration was assayed by using Transwell plates (Costar No. 3422) with a polycarbonate membrane (8-m pores) separating the upper and lower chambers. Membranes were coated with collagen type I (100 g/ml) in PBS overnight at ϩ4°C. Thirty thousand cells in MCDB 104 (100 l) were added to the upper chamber and MCDB 104 (Statens veterina¨rmedicinska anstalt, Uppsala) (600 l) to the lower. Where indicated, 10 ng/ml PDGF-BB or 10% FCS was added to the lower or both chambers to evaluate chemotactic as well as chemokinetic responses. The cells were allowed to migrate for4hat37°C in 5% CO2. After washing, the cells were fixed in methanol for 10 min at room temperature and then stained with Mayer’s Hematoxyline. Cells on the upper surface of the membrane were removed and the membranes were FIG. 1. ␣11 integrin mRNA distribution at 6 weeks of gestation. cut out and mounted on glass slides. Cells that had migrated were Composite of in situ hybridization for ␣11 integrin mRNA on a quantified by counting the transmigrated cells in seven measuring sagital section of a human embryo from 6 weeks of gestation. ␣11 fields of the membrane. integrin mRNA is strongly expressed around cartilage in ribs and in the hindlimb. Arrows point to signal around ribs and forming cartilage in limbs. Asterisk marks nonspecific signal in the eye, also present in control slide hybridized with sense probe. Collagen Gel Contraction
The collagen gel contraction assay has been described previously (Gullberg et al., 1990). In summary, 96-well dishes (Nunc, Roskilde, Denmark) were coated with sterile BSA (2% in PBS) overnight at 4°C and then washed in sterile PBS. A collagen RESULTS solution was made up from five parts double concentrated MCDB 104, one part 0.2 M Hepes, pH 8.0, and four parts collagen type I ␣ (Vitrogen 100, 3 mg/ml). Trypsinized cells were washed in MCDB 11 Protein and mRNA Expression Patterns in 104 containing 0.02% FBS and diluted to 1,000,000 cells/ml. Cells Human Embryos were mixed with the collagen solution to a final concentration of Six-week human embryo. In situ hybridization of sec- 100,000 cells/ml and 1.1 mg collagen/ml. To each well, 100 l cell/collagen suspension was added and gels were allowed to form tions of 6 weeks of gestation (wg) human embryo showed a for 1 h at 37°C. MCDB 104 (100 l), with or without modulating distinct signal (Fig. 1). In the section shown in Fig. 1, the factors, was ejected into the wells to detach the gels. The gels were ␣11 in situ signal appeared around cartilage anlagen includ- further incubated at 37°C. Gel diameters were measured under the ing forming ribs and the distal parts of the limbs (Fig. 1). At microscope at indicated time points. 6 weeks, primary myogenesis is ongoing and immature
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 120 Tiger et al.
FIG. 2. Distribution of ␣11 mRNA transcripts at 8 weeks of gestation. In situ hybridization of ␣11 mRNA on sagital sections from a human embryo at 8 weeks of gestation. Low-resolution autoradiogram highlights areas around forming cartilage (A). Boxed areas in (A) are shown in (B) and (C). ␣11 mRNA is transcribed at high levels around ribs (B), and in intervertebral discs (C). Bars in (B) and (C) represent 800 m. The arrow depicts the rostro-caudal orientation of the embryo. primary myotubes are present, but no ␣11 transcripts were Presence of ␣111 Integrin on Mesenchymal Cells found in muscle tissues. in Vitro Eight-week human embryo. In the sections from 8 wg An ␣11 integrin cytoplasmic tail-specific antibody (Vel- ␣ embryos (Fig. 2), 11 mRNA was transcribed at high levels ling et al., 1999) was used to assay a number of human in areas around forming ribs (Figs. 2A and 2B), around primary cells and cell lines for ␣11 expression. 1518 human forming vertebrae, and in the intervertebral discs (Figs. 2A foreskin fibroblasts (Fig. 5A) and to a lesser degree kerato- and 2C). As in the 6-week embryo section, no ␣11 transcrip- cytes and HT1080 fibrosarcoma cells (data not shown) tion could be detected in myogenic cells. expressed the ␣111 integrin, giving further evidence for an Fig. 3 shows a composite of a section from an 8-week-old expression of ␣11 largely restricted to mesodermally de- embryo immunostained for ␣11 integrin. The strongest rived cells. In the human satellite cells XXVI, ␣111isthe sites of ␣11 expression were found in cells surrounding dominating collagen-binding integrin (Velling et al., 1999), forming ribs, vertebrae, and in intervertebral discs. A sepa- whereas, in the 1518 fibroblasts, ␣21 integrin was ex- rate embryo section showed a strong signal for ␣11 in pressed at the highest level (Fig. 5A). Interestingly, 16 h keratocytes of the cornea (Fig. 4A). Closer examination treatment with TGF-1 (2 ng/ml) did not cause induction of showed that around ribs ␣11-positive cells appeared as ␣11 protein in 1518 fibroblasts or HT1080 fibrosarcoma aligned, stretched, fibroblast-like cells in close proximity to cells (data not shown). Human foreskin fibroblasts localized ␣ forming cartilage (Fig. 4B). To determine ␣11 expression in 11 to focal contacts on collagen throughout the cell, ␣ ␣ relation to cartilage, sections were double stained with similar to 2 integrin chain, whereas 1 integrin was more antibodies to collagen II. Cells in areas around collagen restricted and confined to cell edges (Fig. 5B). In long-term ␣ II-positive cartilage in the forelimb expressed ␣11 (Fig. 4C). cultures, prominent 11 localization to matrix contacts on the dorsal surface of cells was seen (Fig. 5B). In the intervertebral discs, ␣11-positive cells were not in contact with the cartilage-produced collagen II (Fig. 4D). A ␣  ␣  mutually exclusive pattern of ␣11 integrin chain and colla- Generation of Cell Lines Expressing 2 1or 11 1 gen II expression at sites of cartilage formation was thus Integrin Collagen Receptors observed and ␣11 integrin chain was mainly found in ␣111-mediated cell–collagen interactions were studied mesenchymal cells outside the forming cartilage. by using the well-characterized C2C12 mouse satellite cell
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 121
ysis of metabolically labeled ␣2 (C2C12-␣2ϩ) and ␣11 (C2C12-␣11ϩ) expressing cell lines are shown in Fig. 6, lanes 3–11. The incorporation of 35S Met and 35S Cys is predicted to be similar since ␣2 integrin contains 18 cys- teines and 22 methionines and ␣11 contains 20 cysteines and 18 methionines. Expression levels of ␣2 and ␣11 on the respective clones were estimated by immunoprecipitation with a 1 antibody and analysis of unreduced and reduced samples by SDS–polyacrylamide gel electrophoresis. Slightly less ␣11 chain in comparison with ␣2 chain copre- cipitated with the 1 chain (Fig. 6, lanes 8 and 10).
Comparison of Cell Adhesion in Cell Lines Expressing ␣111 Integrin In all further studies, cells expressing ␣111 as sole collagen receptor (C2C12-␣11ϩ) were compared with cell lines expressing ␣21 as the only collagen receptor (C2C12- ␣2ϩ). Efficient cell adhesion to collagen I-coated plates was reached at a ligand concentration of 1 g/ml for both C2C12-␣2ϩ and C2C12-␣11ϩ cells (Figs. 7A and 7B). Both C2C12-␣2ϩ and C2C12-␣11ϩ cells adhered considerably better to collagen I, compared with collagen IV, especially when coated at concentrations below 1 g/ml (Figs. 7A and 7B). Untransfected C2C12 cells failed to bind to either collagen (Fig. 7B). Test of ion-dependence showed maxi- mum cell adhesion to collagen I of C2C12-␣2ϩ and C2C12- ␣11ϩ cells in the presence of 1 mM Mg2ϩ, whereas Ca2ϩ failed to support adhesion (Figs. 7C and 7D). Analysis of C2C12-␣11 cells on collagen I and collagen IV revealed prominent positive protrusions and focal contacts in freshly adhered cells, which were not seen in cells cultured on fibronectin (data not shown).
Ability of ␣111 to Mediate Collagen Gel Contraction and Cell Migration ␣ FIG. 3. 11 integrin protein distribution at 8 weeks of gestation. To determine whether ␣111 is involved in cell migra- Composite of immunohistochemical staining of sagital section of tion and is able to mediate reorganization of collagen human embryo at 8 wg. Note high levels of ␣11 protein around matrices, C2C12, C2C12-␣2ϩ, and C2C12-␣11ϩ cells were vertebrae (arrows), in intervertebral disc (asterisks), around ribs (thin arrows), and around forming cartilage in the distal limb tested in chemotaxis and collagen gel contraction assays. (arrowhead). We have previously determined that fetal bovine serum and PDGF can induce 1 integrin-mediated collagen gel con- traction (Gullberg et al., 1990). In this first analysis of the ability of ␣111 to mediate cell migration and collagen gel line. C2C12 cells did not adhere to, or migrate on, collagen contraction, we chose to limit our analysis to these re- I (see Figs. 7 and 9 below). Immunoprecipitation with a agents. polyclonal antibody to 1 integrin chain and SDS–PAGE C2C12-␣111 cells mediated collagen gel contraction analysis of immunoprecipitates from C2C12 cells did not almost as efficiently as C2C12-␣2ϩ cells (Figs. 8B and 8C). identify either the ␣1or␣11 chains nor any other ␣-chain Interestingly, the parental cell line C2C12 lacking all col- showing an upward shift in SDS–PAGE upon reduction (a lagen receptors also contracted the collagen gels (Fig. 8A), feature of I domain ␣ chains) (Fig. 6, lane 1, and data not although this contraction was delayed and weaker in com- shown). In agreement with the cell attachment data, we parison with C2C12-␣2ϩ and C2C12-␣11ϩ cells. Both thus failed to detect endogenous collagen binding integrins PDGF-BB and serum stimulated ␣21- and ␣111-mediated in C2C12 cells. Full-length human ␣2 and ␣11 integrin collagen contraction, with PDGF-BB consistently being a construct were stably transfected into C2C12 cells and high slightly more potent stimulator of ␣111-mediated contrac- expressing clones were selected. Immunoprecipitation anal- tion in comparison with 10% fetal bovine serum (Fig. 8C).
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 122 Tiger et al.
FIG. 4. ␣11 protein cellular distribution in the eye and in cartilaginous tissues at 8 weeks of gestation. Immunohistochemistry of integrin ␣11 in selected areas from a sagital human embryo section at 8 weeks of gestation. Composite of pictures of cornea (A); note strong staining confined to keratocytes in the cornea. ␣11-positive signal in fibroblastic cells (arrow) around ribs (denoted with r) (B). In double immunofluorescence of the forelimb of a human embryo at 8 wg, ␣11 (red) is detected in areas around forming cartilage, whereas antibodies to collagen II (green) outline the cartilage itself (C). In double immunohistochemistry of a sagital section of the vertebrae, ␣11 (red) is strongly expressed in the outer annulus fibrosus which lacks collagen II. Collagen II expression (green) is high in the cartilage of the vertebrae (D). Bars represent 400 m.
In a chemotaxis assay, ␣111 mediated a migratory chymal cells. Morphologically, it is difficult to distinguish response to PDGF-BB and serum in a collagen-coated filter. perichondrium from surrounding connective tissue. The In the random migration assays, the C2C12-␣11ϩ cells interstitial collagens I/III are expressed in surrounding mes- displayed a stronger chemotactic response to PDGF com- enchyme and perichondrium (Aszodi et al., 1998; Cheah et pared to 10% FBS. The C2C12-␣2ϩ cells, on the other hand, al., 1991; Hamada et al., 1995). Based on the nonoverlapping responded more strongly to serum than to PDGF (Fig. 9). staining pattern of collagen II and ␣11, it appears that ␣111 is expressed in the region just outside the cartilage in perichondrium and/or the mesenchymal cells outside the DISCUSSION perichondrium. This localization is compatible with a role of ␣111 in cartilage repair (Long and Linsenmayer, 1998). ␣ 11 Expression in Mesenchymal Nonmuscle Cells The basic helix–loop–helix transcription factor scleraxis in Vivo and in Vitro is expressed in forming mesoderm and later prefigures the The sites of the strongest ␣11 expression in human forming skeleton (Cserjesi et al., 1995). Inactivation of the embryonic tissues are areas adjacent to forming cartilage. scleraxis gene has revealed a dual role in mesoderm forma- ␣11 protein is prominently expressed in cells around ribs, tion and chondrogenesis in axial cartilage and ribs (Brown et vertebrae, and in intervertebral discs. Hyaline cartilage is al., 1999). It has been suggested that some of the down- surrounded by perichondrium, composed of an inner layer stream target genes for scleraxis are cell adhesion receptors of cells with chondrogenic potential which take part in (Brown et al., 1999). The expression data suggest that ␣11 is appositional growth. Outer perichondrium is composed of expressed in a manner compatible with regulation by scler- cells in a transition zone adjacent to surrounding mesen- axis. The presence of numerous E-boxes containing core
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 123
FIG. 5. Expression of ␣11 integrin in cultured cells. (A) Cultures of differentiated human satellite cells XXVI and subconfluent 1518 human fibroblasts were metabolically labeled, proteins were immunoprecipitated with antibodies, separated on a 6% SDS–PAGE gel under nonreducing conditions, and visualized by fluorography. The antibodies used were directed to integrin subunits 1 (lanes 1, 5), ␣1 (lanes 2, 6), ␣11 (lanes 3, 7), and ␣2 (lanes 4, 8) subunits. Positions of different integrin chains are marked. (B) 1518 fibroblasts were seeded on collagen-coated coverslips and allowed to adhere for 1 h (a–c) or allowed to grow in the presence of 10% fetal bovine serum for several days (d), fixed with methanol, and stained with antibodies to integrin ␣1 (a), ␣2 (b), and ␣11(c, d) chains. Note the localization of ␣11 chain to focal contacts (c). When allowed to grow in culture for several days, ␣11 assumed matrix contact-like distribution on the dorsal surface (d). Bar represents 20 m.
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 124 Tiger et al.
␣11 chain is also highly expressed in keratocytes of the embryonic cornea. In cornea, like in the intervertebral disc, collagens are organized into very precise bundles in a multilayer arrangement (Meek and Fullwood, 2001). In immunoprecipitation assays, adult human corneal fibro- blasts as well as skin fibroblasts and fibrosarcoma cells expressed the ␣11 subunit, indicating an expression re- stricted to mesenchymal cells also in vitro. In summary, ␣111 is highly pressed in mesenchymal cells in vitro, and in vivo on fibroblasts at sites where collagens are arranged in a highly ordered manner. It is still unknown how the organization of collagens in cornea and intervertebral disc occurs, but it has been suggested that the organization of collagen matrices during development is FIG. 6. Integrin levels in stably transfected cell lines. Untrans- driven by cellular interactions (Harris et al., 1981; Stopak fected (C2C12), ␣2-transfected (C2C12-␣2ϩ), and ␣11-transfected and Harris, 1982). In the intervertebral disc, arranged sheets (C2C12-␣11ϩ) cells were metabolically labeled, immunoprecipi- tated with antibodies to integrin 1 chain (1) or ␣11 chain (␣11), of annulus fibroblasts precede the ordered collagen bundles and analyzed by SDS–PAGE followed by autoradiography. Untrans- (Hayes et al., 1999). In corneal fibroblasts, cytoskeletal fected cells lack characteristic 180-kDa integrin ␣1 chain and ␣11 stress fibers are needed for matrix reorganization in vitro chain (lanes 1, 2). ␣2-transfeccted cells analyzed under nonreducing (Mar et al., 2001). We suggest that ␣111 is involved in the (lanes 3, 4) and reducing (lanes 8, 9) conditions. ␣11-transfected ordered collagen matrix organization in both these tissues. cells analyzed under nonreducing (lanes 5, 6) and reducing (lanes 10, 11) conditions. Note that nonspecific band at 240 kDa in lane 8 is also present in precipitate formed using preimmune serum. Lack of ␣11 Expression in Embryonic Myogenic Cells in Vivo We originally identified ␣11 in myotube cultures from consensus sequences for scleraxis binding in the ␣11 pro- human fetal muscle cells and satellite cells (Gullberg et al., moter, support this notion (C. Bergman and D. Gullberg, 1995b; Velling et al., 1999). During in vitro myogenic unpublished observations). differentiation, ␣11 mRNA and ␣11 protein are upregulated The intervertebral disc is a composite structure com- in these myogenic cells (Velling et al., 1999) in a manner posed of an outer annulus fibrosus of sclerotomal origin, similar to ␣7 during in vitro myogenesis of rodent cells allowing twisting of the spine, and an inner nucleus pul- (Song et al., 1992). ␣11mRNA can furthermore be detected posa, derived from notochord, that resists compressive in adult skeletal muscle by Northern blotting. Collectively, forces (Humzah and Soames, 1988; Rufai et al., 1995). these data led us to suggest that ␣11 is involved in human Recent studies of mice devoid of collagen II have clarified myogenesis. However, the lack, or low expression, of ␣11 the mechanism whereby intervertebral discs form (Aszodi mRNA and protein in developing human embryonic et al., 1998). The growing cartilage of vertebrae causes the muscle in vivo, described in this study, does not support a notochord to bulge into primordial intervertebral discs. In role for ␣11 during early myogenesis. The integrin ␣ the absence of cartilage, no bulging of the notochord occurs chain(s) present in embryonic and fetal human muscle, and no intervertebral discs form. The annulus fibrosus is which thus lacks ␣7 (Cohn et al., 1999) and ␣11 integrin composed of an outer part containing annulus fibroblasts chains, remain to be identified. Possible candidates include and lamellae of ordered collagen I/III bundles anchored to ␣v (Hirsch et al., 1994) and/or ␣5 (Gullberg et al., 1995a; cartilage endplates of vertebral bodies. The inner part is Roman and McDonald, 1992; Sastry et al., 1996; Taverna et fibrocartilaginous, contains collagen II, and is derived from al., 1998) integrin chains. cartilage. After birth, the outer annulus differentiates into We cannot completely exclude that part of the ␣11 fibrocartilage (Hayes et al., 1999). In this study, the inter- expression in the embryonic perichondrial regions repre- vertebral disc ␣11 staining appears in what looks like outer sent muscle attachment points. However, the remaining annulus fibrosus, since the staining does not coincide with skeletal muscle tissue, as well as cardiac tissue, lack ␣11 that of collagen II. Annulus fibroblasts are in contact with a expression at the studied developmental stages. We have matrix containing collagens I and III as well as fibronectin previously observed also that the ␣1 integrin chain on and have previously been shown to express ␣51 integrin human fetal muscle cells is upregulated by the in vitro (Hayes et al., 1999). ␣111 appears to be a major collagen culture conditions (Gullberg et al., 1995b). In the case of ␣1, binding integrin on annulus fibroblasts in the immature this upregulation might be related to the reported presence embryonic intervertebral disc. It will be interesting to of binding sites for serum response elements within the determine whether ␣111 is needed for the integrity of the integrin ␣1 promoter (Obata et al., 1997). Another possibil- adult intervertebral disc and whether ␣111 is important in ity is that ␣11 expression in muscle could be spatially and repair situations following disc injury. temporally regulated in a complex manner so that ␣11 is
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 125
FIG. 7. Collagen-specificity and cation-dependence of ␣111-mediated cell adhesion. C2C12 cells transfected with ␣11 integrin cDNA (C2C12-␣11ϩ)or␣2 integrin cDNA (C2C12-␣2ϩ) were assayed for attachment to collagen type I and IV (A, B). (B) The untransfected C2C12 cells are included as a control. Divalent cation dependence of adhesion was assayed on collagen type I coated at 4 g/ml using C2C12-␣11ϩ (C) and C2C12-␣2ϩ (D) cells. Data are presented as relative binding to fibronectin and BSA, where saturated binding to fibronectin represents 100% and background binding to the BSA-blocking represents 0%. Values are given as mean values of triplicate data with error bars denoting standard deviation.
only expressed in certain muscle cells at different times collagen-binding integrin, or alternatively, affect the activ- during development. Yet another alternative is that ␣11 in ity of other collagen-binding integrins. In metabolically muscle is largely involved in repair situations occurring labeled ␣11-transfected cells, we were, however, unable to during muscle regeneration. The induction during in vitro detect ␣1, ␣2, or ␣10 integrin chains. The availability of culture could reflect a sensitivity to growth factors present function-blocking ␣11 integrin antibodies will finally re- in serum, mimicking factors present during muscle regen- solve this matter. Comparison of the amount of ␣2 and ␣11 eration. chains coprecipitating with 1 chain revealed somewhat higher levels of ␣2 chain coprecipitating with 1 and direct ␣  ␣  In Vitro Analysis of ␣111 Integrin Function quantitative comparisons of 2 1 and 11 1 activities in the different assays is thus not possible. Ability to contract collagen matrices. When compar- Interestingly, although the introduction of either the ␣2 ing ␣2 and ␣11 expression levels in the transfected C2C12 integrin chain or ␣11 integrin chain into C2C12 cells cells by immunoprecipitation of metabolically labeled in- strongly enhanced collagen gel contraction, the untrans- tegrins, a number of factors have to be taken into account, fected C2C12 cells were able to autocontract in the absence including heterodimer stability and 1 chain availability. of added factors and the absence of detectable integrin We have no data to indicate that ␣2or␣11 expression receptors for monomeric collagen I. In contrast, cell adhe- depletes 1 chain from other integrin heterodimers, since sion and cell migration on collagen I was strictly dependent cell adhesion to fibronectin and different laminin variants on the presence of 1 integrin collagen receptors. It is likely was unaffected in both ␣2- and ␣11-transfected cells (data that during the longer time period used for the collagen not shown). Although unlikely, it is formally possible that contraction assay, 1 integrin-independent mechanisms for transfection of ␣11 could act indirectly by inducing another collagen gel contraction are activated. The ability to con-
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 126 Tiger et al.
functions. In these situations, 1 integrin receptors inter- acting directly with the collagen fibrils without inducing collagenases are likely to be important. PDGF and FBS have been shown to induce collagen gel contraction by using a number of cell lines. It will be interesting to determine the molecular mechanism whereby FBS and PDGF mediate these actions using ␣2- and ␣11-transfected cells. Another potential candidate receptor involved in the 1 integrin-independent collagen gel con- traction is the discoidin domain receptor DDR1 (Vogel et al., 1997). A characteristic feature of DDR1 is the slow kinetics by which collagens induce the tyrosine kinase activity of these receptors. Developing mice express DDR1 binding activity in forming cartilage and mice devoid of DDR1 are smaller in size (Vogel et al., 2001). Interestingly, low levels of DDR1 expression in C2C12 have been noted and inhibition of DDR1 with a dominant negative DDR1 mutation inhibited differentiation (Vogel et al., 2000). Con- sidering the slow kinetics of DDR1 kinase activity, it is thus possible that DDR1 mediate the 1 integrin- independent contraction of collagen gels in C2C12 cells. Ligand specificity of ␣111. Recent studies have indi- cated that ␣11 and ␣21 integrins both recognize the GFOGER motif in collagen type I (Knight et al., 2000) whereas the binding of ␣11 to collagen IV occurs by a different mechanism (Eble et al., 1993). Comparison of
FIG. 8. ␣111-mediated collagen gel contraction. Collagen gels containing cells were formed in microtiter plates as described in Experimental Procedures. Time-dependent contraction was mea- sured under three conditions, without any stimuli (open symbols), with 25 ng/ml PDGF-bb (gray symbols), and with 10% FBS (black symbols). (A) C2C12 (squares); (B) C2C12-␣2ϩ (circles); (C) C2C12- ␣11ϩ (diamonds). Eight gels per cell line were measured if soluble factors were added, otherwise 16. The results shown are from one representative experiment. Each point represents the mean of the 8 or 16 samples and the error bars show standard deviation within these. FIG. 9. ␣111-mediated cell migration. C2C12 cells, C2C12-␣2ϩ cells, and C2C12-␣11ϩ cells were allowed to migrate through a collagen-coated membrane in the absence of stimulating facors (open bars), toward a gradient of PDGF-BB (10 ng/ml in the lower tract collagen gels in some situations can be accomplished chamber, dotted bars), with PDGF-BB on both sides of the mem- brane (10 ng/ml in both chambers, black bars), toward a gradient of by induction of collagenases and exposure of RGD binding ␣  FBS (10% in the lower chamber, bricked bars), and with FBS on sites to be used by v 3 (Jones et al., 1997). This mecha- both sides of the membrane (10% in both chambers, gray bars). The nism might be of importance in provisional wound matri- results shown are from one represantative experiment performed ces with high ECM turnover. In other locations, like the with duplicate samples. Each bar represents the mean of two cornea and intervertebral disc, it might be important to membranes with error bars showing total spreading between the keep the collagen fibrils intact to ensure normal tissue two.
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 127
␣11 and ␣21 integrin binding to collagen I and collagen and the formation of intervertebral discs. J. Cell Biol. 143, IV has suggested that ␣11 binds collagen IV with higher 1399–1412. affinity than collagen I whereas ␣21 displays a somewhat Brakebusch, C., Hirsch, E., Potocnik, A., and Fa¨ssler, R. (1997).  higher affinity for collagen I then collagen IV (Dickeson et Genetic analysis of 1 integrin function: Confirmed, new and al., 1999; Kapyla et al., 2000; Kern et al., 1994; Tuckwell et revised roles for a crucial family of cell adhesion molecules. J. Cell Sci. 110, 2895–2904. al., 1995). In support of this, ␣1 integrin null fibroblasts Brown, D., Wagner, D., Li, X., Richardson, J. A., and Olson, E. N. show a reduced adhesion to collagen IV compared with (1999). Dual role of the basic helix–loop–helix transcription wild-type fibroblasts, whereas the adhesion to collagen I is factor scleraxis in mesoderm formation and chondrogenesis similar in both types of fibroblasts (Gardner et al., 1996). during mouse embryogenesis. Development 126, 4317–4329. This can be explained by the presence of other collagen I Camper, L., Hellman, U., and Lundgren-Akerlund, E. (1998). Isola- receptors, such as ␣111, on the ␣1 integrin null fibroblasts. tion, cloning, and sequence analysis of the integrin subunit ␣10, It will be important to determine the collagen recognition a 1-associated collagen binding integrin expressed on chondro- motif of ␣111. cytes. J. Biol. Chem. 273, 20383–20389. Cell migration. A functional link between the ECM Carter, W. G., Ryan, M. C., and Gahr, P. J. (1991). Epiligrin, a new and the cytoskeleton is important for successful cell migra- cell adhesion ligand for ␣31 in epithelial basement membranes. tion. Recent analysis of 1 integrin dynamics in stationary Cell 65, 599–610. live cells using GFP-tagged integrins has shown that in Carver, W., Molano, I., Reaves, T. A., Borg, T. K., and Terracio, L. (1995). Role of the ␣11 integrin complex in collagen gel con- stationary cells the intracellular integrin–cytoskeleton traction in vitro by fibroblasts. J. Cell Physiol. 165, 425–437. linkage in focal contacts is static whereas the extracellular Cheah, K. S., Lau, E. T., Au, P. K., and Tam, P. P. (1991). Expression integrin–ECM contacts are dynamic (Smilenov et al., 1999; of the mouse ␣1(II) collagen gene is not restricted to cartilage Zamir et al., 2000). In migratory cells, on the other hand, during development. Development 111, 945–953. GFP-tagged integrins formed a firm attachment with the Cohn, R. D., Mayer, U., Saher, G., Herrmann, R., van der Flier, A., ECM at the leading edge. During embryonic development, Sonnenberg, A., Sorokin, L., and Voit, T. (1999). Secondary mesenchymal cells encounter a matrix rich in interstitial reduction of ␣7B integrin in laminin ␣2 deficient congenital collagens and fibronectin. In addition to being able to muscular dystrophy supports an additional transmembrane link interact and remodel this matrix, it is essential that cells in skeletal muscle. J. Neurol. Sci. 163, 140–152. can move through this matrix. Lack of fibronectin results in Cooke, M. E., Sakai, T., and Mosher, D. F. (2000). Contraction of ␣  ␣  prominent mesodermal defects (George et al., 1993). ␣111 collagen matrices mediated by 2 1A and v 3 integrins. J. Cell Sci. 113, 2375–2383. appears to be one of the receptors well suited for the task of Cserjesi, P., Brown, D., Ligon, K. L., Lyons, G. E., Copeland, N. G., moving through a collagen-rich ECM. In both collagen gel Gilbert, D. J., Jenkins, N. A., and Olson, E. N. (1995). Scleraxis: A ␣  contraction and cell migration, 11 1 was found to be more basic helix–loop–helix protein that prefigures skeletal formation sensitive to PDGF than FBS, whereas the opposite was true during mouse embryogenesis. Development 121, 1099–1110. for ␣21. This is the first indication that ␣21 and ␣111 Dickeson, S. K., Mathis, N. L., Rahman, M., Bergelson, J. M., and use different signaling pathways. Santoro, S. A. (1999). Determinants of ligand binding specificity Gene inactivation through homologous recombination in of the ␣11 and ␣21 integrins. J. Biol. Chem. 274, 32182–32191. transgenic mice should shed light on the possible functions Dickeson, S. K., and Santoro, S. A. (1998). Ligand recognition by the of ␣11 chain in cell adhesion, cell migration, and matrix I domain-containing integrins. Cell. Mol. Life Sci. 54, 556–566. reorganization in mesenchymal cell lineages. Duband, J. L., Belkin, A. M., Syfrig, J., Thiery, J. P., and Koteliansky, V. E. (1992). Expression of ␣1 integrin, a laminin-collagen recep- tor, during myogenesis and neurogenesis in the avian embryo. Development 116, 585–600. ACKNOWLEDGMENTS Eble, J. A., Golbik, R., Mann, K., and Kuhn, K. (1993). The ␣11 integrin recognition site of the basement membrane collagen ␣ We thank Y. Takada for the generous gift of full length human 2 molecule [␣1(IV)]2 ␣2(IV). EMBO J. 12, 4795–4802. cDNA, Dr. P. Wernhoff for antibodies, Helen Blau, K. Rubin, and Gardner, H., Broberg, A., Pozzi, A., Laato, M., and Heino, J. (1999). Nathalie Isnard for cells. This work was supported by grants from Absence of integrin ␣11 in the mouse causes loss of feedback Natural Science Research council (B961, to D.G.), Swedish Cancer regulation of collagen synthesis in normal and wounded dermis. Foundation (2729-300-11x), Go¨ran Gustafssons foundation (to J. Cell Sci. 112, 263–272. D.G.), and Konung Gustaf V:s fond (to D.G.). Gardner, H., Kreidberg, J., Koteliansky, V., and Jaenisch, R. (1996). Deletion of integrin ␣1 by homologous recombination permits normal murine development but gives rise to a specific deficit in REFERENCES cell adhesion. Dev. Biol. 175, 301–313. George, E. L., Georges-Labouesse, E. N., Patel-King, R. S., Rayburn, Agrez, M. V., Bates, R. C., Boyd, A. W., and Burns, G. F. (1991). H., and Hynes, R. O. (1993). Defects in mesoderm, neural tube Arg-Gly-Asp-containing peptides expose novel collagen recep- and vascular development in mouse embryos lacking fibronec- tors on fibroblasts: Implications for wound healing. Cell Regul. 2, tin. Development 119, 1079–1091. 1035–1044. Golbik, R., Eble, J. A., Ries, A., and Kuhn, K. (2000). The spatial Aszodi, A., Chan, D., Hunziker, E., Bateman, J. F., and Fassler, R. orientation of the essential amino acid residues arginine and (1998). Collagen II is essential for the removal of the notochord aspartate within the ␣11 integrin recognition site of collagen IV
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 128 Tiger et al.
has been resolved using fluorescence resonance energy transfer. J. Kapyla, J., Ivaska, J., Riikonen, R., Nykvist, P., Pentikainen, O., Mol. Biol. 297, 501–509. Johnson, M., and Heino, J. (2000). Integrin ␣2 I domain recognizes Gotwals, P. J., Chi-Rosso, G., Lindner, V., Yang, J., Ling, L., Fawell, type I and type IV collagens by different mechanisms. J. Biol. S. E., and Koteliansky, V. E. (1996). The ␣11 integrin is ex- Chem. 275, 3348–3354. pressed during neointima formation in rat arteries and mediates Keely, P. J., Wu, J. E., and Santoro, S. A. (1995). The spatial and collagen matrix reorganization. J. Clin. Invest. 97, 2469–2477. temporal expression of the ␣2 1 integrin and its ligands, Gullberg, D., Sjoberg, G., Velling, T., and Sejersen, T. (1995a). collagen I, collagen IV, and laminin, suggest important roles in Analysis of fibronectin and vitronectin receptors on human fetal mouse mammary morphogenesis. Differentiation 59, 1–13. skeletal muscle cells upon differentiation. Exp. Cell Res. 220, Kern, A., Briesewitz, R., Bank, I., and Marcantonio, E. E. (1994). The 112–123. role of the I domain in ligand binding of the human integrin Gullberg, D., Terracio, L., Borg, T. K., and Rubin, K. (1989). ␣11. J. Biol. Chem. 269, 22811–22816. Identification of integrin-like matrix receptors with affinity for Klareskog, L., Johnell, O., Hulth, A., Holmdahl, R., and Rubin, K. interstitial collagens. J. Biol. Chem. 264, 12686–12694. (1986). Reactivity of monoclonal anti-type II collagen antibodies Gullberg, D., Tingstrom, A., Thuresson, A. C., Olsson, L., Terracio, with cartilage and synovial tissue in rheumatoid arthritis and L., Borg, T. K., and Rubin, K. (1990). 1 integrin-mediated osteoarthritis. Arthritis Rheum. 29, 730–738. collagen gel contraction is stimulated by PDGF. Exp. Cell Res. Klein, C. E., Dressel, D., Steinmayer, T., Mauch, C., Eckes, B., 186, 264–272. Krieg, T., Bankert, R. B., and Weber, L. (1991). Integrin ␣21is Gullberg, D., Velling, T., Sjoberg, G., and Sejersen, T. (1995b). upregulated in fibroblasts and highly aggressive melanoma cells Up-regulation of a novel integrin alpha-chain (␣mt) on human in three-dimensional collagen lattices and mediates the reorga- fetal myotubes. Dev. Dyn. 204, 57–65. nization of collagen I fibrils. J. Cell Biol. 115, 1427–1436. Hamada, K., Okawara, Y., Fryer, J. N., Tomonaga, A., and Fukuda, Knight, C. G., Morton, L. F., Onley, D. J., Peachey, A. R., Messent, H. (1995). The detection of the mRNAs of procollagen types I, II A. J., Smethurst, P. A., Tuckwell, D. S., Farndale, R. W., and and III in human fetal fingers by in situ hybridization using Barnes, M. J. (1998). Identification in collagen type I of an integrin digoxigenin-labelled oligonucleotide probes. Histochem. J. 27, ␣21-binding site containing an essential GER sequence. J. Biol. 309–317. Chem. 273, 33287–33294. Harris, A. K., Stopak, D., and Wild, P. (1981). Fibroblast traction as Knight, C. G., Morton, L. F., Peachey, A. R., Tuckwell, D. S., a mechansim for collagen morphogenesis. Nature 290, 249–251. Farndale, R. W., and Barnes, M. J. (2000). The collagen-binding Hayes, A. J., Benjamin, M., and Ralphs, J. R. (1999). Role of actin A-domains of integrins ␣11 and ␣21 recognize the same stress fibres in the development of the intervertebral disc: specific amino acid sequence, GFOGER, in native (triple-helical) Cytoskeletal control of extracellular matrix assembly. Dev. Dyn. collagens. J. Biol. Chem. 275, 35–40. 215, 179–189. Long, F., and Linsenmayer, T. F. (1998). Regulation of growth Hirsch, E., Gullberg, D., and Tarone, G. (1994). ␣v integrin subunit region cartilage proliferation and differentiation by perichon- is predominantly located in nervous tissue and skeletal muscle drium. Development 125, 1067–1073. during mouse development. Dev. Dyn. 201, 108–120. Mar, P. K., Roy, P., Yin, H. L., Cavanagh, H. D., and Jester, J. V. Hodivala-Dilke, K. M., DiPersio, C. M., Kreidberg, J. A., and Hynes, (2001). Stress fiber formation is required for matrix reorganiza- R. O. (1998). Novel roles for ␣31 integrin as a regulator of tion in a corneal myofibroblast cell line. Exp. Eye Res. 72, cytoskeletal assembly and as a trans-dominant inhibitor of 455–466. integrin receptor function in mouse keratinocytes. J. Cell Biol. Meek, K., and Fullwood, N. (2001). Corneal and scleral collagens. 142, 1357–1369. Micron 32, 261–272. Humzah, M., and Soames, R. (1988). Human intervertebral disc: Nunohiro, T., Ashizawa, N., Graf, K., Hsueh, W. A., and Yano, K. Structure and function. Anat. Rec. 220, 337–356. (1999). Angiotensin II promotes integrin-mediated collagen gel Hynes, R. O. (1992). Integrins: Versatility, modulation, and signal- contraction by adult rat cardiac fibroblasts. Jpn. Heart J. 40, ing in cell adhesion. Cell 69, 11–25. 461–469. Hynes, R. O. (1996). Targeted mutations in cell adhesion genes: Nykvist, P., Tu, H., Ivaska, J., Kapyla, J., Pihlajaniemi, T., and What have we learned from them? Dev. Biol. 180, 402–412. Heino, J. (2000). Distinct recognition of collagen subtypes by Jenkins, G., Redwood, K. L., Meadows, L., and Green, M. R. (1999). ␣11 and ␣21 integrins. ␣11 mediates cell adhesion to type Effect of gel re-organization and tensional forces on ␣21 integrin XIII collagen. J. Biol. Chem. 275, 8255–8261. levels in dermal fibroblasts. Eur. J. Biochem. 263, 93–103. Obata, H., Hayashi, K., Nishida, W., Momiyama, T., Uchida, A., Jones, P. L., Crack, J., and Rabinovitch, M. (1997). Regulation of Ochi, T., and Sobue, K. (1997). Smooth muscle cell phenotype- tenascin-C, a vascular smooth muscle cell survival factor that dependent transcriptional regulation of the ␣1 integrin gene. interacts with the ␣v3 integrin to promote epidermal growth J. Biol. Chem. 272, 26643–26651. factor receptor phosphorylation and growth. J. Cell Biol. 139, Pozzi, A., Moberg, P. E., Miles, L. A., Wagner, S., Soloway, P., and 279–293. Gardner, H. A. (2000). Elevated matrix metalloprotease and Kagami, S., Kondo, S., Loster, K., Reutter, W., Kuhara, T., Yasu- angiostatin levels in integrin ␣1 knockout mice cause reduced tomo, K., and Kuroda, Y. (1999). ␣11 integrin-mediated collagen tumor vascularization. Proc. Natl. Acad. Sci. USA 97, 2202– matrix remodeling by rat mesangial cells is differentially regu- 2207. lated by transforming growth factor-beta and platelet-derived Pozzi, A., Wary, K. K., Giancotti, F. G., and Gardner, H. A. (1998). growth factor-BB. J. Am. Soc. Nephrol. 10, 779–789. Integrin ␣11 mediates a unique collagen-dependent prolifera- Kamata, T., Puzon, W., and Takada, Y. (1994). Identification of tion pathway in vivo. J. Cell Biol. 142, 587–594. putative ligand binding sites within I domain of integrin Racine-Samson, L., Rockey, D. C., and Bissell, D. M. (1997). The ␣21(VLA-2, CD49b/CD29). J. Biol. Chem. 269, 9659–9663. role of ␣11 integrin in wound contraction. A quantitative
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. ␣111 Integrin Is an Interstitial Collagen Receptor 129
analysis of liver myofibroblasts in vivo and in primary culture. of collagens but is not a general receptor for the collagenous J. Biol. Chem. 272, 30911–30917. motif. Eur. J. Biochem. 241, 732–739. Roman, J., and McDonald, J. A. (1992). Expression of fibronectin, Velling, T., Kusche-Gullberg, M., Sejersen, T., and Gullberg, D. the integrin ␣5, and ␣-smooth muscle actin in heart and lung (1999). cDNA cloning and chromosomal localization of human development. Am. J. Respir. Cell Mol. Biol. 6, 472–480. ␣11 integrin. A collagen-binding, I domain-containing, 1- Rufai, A., Benjamin, M., and Ralphs, J. R. (1995). The development associated integrin ␣-chain present in muscle tissues. J. Biol. of fibrocartilage in the rat intervertebral disc. Anat. Embryol. Chem. 274, 25735–25742. 192, 53–62. Vogel, W., Brakebusch, C., Fassler, R., Alves, F., Ruggiero, F., and Sastry, S. K., Lakonishok, M., Thomas, D. A., Muschler, J., and Pawson, T. (2000). Discoidin domain receptor 1 is activated ␣ Horwitz, A. F. (1996). Integrin subunit ratios, cytoplasmic independently of 1 integrin. J. Biol. Chem. 275, 5779–5784. domains, and growth factor synergy regulate muscle proliferation Vogel, W., Gish, G. D., Alves, F., and Pawson, T. (1997). The and differentiation. J. Cell Biol. 133, 169–184. discoidin domain receptor tyrosine kinases are activated by Sheppard, D. (2000). In vivo functions of integrins: Lessons from collagen. Mol. Cell 1, 13–23. null mutations in mice. Matrix Biol. 19, 203–209. Vogel, W. F., Aszodi, A., Alves, F., and Pawson, T. (2001). Discoidin Skerjanc, I. S., Slack, R. S., and McBurney, M. W. (1994). Cellular domain receptor 1 tyrosine kinase has an essential role in aggregation enhances MyoD-directed skeletal myogenesis in mammary gland development. Mol. Cell. Biol. 21, 2906–2917. embryonal carcinoma cells. Mol. Cell. Biol. 14, 8451–8459. Voigt, S., Gossrau, R., Baum, O., Loster, K., Hofmann, W., and Smilenov, L. B., Mikhailov, A., Pelham, R. J., Marcantonio, E. E., Reutter, W. (1995). Distribution and quantification of ␣1-integrin and Gundersen, G. G. (1999). Focal adhesion motility revealed in subunit in rat organs. Histochem. J. 27, 123–132. stationary fibroblasts. Science 286, 1172–1174. Song, W. K., Wang, W., Foster, R. F., Bielser, D. A., and Kaufman, Wayner, E. A., and Carter, W. G. (1987). Identification of multiple S. J. (1992). H36-␣7 is a novel integrin ␣ chain that is develop- cell adhesion receptors for collagen and fibronectin in human mentally regulated during skeletal myogenesis. J. Cell Biol. 117, fibrosarcoma cells possessing unique alpha and common beta 643–657. subunits. J. Cell Biol. 105, 1873–1884. Stopak, D., and Harris, A. (1982). Connective tissue morphogenesis Wu, J. E., and Santoro, S. A. (1994). Complex patterns of expression by fibroblast traction. Dev. Biol. 90, 383–398. suggest extensive roles for the ␣21 integrin in murine develop- Takebe, Y., Seiki, M., Fujisawa, J., Hoy, P., Yokota, K., Arai, K., ment. Dev. Dyn. 199, 292–314. Yoshida, M., and Arai, N. (1988). SR alpha promoter: An efficient Xu, Y., Gurusiddappa, S., Rich, R. L., Owens, R. T., Keene, D. R., and versatile mammalian cDNA expression system composed of Mayne, R., Hook, A., and Hook, M. (2000). Multiple binding sites the simian virus 40 early promoter and the R-U5 segment of in collagen type I for the integrins ␣11, and ␣21. J. Biol. Chem. human T-cell leukemia virus type 1 long terminal repeat. Mol. 275, 38981–38989. Cell. Biol. 8, 466–472. Zamir, E., Katz, M., Posen, Y., Erez, N., Yamada, K. M., Katz, B. Z., Taverna, D., Disatnik, M. H., Rayburn, H., Bronson, R. T., Yang, J., Lin, S., Lin, D. C., Bershadsky, A., Kam, Z., and Geiger, B. (2000). Rando, T. A., and Hynes, R. O. (1998). Dystrophic muscle in mice Dynamics and segregation of cell-matrix adhesions in cultured chimeric for expression of ␣5 integrin. J. Cell Biol. 143, 849–859. fibroblasts. Nat. Cell Biol. 2, 19119–19126. Tuckwell, D., Calderwood, D. A., Green, L. J., and Humphries, M. J. (1995). Integrin ␣2 I-domain is a binding site for collagens. J. Cell Received for publication April 23, 2001 Sci. 108, 1629–1637. Revised June 11, 2001 Tuckwell, D. S., Reid, K. B. M., Barnes, M. J., and Humphries, M. J. Accepted June 11, 2001 (1996). The A-domain of integrin ␣2 binds specifically to a range Published online July 31, 2001
Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved.