Expression and ligand binding of α2β1 integrin on breast carcinoma cells

  • Published on

  • View

  • Download

Embed Size (px)


<ul><li><p>Clin. Exp. Metastasis, 1995, 13, 223-235 </p><p>Expression and ligand binding of integrin on breast carcinoma cells </p><p>Michio Maemura, Steven K. Akiyama*, Virgil L. Woods Jr.t and Robert B. Dickson </p><p>Lombardi Cancer Research Center, Georgetown University, Washington, DC, USA. * Laboratory of Developmental Biology, National Institute of Dental Research, Bethesda, MD, USA and ~f Division of Rheumatic Diseases, School of Medicine, University of California, San Diego, CA, USA </p><p>(Received 7 February 1995; accepted 16 March 1995) </p><p>We examined the expression and ligand specificity of the ~2/31 integrin on human mammary epithelial cells (HMEC) and a panel of breast carcinoma cell lines in vitro. We found that the ~2/31 integrin was universally, hut quite variably expressed on these cells by FACS analysis. No significant correlation was observed between its expression and other known cellular phenotypes. Substrate attachment assays using blocking antibodies demonstrated that ~2/31 integrin served as a receptor for collagen on HMEC and almost all breast carcinoma cells. However, its contribution to laminin binding of these cells appeared to be related to cellular differentiation as evaluated by sex steroid receptor status and by markers of epithelial-mesenchymal transition, i.e. loss of E-cadherin and expression of vimentin. Two different populations of non-malignant immortalized HMEC (184AIN4 and MCF-10A) contained cells capable of using ~2/31 integrin as a laminin receptor. Breast cancer cell lines positive for estrogen receptor (ER) and E-cadherin (MCF-7, T47D, ZR75-1) could also use ~2/31 integrin as a laminin receptor. Conversely, ~2/31 integrin appeared to be incapable of binding to laminin or to he a very minor receptor for laminin on metastatic ER-negative breast carcinoma cells that expressed vimentin (MDA-MB 231, MDA-MB 435, and MDA-MB 436). These findings suggest that the ligand specificity of c~2/31 integrin, i.e. its function as a laminin receptor, may be regulated during the malignant progression of breast carcinoma cells. A reduced contribution of ~2/31 integrin to the cellular laminin binding appears to be associated with an increased malignant phenotype and with an epithelial-mesenchymal transition of breast carcinoma cells. </p><p>Keywords: breast carcinoma cell, collagen, integrin, laminin, metastasis </p><p>Introduction </p><p>Integrins constitute a family of cell surface glycoproteins which mediate cellular adhesion [1 3]. Diverse cellular interactions with extraceilular matrix (ECM) and adjacent cells are involved in embryonic development, maintenance of normal tissue architecture, inflammatory responses, wound healing, and metastasis; integrins play a crucial role in these cell-ECM and cell-cell </p><p>Address correspondence to: R. B. Dickson, Lombardi Cancer Research Center, Georgetown University, 3970 Reservoir Road NW, Washington, DC 20007, USA. Tel: (+1) 202 687 3770; Fax:(+ 1) 202 687 7505. </p><p>interactions [1]. All integrins are heterodimeric molecules consisting of non-covalently associated and /3 subunits [1-4]. lntegrins were originally divided into three subfamilies based on their/3 subunit [4] and diverse combinations of at least 14 ~ and 8 /3 subunits have been reported to form multiple integrins so far [3]. </p><p>It is well known that the invasive and metastatic processes of tumor cells involve multiple steps [5-7]; tumor cells must interact with neighboring cells, basement membrane, and ECM components. These interactions are believed to be mediated by several adhesion molecules, including integrins [5, 8, 9]. </p><p>~', 1995 Rapid Communications ~fOx[ord Lid Clinical &amp; Experimental Metastasis Vol 13 No 4 223 </p></li><li><p>M. Maemura et al. </p><p>Indeed, many immunohistochemical studies have shown that altered expression of the fll integrin family occurs in malignant tumors [10-14]. Some transformed [15-18] or metastatic cells [19, 20] are also reported to exhibit its altered expression compared with their parental cells in vitro. Thus, members of/31 integrin family are of interest with respect to malignant progression [21, 22]. Among the ill integrin family, altered expression of ct2ill integrin has been frequently reported in malignant tumors [10, 13, 14, 23-25]. In breast cancer, expression of ~t2ill integrin is reported to be reduced compared with the normal mammary gland [26-30]. Its expression is correlated with differentiation [26-28, 30] and ER expression [26, 30] of breast cancer. In mammary epithelial cells, ~2ill integrin has been found to mediate their collagen-induced morphogenesis [31]. Therefore, ~2ill integrin appears to play a critical role in development and maintenance of regulatory tissue architecture in the mammary gland. </p><p>~t2ill integrin, universally expressed in many types of tissues [32], was originally reported to be a receptor for collagen [33]. In fibroblasts and platelets, ~2ill integrin binds to collagen but not to laminin [34, 35]. However, in other types of cells, such as endothelial cells, ~2ill integrin serves as a receptor for both collagen and laminin [35, 36]. Thus, its binding specificity may be modulated by unknown cell-type specific factors. Although one human melanoma cell line (LOX) can bind to both laminin and collagen via ~2ill integrin [34], some human melanoma cells fail to bind to laminin [37]. These findings indicate that, even among the cells of the same origin, the ligand specificity of ct2fll integrin might vary. Furthermore, the function of ~2fll integrin has been recently shown not to he restricted to mediating cell-ECM interactions. Immunocytostaining studies have demonstrated that ~2ill integrin is expressed in cell-cell contact borders of endothelial cells and some epithelial cells [38, 39], suggesting it may also mediate cell-cell interactions. In this context, the study of a role of a given integrin in malignant progression requires not only evaluation of the level of its expression, but also that of its functional status, such as its ligand specificity and distribution on the malignant cells. Previous studies have focused on the expression levels of ~t2ill integrin. However, its ligand specificity has not yet been addressed on breast carcinoma cells. </p><p>In the present study, we examine both expression and ligand binding of ~2ill integrin on breast carcinoma cells and HMEC in vitro. Our results suggest that the contribution of ct2ill integrin to the laminin binding, but not its expression level, appears </p><p>to be correlated with some characteristics of breast carcinoma cells, i.e. their ER or progesterone receptor (PR) status, tumorigenicity, metastatic potential, and transition from epithelial to mesenchymal phenotype. </p><p>Materials and methods </p><p>Cell culture MCF-7 cells were obtained from Dr Marvin Rich (Michigan Cancer Foundation, Detroit, MI, USA). MDA-MB 435 cells were kindly supplied by Dr Janet Price (M.D. Anderson Cancer Center, Houston, TX, USA) and 184A1N4 cells were provided by Dr Martha Stampfer (Tissue Culture Unit, University of California San Francisco, Berkeley, CA, USA). MCF-10A cells were obtained from Dr Herbert Soule (Michigan Cancer Foundation, Detroit, MI, USA). All other cells were obtained from the American Type Culture Collection (Rockville, MD, USA). All cell lines, except 184A1N4 and MCF-10A cells, were maintained in phenol red-containing Richter's modified minimal essential medium (IMEM, Gibco BRL, Grand Island, NY, USA) supplemented with 10% fetal calf serum (FCS, Biofluids, Rockville, MD, USA). Prior to the assay, these cells were cultured in phenol red-free IMEM supplemented with 5% charcoal-stripped FCS for at least 3 days. 184A1N4 cells were maintained in IMEM supplemented with 0.5% FCS, 0.5#g/ml hydrocortisone (Sigma, St Louis, MO, USA), 5 ~tg/ml insulin (Biofluids), and 10ng/ml EGF (UBI, Lake Placid, NY, USA). MCF-10A cells were maintained in DMEM/HAM's F12 (Biofluids) supplemented with 5% horse serum (Biofluids), 10/~g/ml insulin, 0.5 #g/ml hydrocortisone, 10 ng/ml EGF, and 1 #l/ml of cholera toxin (Sigma). </p><p>Antibodies against integrins Rat anti-ill monoclonal antibody (mAb), mAb 13 [40], and mouse anti-or2 monoclonal antibody, mAb 12F1 [41], have been previously described. Mouse anti-~2 monoclonal antibodies, mAb 6F1 [42], were provided by Dr Barry Collen (State University of New York, Stony Brook, NY, USA). These monoclonal antibodies were used in the form of purified IgG. </p><p>FACS analysis For FACS analysis, cells cultured on a T75 flask were harvested with 0.05% EDTA in PBS, washed twice with PBS, resuspended at 4 105 cells/100/~1 in PBS, and then incubated with anti-~2 (mAb 12F1) or anti-ill (mAb 13) monoclonal antibodies (20#g/ml final concentration) for 45 min on ice. After washing four </p><p>224 Clinical &amp; Experimental Metastasis Vol 13 No 4 </p></li><li><p>times with PBS, the cells were incubated with affinity purified fluorescein-conjugated secondary antibodies (goat anti-rat or anti-mouse IgG, Boehringer Mannheim, Indianapolis, IN, USA) for 45 min on ice. The cells were then washed four times with PBS and filtered through a nylon mesh to eliminate cell clumps. Fluorescence signals from 10000 cells were acquired on a FACStar PLUs flow cytometer (Becton Dickinson, Mountain View, CA, USA) to quantify cell surface staining. </p><p>Attachment assay Attachment assays were performed essentially as described [43]. Briefly, prior to the assay, 50 #1/well of diluted laminin purified from EHS tumor (Sigma) or type I collagen (COL I) purified from rat tail (Sigma) was added to triplicate wells in a 96-well tissue culture plate followed by incubation for 60 min at 37C to precoat the wells with the substrate. 50/d of 3% bovine serum albumin (BSA) in phosphate buffered saline (PBS) was added to each well to block non-specific binding sites and the plate incubated for a further 30min at 37C. The solution was then removed from the wells, which were washed three times with PBS. Meanwhile, cells cultured on a T75 flask were harvested with 0.05% EDTA in PBS, washed twice with 0.1% BSA/IMEM, and resuspended at 2.5 105cells/ml in 0.1% BSA/IMEM (containing 2.8 mM Ca 2+ and 1.0mM Mg2+). The cell suspension was incubated with/without blocking antibodies (mAb 6F1 or 13) for 60min at 37C before the cells (2.5 x 104/100pl) were added to each precoated well and incubated for 20min at 37C to allow for adherence. The supernatant containing unattached cells was then removed and the attached cells were stained with 50/d of 0.05% crystal violet in 25% methanol for 15 min. The plate was gently rinsed three times by immersion in water and allowed to dry at room temperature. The incorporated dye was dissolved in 100pl/well 0.1M sodium citrate in 50% ethanol and measured by reading absorbance at 540nm with a Dynatec MR700 ELISA reader (Dynatech laboratories, Chantilly, VA, USA). At least three assays were performed for each cell line and representative data are shown. </p><p>Results </p><p>~2fll integrin expression on breast carcinoma cells We examined the expression of the ~2fll integrin on several breast carcinoma cell lines and HMECs in vitro using FACS. Indirect immunofluorescence staining with mAb 12F1, which can recognize the ~2 </p><p>Expression and ligand bindin9 of ~2fll integrin </p><p>subunit of the ~2fll integrin [41], demonstrated that all breast carcinoma cells and HMECs tested expressed the ~2 subunit on their surfaces (Figure 1). Its expression levels on immortalized HMECs, 184A1N4 [44] and MCF-10A [45] cells, and on ER-positive breast carcinoma cell lines, MCF-7, T47D, and ZR75-1 [46], are comparable. Almost all othe~ cell lines tested, which are known as ER-negative breast carcinoma cells, expressed relatively low levels of the ~2 subunit. Although ER-negative breast carcinoma cells therefore generally appear to express lower levels of the ct2 subunit on their surface, its highest level was observed on one of the ER-negative breast carcinoma cell lines, MDA-MB 453. </p><p>We also studied the cell surface expression of the fll subunit of the c~2fll integrin as well as other fll integrin heterodimers. As shown in Figure 1, the fll subunit was universally expressed on all cell lines tested and its cell surface expression was essentially parallel to that of ~2 subunit on individual cell lines. However, the range and the pattern of fll expression levels were quite variable compared with those of c~2 expression level; HMECs and MDA-MB 453 cells expressed high levels of the fll subunit on their surface, while three ER-positive breast carcinoma cell lines presented lower levels of expression. Although some ER-negative breast carcinoma cell lines (MDA-MB 468, MDA-MB 231, and MDA-MB 435) exhibited fll expression levels comparable with those seen on ER-positive lines, other ER-negative cells expressed only low levels of this subunit. Thus, we failed to find any correlation between the ER status of breast carcinoma cells and expression levels of the fll subunit on their surfaces. </p><p>The ~2 subunit is reported to associate exclusively with the fll subunit [3]. To confirm this finding in breast carcinoma cells, we performed an immuno- precipitation followed by Western blotting with anti-~2 and anti-ill antibodies on several breast carcinoma cell lines and HMEC (data not shown). We found that all precipitated pellets with mAb 12F1 (anti-c~2) contained immunoreactive fll subunit and vice versa. However, we failed to find any immunoreactivity of c~2 subunit in the supernatants after the reaction of the immunoprecipitation with mAb 13 (data not shown). These results confirm that on breast carcinoma cells and HMECs the ~2 subunit associates with the fll subunit and that this is an unique partner of ~2 subunit. </p><p>Effects of blockin 9 antibodies on the attachment of breast carc&amp;oma cells To study the role of the ~2fll integrin in the attachment of breast carcinoma cells to defined substrates, we </p><p>Clinical &amp; Experimental Metastasis Vol 13 No 4 225 </p></li><li><p>M. Maemura et al. </p><p>Figure 1. Expression of ~2fll integrin on breast carcinoma cells and HMEC. The cells (4x 105 ) were prepared as described in Materials and methods and resuspended in 100/11 of PBS with 2/~g of mAb 13 (anti-ill, hatched bar) or 2/~g ofmAb 12F1 (anti-~2, open bar). The cell suspension was incubated for 45min on ice followed by four washes with 0.05% Tween20/ PBS. The cells were then incubated with fluorescein- conjugated goat anti-rat lgG (1:320) or anti-mouse IgG (1:160) for 45min on ice. After extensive washing, 10 000 cells were analysed by FACS. Specific fluorescence inten- sity = mean intensity with mAb - - mean intensity without mAb; mean intensity with- out mAb was less than 10. 1, 184A1N4; 2, MCF-10; 3, MCF-7; 4, T47D; 5, ZR75-1; 6, SKBr3; 7, BT474; 8, MDA- MB 468; 9, MDA-MB 453; 10, MDA-MB 231; 11, MDA-MB 436; 12, MDA-MB 435. </p><p>100 </p><p>&gt;, 80 </p><p>/) c-. </p><p>.::: 6o </p><p>._&gt; </p><p> ~ 40 </p><p>20 </p><p>0 1 2 3 4 5 6 7 8 9 10 11 12 </p><p>performed attachment assays in the presence or absence of anti-ill antibody, mAb 13, or anti-e2 antibody, mAb 6F1, both of which were known to inhibit e2ill-ligand binding [40, 42]. As show...</p></li></ul>


View more >