Background
Breast cancer is one of the most common cancers and continues to rank as one of the top causes of death in women [
1]. The high mortality rate associated with breast cancer is directly related to its ability to readily metastasize. Histological type, size of tumor, metastasis, epidermal growth factor receptor 2 (ErbB2) expression and lymph node involvement are key factors used to assess prognosis and probability of response to systemic therapies [
2]. However, breast cancer patients undergoing treatment continue to have different clinical outcomes, despite having similar clinical diagnostic and prognostic profiles. These differences in outcomes underscore the heterogeneity of the disease, and the limitation of using a mainly morphology-based classification system for breast cancer [
3]. To improve the classification of breast cancers and the use of breast cancer therapeutics, investigations into the biological mechanisms underlying breast cancer have identified new and more accurate biological markers and factors of breast cancer. Currently, cathepsin D, estrogen receptors, ErbB2, integrins, p53, urokinase plasminogen activator (uPA), uPA inhibitor-1 and urokinase receptor (uPAR) have been validated as biological prognostic markers in breast cancer [
4]. Amongst these factors, integrins are a family of cell adhesion receptors that are implicated in the establishment, metastasis and progression of many cancers [
5‐
9].
Integrins meditate cell adhesion to the cell-extracellular matrix (ECM), a fundamental cellular process that not only regulates cell growth, differentiation, and death, but also regulates malignant cell growth, metastasis and cancer-induced angiogenesis [
8,
10,
11]. Integrins participate in these cellular processes by providing a dynamic physical linkage between the ECM and the actin cytoskeleton. Engagement of integrins with ECM ligands triggers integrin clustering, and the formation, disassembly and reorganization of actin filaments, stress fibers and focal adhesion complexes [
7,
12]. This dynamic reorganization of these cellular structures allows integrins to function as regulators of cell shape and cellular processes requiring cellular reshaping such as cell adhesion, cell migration and cell division. Integrin clustering and focal adhesions also elicit the activation of a number of intracellular signaling pathways to regulate cytoskeletal and ECM assembly, cell migration, proliferation, differentiation and death [
7,
12]. As the cytoplasmic domain of integrins lacks an actin binding domain and is devoid of enzymatic activity, all these effects are mediated by integrin associated molecules. The integrin associated adhesion proteins that participate in this integrin-actin linkage include the cytoskeletal proteins α-actinin, talin, and skelemin, and the kinases involved in integrin signaling include C-terminal Src kinase, focal adhesion kinase (FAK), integrin linked kinase, and Src [
8]. FAK is a non-receptor protein tyrosine kinase that plays an important role in the localization of integrins to focal adhesions and the assembly of integrin-signaling molecules [
12]. It is involved in anchorage dependent survival signaling and cell adhesion induces FAK autophosphorylation at tyrosine 397 (Y397), which creates a binding site for Src, C-terminal Src kinase, GRB7, phosphatidyl inositol 3 kinase, and phospholipase Cγ. Subsequently, Src phosphorylates FAK at a number of tyrosines including Y925 that serves as binding site for GRB2, which links integrins to the MAP kinase pathway [
12]. Integrin signaling through Src can also be FAK-independent as Src also binds constitutively and directly to β
3, and clustering of β
3 integrins induces autophosphorylation and activation of Src [
13]. The dynamics of integrin signaling is further complicated by its cross-talk with other receptors, including the breast cancer marker, uPAR, and vascular endothelial cell growth factor receptor (VEGFR) [
11,
14].
In this study a series of experiments were performed to better understand the role of integrin-associated proteins and structures, and integrin signaling pathways in breast cancer. A non-breast cancer line, Hek-293, and three breast cancer lines of differing metastatic and invasive capacities were used: MDA-MB-435 that are estrogen receptor-negative and highly metastatic; MDA-MB-231 that are estrogen receptor-negative and highly invasive; and, MCF7 that are estrogen receptor-positive and non-metastatic [
15‐
17]. We determined the levels of integrins expressed by each cell line, and the capacity of a cell agonist to stimulated cell adhesion to integrin ligands and to induce intracellular signaling. We also assessed the capacity of various ECM ligands to induce heterogeneity into the formation and distribution of integrin-associated structures and proteins within the cells. Finally, we determined the levels of uPAR and VEGFR expressed by the cell lines and the capacity of cell adhesion to induce intracellular signaling via integrin-linked Src and MAPK pathways.
Methods
Antibodies, Reagents, Chemicals
Antibodies against β3 (sc-6627), Bcl2 (sc-509), c-Src (sc-8056), ERK (sc-94), FAK (sc-557), pFAK(Y397) (sc-11765), pFAK(Y861) (sc-16663), pErbB2(Y1221/Y1222) (6B12), VEGF (sc-80435), VEGFR2 (sc-57136), uPAR (sc-13522), talin (sc-7534) and HRP secondary antibodies were obtained from Santa Cruz (Santa Cruz Biotechnology, Inc., Santa Cruz, CA); β1 (MAB2253), β6 (MAB2076Z), αvβ3 (LM609), αvβ5 (MAB2019Z) and αvβ6 (MAB2074Z) from Millipore (Millipore Canada Ltd., Etobicoke, ON); β3 (MHCD6100) from Invitrogen (Invitrogen Canada Inc., Burlington, ON); β5 (B5-IVF2) from Abcam (Abcam Inc., Cambridge, MA); MEK, pMEK (S217/S221); c-Src (36D10), pSrc(Y416) (100F9), pSrc(Y527) (2105), pMEK1/2 (9121) and pERK (197G2) from Cell Signaling (New England Biolabs Ltd., Pickering, ON); and, uPAR (MAB807) antibody from R&D (R&D Systems, Inc., Minneapolis, MN). Collagen (type I and IV), fibronectin (FN), vitronectin (VN), fibrinogen (Fg) and an antibody against vinculin (hVIN-1) were obtained from Sigma (Sigma Chemical Co., St. Louis, MO).
Cells and Cell culture
All the cell lines were from ATCC. MDA-MB-435, MDA-MB-231, and Hek-293 cells were cultured in RMPI 1640, and MCF7 cells in F-12 containing 10% fetal calf serum and 100 U/ml penicillin and 100 μg/ml streptomycin. All cells were grown as monolayers on tissue culture plates at 37°C in a humidified incubator with 5% CO2 and 95% air. Cells were subcultured at 80-95% confluence using 0.25% trypsin (w/v)/5 mM EDTA to detach cells.
Flow cytometry
Cells were grown in 100 mm tissue culture plates to 90-95% confluence and harvested with 2% EGTA. For measurement of integrin expression, once harvested all samples were maintained at 4°C to maintain the expression of integrins on the cell surface. Thus, cells were washed and re-suspended in 4°C Tyrode-Hepes Buffer containing 1 mM CaCl2, 1 mM MgCl2, 5.5 mM Glucose and 1 mg/ml BSA. Cells were incubated with primary antibodies for one hour at 4°C, washed three times with ice-cold Tyrode-Hepes Buffer and incubated with PE or Alexa Fluor-488 labeled secondary antibody for another one hour at 4°C. Cells were washed, re-suspended in 0.5 ml of ice-cold Tyrode-Hepes Buffer and kept on ice until analyzed by flow cytometry. Isotype-matched monoclonal antibodies were used as controls. For phorbol 12-myristate 13-acetate (PMA) treatment, cells were grown for 16 hours in media containing 1% fetal calf serum and then the cells were treated with 150 nM PMA for two hours. For mock treatment, the cells were incubated with the same concentration of DMSO as was present in the PMA samples. Data was analyzed using Flowjo program.
Adhesion Assay
Adhesion assays were performed as previously described with minor modifications [
18,
19]. Briefly, 96-well plates were coated with 20 μg/ml of collagen, FN, Fg or VN overnight at 4°C. The wells were blocked with 2% BSA and washed with PBS. MDA-MB-435, MDA-MB-231, MCF7 or Hek-293 cells were suspended in serum free media, with or without the addition of 150 nm PMA. The cells were then transferred to the wells (2 × 10
5 cells/well) and incubated for one hour at 37°C. Unattached cells were removed by washing with PBS and the cells were then incubated in staining solution (20% methanol, 1% formaldehyde and 0.5% crystal violet in H
2O) for 30 min. Plates were washed, lyzed in 0.5% Triton X-100, and adhered cells quantitated by measuring light absorbance at 590 nm.
Western blotting
Cells were grown to 90-95% confluence, washed with ice-cold PBS and lyzed in 500 μl of RIPA buffer (50 mM Tris, pH 8, 150 mM NaCl, 0.1% SDS, 0.5% Na deoxycholic Acid, 1% NP-40 or IGEPAL, 10 μg/ml aprotinin and 10 μg/ml leupeptin), and using a 25 gauge needle. Cell extracts were centrifuged and supernatants kept at -20°C. Equal amounts of protein (24 μg/well) were electrophoretically separated in SDS polyacrylamide gels and proteins were transferred to a nitrocellulose membrane. Membranes were blocked with 5% skim milk and probed with primary antibodies, followed by incubation with HRP-labeled secondary antibodies. Western blots were visualized by an enhanced chemiluminescence detection system according to the manufacturer's protocol (Amersham Life Sciences, Arlington Heights, IL).
Immunofluorescence
Falcon 4-well CultureSlides were treated with 1% SDS, rinsed with PBS and then coated overnight at 4°C with 20 μg/ml of collagen, FN, Fg or VN. Cells were seeded and grown overnight on different ligand-coated chamber cells. Cells were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.2% (v/v) Triton X-100, washed and then blocked with 1% BSA. Filamentous actin (F-actin) was stained using Alexa Fluor 594 phalloidin (Invitrogen, San Diego, CA) for 30 min at a dilution of 1:40. Focal adhesions were stained using an antibody to vinculin (Sigma Chemical Co., St. Louis, MO), or to talin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at a dilution of 1:100 and a fluorescein-conjugated secondary antibody.
Discussion
Integrins play an important role in cell anchorage, migration, differentiation and death [
5,
7], and their upregulated expression in human cancers frequently indicates poor prognosis. Although breast cancer is a heterogeneous form of cancer, α
v integrins as well as other proteins have been identified as prognostic markers. In the present study, using two metastatic (MDA-MB-435 and MDA-MB-231) and a non-metastatic (MCF7) breast cancer cell line, we demonstrated that α
v integrin expression varies between the cell lines (Figure
2A). This variation may partially account for the heterogeneity that is found in breast cancer. In comparison to the non-breast cancer Hek-293 cells, all the cancer cells expressed higher but varying levels of β
5, α
vβ
5 and α
vβ
6. Normal epidermal cells express α
vβ
5 but after transforming into squamous carcinomas, the expression of α
vβ
5 is down-regulated and α
vβ
6 up-regulated that protects the cancer from undergoing anoikis [
24]. Thus, differences in α
vβ
5 and α
vβ
6 expressions may account for some of the heterogeneity in the phenotypes of breast cancers. Furthermore, we found that only MDA-MB-435 cells expressed high levels of β
3 and α
vβ
3. In vivo studies reveal that αvβ3 is also involved in enhanced metastasis of breast cancer to bone [
25]. The high levels of β
3 and α
vβ
3 in metastatic MDA-MB-435 cells is in keeping with β
3 being an important mediator of melanoma cell invasion and migration and with α
vβ
3 as a prognostic indicator in breast cancer [
4,
5,
26,
27]. However, as MDA-MB-231 and MCF7 cells did not express α
vβ
3, α
vβ
3 should not be viewed as a universal prognostic indicator for all forms of breast cancer. Rather, it should be used as an indicator where the use of anti-α
vβ
3 therapeutics is warranted.
Integrins, play a significant role in the acquisition and maintenance of neoplastic phenotype by preventing apoptosis and maintaining cell proliferation, and integrin expression profile can dramatically change upon the normal-to-neoplastic transition [
6]. However, we found that short term (one to two hours) of adhesion onto FN or Fg had minimal effect on integrin expression in MDA-MB-432, MDA-MB-231 and MCF7 cells. Thus, it is likely that changes in integrin expression profile during cancer cell metastasis may either require more time or may also require the activity of matrix-degrading proteases, such as uPA and matrix metalloprotease 2, to modify the surrounding tissue [
5].
In nonmalignant and cancer cells, integrin-mediated adhesion of unstimulated cells is usually low and can be upregulated by the addition of a cell agonist, such as PMA [
18,
19]. In this study, we found that the adhesion of unstimulated breast cancer and Hek-293 cells was already upregulated, and that level of uPAR expressed by the cells (Figure
6) was likely not sufficient enough to upregulate cell adhesion. However, all cell lines when adhered and proliferating constitutively expressed activated pSrc (Figure
4C, lane 1), which may have been influenced by uPAR-integrin interaction, or in MDA-MB-435 and Hek-293 cells, partially a result of Src signaling following its direct binding to β
3 [
13,
14]. Adhesion to VN is mediated by uPAR [
14] and by a number of integrins including α
vβ
1, α
IIbβ
3, α
vβ
3, α
vβ
5, α
vβ
6 and α
vβ
8 [
28]. Similarly, other integrins also share common ligands, which likely accounts for why we did not observe a strong preference for one ECM ligand. In addition, non-integrin adhesion receptors also contributed to cell anchorage as all cells, except MDA-MB-231, adhered to BSA.
The formation of focal complexes, focal adhesion and other integrin-related cellular structures has a profound effect on cell shape and numerous cellular processes that govern the biology of a cell [
12]. Our vinculin and talin staining produced similar results which agree with the role of vinculin in controlling focal adhesion formation by directly interacting with talin [
21]. F-actin and focal adhesion staining demonstrated that the non-breast cancer cell line, Hek-293, was nearly devoid of integrin-associated structures in comparison to the breast cancer lines (Figure
5). We also observed that a two hour PMA treatment induced stress fiber perturbations in all cell lines, and resulted in a loss of focal adhesions in MDA-MB-435 cells. These results are consistent with previous findings that PMA-mediated F-actin reorganization and redistribution is closely linked with cell transformation [
29]. We also concluded that some of the heterogeneity of breast cancer can be explained by variations in the level of integrin-associated F-actin structures between different breast cancers. MDA-MB-435 cells contained numerous well defined stress fibers that protruded into the cell interior and formed numerous focal adhesions. These features readily differentiated MDA-MB-435 cells from the other breast cancer cells. It also appears that MDA-MB-435 focal adhesions were signaling effectively as evident with the correlated transient increases in pFAK, pSrc(Y416) and pERK following PMA treatment (Figure
4), and in the adhesion-induced activation of pFAK and pMEK (Figure
7).
The integrin co-receptors, uPAR and VEGFR, play important roles in the progression of cancers [
11,
14]. All the breast cancer cell lines and Hek-293 cells expressed uPAR but only MCF7 cells expressed high levels of VEGFR. The expression of uPAR by all the cancer lines, is in keeping with uPA/uPAR being a prognostic marker of breast cancer. uPAR participates in many cellular processes by interacting with β
1 and β
3 integrins and modulate their signaling, by serving as a binding site for VN and by inducing cytoskeletal reorganization [
14,
30]. The delivery of an adequate supply of blood to malignant tumors is required for their rapid expansion as they must receive nutrients and oxygen imposed by tumor growth [
11]. Many cancers meet their blood supply demands by inducing angiogenesis, and there is increasing evidence implicating integrin signaling, generated by interactions with ECM proteins and with VEGFR, as a major modulator of cancer-induced angiogenesis [
4,
11]. The high expression of VEGFR by the non-metastatic MCF7 cells, may indicate a critical role for angiogenesis in the progression of MCF7 breast cancers. In MDA-MB-435 and MDA-MB-231 metastatic tumors, uPAR-mediated degradation and remodeling of the ECM to facilitate metastasis [
14], is likely of more importance than VEGFR-mediated angiogenesis in the progression of these cancers.
Breast carcinomas have been reported to contain higher MAPK activity than benign breast tissue, and there is a positive correlation between ERK activation and shorter relapse-free survival period [
31,
32]. Other studies reported a positive correlation between ERK activation and a less aggressive disease and better survival rates [
33]. The magnitude and temporal organization of ERK activity also correlates with specific biological responses [
34,
35]. In intestinal cells, transient ERK activity results in cell growth, while a strong and sustained ERK activity leads to cell cycle arrest [
35]. In our study, we identified marked differences in the regulation of MAPK signaling and ERK activation within the cancer lines. The levels of pMEK and pERK in adhered MDA-MB-435 and MCF7 cells were transient, reaching a maximum within two hours of PMA treatment, while pMEK levels in MDA-MB-231 cells remained constitutively high and pERK levels continued to increase. Furthermore, in contrast to MDA-MB-231 cells in which pMEK levels were adhesion-independent and pERK levels were adhesion-dependent, pMEK levels were adhesion-dependent and pERK levels were adhesion-independent in MDA-MB-435 cells. We speculate that differences in the activity of phosphatases within the cell lines accounted for the different pERK levels, and that alterations in the regulation of phosphatase activity between various breast cancers contributes to variations in their phenotypes. Furthermore, our data supports a relationship between pERK and the metastatic capacity of the cells, as adhered metastatic MDA-MB-435 and MDA-MB-231 cells contained elevated pERK levels compared to non-metastatic MCF7 and Hek-293 cells (Figure
7B).
The autophosphorylation of FAK at Y397, serves as binding site for Src-family protein kinases which following further activation, phosphorylates a variety of substrates such as paxillin, and activates a number of protein kinase cascades [
12,
36]. The expression of Src correlates with metastatic activity of breast cancers, and integrin signaling through Src can be FAK-mediated or FAK-independent as Src in cancers expressing β
3 integrins [
13,
37]. In our studies, all proliferating cells expressed activated pSrc(Y416) but only metastatic MDA-MB-435 cells showed an induction of pSrc levels following PMA stimulation. As this was the only breast cancer to express α
vβ
3, we believe that FAK-independent activation of Src by α
vβ
3 contributes to the metastatic phenotype of MDA-MB-435 breast cancers.
The ability of metastatic cells to loosen their adhesion to the ECM and acquire a migratory phenotype that enables the cancer to move through and expand into other tissues are processes regulated by FAK-Src signaling [
36]. High FAK expression occurs in cancers, including breast cancers, and FAK expression is correlated with a highly malignant and metastatic phenotype [
38‐
40]. Our own observations are consistent with these previous studies, with the breast cancers containing higher levels of FAK than Hek-293 cells. In addition, pFAK levels were markedly elevated in MDA-MB-231 cells, which may reflect the invasive phenotype of this cancer [
15]. The higher levels of pFAK in MDA-MB-231 may contribute to focal adhesion turnover and reorganization, resulting in fewer stable focal adhesions and fewer contacts between integrins and actin stress fibers. This speculation is supported by our observation that MDA-MB-231 cells formed the fewest focal adhesions of the three breast cancers, which may allow for them to more readily disengage from the ECM. Their capacity to remodel and degrade ECM, partially using uPAR-mediated processes, would then facilitate their migration and invasion into other tissues. Other studies have demonstrated that FAK-mediated signaling to ERK does not follow a single linear pathway [
36]. FAK enhances the phosphorylation of MEK1 at Ser-298 facilitating ERK2 activation [
41]. Thus, FAK signaling can potentially affect the tumorogenic, metastatic, and invasiveness of breast cancers by modulating Src and MAPK signaling.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
The study was designed by AT, XL, YL and TAH. All authors have read and approved the final manuscript. TH ensured funding. AT performed data collection, statistical analyses and interpretation of the cell adhesion, PMA treatment and adhesion-induced signaling results. AT collected integrin expression data by western analysis and wrote first draft of manuscript. XL collected integrin expression data by immunocytochemistry analysis and performed data collection, statistical analyses and interpretation of the immunocytochemistry results. YL performed FACS integrin data collection, statistical analyses and interpreted the results. YL also collected pFAK data for Figure
3 and data for Figure
5. TH wrote all manuscript revisions, designed most experimental approaches taken, and performed statistical analyses and interpretation of all the results.
All authors have read and approved the final manuscript.