|
|
Neoplasia. 2008 November; 10(11): 1240–1252. | PMCID: PMC2570600 |
Copyright © 2008 Neoplasia Press, Inc. All rights reserved Mammary Tumors Initiated by Constitutive Cdk2 Activation Contain an Invasive Basal-like Component 1Patrick E Corsino,* Bradley J Davis,* Peter H Nörgaard,† Nicole N Teoh Parker,* Mary Law,* William Dunn,‡ and Brian K Law* *Department of Pharmacology and Therapeutics and the Shands Cancer Center, University of Florida, Gainesville, FL, 32610, USA †Department of Pathology, Copenhagen University Hospital Herlev, DK-2730 Herlev, Denmark ‡Department of Anatomy and Cell Biology and the Shands Cancer Center, University of Florida, Gainesville, FL, 32610, USA Received June 20, 2008; Revised August 13, 2008; Accepted August 14, 2008. The basal-like subtype of breast cancer is associated with invasiveness, high rates of postsurgical recurrence, and poor prognosis. Aside from inactivation of the BRCA1 tumor-suppressor gene, little is known concerning the mechanisms that cause basal breast cancer or the mechanisms responsible for its invasiveness. Here, we show that the heterogeneous mouse mammary tumor virus-cyclin D1-Cdk2 (MMTV-D1K2) transgenic mouse mammary tumors contain regions of spindle-shaped cells expressing both luminal and myoepithelial markers. Cell lines cultured from these tumors exhibit the same luminal/myoepithelial mixed-lineage phenotype that is associated with human basal-like breast cancer and express a number of myoepithelial markers including cytokeratin 14, P-cadherin, α smooth muscle actin, and nestin. The MMTV-D1K2 tumor-derived cell lines form highly invasive tumors when injected into mouse mammary glands. Invasion is associated with E-cadherin localization to the cytoplasm or loss of E-cadherin expression. Cytoplasmic E-cadherin correlates with lack of colony formation in vitro and β-catenin and p120ctn localization to the cytoplasm. The data suggest that the invasiveness of these cell lines results from a combination of factors including mislocalization of E-cadherin, β-catenin, and p120ctn to the cytoplasm. Nestin expression and E-cadherin mislocalization were also observed in human basal-like breast cancer cell lines, suggesting that these results are relevant to human tumors. Together, these results suggest that abnormal Cdk2 activation may contribute to the formation of basal-like breast cancers. Microarray analyses have recently allowed breast tumors to be categorized as luminal, basal-like, normal-like, or Her2-positive based on distinct gene expression profiles, morphologic characteristics, prognostic outcomes, and responsiveness to currently available therapeutic approaches [ 1,2]. The basal-like subtype represents approximately 20% of human breast cancers overall but 39% of breast tumors in premenopausal African American women [ 3]. These tumors are associated with a high rate of recurrence and poor outcome [ 2]. The basal-like subtype of cancers is also termed triple negative because these tumors typically lack estrogen receptor (ER), progesterone receptor, and Her2 overexpression but generally express a subset of myoepithelial markers, including cytokeratin 14 (CK14), CK5, α smooth muscle actin (αSMA), nestin, or p63 (reviewed in [ 4–6]). Basal-like tumors lack responsiveness to tamoxifen and aromatase inhibitors that target ER-positive luminal tumors and herceptin that targets Her2-positive tumors. The mouse basal-like breast cancer models described to date involve genetic deletion of the BRCA1 and p53 tumor-suppressor genes [ 7,8]. Tumors initiated by BRCA1 inactivation in mice express the progesterone receptor [ 9] and overexpress Her2 [ 10] and thus do not fit the triple negative clinical definition of basal breast cancer. Therefore, it is likely that additional genetic lesions contribute to the formation of sporadic human basal-like breast cancers. Microarray studies have suggested several candidate “drivers” of basal breast cancer including epidermal growth factor receptor (EGFR), c-Kit, c-Met, and cyclin E. However, none of these genes have yet been demonstrated to specifically induce basal-like breast cancer when overexpressed. Interestingly, human basal-like breast tumors frequently exhibit p16 overexpression, low levels of Rb and cyclin D1 expression, and high levels of cyclin E expression [ 11]. Based on these observations, it was proposed that Rb inactivation is mechanistically linked to the basal-like subtype [ 11]. Together, these results suggest that basal-like tumors may have low levels of Cdk4/Cdk6 activity but perhaps high levels of Cdk2 activity. We previously described a novel mouse transgenic model of breast cancer in which expression of a cyclin D1-Cdk2 (D1K2) fusion protein [ 12] under the control of the mouse mammary tumor virus (MMTV) promoter/enhancer induces mammary tumorigenesis (MMTV-D1K2 animals) [ 13]. Mammary tumors from these animals exhibit Rb hyperphosphorylation, high levels of Cdk2 activity, and up-regulation of E2F-dependent transcription [ 13]. Thus, MMTV-D1K2 tumors exhibit functional inactivation of Rb tumor-suppressor activity. MMTV-D1K2 tumors are heterogeneous and induce a desmoplastic reaction associated with transforming growth factor beta (TGFβ) secretion by the cancer cells. As mentioned previously [ 13], some of the cancer cell lines derived from the MMTV-D1K2 tumors exhibit the morphologic features of myoepithelial cells. Here, we report a more extensive characterization of MMTV-D1K2 cell lines and demonstrate that these cells express protein markers associated with the basal/myoepithelial lineage. E-cadherin is a potent invasion suppressor expressed in nontransformed mammary epithelial cells [ 14]. The MMTV-D1K2 cell lines exhibit decreased or mislocalized E-cadherin expression in culture. Introduction of cell lines derived from MMTV-D1K2 tumors into the mammary glands of wild type syngeneic mice results in the formation of invasive tumors composed of spindle-shaped cells that exhibit E-cadherin mislocalization to the cytoplasm and the expression of basal/myoepithelial markers. Morphologic and immunohistochemical analyses of the primary tumors demonstrate a biphasic morphology characteristic of adenomyoepithelial-type carcinoma with populations of spindle-shaped cells. These spindle-shaped cells exhibit E-cadherin down-regulation and localization to the cytoplasm and expression of the myoepithelial marker αSMA. These studies indicate the presence of a subpopulation of invasive basal-like breast cancer cells in the primary MMTV-D1K2 tumors. In vitro analysis of multiple clonal cell lines derived from MMTV-D1K2 tumors demonstrate the expression of various subsets of myoepithelial and luminal epithelial markers, a finding consistent with the “mixed-lineage” properties of human basal breast cancers [ 15–17]. In all of the cell lines isolated, E-cadherin expression is either low and/or mislocalized to the cytoplasm. E-cadherin mislocalization is associated with the inability of the cells to form colonies with normal cell-cell contacts in culture and correlates with the lack of β-catenin and p120 ctn staining at cell-cell junctions. In some cell lines, decreased E-cadherin expression is associated with an increase in N- or P-cadherin expression. This “cadherin switch” from E- to N-cadherin expression is associated with increased invasiveness and part of the epithelial-to-mesenchymal transition (EMT) that occurs during the progression of some tumor types [ 18]. Epithelial-tomesenchymal transition has been shown to occur in the basal-like category of breast cancers [ 19]. Isolation of Tumor Cell Lines MMTV-D1K2 cancer cell lines were isolated essentially as described [ 13]. When differential trypsinization was performed, the cells remaining adherent to the flasks and the detached cells were retained. This is critical because several of the basal breast cancer cell lines adhere very loosely to tissue culture flasks or flasks coated with rat tail collagen. The loosely adherent myoepithelial-like cancer cells were separated from myofibroblasts by taking advantage of their differential rates of adhesion. The cancer cells were cloned by limiting dilution. All cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (35-011-CV; Mediatech, Inc., Manassas, VA). Preparation and Analysis of Tumor and Cell Extracts by Immunoblot Cell extracts were prepared as described [ 13], and immunoblot analysis was performed using antibodies from Santa Cruz Biotechnology, Inc. [Santa Cruz, CA; N-cadherin (sc-7939), P-cadherin (sc-7893), vimentin (sc-32322), nestin (sc-23927), β-catenin (sc-7199), zyxin (sc-6437), p130 (sc-317), Cdk2 (sc-163), actin (1616), and p53 (sc-100)] and Sigma-Aldrich [St. Louis, MO; Flag, M2 (F-3165) and αSMA (A-2547)]. Antibodies specific for EGFR (#2232) were obtained from Cell Signaling Technology, Inc. (Danvers, MA). p53 antibodies were also obtained from Oncogene Science (Cambridge, MA). Antibodies specific for CK14 (MS-115) and Her2/ neu (MS-730) were obtained from LabVision/Neomarkers, Inc. (Fremont, CA). p120 catenin (610133) antibody was obtained from BD Transduction Laboratories (San Diego, CA). Tumor samples contain large amounts of immunoglobulin, which interferes with subsequent immunoblot and immunoprecipitation assays. Tumor-associated immunoglobulin was removed by preclearing aliquots of tumor lysate containing 1 mg of protein with 100 µl/tube packed Protein G-Sepharose (10-1242; Invitrogen, Carlsbad, CA). The supernatants were retained for subsequent analyses. Immunofluorescence Microscopy Cells were plated onto glass coverslips in six-well plates. After a 24-hour incubation, cells were fixed with 1% paraformaldehyde in phosphate-buffered saline (PBS) for 20 minutes, followed by a 10-minute incubation with quench solution (50 mMammonium chloride + 0.5% Triton X-100 in PBS). The cells were then blocked for 1 hour with an antibody buffer (10% goat serum + 0.5% Triton X-100 in PBS). Primary staining was performed using the following antibodies at a 1:100 dilution in antibody buffer for 2 hours: CK14 (MS-115; Neomarkers, Inc.); E-cadherin (610181) and p120 catenin (610133) from BD Biosciences Pharmingen; E-cadherin (24E10) from Cell Signaling Technology, Inc.; and zyxin (sc-6437), β-catenin (sc-7199), and N-cadherin (sc-7939) from Santa Cruz Biotechnology, Inc. After four washes with PBS, cells were incubated with secondary antibody for 1 hour using either goat anti-Rabbit Fluor 488 ( A11008; Invitrogen Molecular Probes), goat anti-Mouse Fluor 488 (A11001; Invitrogen Molecular Probes), rabbit anti-Goat Fluor 488 ( A11078; Invitrogen Molecular Probes), or goat anti-Mouse Cy3 (81–6515; Zymed, Carlsbad, CA), at either a 1:200 dilution in antibody buffer for single staining or a 1:300 dilution for double staining. After four washes with PBS, coverslips were mounted onto slides with Vectashield + 4′,6-diamidino-2-phenylindole (DAPI; H-1200; Vector Laboratories, Burlingame, CA). Actin was visualized using Texas Red-X Phalloidin (T7471; Invitrogen Molecular Probes) added during the secondary staining step at a dilution of five units per slide. Images were captured using an upright microscope (Axioplan2; Zeiss, Thornwood, NY) and visualized using Openlab 5.3.0 Improvision software. Orthotopic Tumor Growth Studies Cells in log growth phase were collected by trypsin digestion, suspended in 10% FBS-DMEM, and washed three times with Hank's balanced salt solution (HBSS) (21-020-CV; Mediatech, Inc.). The cells were counted and diluted to a concentration of 107 cells/ml in HBSS. The cell suspension (100 µl) was injected into the #4 mammary glands of adult wild type female FvB mice just beneath the surface of the nipple. Three mice were injected with each polyclonal cell line, and tumor formation occurred from 2 to 6 weeks in all of the injected animals. Tumors were excised at a small size (2–6 mm in diameter) so that tumor invasion into the surrounding stroma could be observed. Immunohistochemical Analysis of Tumor Tissue Sections Two-micrometer serial sections of paraformaldehyde-fixed, paraffin-embedded tumor tissue were dewaxed in Tissue-Clear (Sakura Finetek Europe, Zouterwoude, The Netherlands) and hydrated through a series of diluted ethanol followed by antigen retrieval in 10 mM Tris, pH 9.0, 0.5 mM EGTA solution using microwave oven treatment (15 minutes). Immunostaining was performed with a commercially available kit (Animal Research Kit, ARK; DakoCytomation, Glostrup, Denmark) in accordance with the manufacturer's instructions. Additional blocking of endogenous biotin was performed with the DAKO Biotin Blocking System (DakoCytomation) in accordance with the manufacturer's instructions. The following antibodies were used: monoclonal mouse anti-human SMA, clone 1A4, dilution 1:200 and monoclonal mouse anti-human E-cadherin, clone NCH-38, dilution 1:25 (both from DakoCytomation); and monoclonal mouse anti-human keratin 14 Ab-1, clone LL002, dilution 1:400 (LabVision/NeoMarkers Inc.). Images were captured using a microscope (BX51; Olympus, Center Valley, PA) equipped with a Color View camera using AnalySIS getIT version 5.0 (Soft Imaging System, Munster, Germany). The tumor analyzed by immunohistochemistry in is from a 58-year-old white woman diagnosed with invasive ductal carcinoma, grade III. Morphologically, the tumor is described as a “ring carcinoma” because of its necrotic center. MMTV-D1K2 Hypercellular Lesions Exhibit an Invasive Phenotype We have previously described tumors arising in transgenic mice in which a cyclin D1-Cdk2 fusion protein [ 12] is driven by the MMTV promoter/enhancer, termed MMTV-D1K2 animals [ 13]. Tumors arising in these animals are heterogeneous and contain ductal structures surrounded by spindle-shaped cells. The identity of these spindle-shaped cells is unclear. However, these cells are of interest because they seem to invade into the surrounding mammary fat pad (). This is in contrast to MMTV- neu tumors or tumors derived from MMTV- neu tumor cells that show well-demarcated boundaries (). We isolated cell lines from different MMTV-D1K2 tumors to more fully characterize the cell types composing them. Several of the cell lines, including D1K2-T2 and D1K2-T4, display features in culture similar to that of primary myoepithelial cells [ 20–23] including an elongated morphology, multiple cellular extensions or processes, and a relative lack of cell-cell adhesion (). The BT549 human basal-like breast cancer cell line [ 24] exhibits a similar morphology. In contrast, nontransformed mouse mammary epithelial NMuMG cells have a cuboidal shape and form distinct colonies that exhibit a “cobblestone” morphology. MMTV-D1K2 Tumor Cells Display Characteristics Consistent with Basal-like Breast Cancer Because the MMTV-D1K2 tumor-derived cell lines exhibit a myoepithelial morphology, we examined whether they expressed myoepithelial markers. Immunoblot analysis indicates that the cell lines derived from MMTV-D1K2 tumors express varying amounts of several basal/myoepithelial markers including P-cadherin, EGFR, CK14, αSMA, and nestin (). In contrast, the neuT cell line derived from an MMTV- neu tumor [ 13] expresses low or undetectable levels of these basal markers. Several of the D1K2 cell lines also express the luminal epithelial marker E-cadherin. Because the polyclonal D1K2 cell lines exhibit a variety of cell morphologies, clonal cell lines were isolated. Specifically, the D1K2-T2,CL1 and D1K2-T2,CL6 clonal lines exhibit different morphologies in culture and different expression profiles in immunoblot analyses. This observation indicates that the cancer cell lines isolated are heterogeneous and is consistent with the heterogeneous nature of the primary tumors [ 13]. The D1K2-T4 cell line does not exhibit detectable transgene expression (), although the DNA from this cell line tests positive for the presence of the transgene by polymerase chain reaction analysis (not shown). Thus, expression of the cyclin D1-Cdk2 fusion protein may not be required for maintenance of the basal-like/myoepithelial phenotype. The D1K2 tumor cell lines express high levels of the intermediate filament protein nestin (), which was shown to be a basal breast cancer/myoepithelial marker [ 4]. Quantitative proteomic analyses of protein extracts from the invasive D1K2 tumor cell lines and the BT549 cell line indicate that these cells express higher levels of the structural protein zyxin than the neuT and mouse mammary gland NMuMG cell lines (proteomic analyses will be described in detail elsewhere). This was verified by immunoblot (). The cell lines isolated from the MMTV-D1K2 tumors might represent a small fraction of the cells present in the primary tumors and therefore may not be typical of the overall tumor composition. Immunoblot analyses of the primary tumors () demonstrated that the tumor from which the D1K2-T1 cell line was derived expressed relatively high levels of the D1K2 fusion protein as measured by staining with Flag antibody (Flag (D1K2)). The tumor of origin of the D1K2-T1 cells also displayed low Her2 expression, high levels of E-cadherin expression, and low levels of αSMA staining. The tumor from which the D1K2-T4 cell line was derived did not express the cyclin D1-Cdk2 fusion protein and expressed low levels of E-cadherin, low levels of Her2, and high levels of αSMA. MMTV-neu tumors exhibit high Her2 expression that correlates with high E-cadherin expression. In contrast, αSMA expression is inversely related to Her2/neu transgene expression and E-cadherin expression. Collectively, these results suggest that the cell lines isolated for analysis have protein expression patterns similar to the primary tumors. The lack of expression of the cyclin D1-Cdk2 transgene product in the D1K2-T4 cell line and in the primary tumor from which it was derived was unexpected because these cells resemble the other MMTV-D1K2 cell lines in their morphology and expression pattern of luminal and basal markers. Immunoblot analyses were performed to examine the possibility that other molecular changes occurred that might substitute for D1K2 expression. The results showed that the D1K2-T4 cell line exhibited several features that could render D1K2 expression dispensable including cyclin A overexpression, (presumably mutant) p53 overexpression, and low levels of Rb and p21 expression (). MMTV-D1K2 Tumor Lines Display Mixed Luminal/Myoepithelial Character In the normal mammary gland, E-cadherin is expressed in luminal epithelial cells, whereas αSMA and CK14 are expressed in myoepithelial cells. The observation that luminal proteins such as E-cadherin and myoepithelial markers such as αSMA and CK14 are expressed in the same polyclonal cell population could be explained by the presence of cell subpopulations that each express different subsets of markers. Clonal cell lines were derived from the MMTV-D1K2 tumor cell lines to determine whether these luminal and myoepithelial markers are expressed in the same cells. Immunoblot analysis indicates that in multiple cases the clonal cell lines express luminal markers such as E-cadherin and CK19 and also express basal/myoepithelial markers such as P-cadherin, EGFR, CK14, αSMA, and nestin (). Interestingly, two of the four D1K2-T4 subclones exhibit p53 overexpression. Immunofluorescence microscopy experiments demonstrated that the D1K2-T1,CL1 cell line expresses both CK14 and E-cadherin uniformly in all of the cells (). Together, the results in , A and B, indicate that the cell lines derived from the MMTV-D1K2 tumors exhibit a mixed luminal/myoepithelial protein expression pattern at the single-cell level. A number of human mammary carcinoma cell lines have been analyzed in microarray experiments and classified into luminal or basal-like subgroups [ 24]. Interestingly, analysis of some of these mammary carcinoma cell lines by immunoblot suggests that human breast carcinomas may also frequently exhibit a luminal/myoepithelial mixed-lineage phenotype (). The T47D, MCF7, and MDA-MB-361 cell lines, classified as luminal by microarray analysis, express the myoepithelial marker P-cadherin. Likewise, the MDA-MB-468 and BT549 cell lines, classified as basal by microarray analysis, express the luminal marker E-cadherin (). In contrast, the nontransformed mouse mammary gland NMuMG cell line does not express P-cadherin, and the nontransformed human basal-like cell lines MCF10A and HBL100 do not express E-cadherin. p53 overexpression in the T47D, MDA-MB-468, BT549, MDA-MB-231, and MDA-MB-435s carcinoma cell lines is expected because these cells harbor mutant p53 alleles [ 25–28]. MMTV-D1K2 Tumor-Derived Cell Lines Form Invasive Tumors In Vivo Basal-like tumors are often invasive; therefore, we examined whether the MMTV-D1K2 tumor-derived cell lines would form invasive tumors in vivo. The polyclonal D1K2-T1, D1K2-T2, D1K2-T4, and D1K2-T5 cell lines were injected into the mammary glands of three wild type female FvB mice. All of the mice formed tumors from 2 to 6 weeks after injection. The tumors exhibited invasion into the surrounding mammary fat pad and muscle (). Invasion of individual cancer cells between adipocytes and muscle fibers was observed at the tumor/stroma interface. The infiltrative tumors exhibited inclusions of adipocytes. The disorganized and poorly differentiated appearance of the tumors is similar to that of the cells in hypercellular lesions in the MMTV-D1K2 transgenic mammary glands that exhibit invasion into the fat pad (). The blue Trichrome staining at the invading edge of the tumors (; D1K2-T1 and D1K2-T2) coincides with cells of fibroblastic morphology and is indicative of desmoplasia and fibroblast infiltration into the tumors. The MMTV-D1K2 cell line-derived tumors were negative for ER and progesterone receptor expression and Her2 overexpression by immunohistochemistry (data not shown), consistent with the expression pattern of the cell lines in vitro (, A and C). The finding that some of the D1K2 cell lines express high levels of E-cadherin in cell culture ( and ) seems inconsistent with the poor ability of the cells to form distinct colonies in vitro () and with the invasiveness of the cell lines observed in vivo (). Immunohistochemistry experiments examining the localization of E-cadherin within the tumors demonstrated that the noninvasive tumors derived from neuT cells exhibit strong E-cadherin staining at cell-cell junctions () and lack expression of αSMA and CK14 (not shown). In contrast, tumors derived from the D1K2-T1, D1K2-T4, and D1K2-T5 cell lines exhibit weak, diffuse cytoplasmic E-cadherin staining, suggesting that E-cadherin localization to the cytoplasm contributes to the invasiveness of these cell lines. The D1K2-T1, D1K2-T2, and D1K2-T4 cell line-derived tumors also express variable levels of CK14 (). The positive staining of the tumors for both E-cadherin and CK14 suggests that the tumors, like the initiating cell lines, exhibit a mixed luminal/myoepithelial phenotype. We next examined whether populations of cancer cells with a mixed-lineage phenotype were present in the primary MMTV-D1K2 tumors. The primary tumor from which the D1K2-T1 cell line was derived has a biphasic morphology characteristic of adenomyoepithelial-type carcinoma, with both glandular structures and clusters of less differentiated spindle-shaped cells (). Immunohistochemical staining of serial tumor sections showed that these spindle-shaped cells stain positively for E-cadherin diffusely localized to the cytoplasm and show positive staining for αSMA. In contrast, the surrounding glandular structures are composed of cells that express E-cadherin at their cell-cell junctions and lack αSMA expression. This observation suggests that the primary MMTV-D1K2 tumors contain both luminal/myoepithelial mixed-lineage cell populations and cells that exhibit a luminal-like staining pattern. Thus, it seems likely that the basal-like cancer cell lines isolated from the MMTV-D1K2 tumors represent the spindle-shaped subpopulation of cells present in the primary tumors. MMTV-D1K2 Tumors Resemble Human Basal-like Breast Cancers MMTV-D1K2 cancer cells resemble human basal-like breast cancer cell lines in terms of their morphology, protein expression patterns, and invasiveness in vivo. Therefore, we examined how the morphology of the MMTV-D1K2 tumors compares with that of human basal-like (triple negative) breast tumors. The representative triple-negative breast tumor shown is from a 58-year-old white female with grade III invasive ductal carcinoma. This tumor was verified to lack ER and Her2 overexpression in two separate immunohistochemical analyses (data not shown). The tumor contains a necrotic core and an expanding boundary that invades into the surrounding mammary fat pad and contains inclusions of adipocytes (). Immunohistochemical staining shows that the tumor cells express αSMA, E-cadherin, and CK14. Red Van Gieson's staining indicates that the tumor contains extensive collagen deposits (fibrosis) interspersed between clusters of cancer cells. This fibrosis colocalizes with tumor-associated fibroblasts. In many areas of the tumor, the tumor-associated fibroblasts make up a larger fraction of the tumor volume than the cancer cells. We also examined whether the MDA-MB-231 and MDA-MB-436 cell lines, which exhibit basal-like expression profiles [ 24], form tumors that display features similar to human basal-like breast cancers and similar to tumors formed from MMTV-D1K2 cancer cell lines. Ten million cells of each line were injected orthotopically into three adult female athymic nude mice. The animals developed tumors from 3 to 6 weeks after injection. Tumors were excised when they reached 2 to 6 mm in diameter so that invasion into the surrounding stroma could be visualized. shows that MDA-MB-231 and MDA-MB-436 tumors grown as xenografts in athymic nude mice form invasive tumors that contain inclusions of adipocytes, exhibit fibroblast accumulation and fibrosis at the tumor/stroma boundary, and develop a fibrotic necrotic core when tumors grow larger than approximately 3 mm in diameter. The invasive growth pattern of the MDA-MB-231 cells in vivo is similar to that observed with the MMTV-D1K2 tumor-derived cell lines. MMTV-D1K2 Tumor-Derived Cell Lines Exhibit Extensive Stress Fiber Formation and Cytoplasmic E-cadherin, p120ctn, and β-Catenin Localization Primary myoepithelial cells exhibit constitutive stress fibers in culture [ 22]. Zyxin is a component of focal adhesions and associates with actin stress fibers [ 29]. Immunoblot experiments () indicate that the D1K2 tumor cell lines express higher levels of zyxin and nestin than the neuT cell line. Therefore, we examined zyxin localization and colocalization with actin in immunofluorescence microscopy studies (). The neuT cells grow as colonies and exhibit cortical actin staining but no detectable stress fibers. Zyxin is present primarily at the outer border of the neuT colonies. In contrast, the D1K2 cancer cell lines exhibit constitutive stress fiber formation. Zyxin colocalizes with the ends of actin stress fibers and is present around the periphery of each of the cells that is not part of a colony. It is unclear from the studies in whether the cytoplasmic E-cadherin localization observed in the tumors formed from the MMTV-D1K2 cancer cell lines results from environmental influences in the tumor milieu or is caused by intrinsic properties of the cancer cells themselves. Immunofluorescence microscopy studies indicate that the noninvasive neuT cancer cells form colonies in culture and that E-cadherin and the E-cadherin-associated protein p120 ctn localize to cell-cell contacts in these colonies (). In contrast, in several of the D1K2 tumor cell lines, E-cadherin and p120 ctn are largely localized to the cytoplasm. This is not specific to the D1K2 cancer cell lines because the BT549, MDA-MB-435s (), MDA-MB-231, and MDA-MB-436 (data not shown) human basal-like breast cancer cell lines also exhibit punctate cytoplasmic E-cadherin localization and diffuse cytoplasmic p120 ctn localization. E-cadherin also functions to sequester β-catenin by localizing it to adherens junctions and preventing it from translocating to the nucleus and functioning as a transcriptional coactivator. E-cadherin localization to the cytoplasm (or loss of E-cadherin expression as observed in the D1K2-T5,CL1 cell line) correlates with decreased junctional β-catenin staining in the D1K2 and BT549 cell lines (). These observations are significant because E-cadherin is thought to function as a tumor suppressor by mediating cell-cell adhesion and by restraining the proinvasive oncogenic effects of β-catenin [ 30–32] and p120 ctn [ 14,33,34]. MMTV-D1K2 Tumors Cyclin D1 is overexpressed in approximately 40% to 50% of human breast cancers [ 35,36], but cyclin D1 overexpression typically occurs in luminal tumors rather than in basal-like breast cancers [ 11]. Cyclin E overexpression has been noted in basal-like breast cancers [ 37]. Cyclin E overexpression in breast cancers correlates with ER negativity and poor prognosis, whereas cyclin D1 overexpression correlates with ER expression and a favorable outcome [ 38]. Cyclin E potently activates Cdk2, and the cyclin D1-Cdk2 fusion protein functions as a constitutively active form of Cdk2 [ 12,13]. Whether any type of constitutive Cdk2 activation is sufficient to induce basal breast cancer formation requires further study. The animal models of basal-like breast cancer constructed to date involve genetic inactivation of BRCA1 and p53 [ 7,39]. It is unknown whether the cyclin D1-Cdk2 fusion protein induces the formation of basallike cancer cells through a mechanism distinct from BRCA1 and p53 deletion or whether expression of the cyclin D1-Cdk2 fusion protein is functionally equivalent to BRCA1 and p53 deletion. Interestingly, Cdk2 has recently been shown to inhibit the ubiquitin ligase activity of the BARD1/BRCA1 complex [ 40], and the BARD1/BRCA1 ubiquitin ligase complex seems to mediate the tumor-suppressive functions of the BRCA1 gene [ 41]. p53 function is frequently lost in breast cancers, and p53 inactivation is thought to contribute to cell invasiveness [ 42]. p53 also suppresses tumorigenesis in part by inducing expression of the Cdk inhibitor p21. p21-mediated inhibition of proliferation plays a critical role in suppressing tumorigenesis in some contexts [ 43]. We have shown previously that D1K2 can function to sequester p21 and p27 [ 12,13]; therefore, D1K2 may partially override p53 function, with respect to p21. “Mixed-Lineage” Characteristics of MMTV-D1K2 Tumor Cell Lines Previously, studies of transgenic mouse breast cancer models (reviewed in Sutherland and Musgrove [ 44]) showed that expression of cyclin D1 in the mammary gland under the control of the MMTV promoter resulted in adenocarcinomas in 75% of the mice, although some squamous differentiation was observed [ 45]. MMTV-cyclin D2 transgenic mice also develop adenocarcinomas albeit at a lower frequency (19%) [ 46]. In contrast, MMTV-cyclin D3 mice form primarily squamous cell carcinomas [ 47]. These studies indicate that the type of cell cycle deregulation that drives tumor formation can influence the differentiation status of the resulting tumors. The MMTV-cyclin D1-Cdk2 transgenic mouse tumors described here are morphologically heterogeneous, including metaplastic, adenosquamous, and adenomyoepithelial-type carcinomas [ 13]. The spindle cell myoepithelial-like component of these tumors stains positively for E-cadherin and αSMA and exhibits E-cadherin mislocalization to the cytoplasm. Cell lines isolated from these tumors exhibit several similarities with human basal-like breast cancers including the following: 1) the expression of myoepithelial markers such as αSMA, nestin, CK14, and EGFR; 2) lack of Her2 overexpression and lack of ERα expression; 3) the expression of subsets of luminal markers including E-cadherin, CK18 (not shown), and CK19, consistent with the luminal/myoepithelial “mixed-lineage” nature of human basal-like breast cancers [ 15,16,48]; 4) the cells exhibit myoepithelial-like morphology and cytoskeletal features in vitro; 5) the cell lines form invasive tumors with spindle morphology in wild type mouse mammary fat pads in vivo; and 6) the cells seem to have undergone EMT that has recently been shown to occur in the basal-like subtype of breast tumors [ 19]. MMTV-cyclin D1-Cdk2 transgene expression in a mammary stem-like cell might explain the observed heterogeneity in tumor morphology. Alternatively, cyclin D1-Cdk2 expression may inhibit lineage specification or block differentiation. Both of these hypotheses are consistent with the observation that the markers CK14 and nestin expressed in the MMTV-cyclin D1-Cdk2 tumor cell lines are associated with relatively undifferentiated cell populations [ 17,49–53]. MMTV-D1K2 Invasiveness Several non-mutually exclusive mechanisms could contribute to the invasiveness of MMTV-D1K2 tumors including E-cadherin mislocalization/down-regulation and expression of proteins previously correlated with invasiveness such as nestin [ 54] and zyxin [ 55]. The role of Cdk2 activation in these processes is unclear. Chronic treatment with the Cdk2 inhibitor roscovitine did not revert the MMTV-D1K2 cell lines to a luminal-like morphology or increase the formation of adherens junctions in culture (data not shown). It is possible that Cdk2 activation directly influences cell invasiveness, perhaps by functionally inactivating Rb. Recently, it has been shown that knock-down of Rb expression by short-interfering RNA decreases E-cadherin expression and induces EMT [ 56]. However, it is also possible that Cdk2 activation induces irreversible changes to the cells that result in increased invasiveness. The mixed-lineage nature of the MMTV-D1K2 cells may result from an alteration in the normal differentiation program that causes increased invasiveness. The observation that invasive human basal-like breast cancer cell lines exhibit the same E-cadherin down-regulation/mislocalization and mixed-lineage expression pattern is consistent with this hypothesis. We have shown previously that MMTV-D1K2 cell lines exhibit aneuploidy [ 13]. Cdk2 activation may induce genetic instability that results in subpopulations of tumor cells with increased invasive properties. The observation that the D1K2-T4 cells do not express D1K2, but exhibit E-cadherin mislocalization, a mixed-lineage phenotype, and invasiveness in vivo, raises the possibility that D1K2 initiates the formation of a subpopulation of invasive cells but is not required for its maintenance. Intriguingly, the D1K2-T4 cells exhibit alterations that might mimic D1K2 function and mediate invasiveness in its absence, including p53 mutation [ 42] and decreased Rb expression [ 56]. Another potential explanation for the E-cadherin down-regulation and cytoplasmic localization is the production of TGFβ by the MMTV-D1K2 cancer cells. We have previously shown that the MMTV-D1K2 cells secrete TGFβ [ 13]. Transforming growth factor β is capable of inducing EMT associated with E-cadherin downregulation/cytoplasmic localization [ 57]. Transforming growth factor β signaling has been associated with basal-like breast cancers [ 58–60], and EMT has been specifically associated with basal-like tumors [ 19]. Chronic treatment with a TGFβ type I receptor kinase inhibitor did not convert the MMTV-D1K2 cell lines to a luminal morphology in culture (data not shown). However, autocrine TGFβ might function to induce an aberrant luminal/myoepithelial mixed-lineage differentiation status that cannot be reversed by blocking TGFβ signaling. The previously observed tumor-associated fibrosis (desmoplasia) in the MMTV-D1K2 tumors [ 13] is recapitulated in the tumors generated from the MMTV-D1K2 tumor-derived cell lines (). This suggests that the desmoplasia is caused by the cancer cells themselves and rules out possible effects caused by transgene expression in the stroma. Desmoplasia initiated by the MMTV-D1K2 cancer cells could result from TGFα production because TGFα is well known to induce fibrosis and desmoplasia [ 61,62]. In summary, the evidence presented here suggests that constitutively active Cdk2 in the form of a cyclin D1-Cdk2 fusion protein induces tumors that contain an invasive component that exhibits multiple features in common with human basal-like tumors and tumor-derived cell lines. Current efforts are focused on understanding the respective roles of Cdk2 hyperactivation, genetic instability, and TGFβ production in the formation of the invasive basal-like cancer cells in the MMTV-D1K2 tumors. It is hoped that these studies will yield insights into the mechanisms responsible for the invasiveness of human breast tumors. The authors thank Harold L. Moses' laboratory (Vanderbilt-Ingram Cancer Center, Nashville, TN) for supplying the neuT tumor cell line. The authors thank Scott McClung of the University of Florida ICBR Proteomics core facility for performing proteomic analyses of cancer cell line extracts. The technical assistance of Kirsten Jensen and Lise-lotte Thyme in performing immunohistochemical staining is greatly appreciated. | | Cdk | cyclin-dependent kinase | | | CK | cytokeratin | | | CK14 | cytokeratin 14 | | | DAPI | 4′,6-diamidino-2-phenylindole | | | D1K2 | cyclin D1-Cdk2 fusion protein | | | EGFR | epidermal growth factor receptor | | | EMT | epithelial-to-mesenchymal transition | | | H&E | hematoxylin and eosin | | | MMTV | mouse mammary tumor virus | | | αSMA or SMA | alpha smooth muscle actin | | | TGFβ | transforming growth factor beta |
1. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA. 2001;98:10869–10874. [PubMed] 2. Rakha EA, El-Rehim DA, Paish C, Green AR, Lee AH, Robertson JF, Blamey RW, Macmillan D, Ellis IO. Basal phenotype identifies a poor prognostic subgroup of breast cancer of clinical importance. Eur J Cancer. 2006;42:3149–3156. [PubMed] 3. Carey LA, Perou CM, Livasy CA, Dressler LG, Cowan D, Conway K, Karaca G, Troester MA, Tse CK, Edmiston S, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA. 2006;295:2492–2502. [PubMed] 4. Li H, Cherukuri P, Li N, Cowling V, Spinella M, Cole M, Godwin AK, Wells W, DiRenzo J. Nestin is expressed in the basal/myoepithelial layer of the mammary gland and is a selective marker of basal epithelial breast tumors. Cancer Res. 2007;67:501–510. [PubMed] 5. Abd El-Rehim DM, Pinder SE, Paish CE, Bell J, Blamey RW, Robertson JF, Nicholson RI, Ellis IO. Expression of luminal and basal cytokeratins in human breast carcinoma. J Pathol. 2004;203:661–671. [PubMed] 6. Dabbs DJ, Chivukula M, Carter G, Bhargava R. Basal phenotype of ductal carcinoma in situ: recognition and immunohistologic profile. Mod Pathol. 2006;19:1506–1511. [PubMed] 7. McCarthy A, Savage K, Gabriel A, Naceur C, Reis-Filho JS, Ashworth A. A mouse model of basal-like breast carcinoma with metaplastic elements. J Pathol. 2007;211:389–398. [PubMed] 8. Weaver Z, Montagna C, Xu X, Howard T, Gadina M, Brodie SG, Deng CX, Ried T. Mammary tumors in mice conditionally mutant for Brca1 exhibit gross genomic instability and centrosome amplification yet display a recurring distribution of genomic imbalances that is similar to human breast cancer. Oncogene. 2002;21:5097–5107. [PubMed] 9. Poole AJ, Li Y, Kim Y, Lin SC, Lee WH, Lee EY. Prevention of Brca1-mediated mammary tumorigenesis in mice by a progesterone antagonist. Science. 2006;314:1467–1470. [PubMed] 10. Brodie SG, Xu X, Qiao W, Li WM, Cao L, Deng CX. Multiple genetic changes are associated with mammary tumorigenesis in Brca1 conditional knockout mice. Oncogene. 2001;20:7514–7523. [PubMed] 11. Gauthier ML, Berman HK, Miller C, Kozakeiwicz K, Chew K, Moore D, Rabban J, Chen YY, Kerlikowske K, Tlsty TD. Abrogated response to cellular stress identifies DCIS associated with subsequent tumor events and defines basal-like breast tumors. Cancer Cell. 2007;12:479–491. [PubMed] 12. Chytil A, Waltner-Law M, West R, Friedman D, Aakre M, Barker D, Law B. Construction of a cyclin D1-Cdk2 fusion protein to model the biological functions of cyclin D1-Cdk2 complexes. J Biol Chem. 2004;279:47688–47698. [PubMed] 13. Corsino P, Davis B, Law M, Chytil A, Forrester E, Norgaard P, Teoh N, Law B. Tumors initiated by constitutive Cdk2 activation exhibit transforming growth factor beta resistance and acquire paracrine mitogenic stimulation during progression. Cancer Res. 2007;67:3135–3144. [PubMed] 14. Sarrio D, Perez-Mies B, Hardisson D, Moreno-Bueno G, Suarez A, Cano A, Martin-Perez J, Gamallo C, Palacios J. Cytoplasmic localization of p120ctn and E-cadherin loss characterize lobular breast carcinoma from preinvasive to metastatic lesions. Oncogene. 2004;23:3272–3283. [PubMed] 15. Livasy CA, Karaca G, Nanda R, Tretiakova MS, Olopade OI, Moore DT, Perou CM. Phenotypic evaluation of the basal-like subtype of invasive breast carcinoma. Mod Pathol. 2006;19:264–271. [PubMed] 16. Kim MJ, Ro JY, Ahn SH, Kim HH, Kim SB, Gong G. Clinicopathologic significance of the basal-like subtype of breast cancer: a comparison with hormone receptor and Her2/neu-overexpressing phenotypes. Hum Pathol. 2006;37:1217–1226. [PubMed] 17. Laakso M, Loman N, Borg A, Isola J. Cytokeratin 5/14-positive breast cancer: true basal phenotype confined to BRCA1 tumors. Mod Pathol. 2005;18:1321–1328. [PubMed] 18. Maeda M, Johnson KR, Wheelock MJ. Cadherin switching: essential for behavioral but not morphological changes during an epithelium-tomesenchyme transition. J Cell Sci. 2005;118:873–887. [PubMed] 19. Sarrio D, Rodriguez-Pinilla SM, Hardisson D, Cano A, Moreno-Bueno G, Palacios J. Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer Res. 2008;68:989–997. [PubMed] 20. Warburton MJ, Ormerod EJ, Monaghan P, Ferns S, Rudland PS. Characterization of a myoepithelial cell line derived from a neonatal rat mammary gland. J Cell Biol. 1981;91:827–836. [PubMed] 21. Zavizion B, Politis I, Gorewit RC. Bovine mammary myoepithelial cells: 2. Interactions with epithelial cells in vitro. J Dairy Sci. 1992;75:3381–3393. [PubMed] 22. Zavizion B, Politis I, Gorewit RC. Bovine mammary myoepithelial cells: 1. Isolation, culture, and characterization. J Dairy Sci. 1992;75:3367–3380. [PubMed] 23. Zavizion B, van Duffelen M, Schaeffer W, Politis I. Establishment and characterization of a bovine mammary myoepithelial cell line. In Vitro Cell Dev Biol Anim. 1996;32:149–158. [PubMed] 24. Neve RM, Chin K, Fridlyand J, Yeh J, Baehner FL, Fevr T, Clark L, Bayani N, Coppe JP, Tong F, et al. A collection of breast cancer cell lines for the study of functionally distinct cancer subtypes. Cancer Cell. 2006;10:515–527. [PubMed] 25. Katayose D, Gudas J, Nguyen H, Srivastava S, Cowan KH, Seth P. Cytotoxic effects of adenovirus-mediated wild-type p53 protein expression in normal and tumor mammary epithelial cells. Clin Cancer Res. 1995;1:889–897. [PubMed] 26. Wang NP, To H, Lee WH, Lee EY. Tumor suppressor activity of RB and p53 genes in human breast carcinoma cells. Oncogene. 1993;8:279–288. [PubMed] 27. Casey G, Lo-Hsueh M, Lopez ME, Vogelstein B, Stanbridge EJ. Growth suppression of human breast cancer cells by the introduction of a wild-type p53 gene. Oncogene. 1991;6:1791–1797. [PubMed] 28. Schafer JM, Lee ES, O'Regan RM, Yao K, Jordan VC. Rapid development of tamoxifen-stimulated mutant p53 breast tumors (T47D) in athymic mice. Clin Cancer Res. 2000;6:4373–4380. [PubMed] 29. Crawford AW, Michelsen JW, Beckerle MC. An interaction between zyxin and alpha-actinin. J Cell Biol. 1992;116:1381–1393. [PubMed] 30. Maruyama K, Ochiai A, Nakamura S, Baba S, Hirohashi S. Dysfunction of E-cadherin-catenin system in invasion and metastasis of colorectal cancer. Nippon Geka Gakkai Zasshi. 1998;99:402–408. [PubMed] 31. Nollet F, Berx G, van Roy F. The role of the E-cadherin/catenin adhesion complex in the development and progression of cancer. Mol Cell Biol Res Commun. 1999;2:77–85. [PubMed] 32. Wong AS, Gumbiner BM. Adhesion-independent mechanism for suppression of tumor cell invasion by E-cadherin. J Cell Biol. 2003;161:1191–1203. [PubMed] 33. Bellovin DI, Bates RC, Muzikansky A, Rimm DL, Mercurio AM. Altered localization of p120 catenin during epithelial to mesenchymal transition of colon carcinoma is prognostic for aggressive disease. Cancer Res. 2005;65:10938–10945. [PubMed] 34. Shibata T, Kokubu A, Sekine S, Kanai Y, Hirohashi S. Cytoplasmic p120ctn regulates the invasive phenotypes of E-cadherin-deficient breast cancer. Am J Pathol. 2004;164:2269–2278. [PubMed] 35. Buckley MF, Sweeney KJ, Hamilton JA, Sini RL, Manning DL, Nicholson RI, deFazio A, Watts CK, Musgrove EA, Sutherland RL. Expression and amplification of cyclin genes in human breast cancer. Oncogene. 1993;8:2127–2133. [PubMed] 36. Sutherland RL, Musgrove EA. Cyclin D1 and mammary carcinoma: new insights from transgenic mouse models. Breast Cancer Res. 2002;4:14–17. [PubMed] 37. Foulkes WD, Brunet JS, Stefansson IM, Straume O, Chappuis PO, Begin LR, Hamel N, Goffin JR, Wong N, Trudel M, et al. The prognostic implication of the basal-like (cyclin E high/p27 low/p53+/glomeruloid-microvascularproliferation+) phenotype of BRCA1-related breast cancer. Cancer Res. 2004;64:830–835. [PubMed] 38. Loden M, Stighall M, Nielsen NH, Roos G, Emdin SO, Ostlund H, Landberg G. The cyclin D1 high and cyclin E high subgroups of breast cancer: separate pathways in tumorogenesis based on pattern of genetic aberrations and inactivation of the pRb node. Oncogene. 2002;21:4680–4690. [PubMed] 39. Brodie SG, Deng CX. BRCA1-associated tumorigenesis: what have we learned from knockout mice? Trends Genet. 2001;17:S18–S22. [PubMed] 40. Hayami R, Sato K, Wu W, Nishikawa T, Hiroi J, Ohtani-Kaneko R, Fukuda M, Ohta T. Down-regulation of BRCA1-BARD1 ubiquitin ligase by CDK2. Cancer Res. 2005;65:6–10. [PubMed] 41. Shakya R, Szabolcs M, McCarthy E, Ospina E, Basso K, Nandula S, Murty V, Baer R, Ludwig T. The basal-like mammary carcinomas induced by Brca1 or Bard1 inactivation implicate the BRCA1/BARD1 heterodimer in tumor suppression. Proc Natl Acad Sci USA. 2008;105:7040–7045. [PubMed] 42. Junk DJ, Vrba L, Watts GS, Oshiro MM, Martinez JD, Futscher BW. Different mutant/wild-type p53 combinations cause a spectrum of increased invasive potential in nonmalignant immortalized human mammary epithelial cells. Neoplasia. 2008;10:450–461. [PubMed] 43. Willenbring H, Sharma AD, Vogel A, Lee AY, Rothfuss A, Wang Z, Finegold M, Grompe M. Loss of p21 permits carcinogenesis from chronically damaged liver and kidney epithelial cells despite unchecked apoptosis. Cancer Cell. 2008;14:59–67. [PubMed] 44. Sutherland RL, Musgrove EA. Cyclins and breast cancer. J Mammary Gland Biol Neoplasia. 2004;9:95–104. [PubMed] 45. Wang TC, Cardiff RD, Zukerberg L, Lees E, Arnold A, Schmidt EV. Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature. 1994;369:669–671. [PubMed] 46. Kong G, Chua SS, Yijun Y, Kittrell F, Moraes RC, Medina D, Said TK. Functional analysis of cyclin D2 and p27(Kip1) in cyclin D2 transgenic mouse mammary gland during development. Oncogene. 2002;21:7214–7225. [PubMed] 47. Pirkmaier A, Dow R, Ganiatsas S, Waring P, Warren K, Thompson A, Hendley J, Germain D. Alternative mammary oncogenic pathways are induced by D-type cyclins; MMTV-cyclin D3 transgenic mice develop squamous cell carcinoma. Oncogene. 2003;22:4425–4433. [PubMed] 48. Laakso M, Tanner M, Nilsson J, Wiklund T, Erikstein B, Kellokumpu-Lehtinen P, Malmstrom P, Wilking N, Bergh J, Isola J. Basoluminal carcinoma: a new biologically and prognostically distinct entity between basal and luminal breast cancer. Clin Cancer Res. 2006;12:4185–4191. [PubMed] 49. Asano-Miyoshi M, Hamamichi R, Emori Y. Cytokeratin 14 is expressed in immature cells in rat taste buds. J Mol Histol. 2008;39:193–199. [PubMed] 50. Wu PC, Lai VC, Fang JW, Gerber MA, Lai CL, Lau JY. Hepatocellular carcinoma expressing both hepatocellular and biliary markers also expresses cytokeratin 14, a marker of bipotential progenitor cells. J Hepatol. 1999;31:965–966. [PubMed] 51. Li L, Mignone J, Yang M, Matic M, Penman S, Enikolopov G, Hoffman RM. Nestin expression in hair follicle sheath progenitor cells. Proc Natl Acad Sci USA. 2003;100:9958–9961. [PubMed] 52. Carriere C, Seeley ES, Goetze T, Longnecker DS, Korc M. The nestin progenitor lineage is the compartment of origin for pancreatic intraepithelial neoplasia. Proc Natl Acad Sci USA. 2007;104:4437–4442. [PubMed] 53. Wiese C, Rolletschek A, Kania G, Blyszczuk P, Tarasov KV, Tarasova Y, Wersto RP, Boheler KR, Wobus AM. Nestin expression—a property of multilineage progenitor cells? Cell Mol Life Sci. 2004;61:2510–2522. [PubMed] 54. Kleeberger W, Bova GS, Nielsen ME, Herawi M, Chuang AY, Epstein JI, Berman DM. Roles for the stem cell associated intermediate filament nestin in prostate cancer migration and metastasis. Cancer Res. 2007;67:9199–9206. [PubMed] 55. Sy SM, Lai PB, Pang E, Wong NL, To KF, Johnson PJ, Wong N. Novel identification of zyxin upregulations in the motile phenotype of hepatocellular carcinoma. Mod Pathol. 2006;19:1108–1116. [PubMed] 56. Arima Y, Inoue Y, Shibata T, Hayashi H, Nagano O, Saya H, Taya Y. Rb depletion results in deregulation of E-cadherin and induction of cellular phenotypic changes that are characteristic of the epithelial-to-mesenchymal transition. Cancer Res. 2008;68:5104–5112. [PubMed] 57. Xie L, Law BK, Chytil AM, Brown KA, Aakre ME, Moses HL. Activation of the Erk pathway is required for TGF-beta1-induced EMT in vitro. Neoplasia. 2004;6:603–610. [PubMed] 58. Shipitsin M, Campbell LL, Argani P, Weremowicz S, Bloushtain-Qimron N, Yao J, Nikolskaya T, Serebryiskaya T, Beroukhim R, Hu M, et al. Molecular definition of breast tumor heterogeneity. Cancer Cell. 2007;11:259–273. [PubMed] 59. Sheridan C, Kishimoto H, Fuchs RK, Mehrotra S, Bhat-Nakshatri P, Turner CH, Goulet R, Jr, Badve S, Nakshatri H. CD44+/CD24- breast cancer cells exhibit enhanced invasive properties: an early step necessary for metastasis. Breast Cancer Res. 2006;8:R59. [PubMed] 60. Honeth G, Bendahl PO, Ringner M, Saal LH, Gruvberger-Saal SK, Lovgren K, Grabau D, Ferno M, Borg A, Hegardt C. The CD44+/CD24- phenotype is enriched in basal-like breast tumors. Breast Cancer Res. 2008;10:R53. [PubMed] 61. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, Wakefield LM, Heine UI, Liotta LA, Falanga V, Kehrl JH, et al. Transforming growth factor type beta: rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Natl Acad Sci USA. 1986;83:4167–4171. [PubMed] 62. Lohr M, Schmidt C, Ringel J, Kluth M, Muller P, Nizze H, Jesnowski R. Transforming growth factor-beta1 induces desmoplasia in an experimental model of human pancreatic carcinoma. Cancer Res. 2001;61:550–555. [PubMed]
|
This article has been 
