Tumor cell invasion into the surrounding stroma requires increased cell motility and extensive remodeling of the extracellular matrix. Endo180 (CD280, MRC2, urokinase-type plasminogen activator receptor-associated protein) is a recycling endocytic receptor that functions in both these cellular activities by promoting cell migration and uptake of collagens for intracellular degradation. In the normal breast, Endo180 is predominantly expressed by stromal fibroblasts. The contrary observation that Endo180 is expressed on epithelial tumor cell lines that display a high invasive capacity suggested that up-regulation of this receptor may be an associated and functional component in the acquisition of a more aggressive phenotype by tumor cells in vivo. Here, we show that high levels of Endo180 are found in a subset of basal-like breast cancers and that this expression is an independent prognostic marker for shorter disease-free survival. Two potential mechanisms for Endo180 up-regulation were uncovered. First, it was shown that Endo180 can be transcriptionally up-regulated in vitro following transforming growth factor-β treatment of breast cancer cells. Second, a proportion of Endo180+ tumors were shown to have Endo180 gene copy number gains and amplifications. To investigate the functional consequence of Endo180 up-regulation, MCF7 cells transfected with Endo180 were inoculated into immunocompromised mice. Expression of wild-type Endo180, but not an internalization-defective Endo180 mutant, resulted in enhanced tumor growth together with a reduction in tumor collagen content. Together, these data argue that elevated expression of this receptor in tumor cells could have important consequences in subsets of basal-like carcinomas for which there is a current lack of effective treatment. [Cancer Res 2007;67(21):10230–11]
- breast cancer
- metaplastic breast carcinomas
In the multistep model of cancer progression, the acquisition of invasive properties by a subset of cells within the tumor represents an essential prerequisite for metastasis to occur. A key characteristic of such invasive tumor cells is their transformation to a mesenchymal-like morphology that is accompanied by an up-regulation of proteins involved in remodeling of the extracellular matrix and the promotion of cell migration. Endo180 (CD280, urokinase-type plasminogen activator receptor-associated protein, MRC2) is a constitutively recycling endocytic receptor ( 1, 2) that is directly linked to both of these processes. First, Endo180 is a novel collagen receptor mediating binding and uptake of ligand for delivery to intracellular degradative organelles ( 3– 9). Second, overexpression of Endo180 is promigratory, whereas its genetic ablation or silencing impairs the migratory process ( 4, 5, 10), and it has recently been shown that Endo180 signals from endosomal compartments to generate spatially localized Rho-ROCK-MLC2–based contractility at sites of adhesion disassembly ( 11).
In normal adult tissue, Endo180 expression is predominantly restricted to stromal fibroblasts ( 1, 7, 12). In vitro, high levels of expression (∼1 × 106 receptors per cell) are found in primary human ( 1, 13) and mouse embryonic fibroblasts ( 4, 5) as well as immortalized fibroblast cell lines ( 1). By contrast, little or no expression of Endo180 is found on a range of epithelioid cell lines or on epithelial cells in vivo ( 1, 7, 12). Together, these data have led to the proposal that Endo180 functions in mesenchymal cells as a promigratory receptor with a physiologic role in the uptake and clearance of extracellular collagens. This proposal is supported by recent in vivo studies in which mice with a targeted deletion in the Endo180 gene, crossed with tumor-prone animals, resulted in a decreased tumor burden that was attributed to the failure of surrounding stromal cells to clear collagens from the invasive front ( 7). Although the majority of studies to date have focused on the physiologic and pathologic role of Endo180 in stromal cells, we have noted that high levels of Endo180 receptor expression are found on a wide range of the more invasive epithelial tumor cell lines. In light of this, we have investigated whether this in vitro expression might reflect an up-regulation of Endo180 in tumor cells in vivo and the role that such expression could play in promoting tumor progression.
Materials and Methods
Antibodies and cells. The anti-human Endo180 monoclonal antibody (mAb) A5/158 was generated by immunizing mice with AG1523 human fibroblasts and has been described elsewhere ( 1, 13). Where indicated, mAb A5/158 directly conjugated to Alexa Fluor 555 (Molecular Probes) according to the manufacturer's instructions. To generate anti-Endo180 mAb 39.10, full-length purified human Endo180 ( 1) was used to immunize a BALB/c mouse and hybridomas were screened for their ability to immunoprecipitate Endo180 from AG1523 human fibroblasts. Specificity of the antibody was confirmed by Western blot analysis of Endo180− cell lines transfected with vector alone or wild-type (WT) human Endo180 and Western blot analysis of purified Endo180-Fc constructs. The following collagen antibodies were purchased from DPC Biermann: rabbit anti-collagen I, anti-collagen IV, and anti-collagen V. Unconjugated E-cadherin antibodies HECD-1 (mouse IgG1) and clone 36 (mouse IgG2A) were purchased from Abcam and BD Biosciences, respectively. Smooth muscle actin antibody was purchased from Sigma-Aldrich (1:5,000). The pan-cytokeratin antibody was purchased from Dako. Anti-tubulin antibody was purchased from Sigma-Aldrich. Alexa Fluor 488–conjugated and Alexa Fluor 555–conjugated secondary antibodies and Oregon Green 488 gelatin were purchased from Molecular Probes. MCF7 cells expressing WT Endo180 and Endo180(Ala1468/Ala1469) or transfected with vector alone have been previously described ( 3). Briefly, transfected cells were cultured in DMEM with 10% FCS, 10 μg/mL insulin, and 0.5 mg/mL G418. Populations of cells expressed with WT Endo180 or Endo180(Ala1468/Ala1469) were selected by fluorescent-activated cell sorting and maintained in the same culture medium. For growth factor treatment of MCF7 cells, cells were cultured in 5% FCS for 24 h before incubation with the following factors for 48 to 72 h: fibroblast growth factor-1 (FGF-1; 5 ng/mL), hepatocyte growth factor (HGF; 5 ng/mL), insulin-like growth factor-II (IGF-II; 50 ng/mL), platelet-derived growth factor-BB (PDGF-BB; 50 ng/mL), and transforming growth factor-β1 (TGF-β1; 5 ng/mL). All factors were purchased from R&D Systems.
Real-time quantitative PCR. Cells were lysed in Trizol (Invitrogen) and stored at −20°C, and RNA was extracted according to the manufacturer's instructions. DNase digestion was done using RNase-Free DNase Set (Qiagen) and the RNA was cleaned up using RNeasy MinElute kit (Qiagen). cDNA synthesis was carried out using Omniscript RT kit (Qiagen) and RNaseOUT inhibitor (Invitrogen). Up to 75 ng of RNA per reaction were transcribed into cDNA using an oligo(dT)n primer (Qiagen). cDNA (1 μL) was used per quantitative real-time RT-PCR. Each analysis reaction was done in triplicate. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous control throughout all experimental analysis. Gene expression analysis was done using Taqman Gene Expression Assays on an ABI Prism 7900HT sequence detection system (Applied Biosystems). Analysis was done using the ΔΔCt method, which determines fold changes in gene expression relative to a comparator sample (untreated MCF7 cells). Significant deviation of the mean value of each sample group from a fold difference of 1 (no change compared with the comparator sample) was tested using a t test on log10 transformed data. Quantitative real-time RT-PCR primers were purchased from Applied Biosystems. The Assays-on-Demand references for Endo180 and GAPDH are Hs00195862_m1 and 4310884E, respectively.
Immunofluorescent staining of tumors. Collection of fresh breast tumor material and normal breast tissue from reduction mammoplasties was approved by the Research Ethics Committee of the Royal Marsden NHS Trust. Tumors were graded according to a modified Bloom-Richardson system ( 14) and classified according to internationally accepted criteria ( 15). All tumors were negative for estrogen (ER), progesterone (PgR), and HER2 receptors, six were classified as grade 3 invasive ductal carcinoma (IDC) of no special type, and two were classified as metaplastic breast carcinomas. All cases were reviewed by J.S. Reis-Filho to corroborate the diagnosis. Cryosections (7 μm) were cut from normal adult human breast, human breast tumors, or MCF7 xenograft tumors and stored at −80°C. Sections were thawed, fixed in 4% paraformaldehyde for 20 min, and then stained as follows. For human breast, sections were stained with anti-Endo180 mAb A5/158 and pan-cytokeratin antibody (20 μg/mL) followed by Alexa Fluor 488 anti-mouse Ig and Alexa Fluor 555 anti-rabbit Ig. For tumor xenografts, sections were stained with Alexa Fluor 555–conjugated anti-Endo180 mAb A5/158 followed by anti-collagen I antibody and Alexa Fluor 488 anti-rabbit Ig. Nuclei were counterstained with TO-PRO-3 (Molecular Probes) and sections were mounted in Vectashield H-1000 (Vector Laboratories, Inc.). Images were collected sequentially in three channels on a Leica TCS SP2 confocal microscope.
Tissue microarrays. The tissue microarrays (TMA) containing 245 and 880 invasive breast carcinomas were generated in the Breakthrough Breast Cancer Research Centre and in the University of Nottingham, respectively. Full details of the TMA characterization and the cohort of patients are described in the Supplementary Methods and elsewhere ( 16– 20). For Endo180 staining, the TMAs were dewaxed in xylene, taken through ethanol (99.7–100%, v/v), and subjected to high-temperature antigen retrieval (18 min of microwaving) in target retrieval citrate buffer (pH 6; Dako, DakoCytomation). Slides were allowed to cool for 20 min at room temperature and then incubated with anti-Endo180 mAb 39.10 (40 μg/mL) for 30 min at room temperature. Detection was achieved with the Dako Envision/HRP system. Positive and negative controls were included in each slide run.
Statistical analysis was done using Statistical Package for the Social Sciences 13.0 statistical software (SPSS, Inc.). Median follow-up was defined as the median follow-up for those patients still alive and disease-free at the latest hospital visit. Previously validated cutoff values for the different biomarkers included in this study were chosen a priori ( 16– 20). Standard cutoff values were used for established prognostic factors and were the same as for previously published patient series ( 17). All factors were used as dichotomous covariates in the statistical analysis with the exception of grade, Nottingham prognostic index (NPI), and the classification into Nielsen groups ( 21), which were categorized into three groups. These studies were approved by the Research Ethics Committee of the Royal Marsden NHS Trust under the title “Breast cancer invasion and metastasis” and by the Nottingham Research Ethics Committee 2 under the title “Development of a molecular genetic classification of breast cancer”.
The associations between the Endo180 expression and clinicopathologic variables were evaluated by the χ2 test. Confidence intervals of 95% were adopted. A two-sided P value of <0.05 was considered statistically significant. Survival curves were calculated by the Kaplan-Meier method. Differences in survival based on Endo180 expression were estimated using the log-rank test. Multivariate Cox regression analysis was used to evaluate any independent prognostic effect of the variable on disease-free survival and the overall survival, which was adjusted by such well-known prognostic factors as tumor grade, lymph node stage, tumor size, and ER status.
Fluorescence in situ hybridization. Dual-color fluorescence in situ hybridization (FISH) was done on 2-μm-thick tissue sections mounted on polylysine-coated slides using a biotin-labeled centromeric probe for chromosome 17 (CEP17; Spot-Light Chr17 centromeric probe, Zymed) and a digoxigenin-labeled in-house–generated probe for human Endo180 gene (MRC2) as previously described ( 22). Full details are provided in the Supplementary Methods. FISH scoring was done by counting fluorescence signals in at least 50 morphologically unequivocal, invasive neoplastic cell nuclei for each case. Overlapping nuclei were not scored. For each specimen, the ratio of MRC2 signals to CEP17 signals was calculated. Conservative thresholds to define MRC2 gene copy number gains and amplifications were used. A tumor was considered to harbor MRC2 copy number gains when MRC2 to CEP17 ratios were >1.5 but ≤2 and amplified when MRC2 to CEP17 ratios were >2.0.
Tumor xenografts. Ninety-day release estradiol pellets (0.36 mg; Innovative Research of America) were s.c. implanted on the back of the neck of 6-week athymic female mice (Ncr-nude). Five days later, 3 × 106 cells suspended in 100 μL sterile PBS were s.c. injected on the right flanks. In each experiment, 10 mice were inoculated per cell line. Tumors were measured weekly with calipers and allowed to reach a maximum diameter of no greater than 12 mm. Tumors were excised from the mice at necropsy and cryopreserved for histology and collagen quantification. Mice were allowed food and water ad libitum. All procedures were done in accordance with UK Home Office legislation.
Collagen quantification assay. Full details of the collagen quantification are provided in the Supplementary Methods. In brief, collagen quantification assays were carried out as previously described ( 23) and collagen content was calculated by comparison with a standard curve of different hydroxyproline concentrations using the conversion factor of 1 mg hydroxyproline corresponding to 6.94 mg collagen.
Endo180 expression can be transcriptionally up-regulated in cultured epithelial cells. Although Endo180 expression is predominantly found in vivo and in vitro on cells of mesenchymal origin, it has previously been reported that the breast cancer cell line MDA-MB-231 expresses high levels of Endo180. Moreover, down-regulation of Endo180 expression in these cells results in an impairment in cell migration ( 3, 10) and collagen binding (data not shown). Here, a more detailed examination of a series of colorectal and breast cancer cell lines revealed several lines of epithelial origin with Endo180 expression (MDA-MB-231, MDA-MB-468, Cal51, and BE) and this expression correlated with a more invasive and/or migratory phenotype (Supplementary Table S1). In contrast, the cell lines identified as Endo180−, such as MCF7, T47D, and MCF10A, are characterized as having a largely epithelioid morphology with strong intercellular junctions.
The acquisition of mesenchymal properties by epithelial cells is a well-known phenomenon in developmental biology ( 24), and recently, several studies have sought to investigate similar events in human cancers ( 25– 27). A major factor that has many contributory roles during this process is TGF-β ( 28). Following treatment of MCF7 cells with TGF-β1, immunohistochemical analysis revealed a subpopulation of cells with up-regulated Endo180 expression ( Fig. 1A ) and a corresponding increase in total Endo180 protein was also detected by Western blotting (Supplementary Fig. S1). In contrast, little or no effect on Endo180 expression was observed when cells were treated with HGF, PDGF, IGF-II, or FGF. In parallel cultures, TGF-β1 treatment, but not treatment with other growth factors, resulted in a significant increase in Endo180 mRNA levels as detected by quantitative real-time RT-PCR ( Fig. 1B). Endo180 expressed ectopically in MCF7 cells can bind and internalize exogenously added collagen and gelatin ( 3). To confirm that the endogenous Endo180 expression up-regulated by TGF-β1 in MCF7 cells yields a fully functional collagen binding/internalization receptor, these cells were incubated with Oregon Green 488–labeled gelatin. In the TGF-β1–treated cells, gelatin binding was exclusively restricted to the subpopulation of cells that expressed Endo180 ( Fig. 1C). Moreover, in these cells, the gelatin was clearly present in intracellular compartments, confirming that Endo180 expression in these cells confers the ability to internalize and process collagens. Cells that had been treated with other growth factors and failed to up-regulate Endo180 showed, as expected, no gelatin binding or internalization (data not shown).
Endo180 is expressed by invasive human breast tumors. The up-regulation of Endo180 in TGF-β1–treated MCF7 cells suggested that expression of this receptor might be a feature of tumor cells in vivo that had acquired a more aggressive phenotype. Consequently, we examined the expression of Endo180 in a series of invasive breast carcinomas and normal breast tissue. As shown in Fig. 2A , the epithelial structures of the breast contain an inner luminal layer of polarized cells and an underlying myoepithelial or basal layer in contact with basement membrane. Staining of normal breast tissue with the anti-Endo180 mAb A5/158 revealed strong expression in the majority of intralobular and interlobular fibroblasts and a lower level expression on many myoepithelial cells. Endo180 expression was not detected in luminal epithelial cells ( Fig. 2A; Supplementary Fig. S2). In a previous study using paraffin wax–embedded material, it has been reported that, in normal breast tissue, Endo180 expression is limited and only detected in occasional intralobular fibroblast-like cells and few or no myoepithelial cells ( 12). To investigate this discrepancy, cryosections of normal human breast from two different donors were stained with two independent anti-Endo180 mAbs. In all cases, a similar staining pattern was observed ( Fig. 2B; Supplementary Fig. S2). In addition, staining of paraffin wax–embedded normal breast again revealed staining of stromal fibroblasts and myoepithelial cells ( Fig. 2C). This suggests that the lower level of staining previously reported results from the sample fixation conditions, antigen retrieval methods, and/or the lower affinity of the polyclonal antibody used. It is of note that strong Endo180 expression by fibroblastic cells has been reported in the mouse mammary gland ( 7). Finally, we specifically addressed whether Endo180 was expressed on the angiogenic vasculature as a serial analysis of gene expression screen had identified Endo180 transcripts as being up-regulated in the angiogenic endothelium of colorectal tumors ( 29). No Endo180 expression was detected in the tumor vasculature of any of the breast (Supplementary Fig. S3) or colorectal (data not shown) tumors examined.
Given our observation that Endo180 is expressed in normal myoepithelial cells, we selected a series of high-grade IDCs that were negative for ER, PgR, and HER2 because it is well established that such tumors have a transcriptome that is more related to myoepithelial/basal cells than to luminal epithelial cells ( 21, 30– 34). Tumor cryosections were stained with the anti-Endo180 mAb A5/158 and a pan-cytokeratin antibody, which recognizes all breast epithelial cells ( Fig. 2B). In four of eight hormone receptor–negative tumors examined, a staining pattern similar to that previously reported ( 12) was observed with Endo180 expression restricted to stromal fibroblasts and myoepithelial cells localized to the edge of ductal carcinoma in situ (DCIS) lesions. However, the remaining four tumors all showed high levels of Endo180 expression in cytokeratin+ cells and these tumors were all characterized pathologically as metaplastic breast carcinomas or as IDCs with focal metaplastic elements. To show that this expression of Endo180 in tumor cells was not due to nonspecific reactivity of the A5/158 anti-Endo180 mAb used, parallel sections were stained with the independently derived anti-Endo180 mAb 39.10. As shown in Supplementary Fig. S4, these two mAbs gave an identical staining pattern in both cytokeratin+ tumor cells and stromal fibroblasts.
To extend this analysis, two independent TMAs of invasive breast carcinomas were stained for Endo180 expression. In the Breakthrough Breast Cancer and Nottingham TMAs, Endo180 expression was detected in 3.7% and 5.8% of tumors, respectively ( Table 1 ; Fig. 3A ). On both arrays, expression of Endo180 was significantly correlated with lack of ER expression (P = 0.0001 and 0.013), with expression of basal cytokeratins Ck5/6 (P = 0.0025 and 0.006) and Ck14 (P = 0.0022 and 0.005), and with basal-type breast cancers as defined by Nielsen et al. (P < 0.0001 and P = 0.003; ref. 21). Univariate analysis in the Nottingham cohort revealed that strong expression of Endo180 was associated with a shorter disease-free survival ( Fig. 3B). On multivariate analysis, Endo180 emerged as an independent predictor of disease-free survival ( Fig. 3C).
Amplification of Endo180 in basal-like breast cancers. FISH analysis was done on the eight Endo180+ breast cancers from the Institute of Cancer Research TMA ( Fig. 3A) plus two Endo180+ samples identified by immunostaining ( Fig. 2B). Of the 10 samples analyzed, 2 showed amplification of the Endo180 (MRC2) gene (MRC2 to CEP17 ratios of >2.0) and 3 showed MRC2 gain of copy number (MRC2 to CEP17 ratios of >1.5 but ≤2.0; Fig. 4 ).
Functional analysis of Endo180 in tumor cells. It has previously been shown that expression of Endo180 in MCF7 cells results in the promotion of a migratory phenotype ( 3, 10, 11). This is accompanied by a loss of E-cadherin from cell-cell junctions that can be reversed by treatment with Endo180 small interfering RNA oligonucleotides ( 11). To address whether this functionality of Endo180 has potential relevance during tumor progression, two approaches were taken. First, cryosections of Endo180+/cytokeratin+ tumors ( Fig. 2B) were double labeled for Endo180 and E-cadherin. We observed that down-regulation of E-cadherin expression was associated with an increased in Endo180 expression. For example, in the metaplastic carcinoma 38T, a high level of Endo180 expression corresponds with a lack of E-cadherin expression, whereas in metaplastic tumor 126T there was a graded decrease in E-cadherin staining mirrored by an increase in Endo180 staining ( Fig. 5A ). This latter observation was corroborated by additional staining using alternative E-cadherin and Endo180 mAbs ( Fig. 5A).
Second, the effects of ectopic Endo180 expression on tumor growth were examined. MCF7 cells were transfected with WT Endo180, the Endo180(Ala1468/Ala1469) mutant, or vector alone and selected to generate stable expressing populations ( 3). The Endo180(Ala1468/Ala1469) mutant has a defective endocytosis motif and is therefore expressed at the cell surface but unable to undergo internalization ( 2). As a consequence, a higher proportion of Endo180(Ala1468/Ala1469) mutant is detected at the cell surface when analyzed by flow cytometry ( Fig. 5B). Western blot analysis showed equivalent levels of total receptor expression (data not shown). When inoculated into immunocompromised mice, it was consistently observed that expression of WT Endo180 resulted in enhanced tumor growth compared with the vector alone–transfected cells. In contrast, no growth advantage was observed with the MCF7 cells expressing the Endo180(Ala1468/Ala1469) mutant, which indicates that the physiologic function of Endo180 in promoting tumor growth requires normal receptor recycling ( Fig. 5B). Similar trends of lower xenograft growth rates were found with MDA-MB-231 and MDA-MB-468 cells in which Endo180 expression was down-regulated with lentiviral short hairpin RNA compared with their respective nontarget controls (Supplementary Fig. S5).
It was hypothesized that the ability of WT Endo180 to internalize collagens for delivery to intracellular degradative compartments ( 3– 9) might be important for the protumorigenic effects of Endo180 in MCF7 cell xenografts. To address this, cryosections from the tumor xenografts were stained for collagens and the anti-Endo180 mAb A5/158. mAb A5/158 is specific for human Endo180 and hence will detect transfected Endo180 in the MCF7 cells but not mouse Endo180 expressed on any murine stromal cells infiltrating the xenograft. As expected, heterogeneous staining patterns were observed both between and within individual tumors but it was consistently observed that transfected MCF7 cells retained their expression of Endo180 and that high levels of WT Endo180 expression were associated with decreased collagen I ( Fig. 5C) and collagen V (data not shown) deposition when compared with MCF7-vector alone–transfected and MCF7-Endo180(Ala1468/Ala1469) xenografts. Given the heterogeneous nature of the tumor material, the total collagen in the tumor xenograft tissue was assessed by hydroxyproline quantification. As shown in Fig. 5C, there was a significant reduction of collagen within the MCF7-WT Endo180 tumors when compared with the MCF7-vector alone tumors. Interestingly, MCF7-Endo180(Ala1468/Ala1469) tumors showed a higher total collagen level compared with the MCF7-WT Endo180 or MCF7-vector alone tumors. Due to its inability to be internalized, a higher proportion of the Endo180(Ala1468/Ala1469) mutant is localized to the cell surface ( Fig. 5B), and as a consequence when incubated at 4°C with Oregon Green 488 gelatin, cells expressing this mutant receptor show increased levels of cell surface–associated collagen compared with cells expressing WT Endo180 ( Fig. 5D). This retention of cell surface collagen by the Endo180(Ala1468/Ala1469) mutant may be the reason why enhanced extracellular collagen accumulation is observed in its associated tumor xenografts in vivo.
Expression of Endo180 in a subset of basal-like tumors. In a recent study ( 7), the internalization receptor Endo180 was implicated to play a role in tumor progression. It was shown that crossing mice with a targeted deletion in Endo180 with mice harboring polyomavirus middle T under the control of the mouse mammary tumor virus promoter resulted in a reduced tumor growth in the mammary glands compared with mice expressing WT Endo180. In this study, Endo180 expression was restricted to the stromal fibroblasts and was not detected on the tumor cells. It was therefore concluded that the protumorigenic effect of Endo180 was mediated by the tumor microenvironment. Here, we have investigated the potential involvement of Endo180 in human breast cancers. First we have shown that this receptor is strongly expressed by both the intralobular and interlobular fibroblasts of normal human breast and to a lesser extent by normal breast myoepithelial cells and myoepithelial cells surrounding foci of DCIS. Gene expression profiling has led to the classification of breast cancers into five groups, luminal A, luminal B, basal-like, HER2+, and normal breast-like ( 21, 30– 34), and importantly, these groups have prognostic and predictive implications. Basal-like tumors, which comprise ∼20% of breast cancers, are so named because their transcriptome more closely resembles myoepithelial/basal cells of the normal breast rather than luminal epithelial cells. Typically, basal-like tumors are hormone receptor negative in that they do not express ER, PgR, or HER2 ( 21, 35– 37). Although it had previously been reported that Endo180 was not expressed in the tumor cells of 13 invasive breast carcinomas examined ( 12), the subtype or hormone receptor status of these tumors was not stated. Given that Endo180 is expressed at a low level by the myoepithelial cells of the normal breast, it was reasoned that an examination of Endo180 expression should be more prevalent in tumors with myoepithelial/basal-like differentiation. We first analyzed a series of eight high-grade IDCs with triple-negative receptor status (ER−/PgR−/HER2−) and found strong Endo180 expression in four of them. The positive cancers harbored metaplastic elements, a hallmark feature of basal-like breast cancers ( 38, 39). In addition, in the breast cancer cell lines, Endo180 was preferentially expressed by MDA-MB-231 and MDA-MB-468 cells, which have been shown to have basal-like transcriptomic characteristics ( 40). Together, these data suggest that expression of Endo180 in vitro or in vivo is indicative of tumor cells that have a basal-like phenotype. We subsequently analyzed the expression of Endo180 in two large cohorts of invasive breast cancers. In both cohorts, Endo180 expression was significantly associated with lack of ER, expression of basal cytokeratins, and a basal-like phenotype. Interestingly, where tested, Endo180+ cancers were also significantly more frequently positive for epidermal growth factor receptor (EGFR) and P-cadherin, two known markers of basal-like cancers. In the larger cohort of breast cancers, Endo180 expression was shown to significantly correlate with a shorter disease-free survival and moreover by multivariate analysis to be an independent predictor of shorter disease-free survival. Although Endo180 expression showed a trend for shorter survival in the Breakthrough cohort, this did not reach statistically significant levels (P < 0.1; data not shown). This is not surprising given the low prevalence of Endo180+ tumors and indicates that larger studies will be required to further explore the prognostic significance of Endo180 expression.
Regulation of Endo180 expression. In these studies, two potential mechanisms mediating Endo180 up-regulation in tumor cells were uncovered. Given that genomic gains of chromosome 17q distal to the HER2 amplicon are preferentially found in basal-like breast cancers ( 41), we investigated whether Endo180 expression in basal-like cancers would be driven by Endo180 (MRC2) gene copy number gains/amplification. Of the 10 cases analyzed, MRC2 gene copy number gains and amplifications were found in 3 and 2 cases, respectively, indicating that in a subset of Endo180+ breast cancers these changes in gene copy number may drive Endo180 overexpression. However, as gene amplification was found in only 20% of Endo180+ breast cancers, in the majority of cases, it is likely that Endo180 expression is regulated at the transcriptional level. TGF-β is well recognized to promote the transition of epithelial cells to a more mesenchymal phenotype and hence enhance their invasive properties ( 28, 42, 43). Here, we show that, at least in cultured cancer cells, TGF-β pathway activation can lead to Endo180 up-regulation. Consistent with previous reports that MCF7 cells do not undergo a full epithelial to mesenchymal transition following TGF-β treatment ( 44), this up-regulation was only observed in a subpopulation of cells. However, in this subpopulation, Endo180 expression was associated with the ability to bind and internalize collagen and the loss of an epithelial phenotype (data not shown), features strikingly similar to those observed in Endo180-transfected MCF7 cells. Although these data show that Endo180 expression can be modulated by TGF-β signaling, further studies will be required to assess whether TGF-β promotes up-regulation of Endo180 expression in tumor cells in vivo and the interplay between TGF-β signaling and the transcriptomic program associated with the myoepithelial/basal-like differentiation of tumor cells.
Physiologic function of Endo180 in breast tumor cells. Although these studies support previous reports that the expression of Endo180 in epithelial cells in vitro is sufficient to promote a mesenchymal phenotype ( 11), it was important to assess the potential functional significance of the Endo180 expression that had been observed in breast tumor cells in vivo. Here, we show that ectopic expression of Endo180 in MCF7 cells resulted not only in increased tumor growth but also in a quantifiable decrease in intratumoral collagen content. In contrast, when mice were inoculated with MCF7 cells expressing an internalization-defective Endo180 mutant, there was no growth advantage, and instead, an increased collagen accumulation was observed. These data complement a previous study in which loss of Endo180 in the stromal fibroblasts of tumor-prone mice was shown to restrict tumor growth and increase peritumoral fibrosis ( 7). The results presented here not only directly show a role for Endo180 in collagen matrix turnover in vivo but also provide new evidence that tumor cells expressing this receptor acquire this property. Moreover, the results of this study also clarify that Endo180 internalization is necessary for its biological function in tumor pathology and further support the notion that extracellular matrix remodeling in tumor progression not only requires the action of extracellular proteases but also the uptake and intracellular degradation of collagens. The finding that Endo180 expression is up-regulated by TGF-β stimulation implies that, in addition to the basal-like tumor cells identified here, Endo180 could potentially also be transiently expressed by other tumor cells that are exposed to locally increased levels of stromally derived TGF-β. This up-regulation may serve to promote the migration and matrix degradation of tumor cells at the invasive margins. In support of this suggestion, it was notable that Endo180 expression, at least in the tumor samples analyzed, was inversely correlated with the expression of E-cadherin. Loss of E-cadherin expression or its redistribution in invasive cells serves both to support detachment of epithelial cells and to promote promigratory signaling pathways ( 45). Certainly, these observations warrant further studies to investigate whether up-regulation of endogenous Endo180 is a prerequisite for the acquisition of an invasive phenotype in vivo.
Conclusion. The most significant advances in breast cancer treatment in recent decades have been hormonal therapy to treat ER/PgR-positive tumors and Herceptin to treat HER2-overexpressing tumors. However, treatments for women with triple-negative (ER−/PgR−/HER2−) and basal-like breast cancer, which although show high rates of response to neoadjuvant chemotherapy still have a more aggressive clinical behavior ( 46), remain elusive. Here, we have shown that a subset of basal-like breast tumors shows strong up-regulation of the collagen internalization receptor Endo180, which in some cases is driven by Endo180 copy number gains/amplification, and that Endo180 promotes a protumorigenic mesenchymal phenotype. Further studies examining the prognostic and predictive implications of Endo180 expression and oncogenic roles of this receptor in breast cancers are warranted.
Grant support: Breakthrough Breast Cancer (C.M. Isacke and J.S. Reis-Filho), Association of International Cancer Research (C.M. Isacke), Breast Cancer Campaign (A.R. Green and I.O. Ellis), Cancer Research UK (C.M. Isacke), and The Wellcome Trust (C.M. Isacke).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Emad Rakha (City Hospital, Nottingham, United Kingdom) and Rehab Samaka (Histopathology Department, Menoufia University, Menoufia, Egypt) for scoring the Nottingham TMA, Michelle James for assistance in building the Breakthrough TMA, Katrina Todd for collecting the TMA images, and Niina Pirinen for characterization of the 39.10 antibody.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
D. Wienke and G.C. Davies contributed equally to this work.
Current address for D. Wienke: Oncology Research NCE, Global Preclinical R&D, Merck KGaA, Darmstadt, Germany. Current address for J. Sturge: Prostate Cancer Research Group, Department of Oncology, Division of Surgery, Oncology, Reproductive Biology and Anaesthesia, Faculty of Medicine, Imperial College London, Hammersmith Hospital Campus, Du Cane Road, London, United Kingdom.
- Received September 21, 2006.
- Revision received August 8, 2007.
- Accepted August 28, 2007.
- ©2007 American Association for Cancer Research.