In recent years, it has become increasingly accepted that the deregulation of mechanisms normally involved in developmental processes has tumorigenic effects in adult tissues. One example is the BMP family of growth factors, whose function in cancer physiology has been demonstrated in many tumor types, including breast cancer [9–11]. In spite of this, little is known about BMP target genes in the context of tumors. The transcriptional responses of breast cancer cells to BMP signaling have been studied only minimally. More precisely, the effects of BMP2 and BMP7 treatments on transcription in MCF-7 and MDA-MB-468 breast cancer cell lines, respectively, have been analyzed using cDNA microarrays of limited content (from several hundreds to 14, 500 gene probes) [28–30]. In this study, we have therefore set up an experimental procedure to identify potential BMP target genes in breast cancer by studying the effects of two BMP ligands, BMP4 and BMP7, on genome-wide gene expression. These two BMPs were selected based on their essential role in breast cancer, which we and others have demonstrated in recent years [9–11]. Both ligands are highly expressed in primary breast carcinomas as well as in breast cancer cell lines [12–14, 16, 17]. BMP7 expression was also shown to be associated with early bone metastasis . Additionally, in vitro studies have implicated BMP4 and BMP7 as important regulators of proliferation and migration of breast cancer cells [14, 18, 20, 21, 31].
Our experimental approach allowed multiple types of analyses and revealed interesting insights into how the stimulation of BMP4 and BMP7 signaling influences the transcriptome of breast cancer cell lines. First of all, we showed relatively high numbers of DEPs as a result of BMP treatments, indicating a strong impact of BMP4 and BMP7 on the cell lines studied. Moreover, the transcriptional response to BMP4 was of a clearly higher magnitude than that induced by BMP7. Another aspect of the study was the opportunity to compare the effects of BMP signaling between cell lines. Interestingly, clear differences were seen in the amounts of DEPs, as well as in their expression patterns, as revealed by hierarchical clustering. BMP signaling pathways are regulated in a very complex manner and at many different levels, from the availability of BMP receptors, BMP ligands and BMP antagonists in the extracellular compartment to the presence or absence of various intracellular signal mediators and transcriptional co-activators or co-repressors [32, 33]. Therefore, multiple factors influence the outcome of BMP signaling on the transcriptional level in a given cell. We have previously reported that all six BMP specific receptors (ACVR1, BMPR1A, BMPR1B, ACVR2A, ACVR2B, and BMPR2) are uniformly expressed among the breast cancer cell lines studied here . Similarly, we have shown that SMAD4 is expressed and that phosphorylation of SMAD-1/5/8 is induced in these cell lines after BMP7 and BMP4 treatment [14, 21]. Taken together, the expression profiles of BMP specific receptors or the mediators of the canonical intracellular pathway do not seem to have a major role in explaining the different transcriptional responses in the breast cancer cells. Nevertheless, due to the complexity of BMP signaling regulation, it is easy to understand that different cell lines may have different transcriptional responses to BMP stimulation. This observation highlights the importance of testing multiple cell lines when studying BMP signaling in cancer. An additional finding that could be inferred from our data is that induction of gene transcription, compared with inhibition, was the common response among those genes most frequently regulated by BMP4 and BMP7 in breast cancer cells. Likewise, previous microarray-based transcriptomic analyses of TGF-β and BMP have shown that induction of gene expression is the predominant response of mammalian cells to stimulation by these growth factors [28, 30, 33–35].
After BMP4 and BMP7 stimulation, the microarray analyses identified a large number of differentially expressed genes in our panel of cell lines. To explore the biological functions of these genes, GO enrichment analyses were performed. These revealed very similar results for both BMP ligands, namely, regulation of transcription and developmental processes. It seems, therefore, that the functions most prevalently influenced by BMP signaling in breast cancer cells do not differ remarkably from conventional roles that BMPs possess during development [36, 37].
Synexpression groups are synchronously coexpressed gene sets, particularly apparent during embryonic development and in the response of cells to hormones and growth factors [38, 39]. Our analyses unveiled that treatment of breast cancer cells with either BMP4 or BMP7 resulted in the coordinated expression of a group of genes (clusters A and B, respectively). Most interestingly, a considerable number of the genes in these two synexpression groups were common for the two ligands (group C). Moreover, our data indicated that treatment of a cell line with either BMP4 or BMP7 results in similar transcriptional responses of group C genes. We therefore hypothesize that group C represents molecular responses shared by the BMP4 and BMP7 signaling pathways. This finding prompted us to ask what functions these common genes fulfill in the cell. GO enrichment analysis of the genes in group C revealed that these genes are involved in two main biological processes, regulation of gene expression and regulation of development and morphogenesis. These results support the notion that genes known to regulate development also have functions that are important for the maintenance of cancer cells.
We also studied the temporal patterns of the transcriptional response after BMP treatment. The number of DEPs showed a tendency to increase with time, a trend previously noticed in transcriptome analysis of TGF-β family members in murine mammary epithelial cells and in breast cancer cells [29, 35]. The DEPs could be grouped according to their temporal pattern of expression, varying from early to late responders. These temporal clusters were found in every cell line, and some of them even contained over a thousand gene elements. The next logical step was to explore whether there was a time-dependent shift in the distribution of gene functions. Although GO enrichment results were not obtained for all the probe clusters of all the cell lines, interesting features could be identified, especially in the case of the BMP4 data. Transcriptional regulation in the first 6 hours concentrated most notably on genes involved in developmental processes, metabolic processes, gene expression and signal transduction. Gene expression was also well-represented after 12 hours, while metabolism became by far the most prominent function at this time point. Most interestingly, 24 hours after BMP4 stimulation there was an evident overrepresentation of genes involved in cell proliferation, although this phenomenon was observed exclusively in MDA-MB-361 cells. All in all, the enriched biological functions indeed fluctuated in time and in a logical sequence, with regulation of gene expression and signal transduction leading to changes in metabolism and finally to regulation of cell proliferation, a phenotype relevant for cancer cell physiology. The fact that we did not see enrichment of cell proliferation-associated functions in more than one cell line could be due to differences in the speed of BMP signaling in different cell lines. Even though a longer experiment certainly could have clarified this issue, we concentrated our analysis on the first 24 hours after BMP treatment because we were interested primarily in the identification of BMP target genes.
As mentioned, one of the main goals of this study was to identify potential novel gene targets of BMP signaling relevant in breast cancer. We provided lists of candidate genes that are strongly and rather uniformly regulated by BMP4 or BMP7 throughout the cell lines and time points. Some of them, such as members of the Id family of inhibitors of DNA binding, are well-known BMP target genes [27, 29]. Id proteins are transcription factors that regulate cell growth and differentiation , and all four members of the protein family play crucial roles in various aspects of normal and malignant breast biology . Others are newly linked, in this work, to BMP signaling, and some of these genes have interesting connections with tumor biology, such as PTPRG or DUSP2. A positive feedback regulation where BMP treatment leads to increased expression of BMP antagonists is known to exist. In our study, no consistent expression changes were observed for any of the known BMP antagonists, such as noggin, gremlin, sclerostin and follistatin. Previous studies have shown a wide time window in the induction of e.g. noggin expression in different tissues after BMP treatment, ranging from 1 to 48 hours [42, 43]. Thus it is possible that the feedback effect in the breast cancer cells was not evident at time points analyzed here.
Protein tyrosine phosphatases (PTPs) are key regulators of the cellular protein phosphorylation balance, critical in the control of a wide spectrum of physiological processes such as cell proliferation, differentiation, transformation, transport and locomotion. Subsequently, aberrations in phosphorylation processes play a major role in the pathogenesis of numerous diseases, including cancer [44, 45]. PTPRG is a receptor-type PTP implicated as a candidate tumor suppressor gene in several types of tumors, including breast cancer [46, 47]. In MCF-7 breast cancer cells, PTPRG inhibits proliferation and anchorage-independent growth and reduces tumor formation in a xenograft model [47, 48]. Delayed cell cycle re-entry by increasing the level of cell cycle regulators p21 and p27 could explain the inhibitory effect of PTPRG on cell growth . Based on the above, upregulation of PTPRG in BMP-stimulated cancer cells could contribute to the observed BMP-induced antiproliferative effect [14, 21]. DUSP2 also belongs to the PTP family of phosphatases. It is a mitogen-activated protein kinase (MAPK) phosphatase (MKP) that dephosphorylates both threonine and tyrosine residues within target MAPKs leading to their deactivation. MAPK signaling controls cellular processes such as proliferation, differentiation, migration and apoptosis . Therefore, abnormal MKP activity, and hence anomalous MAPK signaling, has important consequences for processes critical to the development and progression of human cancer. The role of DUSP2 in cancer has been examined in only a few studies, and data are controversial. Overexpression of DUSP2 expression was found in 37 of 39 malignant effusions from serous ovarian carcinoma patients and was associated with poor survival . By contrast, decreased DUSP2 transcript levels were reported in cancerous breast, colon, lung, ovary, kidney and prostate tissues, and reduced DUSP2 protein levels were observed in cervical and colon cancer . Additionally, DUSP2 suppression was associated with tumorigenesis and malignancy in colon cancer, and DUSP2 overexpression induced apoptosis and inhibited tumor growth in HeLa cells in vitro and in xenograft models .