Tumor Microenvironment Induced Gene Expression Changes in Invasive Breast Cancer Cells.

Abstract

Abstract Introduction:Identifying and using gene expression signatures to predict potentially invasive breast cancer cells on the basis of the same has been performed with limited success by a number of different groups. Genes are identified on the basis of their expression with little or no relation to their functional significance in metastasis. We have taken two very different approaches to this problem. First in two rodent models we have identified three molecular pathways based on cellular functions that are altered in the invasive cells and grouped the genes on that basis. Secondly, we followed and separated the metastatic cells through different stages of metastasis viz, invasion, intravasation and distant metastasis formation. Genes altered in these populations are grouped into stably and transiently altered as the cells progress through metastasis.Materials and Methods:Rodent models: Two different rodent models were used in this study. The first model is MTLn3-derived mammary tumors in rats. The second model is MMTV-PyMT transgenic mouse breast tumor model.Isolation of Invasive cells: In vivo invasion assay was used to study the gene expression pattern of invasive subpopulation of carcinoma cells within live primary tumors. In brief, the invasive cells were collected from breast tumor using microneedles containing matrigel and epidermal growth factor. Macrophages were removed from this population by using MACS CD11b Microbeads.Isolation of Primary Tumor Cells, circulating tumor cells and lung metastasis: To isolate the average population of carcinoma cells from primary tumor and lungs, a small piece of tumor or a portion of the lung was minced and filtered with 30μm CellTrics nylon filter to obtain single cell suspension. The filtered cells were then washed 3 times with 3%BSA in PBS and resuspended in Accutase. In order to collect the CTCs, blood was collected in heparinized syringes from the right atrium. RBCs were lysed using BD Pharm Lyse lysing buffer. The residual cells were filtered, washed and resuspended. Fluroscent sorting was performed on all these cell types.Statistical Analysis: Patient data comprising of gene expression data from microarray and clinical data was obtained from the Netherlands Cancer Institute database including 295 breast cancer patients. Genes in the invasion signature were related to metastasis free survival by using the Cox proportional hazards model.Results and Discussion: Genes identified by these two methods were shown to be highly predictive of metastatic events when compared to a public database containing both gene expression and patient metastasis free survival data. Genes belonging to the pathway derived invasion signature and the transiently altered genes offer a better way of predicting invasive cancers than the existing gene signatures which are identified only on the basis of expression. We conclude that the metastatic cells are a special population and genes identified from changes in their gene expression pattern when linked to their biology and tumor microenvironment is much more potent in prediction metastasis free survival in patients than genes chosen on the basis of their level of expression alone. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4160.