Abstract
Our current understanding of cancer-stem cells (CSCs) is that they are slow growing, generally mesenchymallike cells capable of generating tumors. Convincing evidence for the existence of such cells comes from recent lineage tracing experiments. CSCs have been reported as being resistant to conventional drug treatment and have been considered as being responsible for failure of chemotherapy. Recently, several databases aiming the genetic characterization of a large number of cancer cell lines have been made publicly available. In addition to gene expression data, these databases contain cytotoxicity information for all cell lines for a number of drugs as well. It is possible to classify known cell lines derived from a given tumor, based on how similar they are to CSCs, or in other words, to define their stem-ness, using gene-lists that define such cells. Using two such, independently generated, gene lists we found that breast cancer cell lines could be categorized into two distinct groups which we designate CSC-like and non-CSC-like. We then identified drugs to which the two groups were most sensitive to. We also generated sensitivity profiles for all drugs, within one such database, to identify chemotherapeutics with preferential action on breast cancer. We believe this is a straight-forward approach for swiftly identifying drugs that would selectively target a subpopulation of cells for any given tumor type.
Keywords: Cancer, cancer stem cells, chemotherapy, cytotoxicity, database mining, lapatinib, LBW242, TKI258.