The concept that some cancers are sustained by rare cancer stem cells (CSCs) has led to efforts to develop drugs that selectively target these CSCs. Previously, a human pluripotent stem cell (hPSC) line was developed that has characteristics of cancer stem cells, including enhanced self-renewal and low differentiation potential. Now, Mickie Bhatia and colleagues use this neoplastic hPSC line to screen several hundred known compounds for their ability to induce differentiation and overcome neoplastic self-renewal (Cell, published online 24 May 2012; doi:10.1016/j.cell.2012.03.049). The authors set out to identify compounds that would induce differentiation in neoplastic hPSCs but would have little effect on normal hPSCs. Of 590 compounds screened, only thioridazine significantly induced differentiation in neoplastic hPSCs but not in normal hPSCs. The authors tested 2,446 additional compounds, whereby they identified 26 candidates, including thioridazine and 2 other phenothiazine compounds. Of these three compounds, thioridazine still showed the strongest effect on neoplastic hPSCs. The authors conducted xenotransplantation studies in vivo and treated leukemic disease–initiating acute myeloid leukemia (AML) leukemic stem cells (LSCs) with thioridazine. This treatment reduced the leukemic disease potential of these cells.