Unbiased compound screening identifies unexpected drug sensitivities and novel treatment options for gastrointestinal stromal tumors.

Most gastrointestinal stromal tumors (GIST) are caused by oncogenic KIT or platelet-derived growth factor receptor activation, and the small molecule kinase inhibitor imatinib mesylate is an effective first-line therapy for metastatic or unresectable GIST. However, complete remissions are rare and most patients ultimately develop resistance, mostly because of secondary mutations in the driver oncogenic kinase. Hence, there is a need for novel treatment options to delay failure of primary treatment and restore tumor control in patients who progress under therapy with targeted agents. Historic data suggest that GISTs do not respond to classical chemotherapy, but systematic unbiased screening has not been performed. In screening a compound library enriched for U.S. Food and Drug Administration (FDA)-approved chemotherapeutic agents (NCI Approved Oncology Drugs Set II), we discovered that GIST cells display high sensitivity to transcriptional inhibitors and topoisomerase II inhibitors. Mechanistically, these compounds exploited the cells' dependency on continuous KIT expression and/or intrinsic DNA damage response defects, explaining their activity in GIST. Mithramycin A, an indirect inhibitor of the SP1 transcription factor, and mitoxantrone, a topoisomerase II inhibitor, exerted significant antitumor effects in mouse xenograft models of human GIST. Moreover, these compounds were active in patient-derived imatinib-resistant primary GIST cells, achieving efficacy at clinically relevant concentrations. Taken together, our findings reveal that GIST cells have an unexpectedly high and specific sensitivity to certain types of FDA-approved chemotherapeutic agents, with immediate implications for encouraging their clinical exploration.