Targeted therapeutics for high-risk cancers remain an unmet medical need. Here we report the results of a large-scale screen of over 11,000 molecules for their ability to inhibit the survival and growth in vitro of human leukemic cells from multiple sources including patient samples, de novo generated human leukemia models, and established human leukemic cell lines. The responses of cells from de novo models were most similar to those of patient samples, both of which showed striking differences from the cell-line responses. Analysis of differences in subtype-specific therapeutic vulnerabilities made possible by the scale of this screen enabled the identification of new specific modulators of apoptosis, while also highlighting the complex polypharmacology of anti-leukemic small molecules such as shikonin. These findings introduce a new platform for uncovering new therapeutic options for high-risk human leukemia, in addition to reinforcing the importance of the test sample choice for effective drug discovery.

针对高危癌症的靶向治疗仍然是一项未得到满足的医疗需求。在这里，我们报告了超过 11,000 个分子的大规模筛选结果，这些分子具有抑制人类白血病细胞体外存活和生长的能力，这些分子来自多种来源，包括患者样本、从头生成的人类白血病模型和已建立的人类白血病细胞系。来自从头模型的细胞反应与患者样本的反应最相似，两者均显示出与细胞系反应的显着差异。通过这种筛选的规模，分析亚型特异性治疗脆弱性的差异成为可能，从而能够识别细胞凋亡的新特异性调节剂，同时也突出了抗白血病小分子（如紫草素）的复杂多药理学。这些发现为发现高危人类白血病的新治疗方案提供了一个新平台，此外还加强了测试样本选择对有效药物发现的重要性。