AbstractAs the result of intense clinical and basic research, acute promyelocytic leukemia (APL) has progressively evolved from a deadly to a curable disease. Historically, efforts aimed at understanding the molecular bases for therapy response have repeatedly illuminated APL pathogenesis. The classic model attributes this therapeutic success to the transcriptional reactivation elicited by retinoic acid and the resulting overcoming of the differentiation block characteristic of APL blasts. However, in clinical practice, retinoic acid by itself only rarely yields prolonged remissions, even though it induces massive differentiation. In contrast, as a single agent, arsenic trioxide neither directly activates transcription nor triggers terminal differentiation ex vivo, but cures many patients. Here we review the evidence from recent ex vivo and in vivo studies that allow a reassessment of the role of differentiation in APL cure. We discuss alternative models in which PML-RARA degradation and the subsequent loss of APL cell self-renewal play central roles. Rather than therapy aimed at inducing differentiation, targeting cancer cell self-renewal may represent a more effective goal, achievable by a broader range of therapeutic agents.