Acute promyelocytic leukemia (APL) is the first example among the human malignancies that responds to differentiation therapy, in that complete remission (CR) can be achieved in up to 90% of patients by using a differentiation inducer, all-trans retinoic acid (ATRA). The specific chromosomal translocation t (15; 17) in APL has been shown to fuse the gene for the retinoic acid receptor alpha (RAR alpha) with a chromosome 15q locus, PML. Alterations to the RAR alpha and the PML gene structures in the t (15; 17) have been characterized and used as specific molecular marker for diagnosis of the disease. PML/RAR alpha antagonizes wild-type PML and RXR and could block the differentiation pathways mediated by these two regulators. A variant translocation t (11; 17) has also been discovered in a subset of APL which fuses RAR alpha to a new gene, PLZF on chromosome 11q23. Both PML/RAR alpha and PLZF/RAR alpha display the" dominant negative" effect on the wild-type RAR/RXR. However, t (11; 17) APL patients differ from t (15; 17) APL in that they respond poorly to ATRA. Morphologically defined APL cases which however do not have PML/RAR alpha generally show no response to ATRA. Recently it has been shown that PML/RAR alpha can be modulated directly by ATRA. All these data support the idea that PML/RAR alpha is a specific target of ATRA which can overcome the differentiation block imposed by PML/RAR alpha. The ATRA treatment of APL thus further reinforces the concept of differentiation therapy.