Acute promyelocytic leukemia (APL) is an interesting model in cancer research, because it can respond to the dierentiation/apoptosis induction therapy using all-trans retinoic acid (ATRA) and arsenic trioxide (As 2 O 3 ). Over the past 5 years, it has been well demonstrated that As 2 O 3 induces a high complete remission (CR) rate in both primary and relapsed APL patients (around 85*90%). The side eects are mild to moderate in relapsed patients, while severe hepatic lesions have been found in some primary cases. After CR obtained in relapsed patients, chemotherapy in combination with As 2 O 3 as post-remission therapy has given better survival than those treated with As 2 O 3 alone. The eect of As 2 O 3 has been shown to be related to the expression of APLspeci®c PML ± RARa oncoprotein, and there is a synergistic eect between As 2 O 3 and ATRA in an APL mouse model. Cell biology studies have revealed that As 2 O 3 exerts dose-dependent dual eects on APL cells. Apoptosis is evident when cells are treated with 0.5*2.0 mM of As 2 O 3 while partial dierentiation is observed using low concentrations (0.1*0.5 mM) of the drug. The apoptosis-inducing eect is associated with the collapse of mitochondrial transmembrane potentials in a thiol-dependent manner, whereas the mechanisms underlying APL cell dierentiation induced by low dose arsenic remain to be explored. Interestingly, As 2 O 3 over a wide range of concentration (0.1*2.0 mM) induces degradation of a key leukemogenic protein, PML ± RARa, as well as the wild-type PML, thus setting up a good example of targeting therapy for human cancers. Oncogene (2001) 20, 7146 ± 7153.