Remarkable progress has been made in identifying the oncogenic fusion proteins that are produced as a result of specific chromosomal translocations in the human acute leukemias (1). Because these proteins are expressed only by cells within the leukemic clones, they hold great promise as targets for new drugs that will specifically destroy leukemic cells while leaving normal cells unharmed. Both all-trans-retinoic acid (t-RA) and arsenic trioxide (As2O3) have shown this type of activity in patients with acute promyelocytic leukemia (APL) whose blast cells harbor the t (15; 17) chromosomal translocation and express the PML/RARα fusion protein. Each of these agents was identified empirically in innovative clinical trials conducted by investigators at the Shanghai and Harbin Institutes of Hematology in China (2, 3). A plausible biochemical basis for the selective action of t-RA became apparent once the PML/RARα fusion gene was cloned, since the aberrant protein contains the retinoic binding domain of retinoic acid receptor (RAR) α (4–8). Subsequent work showed that t-RA can induce differentiation of leukemic promyelocytes (2, 9–11), and clinical trials combining r-RA with standard cytotoxic drugs have led to improved survival in patients with APL (12–16).

More recently, As2O3 was shown to be active against APL that was resistant to t-RA and other forms of chemotherapy (3); ie, among 15 patients with this disease, 14 achieved complete remission after treatment with arsenic (17). Few clues to the specific activity of As2O3 in APL can be gleaned from an analysis of the structure of the PML/RARα protein. However, studies by Shao et al.(18), published in this issue of the Journal, together with recent reports from Z. Chen’s laboratory (17, 19, 20, 25), afford important insights into the mechanism of action of this new agent.