Homodimerization of Retinoic Acid Receptor alpha through its Fusion Partners Underlies Pathogenesis of Acute Promyelocytic Leukemia.

In almost all cases of acute promyelocytic leukemia (APL), t(15;17)(q22;q12) fuses the promyelocytic leukemia (PML) gene with the retinoic acid receptor alpha (RARA) gene. In rare cases of APL, other genes are fused to RARA including PLZF, NPM, NuMA, and STAT5b. These chimeras are collectively referred to as X-RARA fusions. Common features of the X-RARα fusions include their ability to form dimers through the X domain proteins while retaining the ability to interact with rexinoid receptors (RXRs, the proteins with which RARα normally heterodimerizes). We previously created an artificial X-RARα: p50-RARα (a fusion of the dimerization interface of p50 NFκB with the portion of RARα found in naturally occurring X-RARα proteins). p50-RARα has effects similar to PML-RARα including (i) enhancing repression at retinoic acid response elements, due to impaired release of co-repressors at low levels of all-trans retinoic acid (ATRA), and (ii) inhibiting differentiation of myeloid cell lines. We generated transgenic mice expressing p50-RARα under the control of the MRP8 promoter. p50-RARα had a minimal effect on myelopoiesis and initiated myeloid leukemias at a rate of <5% in the first year of observation. Nevertheless, transduction of p50-RARA transgenic bone marrow with a retrovirus expressing an activated cytokine receptor (β_c^V449E) generated leukemias with features of APL including therapeutic responsiveness to ATRA. Complementing our work with the p50-RARA transgenic mice, retroviral co-transduction of normal bone marrow with β_c^V449E plus either p50-RARA or PML-RARA generated APL-like myeloid leukemias. Although retrovirally generated β_c/p50-RARα and β_c/PML-RARα leukemias were nearly identical, the PML fusion appeared associated with a subtle decrease in differentiation relative to the p50 fusion. In contrast to these X-RARα fusions, preliminary experiments suggest that RCRα (a homodimeric form of RARα in which the dimerization interface of RARα has been replaced by the dimerization domain of the homodimeric COUP-TF transcription factor) does not readily cooperate with β_c^V449E to cause leukemia. This finding lends support to the hypothesis that the decreased binding site selectivity of X-RARα fusions contributes to APL pathogenesis. Furthermore, we have generated a novel X-RARα fusion, F3-RARα, in which three copies of the F36M mutant of the Rapamycin-binding protein have been fused to RARα. This protein mimics the in vitro effects of PML-RARα on transcription & differentiation, and de-dimerization of F3-RARα by rapamycin reverses these effects. Additional experiments in vivo with F3-RARα are expected to further confirm the importance of abnormal transcription factor dimerization in leukemia pathogenesis.