Retinoid therapy transformed response and survival outcomes in acute promyelocytic leukemia (APL), but has demonstrated only modest activity in non-APL forms of acute myeloid leukemia (AML). The retinoic acid receptors (RARs) and retinoid X receptors (RXRs) are ligand-activated transcription factors that influence hematopoietic stem cell self-renewal and differentiation. In normal hematopoiesis, RARA and RXRA are dynamically regulated during myeloid maturation, with highest expression in mature neutrophils. In acute myeloid leukemia (AML) RARA and RXRA have parallel expression among AML subtypes, with the highest expression in M4/M5 myelomonocytic and monocytic subtypes. Using a murine reporter assay, we found that natural RXRA ligands, but not RARA ligands, were present in vivo in primary mouse myelomonocytic leukemia cells (derived with an MLL-AF9 retrovirus) and acted as tumor suppressors. RXR ligands were absent in erythroleukemia (derived with TLS-ERG) or T-cell leukemia (derived with activated Notch1), suggesting the presence of natural RXRA ligands specifically in myelomonocytic leukemia. Moreover, we found that deletion of Rxra and Rxrb was naturally selected in MLL-AF9, but not in Notch1 or TLS-ERG leukemias, and that loss of Rxra and Rxrb accelerated leukemic growth. This suggests that Rxrs act as tumor suppressors, but only when exposed to natural ligands. In MLL-AF9 derived leukemia cells, pharmacologic treatment with single-agent retinoid led to modest growth inhibition due to non-permissive activity of the RARA:RXR heterodimer, whereas concurrent activation of both RARA, with all-trans retinoic acid (a pan-RAR ligand), and RXR, by bexarotene (a pan-RXR ligand), enabled efficient co-repressor release and synergistic leukemic apoptosis. We observed that co-repressors release (SMRT/NCoR) from the RARA:RXRA heterodimer was specifically associated with RARA activation, whereas growth inhibition required RXR binding, and this occurred through apoptosis rather than maturation. Generating a series of RXRA mutant proteins we demonstrated the active contribution of RXRA to the activity of the RARA:RXRA heterodimer, specifically requiring the activation domains (AF1 and AF2) and the ability to recruit co-activator to the RXRA element. Finally, we observed a significant dose-dependent effect of combination ATRA and bexarotene treatment in in vivo in MLL-AF9 leukemic mice with a striking reduction in the tumor burden of treated mice compared to the control cohort. These data provide a strategy for clinical retinoid therapies in leukemias beyond acute promyelocytic leukemia and provide a mechanism for integrating retinoid therapy into future clinical trials.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
