Retinoic acid receptors (RARs) are members of a much larger group of ligand-dependent transcription factors, the so-called nuclear receptors. In mammals, there are three distinct RAR subtypes: RAR, RAR and RAR.1 RARs share a high sequence homology, but show distinct transcriptional properties and expression pattern, suggesting similar but not necessarily overlapping functions. In haematopoietic differentiation, for example, the function of RAR is uniquely required to maintain the stem cell compartment, whereas RAR seems to work primarily at later stages, favouring differentiation.2 In acute promyelocytic leukaemia (APL), chromosomal translocations involve in 100% of cases the -subtype of the RAR gene (RAR), on chromosome 17. These translocations lead to the synthesis of chimeric proteins: in 95% of the cases, the translocation partner is the PML (promyelocytic leukaemia) gene on chromosome 15, leading to the formation of PML–RAR fusion gene.3 Both in vitro and in vivo studies have well defined the oncogenic roles of PML–RAR fusion protein in the pathogenesis of APL, acting as an altered RAR.4 In our previous work, we studied the transcriptional and biological properties of chimeric constructs where PML was fused to RAR or RAR to generate the PML–RAR and PML–RAR fusion proteins (which do not exist in nature). Both chimeric proteins demonstrated similar properties to PML–RAR in vitro.5 Here, we extended this analysis to determine directly whether PML–RAR (as a test case), like PML–RAR, is able to induce APL.