Altered retmoic acid receptors

Structurally and functionally altered retinoic acid receptors have been associated with rare human neoplasms: acute promyelocytic leukemia and hepatoceilular carcinoma. Whereas the retinoic acid receptor 13(RAR13) rearrangement in hepatocellular carcinoma is unique, in acute promyelocytic leukemia (APL), RARU fusion to the promyelocytic leukemia (PML) gene by the t(15;17) translocation is a general feature of the disease. APL is an important model in cancer biology because retinoic acid induces complete remissions in this malignancy, providing the first example of differentiation therapy and of an antineoplastic drug directly targeted at the underlying genetic lesion. The molecular basis of PML./RARa fusion leukemogenesis is discussed with respect to dominant negative inhibition of nuclear receptor and PML functions,-de The, H. Altered retinoic acid receptors. FASEB J. 10,955-960 (1996) Key Word.t: leukemia . dominanl-negalive ‘ transcriplion nuclear bodies GIVEN THE KEY ROLE of retinoids in development, it was to be expected that acquired abnormalities in proteins implicated in the retinoid response pathways should be associated with disorders of either differentiation or proliferation. Remarkably, to date, structural defects were encountered in the nuclear receptors only, and mostly in retinoic acid receptor a (RARa)2 in two types of situations: cell lines that have acquired resistance to the differentiating effects of retinoids, and human cancers. A related situation will be discussed in which introduction of a receptor with dominant negative activity is shown to affect a developmental/differentiation program. SPONTANEOUS RECEPTOR MUTANTS IN DIFFERENTIATION-RESISTANT CELL LINES Identification of the nuclear receptors for retinoic acid (RARs) paved the way for genetic analyses of receptor structure in cell lines that had become resistant to the differentiating effects of retinoids. Apart from the acute promyelocytic leukemia (APL) cell line NB4 (see below), the principal models studied were the HL6O myelomonocytic and the P19 embryonal carcinoma cells. Surprisingly, in both cell lines the mutations mapped to the hormone binding domain of RARa. In a clone of RA-resistant HL6O cells, an RARa mutation that abolished RA binding was found in the hormone binding domain of the receptor (1). As in HL6O cells, RARa is haploid due to the loss of a large fragment of chromosome 17; the RARa function was lost. That such a clone failed to differentiate in the presence of RA implies that endogenous RAR(3 or y cannot substitute for RARx. Nevertheless, forced overexpression of RAR(3 or cx restored RA sensitivity, suggesting that failure to differentiate the parental cells related to the low abundance of these receptors. In the case of the mutant receptor found in P19 cells, a premature stop codon led to a large deletioti in the carboxy-terminal end of the receptor, including the RA binding site and the AF2 domain (2). In contrast to the previous situation, a dominant negative behavior or a very sensitive dosage effect was assumed, as the other RARa gene appears intact. Recent work has shown that F9 cells with a homozygote deletion in the RARa gene still differentiate in the presence of RA, favoring the dominant negative hypothesis (3). These two cases exemplify several themes commonly encountered in retinoid biology: redundancy, dominant negative activities. EXPRESSION OF DOMINANT NEGATIVE RARS Extensive analysis of the functional anatomy of the receptors led to the identification of the receptor domains involved in activating transcription. As expected, receptors (such as RARa4O3) deleted for these functions exhibit a dominant negative phenotype in transient cotransfection assays with RA-responsive reporter genes. When introduced in pluripotent hematopoietic cells, RARa4O3 leads to profound modifications in the differentiation program of these cells. Not only is the spontaneous neutrophil/granulocyte differentiation suppressed, but treatment with GM-CSF, which would normally force granulocytic differentiation, leads in the presence of RARa4O3 to a maturation arrest highly reminiscent of APL (see below). As in APL, this differentiation arrest is reversible upon stimulation with high doses of retinoic acid (RA) (4, 5). Remarkably, RARa4O3 was also shown to immortalize very early progenitor cells that have maintained the abiltAddress correspondence and reprint requests to Dr. de The, at: CNRS UPR 9051 and Service de Biochimie, H#{244}pital St. Louis, 1, av. C Vellefaux, 75475 Paris cedex 10, France. 2Abbreviations: RA, retinoic acid; RAR, retinoic acid receptor; APL, acute promyelocytic leukemia; PML, promyelocytic leukemia. ity to differentiate into lymphoid or erythroid lineages (6). Again, in that case, high doses of retinoic acid relieve the block and allow myeloid differentiation. Note that such dominant negative approach is more powerful than RAR gene inactivation, underlining the functional redundancy between RARs. Indeed, in the hematopoietic differentiation, RARa4O3 allowed the delineation of two distinct RA-sensitive steps unraveled by either vitamin A deprivation or RAR knockouts (although one could argue that by titrating RXRs, these dominant negative constructs could also block the response to other nuclear receptors, such as TR or VDR, believed to play an important role in hematopoietic differentiation). When RARa4O3 is targeted to the skin in vivo, profound abnormalities in epidermal differentiation appear, with inhibition of terminal differentiation and loss of barrier function. Depending on the promoter used in the transgenic mouse, different phenotypes have been described, but in all cases a block in maturation with a shift in the keratin expression pattern were observed. Note in this respect that several keratin genes are primary RA targets (7). These data suggest a role for RA in normal epidermal maturation and in regulating the ordered cascade in keratin gene expression from the basal layer to the stratum corneum. Similarly, expression of a carboxy-terminally truncated RAR’y into keratinocyte cell lines impaired their activity to differentiate and express late markers of squamous differentiation. In this case, the residual transcriptional activity of the A/B domain correlates highly with the degree of terminal differentiation. Thus, other mechanisms that simple competition for DNA binding of AF2 less receptot-s are implicated in these effects on differentiation (8). HUMAN CANCERS WITH STRUCTURALLY