PML-RARA-RXR oligomers mediate retinoid and rexinoid/cAMP cross-talk in acute promyelocytic leukemia cell differentiation

Abstract

PML-RARA was proposed to initiate acute promyelocytic leukemia (APL) through PML-RARA homodimer-triggered repression. Here, we examined the nature of the PML-RARA protein complex and of its DNA targets in APL cells. Using a selection/amplification approach, we demonstrate that PML-RARA targets consist of two AGGTCA elements in an astonishing variety of orientations and spacings, pointing to highly relaxed structural constrains for DNA binding and identifying a major gain of function of this oncogene. PML-RARA-specific response elements were identified, which all conveyed a major transcriptional response to RA only in APL cells. In these cells, we demonstrate that PML-RARA oligomers are complexed to RXR. Directly probing PML-RARA function in APL cells, we found that the differentiation enhancer cyclic AMP (cAMP) boosted transcriptional activation by RA. cAMP also reversed the normal silencing (subordination) of the transactivating function of RXR when bound to RARA or PML-RARA, demonstrating that the alternate rexinoid/cAMP-triggered APL differentiation pathway also activates PML-RARA targets. Finally, cAMP restored both RA-triggered differentiation and PML-RARA transcriptional activation in mutant RA-resistant APL cells. Collectively, our findings directly demonstrate that APL cell differentiation parallels transcriptional activation through PML-RARA-RXR oligomers and that those are functionally targeted by cAMP, identifying this agent as another oncogene-targeted therapy.

Figure 1.

Selection of high affinity binding sites for PML-RARA homodimers. (A) Band shift assays representing the different steps of the selection/amplification procedure with Cos cell–expressed PML–RARA protein. The positive control (β) is the RAREB response element. V represents the initial pool of random oligonucleotides and OLI and DIM point to the two complexes containing PML–RARA oligomers and dimers, respectively. The fifth (5) and sixth (6) round of purification derived from the OLI complex and the sixth (6) round of purification from the DIM complex are shown. Note that the selected sequences are specific for PML–RARA, as they can be supershifted with anti-PML antibodies (PML-Ab). (B) Optimized binding site in a DR5 architecture (DR5) bind PML–RARA and RARA/RXR with a higher affinity than RAREB (β).

Figure 2.

Summary of the binding elements identified by random selection. (A) Nucleotide frequency within the identified core motif. (B) Summary of the various architectures and spacings identified among the 109 sequences with two AGGTCA core motifs. (C) Preference for specific nucleotides adjacent to the AGGTCA core in the various architectures. Comparison between the natural RAREB response element from the RARB promoter and the optimized DR5 site identified here. Nonconserved residues are indicated by x.

Figure 3.

Binding of PML-RAR and PML-RAR–RXR to direct, inverted, or everted optimized repeats. (A) Extracts of Cos cells transfected with PML–RARA were analyzed by band shifting using as probes DRs separated with the indicated number of bases, as well as palindromes (inverted or everted repeats, IR0 and ER8). On the right, controls were the RAREB probe (β), untransfected cells (Ø), or a DR5 where one hexamer was mutated (−). Ratio of bound/nonbound probe is shown above. Bold line, OLI complex; thin line, DIM complex. (B) Same as A, using extracts of Cos cells transfected with PML–RARA and RXRA. Note that a single complex corresponding to PML–RARA-RXRA oligomers (arrow) is observed with most probes. A rapidly migrating complex (arrowhead) is observed with DR1 and DR6 corresponds to RXRA homodimers.

Figure 4.

Identification of PML–RARA-RXR tetramers in APL cells. Binding sites ER8, DR8, and DR2 were presented to increasing amounts of the NB4 nuclear extracts. The gel mobility of the complexes was compared with extracts of Cos cells transfected with RXRA (RXR), PML–RARA (P/R), or both (P/R+RXR). Identity of DNA–protein complexes was tested by supershifting with anti-RARA (α-A), anti-PML (α-P), and anti-RXR (α-X) antibodies. The arrow points to the PML–RARA-RXR oligomers, OLI and DIM, and to the PML–RARA-only complexes.

Figure 5.

Identification of PML–RARA-specific response elements. (A) DRs separated with spacers of the indicated length, as well as ER8 and IR0, were inserted in a TK-Luc vector and tested for induction in the presence of RA. Values reported are averaged from three independent experiments in Cos cells. Shaded bars denote the induction in the presence of cotransfected PML–RARA expression vector, striped bars represent induction by endogenous RARA–RXR complexes. (B) PML–RARA induces repression on several PML–RARA binding sites. Note that RA only relieves repression on DR5 and DR6, but activates on DR8. (C) Reporter constructs were electroporated in NB4 APL cells that express endogenous RARA and PML–RARA proteins. (D) As in C, except that non-APL U937 cells that express RARA, but not PML–RARA, were used.

Figure 6.

Identification of natural PML–RARA-specific RA targets in APL cells. CD37 mRNA expression relative to TBP mRNA in NB4 cells (solid lines) or HL60 (dotted line) were monitored using quantitative RT-PCR after exposure to 1 μM RA for the indicated time.

Figure 7.

Synergistic transcriptional activation by RA and cAMP. (A) NB4 cells were electroporated with the indicated reporters and treated with the following retinoids, as indicated: RA, BMS753 (RARA+), or BMS649 (RXR+), in the presence or absence of 160 μM cAMP for 18 h. (B) Dose response of RA-triggered activation on PML–RARA-specific reporter DR8 in NB4 cells, in the presence or absence of cAMP. (C) Same as in A, except that NB4LR2 cells were tested. (D) Same as in A and C, except that U937 cells were tested.

Figure 8.

cAMP restores RA-triggered differentiation in NB4MR4 cells. NBT reduction (A), CD11b expression (B), or May Grunwald Giemsa staining (C) were measured after a 4-d treatment with the indicated drug (10⁻⁶ M) and/or 200 μM cAMP in NB4 or NB4MR4 cells, as indicated. (D) Transcriptional activation on a DR8 reporter upon electroporation in NB4MR4 cells and treatment with the indicated drugs. Note that activation of PML–RARA-dependent transcription mirrors differentiation induction.