NF-κB Target Genes Activation Via Displacement of HDACs Bound to NF-κB p50 Homodimers by C/EBPα, Its AML Oncoproteins or C/EBPβ.

Abstract 3637

The NF-κB transcription factors are involved in inflammation, proliferation, and apoptosis. The p50:p65 heterodimer, the most common transcription activating form of NF-κB, is retained in the cytosol and translocates into the nucleus only upon stimulation. In contrast, the p50 homodimer is the primary nuclear NF-κB species in unstimulated cells, and recruit HDACs to repress NF-κB-target gene expression. Dysregulation of NF-κB frequently occurs in AML and lymphoma. C/EBPα is mutated in approximately 10% of patients with AML, resulting in expression of C/EBPαp30 or C/EBPαLZ oncoproteins. We showed that C/EBPα or its oncoproteins, including leucine zipper variants (C/EBPαLZ) that cannot bind DNA directly, protects hematopoietic cells from apoptosis. Through interaction with NF-κB p50, C/EBPα or its AML mutants activate several anti-apoptotic genes including BCL2 and FLIP. C/EBPβ or its truncated LIP isoform are over-expressed in Hodgkin lymphoma and breast cancer. Using Ba/F3 cell lines carrying an estrogen receptor fusion of C/EBPα or C/EBPαLZ, we now show that activation of C/EBPα isoforms results in a significant 90% reduction in the association of endogenous HDAC1 or HDAC3 with the FLIP promoter, within 5 hours. In contrast, activation of the basic region mutant C/EBPαBR that cannot bind p50 did not have an effect. Accordingly, ChIP analysis demonstrated that on the FLIP promoter, the expression of C/EBPα or C/EBPαLZ resulted in a 2.5 or 4-fold increase in histone H3 or H4 acetylation, respectively. To study the specific displacement of HDACs from p50 we used Eμ-C/EBPα transgenic mice which express C/EBPα in splenic lymphoid cells, in contrast to control littermates. The ectopic presence of C/EBPα resulted in a significant reduction of endogenous HDAC1 or HDAC3 on the FLIP promoter to 17 or 4% of the expression in wild type mice. In contrast, in p50-/- mice, the C/EBPα transgene had a minimal effect on HDAC1 or a modest reduction of HDAC3 occupancy on the FLIP promoter, to 90 or 50% of levels seen in p50-/- controls. These findings suggest that on the FLIP promoter C/EBPα displaces primarily p50-bound HDACs. HDAC1 or HDAC3 co-ip with p50 in transiently transfected 293T cells. The co-expression of C/EBPα, its C/EBPαLZ or C/EBPαp30 AML mutants, or C/EBPβ interrupted interaction of p50 with HDAC1 and to a lesser extent with HDAC3. In striking correlation to our previous findings, the basic region mutant C/EBPαBR does not interrupt this interaction. As a control, we confirmed that C/EBPα or C/EBPβ do not disrupt the NF-κB p50:p65 co-ip. We have shown that interaction with p50 is required for induction of bcl-2 and FLIP by C/EBPα and that p50 binds and induces the CEBPA gene. Using ChIP analysis, we now show that in U937 AML cells endogenous C/EBPα, C/EBPβ and p50 specifically bind the proximal segment of the nfkb1 (p50) promoter. In addition, C/EBPαLZ expressed from the metallothionein promoter in Ba/F3 cells binds a proximal site in the nfkb1 promoter. C/EBPα or C/EBPαLZ induced p50 RNA expression 3.5-fold, while C/EBPαBR was not effective. Treating the cells with the translation inhibitor cycloheximide did not change p50 induction, suggesting direct activation. C/EBPα, C/EBPαLZ but not C/EBPαBR activated to a similar degree luciferase reporters containing 1800 or 250 bp fragments of the nfkb1 promoter. In addition, C/EBPβ or its truncated LIP isoform activated these reporters, although LIP cannot activate a (CEBP)₂TK-Luc reporter. The murine nfkb1 promoter contains 2 proximal κB sites that are 100% conserved between mouse and human. p50 but not C/EBP avidly binds either of these sites in a gel shift assay. Disruption of either κB site abrogated 65 to 80% of reporter activation, and mutating both resulted in activation of less than 10% by either C/EBPα or C/EBPαLZ. In summary, we show that C/EBPs displace HDAC from NF-κB p50, providing a mechanism whereby C/EBPα or its AML mutants can derepress NF-κB target genes and contribute to their dysregulated expression. In addition, C/EBPα, a C/EBPαLZ AML variant that cannot bind DNA directly, or C/EBPβ bind and activate the nfkb1 promoter through interaction with NF-κB bound to two proximal, highly conserved κB sites. Together these findings lend support to the idea that disrupting the C/EBP:NF-κB p50 complex may offer a complementary approach for targeting NF-κB activity in the nucleus to favor apoptosis in AML or other malignancies in which C/EBP family members are expressed.