C/EBPα, C/EBPα Oncoproteins, or C/EBPβ Preferentially Bind NF-κB p50 Compared with p65 Via Conserved Residues In the C/EBP Basic Region

Abstract 708

Nuclear Factor-κB (NF-κB) is a key mediator of the inflammatory response and also inhibits apoptosis. Activation of NF-κB is prevalent in AML, Hodgkin's lymphoma and subsets of non-Hodgkin's lymphoma, as well as in multiple solid tumors. NF-κB p50 (NF-κB1) and p65 (Rel A) are the most prevalent NF-κB subunits. NF-κB p65 has 30–50 fold higher affinity than p50 for the cytoplasmic protein IκB, and so under basal conditions p50:p65 heterodimers are held in the cytoplasm whereas p50:p50 homodimers are found in the nucleus bound to chromatin where they repress NF-κB target genes via association with HDACs. Canonical NF-κB activation signals lead to IκB proteosome-mediated degradation and translocation of p50:p65 heterodimers into the nucleus to replace inhibitory p50:p50 dimers with activating complexes on chromatin. We postulate an alternative means to activate NF-κB target genes wherein induction or activation of C/EBP proteins allows them to bind chromatin-bound p50 and displace HDACs in cells lacking canonical NF-κB activation. We found that C/EBPα induces bcl-2 and FLIP RNA expression via promoter interaction dependent upon the presence of NF-κB p50, that C/EBPα present in two myeloid leukemia cell lines preferentially binds p50 compared with p65, and that mutation of the C/EBPα basic region prevents bcl-2 induction and weakens interaction with p50 but not p65 (Paz-Priel et al 2005; 2009; Wang et al 2009). Using p50 and p65 variants containing the FLAG epitope at either their N- or C-termini we now demonstrate using reciprocal co-immunoprecipitation that p50 has much higher affinity than p65 for C/EBPα, for N-terminal or leucine zipper mutants of C/EBPα expressed in a subset of AML cases, and for the longer LAP or shorter LIP isoforms of C/EBPβ. The FLAG-tagged p65 proteins retained the ability to bind DNA in a gel shift assay and to activate a reporter containing two κB sites as effectively as untagged p65. Deletion of the p65 trans-activation domain (TAD) did not increase p65 affinity for C/EBPα, indicating that an inhibitory effect of the TAD does not account for specificity of C/EBPα:p50 interaction. Moreover, the increased affinity of the p50 rel homolog domain (RHD) for C/EBPα compared with the p65 RHD suggests that residues that differ between these RHDs rather than those in common make key contacts wih the C/EBP basic region. Germline deletion of the gene encoding p65, by injecting p65(flox/flox); Mx1-CRE mice with pIpC, did not reduce C/EBPα expression, prevent interaction of C/EBPα with the bcl-2 promoter in a ChIP assay, or reduce bcl-2 RNA expression, in contrast to our prior findings with p50 gene deletion. A C/EBPα variant harboring a single point mutation in the basic region, C/EBPα(R300A), has minimal affinity for p50, and further saturating mutagenesis of the C/EBPα basic region identifies an additional block of six amino acids, RERNNI, identical in C/EBPβ and critical for interaction with NF-κB p50. In conclusion, increased affinity of C/EBP proteins for p50 compared with p65 lends support to the hypothesis that C/EBP interaction with p50:p50 complexes provides an alternative means for normal cells or cancer cells to activate NF-κB target genes, for example in response to C/EBP induction or activation. C/EBP interaction with p50:p65 in activated cells may also occur, via p50, to further activate genes regulated by NF-κB. In addition, these findings suggest that targeting the C/EBP:p50 interface rather than C/EBP:p65 interaction will prove more effective in the therapy of inflammatory or malignant conditions, and our mapping of relevant basic region residues and the finding that the p50 RHD retains increased affinity for C/EBPα compared wih the p65 RHD provides a first step towards rational design of such inhibitors.