A Positive Feedback Loop Between C/EBPα and NF-κB p50 Inhibits Both the Intrinsic and Extrinsic Apoptosis Pathways

Resistance to apoptosis is a hallmark of leukemia, representing a major therapeutic obstacle. We previously showed that CCAAT/enhancer binding protein α (C/EBPα) and its AML mutants inhibit apoptosis in hematopoietic cells dependent upon induction of bcl-2 in cooperation with NF-κB p50. We now identify the genes encoding mcl-1, an anti-apoptotic bcl-2 member and FLIP, an inhibitor of the extrinsic apoptosis pathway, as additional direct genetic targets of the C/EBPα:NF-κB complex. In particular, C/EBPα binds to the BCL2, FLIP or MCL1 promoters, as detected by ChIP assay in both Ba/F3 cells and normal murine marrow or spleen cells. In contrast, C/EBPα does not bind to the promoters of these genes in marrow or spleen cells obtained from nfkb1(−/−) mice that lack NF-κB p50. In addition, induction of C/EBPα from the zinc-responsive MT promoter in Ba/F3 cells leads to induction of BCL2, FLIP, and MCL1 mRNAs. Furthermore, C/EBPα(ΔK312), an AML derived leucine zipper (LZ) mutant that does not bind DNA but can interact with NF-κB p50, retains the ability to localize to the BCL2, FLIP, and MCL1 promoters and to induce their corresponding mRNAs. The promoters for each of these genes also bind endogenous NF-κB p50 as assessed by ChIP assay. We previously demonstrated that the ability of C/EBPα to block the intrinsic apoptotic pathway and protect cells from radiation- or cytokine withdrawal-induced apoptosis is dependent upon induction of bcl-2. We now show that in C/EBPα transgenic splenocytes bcl-2 protein is induced only in the presence of NF-κB p50. As expected from the induction of FLIP by C/EBPα, we now also find that transgenic C/EBPα protects wild-type but not nfkb1(−/−) splenocytes from apoptosis induced by FasL activation of the extrinsic apoptotic pathway. In addition, we now demonstrate direct interaction of C/EBPα with NF-κB p50 using highly purified proteins isolated from E. coli. Together these data indicate that BCL2, MCL1, or FLIP are induced cooperatively by C/EBPα or C/EBPα(ΔK312) proteins that tether to the corresponding promoters via bound NF-κB p50. Additionally, we found that C/EBPα or C/EBPα(ΔK312) bind to the nfkb1 promoter and induce its mRNA expression. Conversely, we also demonstrate that immature, lineage-negative marrow cells isolated from nfkb1(−/−) mice have an approximately 3-fold reduced level of C/EBPα and that NF-κB p50 binds the endogenous CEBPA promoter in normal murine marrow cells in the ChIP assay, indicating cross-regulation and a potential positive feedback loop between C/EBPα and NF-κB p50. Indeed, induction of exogenous C/EBPα in Ba/F3 cells increases nuclear localization of NF-κB as assessed by gel shift assay. In summary, C/EBPα and its AML mutants interact directly with NF-κB p50 via their basic regions; through binding to a DNA κB site the complex activates several anti-apoptotic genes including BCL2, MCL1 and FLIP, thereby inhibiting both the extrinsic and intrinsic apoptotic pathways. In addition, through positive feedback NF-κB p50 induces C/EBPα expression, and C/EBPα induces NF-κB p50. These effects are retained by the C/EBPα AML mutants and do not depend on DNA binding, contributing to NF-κB deregulation in malignant blasts. As C/EBPα or mutant C/EBPα variants and activated NF-κB are commonly expressed together in AML stem cells, targeting the C/EBPα:NF-κB p50 interaction with small molecules may provide a novel therapeutic approach to this and other malignancies.