The CEBPA gene is mutated in 10% of acute myeloid leukemia (AML) cases, and wild-type CEBPA levels are reduced in additional AMLs, e.g. as a result of AML1-ETO or flt3ITD mediated inhibition of transcription or bcr-abl mediated inhibition of translation. One consequence of diminished C/EBPα activity or expression is likely inhibition of myeloid differentiation. We now find, via analysis of data from two published microarray studies, that CEBPA and Bcl-2 RNA levels correlate highly in low risk human AMLs (P = 0.003 and 0.001), suggesting that inhibition of apoptosis via induction of bcl-2 by C/EBPα or its mutant variants also contributes to transformation. Lack of correlation in other risk groups may reflect the presence of additional mutations that inhibit apoptosis. Among the low-risk patients, those with AML1-ETO had the lowest, CBFβ-SMMHC intermediate, and PML-RARα the highest levels of CEBPA (P < 0.05). Patients with CEBPA mutations were mainly in the intermediate-risk, normal karyotype group, and within this group the average CEBPA level was higher than those having wild-type alleles (P = 0.0002), suggesting a selective advantage to enable dominant-inhibition of remaining wild-type C/EBPs and/or induction of bcl-2. The two categories of C/EBPα mutant alleles associated with AML are C/EBPα p30, lacking an N-terminal transactivation domain, and C/EBPα LZ, carrying in-frame mutations in the leucine zipper (LZ) that prevent DNA-binding. Each of this oncoproteins induced bcl-2 in hematopoietic cell lines, and wild-type C/EBPα also induced bcl-2 when transduced into normal murine myeloid progenitors or in the splenocytes of H2K-C/EBPalpha]-Eμ transgenic mice. Remarkably, C/EBPα LZ oncoproteins activated the bcl-2 P2 promoter despite lack of DNA-binding, and C/EBPα p30 also activated the promoter. C/EBPα and the C/EBPα oncoproteins cooperated with NF-κB p50, but not p65, to induce bcl-2 promoter transcription. Endogenous C/EBPα preferentially coimmunoprecipitated with p50, versus p65, in myeloid cell extracts. Mutation of residues 297–302 in the C/EBPα basic region (BR) prevented induction of endogenous bcl-2 or the bcl-2 promoter and interaction with p50 but not p65 when co-expressed in 293T cells, whereas C/EBPα p30 or a C/EBPα LZ oncoprotein retain the ability to co-ip with p50. These findings offer an explanation for preferential in-frame rather than out-of-frame mutation of the LZ with sparing of the BR in the C/EBPα LZ oncoproteins that constitute one-third of CEBPA mutations in AML. C/EBPα or its mutant variants may tether to a subset of NF-κB targets genes, including Bcl-2, via p50 to facilitate gene activation. Targeting interaction between the C/EBPα basic region and NF-κB p50 may contribute to the therapy of AMLs expressing either mutant or residual wild-type C/EBPα , by increasing susceptibility to apoptosis. Targetting this interaction may also be relevant to the therapy of other malignancies expressing C/EBPs, not only via down-regulation of bcl-2 but also by preventing their cooperative activation of genes mediating inflammation and angiogenesis, which have recently been shown to contribute to the pathogenesis of several solid tumors in animal models.