As overactivity of pro-survival Bcl-2 proteins promotes neoplasia and enhances resistance of malignant cells to cytotoxic therapies, a promising approach for incurable lymphoid tumors is to directly target the pro-survival proteins. One approach is to mimic their physiological antagonists, the BH3-only proteins. We initially determined which of the diverse BH3-only proteins would be optimal to mimic. The interactions of the BH3 domains of this family with a groove on the Bcl-2-like proteins have been considered promiscuous. However, we found that the interactions between eight BH3 peptides and five Bcl-2-like proteins varied over 10,000 fold in affinity, and that only certain protein pairs associate inside cells (Chen et al

Mol Cell
). Bim and Puma potently engaged all the pro-survival proteins comparably. Bad, however, bound preferentially to Bcl-2, Bcl-xL and Bcl-w. Strikingly, Noxa bound only Mcl-1 and A1. In accord with their complementary binding, Bad and Noxa cooperated to induce potent killing.

We next determined the mechanism of action of putative BH3 mimetic compounds. Because killing by the BH3-only proteins require the action of the essential cell death mediators, Bax and Bak, we initially screened putative BH3 mimetic compounds on cells genetically engineered to lack both Bax and Bak. Only those compounds that mimicked the BH3-only proteins (i.e, inert on BaxBak double null cells, but active in Bax or Bak expressing cells) were evaluated further. Surprisingly, among seven putative BH3 mimetics tested, we found that only the recently described ABT-737 (Abbott Laboratories; Oltersdorf et al,

), required the pro-apoptotic protein Bax or Bak to induce apoptosis. The cytotoxicity of ABT-737 alone was modest in a range of hematopoietic and non-hematopoietic cell types. Further investigations revealed that, like Bad, ABT-737 only targeted Bcl-2, Bcl-xL and Bcl-w. As hematopoietic cells typically express Mcl-1, we attempted to augment the activity of ABT-737 by concomitantly neutralizing Mcl-1. Targeting of Mcl-1 by overexpressing the BH3-only protein Noxa, or Mcl-1 down-regulation by RNAi, cytokine deprivation or cytotoxic agents, allowed ABT-737 to efficiently kill diverse cell types, even when Bcl-2 was over-expressed. We conclude that ABT-737 is a highly selective and specific BH3 mimetic compound that should prove highly efficacious in tumors where Mcl-1 is low, or when combined with agents that down-regulate Mcl-1. This hypothesis is currently being tested in vitro in primary human lymphoid malignancies, and in vivo using immunocompetent murine models of lymphoma. ABT-737 also provides strong proof-of-principle that targeting pro-survival Bcl-2 proteins is feasible, but the optimal utility of such BH3 mimetics depends on a thorough understanding of the pathways to apoptosis.

Author notes

Corresponding author