Abstract
Bcl-2, originally identified as a universal inhibitor of apoptotic cell death, has since been implicated in suppressing autophagy, the cell’s quality control mechanism. Our recent study demonstrates that the anti-autophagic aspect of Bcl-2 can function as a promoter of oncogenic growth, independently of its role in apoptosis signaling. It is likely that the increase in Bcl-2 often seen in breast and other cancers might render cells error-prone by blunting autophagy, while concomitantly keeping damaged cells alive. The outcome of such a ‘double hit’ of Bcl-2 may synergistically promote tumor growth and increase the chance of cancer development and drug resistance.
The initial description of Bcl-2 focused on its control of the intrinsic apoptotic pathway. However, more recent investigations have shown that Bcl-2 also carries out daily jobs in multiple homeostatic pathways including mitochondrial metabolism, ER stability, cell cycle checkpoints and autophagy. Since the loss of homeostatic balance is an underlying mechanism in the pathogenesis of numerous pathologies, including cancer, these properties of Bcl-2 may have considerable untapped potential for unveiling the ‘real face’ of cancer, as can be envisioned in our recent study on the anti-autophagic aspect of Bcl-2.
The Bcl-2 proteins are structurally characterized by a hydrophobic groove formed by the BH1 to BH3 domains, which are responsible for interaction with the BH3-domain-containing Bcl-2 family members, BH3-like proteins such as Beclin 1, as well as BH3-mimetics, like ABT-737. Structural alignments of the BH3 motif of Beclin 1 and the pro-apoptotic Bcl-2 proteins reveal a highly conserved topology and groove contact sites despite their overall sequence variability, leading to the conclusion that Beclin 1 is a putative BH3-only protein. However, no evident apoptogenic effects by Beclin 1 have been found in vivo. Moreover, Bcl-2 complexed with Beclin 1 remains anti-apoptosis competent, suggesting that autophagy and apoptosis may be independently regulated by Bcl-2 in parallel pathways. Along this line, our study indicates that the anti-apoptotic and anti-autophagic aspects of Bcl-2 function are genetically separable, since a deletion of the N-terminal α1 helix abrogates the constraint on apoptosis but not the restriction of autophagy. We thus propose that Bcl-2 antagonizes Beclin 1 in an antiapoptosis-independent manner. Notably, our previous study of the γ-herpesvirus 68 Bcl-2-Beclin 1 interaction showed that an equivalent mutation of the α1 helix in vBcl-2 blocks its association with Beclin 1, but not with Bax and Bak, highlighting a distinct role for the α1 helix in mediating anti-apoptotic and anti-autophagic interactions of viral and cellular Bcl-2s. Since vBcl-2 lacks a long flexible loop between α1 and α2, the N-terminal α1 helix that is outside of, but contiguous with, the hydrophobic cleft may contribute to proper folding of the protein to form a compact anti-autophagic groove. The striking difference in the ligand-binding specificity of cellular Bcl-2 and its viral homolog suggests that they may undergo different conformational changes when bound to distinct BH3 ligands to achieve functional specificity. Thus, our data, in conjunction with recent findings, provide a molecular explanation for the distinct modes of interaction and the parallel regulation of autophagy and apoptosis by Bcl-2 in living cells.
Overexpression of Bcl-2 is common in various human cancers and contributes to increased resistance to chemotherapy and/or aggressive cancer phenotype. Despite what we know regarding the action of apoptosis inhibition by Bcl-2, the mechanisms underlying Bcl-2-mediated oncogenesis are still not clear. Notably, Bcl-2 amplification on its own would not suffice for transforming cells; rather, it drastically complements the actions of other tumorigenic changes like Ras or Myc. If anti-apoptosis features are predominantly the functions of Bcl-2 in cancer as was generally thought, one might intuitively expect that a tumor with low levels of Bcl-2 would be inherently more sensitive to spontaneous or drug-induced apoptosis and have a better prognosis than the Bcl-2-proficient tumors; however, this does not seem to be universally true. In some malignancies Bcl-2 expression has either no discernible impact on prognosis or is even associated with better clinical outcome, highlighting the complexity of Bcl-2 functionalities beyond apoptosis regulation. Intriguingly, when bcl-2 is silenced in some breast cancer and leukemic cells, these cells undergo autophagic cell death, rather than apoptosis, suggesting that the anti-autophagic property of Bcl-2 is, at least in some contexts, eminently exploited in cancer. This view is further reinforced in our recent work, which shows that the anti-autophagic Bcl-2 can promote tumor cell growth independently of its anti-apoptosis activity. Using a mutant of Bcl-2 that no longer blocks apoptosis but still has the ability to attenuate autophagy, we show that increased Bcl-2 expression continues to foster malignant growth of breast cancer cells, challenging the longstanding perception that it is the anti-apoptotic effect of Bcl-2 that dictates tumor growth. The pro-oncogenic action of this mutant is further linked to the inhibition of autophagy, through its effect on Beclin 1, which has been implicated in autophagy. Indeed, the growth-promoting effect of the Bcl-2 mutant was easily satisfied when Beclin 1 is present or overexpressed. Although this is not an unexpected finding, it nevertheless suggests that attenuation of Beclin 1 can provide a novel strategy for Bcl-2 to elicit oncogenic stress, again highlighting the importance of autophagy in countering malignancies.
Considering that autophagy represents, intrinsically, a survival mechanism in response to metabolic stresses, a pertinent question is why cancerous cells go to great lengths to suppress such an important survival mechanism? This question might point to the multifaceted effects of autophagy on homeostasis and its threshold regulation. Autophagy more likely resides in a biologically latent state in the absence of stimuli, although the exact molecular nature of this latency remains unclear. When cells experience various stress conditions, autophagy becomes activated to trigger responses that alleviate stress, take care of the damage (e.g., organelle, protein, and/or genome damage) or else eliminate the affected cells via so-called autophagic cell death. Clearly, such protective autophagic responses do not facilitate tumor expansion. Accordingly, tumor cells that hijack Bcl-2 may prevent this signal reaching the critical threshold before any anti-tumor effects occur, while still satisfying the needs of tumor cell survival. It is important to note that many tumors are able to evade apoptotic death by mutating downstream apoptosis effectors, such as Apaf-1 and caspases, thus bypassing the need for Bcl-2 expression in order to thwart apoptosis, and rending its anti-apoptosis action dispensable. In this respect, the autophagy-restraining activity of Bcl-2 may be even more dominant over its anti-apoptotic effects in oncogenic progression.
An issue worth considering is whether autophagy dysfunction can represent a driving cause of malignancies or serve as a ‘second hit’ in the process of cancer metabolic transformation. This question will be better answered with our growing understanding of the complexity of this pathway and its associated machinery. Nonetheless, our work clearly establishes the importance of autophagy signaling in Bcl-2-associated tumor growth, and highlights the potential value of pharmacological targeting of this pathway in certain human cancers.