A Survivor Hits the Breaks

INTRODUCTION

Bcl-2 is the founding member of a family of proteins that regulate apoptosis by controlling the integrity of the mitochondrial outer membrane. In this issue, Zu, et al. propose another function for Bcl-2; the inhibition of DNA repair by non-homologous end joining.

There is a native American story about a man who petitions the gods for eternal life, which is granted. But in the way of such things he had forgotten to stipulate conditions, and he aged and aged until, shrunken and wrinkled, he became a grasshopper. Aside from the improbable entymological phenomenon, there are parallels here with a paper in this issue (Zu, et al., 2008), a set of observations that raise puzzling and unexpected questions about a form of cellular immortality and its links to DNA repair.

Zu et al. (2008) describe an ability of Bcl-2, the archetypal anti-apoptotic protein to block DNA repair by the process of non-homologous end-joining (NHEJ). The effect is distinct from that of the anti-apoptotic mechanism, which involves interactions with other proapoptotic Bcl-2 family proteins to prevent the permeabilization of the mitochondrial outer membrane and thereby prevent apoptosis via the mitochondrial pathway (Green, 2007). This anti-apoptotic activity involves four “Bcl-2 homology” regions in the molecule, but only two of these are involved in the inhibitory effect on NHEJ (Figure 1).

Figure 1.

Bcl-2 blocks DNA repair by non-homologous end joining (NHEJ). The model proposed by Zu, et al., 2008, suggests that nuclear Bcl-2 binds and sequesters the abundant Ku70 protein to prevent assembly of the NHEJ complex on DNA double strand breaks. This binding and inhibition of NHEJ is dependent on the BH1 and BH4 domains of Bcl-2 (shown in the model in blue, BH4 on the left) and not the BH2 and BH3 domains (shown in RED) which are also required for the anti-apoptotic effects of the protein.

DNA double strand breaks (DSBs) are a particularly deleterious form of DNA damage and if unrepaired can result in cancer-causing mutations or promote aging. DNA DSBs occur as a result of oxidative metabolism, DNA replication, V(D)J recombination during immune system maturation or can arise from exogenous agents such as ionizing radiation. The repair of DNA DSBs is achieved through two specific DNA repair pathways, non-homologous end joining (NHEJ) and homologous recombination (HR) repair (Wyman and Kanaar, 2006). HR is cell cycle dependent and requires a homologous sister chromatid to ensure error-free DNA repair. In contrast, NHEJ facilitates direct repair of damaged non-homologous DNA ends. NHEJ directly modifies the ends of a DSB to be compatible for ligation, and is therefore the predominant pathway for repairing DSBs in non-cycling mammalian cells, where a homologous template is not readily available (Phillips and McKinnon, 2007). This direct processing of DNA ends is an effective way to rejoin DNA breaks and provides the cell with a ready means to maintain cellular homeostasis after DNA damage.

NHEJ involves a set of core components comprising the Ku complex (a heterodimeric protein complex of Ku70 and Ku80), the DNA-dependent protein kinase (DNA-PKCS), Cernunnos/XLF and the DNA ligase IV/Xrcc4 complex, which together coordinate rejoining of broken DNA ends (Wyman and Kanaar, 2006). Mice in which any of these have been inactivated are radiosensitive, immune deficient and prone to cancer. The initial step in NHEJ is recognition of the DNA strand breaks by the Ku complex, a toroidal protein complex that can access and envelope DNA at the break site. This provides a DNA-protein interface that can assemble the other NHEJ components and serves to enhance the efficacy of DNA repair. Thus, the Ku complex is critical for efficient NHEJ and its perturbation can compromise this DNA DSB repair pathway. It is the Ku70 component of this complex that appears to be the target of Bcl-2 in the nucleus to inhibit NHEJ. However, Ku70 is highly abundant, and it remains to be seen if there is sufficient Bcl-2 in the nucleus to effectively sequester this component, or if the complex with Bcl-2 takes on dominant inhibitory activity.

These observations may help to explain some intriguing phenomena, such as why cancers that are caused by constitutive expression of Bcl-2, such as follicular lymphoma, can be treated by radiotherapy. It may also explain the curious expression pattern of Bcl-2 during the development of B cells; pro-B cells express high levels of Bcl-2, but when V(D)J rearrangment commences in the pre-B cell stage, Bcl-2 disappears and does not reappear until the B cells mature. That is, Bcl-2 is gone when NHEJ is required for DNA rearrangement essential to B cell development.

Zu et al. (2008) also suggest that this explains oncogenesis by Bcl-2, which contributes not only to anti-apoptosis, but perhaps also to genomic instability due to inhibition of NHEJ. Knockout of any component of NHEJ can indeed result in spontaneous cancers (in conjunction with another mutation, notably loss of p53 function). Expression of Bcl-2 as a transgene in mice generally does not, unless combined with another oncogene, such as c-Myc. DNA DSBs cause cell cycle arrest and apoptosis via engaging p53, but tumors expressing Bcl-2 tend not to lose p53 function (even if Bcl-2 would block the apoptosis, one would expect that loss of p53 would avoid the cell cycle arrest). These observations are not entirely consistent with a model that predicts, as Zu and colleagues do, that Bcl-2 should promote cancer by both blocking NHEJ and blocking apoptosis downstream of p53.

Therefore, perhaps, while the inhibition of NHEJ by Bcl-2 is a robust phenomenon, it may not contribute to oncogenesis. And a deeper issue is this: why would a mechanism whereby cell survival is linked to inhibition of NHEJ be sustained in the face of selection (especially if it promoted cancer)? We propose another scenario. Suggestive evidence exists that retroviral insertion depends on NHEJ (Izsvak et al., 2004; Skalka and Katz, 2005). If retroviruses were to infect a cell that happens to express Bcl-2, any defensive apoptotic response to the virus would fail and the infection would succeed. By blocking NHEJ in cells expressing Bcl-2, retroviral integration would be reduced. Of course, this is readily testable.

This may not be restricted to retroviruses. Early observations on Bcl-2 suggested that this molecule can cause cell cycle arrest (or delay), although for the most part this was likely to be an artifact of cell survival (as cells that would have died due to limiting growth factors persisted, but could not enter cycle until their metabolic reserves were fully restored (Rathmell, et al., 2000). But one of these observations was intriguing (and unlikely to be explained by the same artifact); fibroblasts that were transduced with an expression construct for Bcl-2 failed to grow (Huang, et al., 1997). The experiment required stable expression of the construct and expression of a selection marker. If Bcl-2 blocked NHEJ involved in integration of the transduced DNA, this would explain these results. And if so, another aspect of these observations becomes particularly intriguing: Bcl-2 mutated at tyrosine 28 (Y28F) did not show this inhibitory effect, while sustaining anti-apoptotic activity. Y28 is in the proximity of one of the regions required for Bcl-2 interaction with Ku70 and inhibition of NHEJ. It will be interesting to determine if this residue is required for the inhibitory effects.

The ability of Bcl-2 to modulate DNA repair may extend beyond NHEJ as it apparently can suppress another DNA repair pathway, mismatch repair (MMR) (Hou et al., 2007). Like NHEJ and the interaction of Bcl-2 with Ku, the BH4 domain was found to be required for Bcl-2 to modulate MMR, in this case by binding the key MMR component MSH6. Is it likely that Bcl-2 not only acts at the mitochondria to prevent apoptosis, but also restricts DNA repair by these different mechanisms? Bcl-2 has also been implicated in other cellular functions as well, including regulating the IP3-receptor controlling calcium efflux from the endoplasmic reticulum, and regulating autophagy through binding of Beclin-1. But this is what Bcl-2 does: sequesters proteins to control intracellular processes. Perhaps we are only seeing a part of a network of biological responses that are integrated by Bcl-2, acting as a node. How these might make sense in a physiological setting is a challenge facing us.

Of course, many more questions arise from these considerations. What regulates this activity of Bcl-2 and its nuclear localization? Do other anti-apoptotic Bcl-2 family members, such as Bcl-xL, Mcl-1, and A1, block NHEJ (and, perhaps viral integration)? Do anti-apoptotic viral homologs of Bcl-2 block apoptosis without interfering with NHEJ? Immortality, it seems, comes at a cost, but understanding why we pay the price will be interesting.