Safety control for apoptotic irreversibility

Historically defined by Kerr in 1971 (1), apoptosis refers to programmed cell death participating in highly regulated processes of tissue remodeling or renewing involved during development, normal cell turnover, and cell elimination following injury. Apoptosis is an elaborated form of cellular suicide whereby cells sacrifice themselves for the well-being of the whole organism by dying in a quiet manner, without undergoing cell lysis. The main challenge of apoptosis is to drive it to completion without endangering the neighboring tissues or the host organisms by toxin release or abnormal cells (with a modified genome), escaping this process. Therefore, there is a need for direct sensors to ensure the correct completion of the apoptotic process. The work of Solier et al. (2) contributes to our understanding of such complex control mechanisms by demonstrating an interplay between the heat shock protein chaperone (HSP90α) chaperone and the DNA-dependent protein kinase (DNA-PK) during apoptosis.

In a previous study, Solier and Pommier (3) showed that, in apoptotic cells, DNA repair proteins such as the ataxia telangiectasia mutated (ATM), the checkpoint control kinase (Chk2), and DNA-PK are found together with the phosphorylated forms of the histones H2AX and H2B at the nuclear periphery inside the nucleus, in structures they referred to as the “apoptotic ring.” Here they demonstrate that the phosphorylated form of HSP90α colocalizes with DNA-PK in the apoptotic ring. A fraction of HSP90α translocates to the nucleus upon stress (4), and tightly interacts with histones (5), inducing a condensed state of chromatin (6). HSP90α is a client of DNA-PK kinase (7, 8), and, as such, its modification at this location could be an indirect consequence of their colocalization. However, the authors clearly validate that HSP90α is a chaperone of DNA-PK, and, although it is modified during apoptosis, it contributes to the stabilization of DNA-PK. By using HSP90α inhibitors and Hsp90α gene silencing, they further demonstrate that HSP90α acts as a cofactor for DNA-PK stability and activity during apoptosis.

A link between DNA damage and apoptosis is supported by a large body of literature. However, the rules that govern this association have remained unresolved. DNA-PK is the central enzyme of nonhomologous end-joining DNA repair, the most efficient repair pathway of DNA double-strand breaks in mammalian cells. The control of its activity by HSP90α during apoptosis highlights the possible role of HSP90α in DNA damage repair. HSP90α inhibitors and protein depletion by RNA silencing strongly inhibit double-strand break repair after irradiation (8, 9). Moreover, HSP90α colocalized with main repair enzymes, at DNA repair foci formed after irradiation (8), and at the apoptotic ring during apoptosis (3), suggesting that it could play a direct role in the regulation of DNA repair. During apoptosis, repair of fragmented DNA is inhibited by several mechanisms, such as specific enzymatic cleavage and/or degradation of the key enzymes of the double strand break repair pathways (53BP1, MDC1, and PARP). For example, the C-terminal apoptotic fragment of PARP-1 loses its DNA-dependent catalytic activity upon cleavage with caspase 3 (10). In such a context of repair inhibition, what could be the consequence of stabilizing DNA-PK? It is likely that DNA-PK participates in several mechanisms other than DNA repair.

DNA-PK phosphorylates numerous cytoplasmic or nuclear proteins that are not directly involved in DNA repair and could play a similar regulatory role in apoptosis by modifying the activity of key enzymes (namely HSP90α) after being activated by DNA fragmentation. Phosphorylation is the most frequently occurring posttranslational modification of HSP90α. Interestingly, many kinases that regulate the HSP90α phosphorylation status, including DNA-PK (2, 11), are HSP90α clients at the same time. In general, HSP90α is phosphorylated at multiple sites located in distinct regions of the dimeric protein. Phosphorylation of HSP90α at Thr22, Thr90, Tyr24, and Tyr300 modulates its ability to chaperone selected clientele (12–15), supporting the notion that dynamic phosphorylation/dephosphorylation events represent a key regulatory mechanism for chaperone function (11). HSPs function as molecular chaperones in regulating cellular homeostasis and promoting cell survival (15, 16). Thus, Thr-7 phosphorylation of HSP90α by DNA-PK could play a major role in disabling its cytoprotection activity that prevents apoptosis (17).

The lesson biologists have repeatedly learned with complex systems is that most of the activities are in a meta-stable equilibrium, with regulatory control loops to prevent accidental events or to shift to irreversible processes if the balance is lost. Repair and apoptosis are two common events in life that have to be tightly controlled to prevent tissue destruction or modification. Apoptosis is triggered when the extent of damage exceeds the repair ability of the cell. Meanwhile, as DNA damage and fragmentation increase considerably during this process, there must be a “safety” control to prevent reversion of the apoptotic process when fragmentation of the genome has been instigated. Essentially, cells surviving a transient and reversible apoptotic response acquire permanent genetic changes and undergo oncogenic transformation (14). The dual activity of HSP90α described by Solier et al. (2), associating its phosphorylation by DNA-PK at the H2AX apoptotic ring and the stabilization of the kinase during this process, could be crucial in ensuring complete HSP90α phosphorylation and consequent inhibition of its cytoprotective activity, given that the apoptotic ring is present (Fig. 1). Therefore, the DNA-PK/HSP90α regulatory loop likely plays a key security role in shifting the apoptotic process to irreversible death and preventing cells with damaged chromosomes from escaping this process.

Fig. 1.

Schema of regulation pathways with positive controls (→) and negative controls (┤) involved in apoptosis.