CAD: BH3-mimetic induced cancer cell senescence

Targeted activation of apoptotic signaling pathways is an actively pursued avenue toward new cancer treatments. One mechanism cancer cells deploy to resist cell death is deregulation of the BCL2 family of proteins that control mitochondrial membrane permeablization (MOMP) during cell death. Prosurvival BCL2 family members sequester the prodeath family members and maintain mitochondrial membrane integrity. To promote apoptosis, BH3-only proteins antagonize the activity of the prosurvival BCL2 members. This leads to MOMP and the release of cytochrome c into the cytoplasm initiating active caspase signaling. Development of small molecule BH3-like mimetics that can bind to and inhibit the activity of the prosurvival BCL2 family members can directly activate apoptosis [28]. Clinical trials have demonstrated the potential of these small molecules in the treatment of cancer, particularly those of hematopoietic origin [28].

Although a promising therapeutic not all cancer cell types display acute sensitivity to BH3-mimetics. For example, those of solid tumor origin frequently demonstrate resistance. Despite this resistance to apoptosis, the cells do react to BH3-mimetics responding with partial MOMP and subapoptotic activation of caspases [12, 29]. This stimulation is sufficient to induce DNA damage through the caspase-mediated cleavage of ICAD and subsequent activation of CAD (Fig. 3). The extent of CAD-induced DNA damage promotes the phosphorylation of histone H2AX and p53-dependent changes in gene expression. Notably, the CAD-induced DNA breaks do not contribute to an apoptotic response rather promote the cells to enter into a p53-dependent senescent state [12]. This emerging concept warrants further investigation, as BH3-mimetics progress toward clinical utility. Further appreciation of the molecular requirements for CAD-induced DNA breaks to promote senescence could facilitate approaches to either improve the efficiency of senescence induction or shift the signal toward apoptosis.

CAD: cell fate in prolonged mitotic arrest and slippage

Prolonged mitotic arrest, induced by antimitotic agents, is a therapeutic anticancer approach that can promote cell death and cell cycle arrest in treated tumors. Caspase-dependent DNA damage is apparent in the arrested cells and predominately localizes to telomeric repeats [13, 30-32]. During mitotic arrest, the balance between prosurvival and prodeath BCL2 family members gradually shifts in favor of the prodeath molecules. This results in leakage of cytochrome c from the mitochondria into the cytoplasm that initiates caspase signaling, the activation of CAD, and telomeric DNA damage. Hain et al. [31] recently explored the significance of the CAD-induced telomeric DNA damage, and demonstrated this promoted a DNA-PK-dependent loss of TRF2 from the telomeres, further signaling dysfunction.

Strikingly, CAD has a pronounced impact on cell fate decisions both during mitotic arrest and in the subsequent interphase following slippage [13, 31, 32]. Topham et al. [32] observed in mitotically arrested cells that CAD-induced DNA damage promoted not only death in mitosis but also appear to contribute to the efficiency of the mitotic arrest (Fig. 3). Following slippage from mitotic arrest, Orth et al. [13] demonstrated the extent of CAD-induced DNA breaks directly impact p53 induction and p53-dependent cell fate choices (Fig. 3). These results support a role for CAD-induced DNA damage as a response to stress that is a signal for cell fate decisions in addition to its role in apoptosis. Of note, some cancer cells respond to antimitotic agents by entering into a senescent state following slippage. Although this scenario remains to be examined, Orth et al. [13] suggested that the CAD-induced DNA damage has the potential to promote entry into senescence following slippage.

CAD: skeletal muscle differentiation

Caspase activity has been implicated as a driver in several cellular differentiation events. This activity is linked to both cell autonomous and noncell autonomous signals that direct differentiation. Skeletal muscle differentiation utilizes transient subapoptotic caspase signaling to promote reorientation of cellular signaling and gene expression to propel development at a cell autonomous level [33-35]. Additionally, caspase-dependent apoptosis that accompanies differentiation has been implicated in providing extracellular signals that contribute to myoblast fusion into myotubes [36]. Though the role of caspase is not completely understood, the elucidation of the molecular basis of this signaling has identified several substrates that transduce the caspase signal to promote differentiation, including cleavage of ICAD and activation of CAD.

Differentiating skeletal muscle cells display transient caspase-dependent DNA nicks [11, 37, 38]. Here, partial cleavage of ICAD is sufficient to activate CAD to nick genomic DNA in the absence of whole genome fragmentation. In response to the DNA nicks, the differentiating cells phosphorylate histone H2AX and activate ssDNA break repair pathways to avoid genome fragmentation [11, 39]. The DNA nicks and subsequent repair have been proposed as a mechanism of genomic reprogramming toward the terminally differentiated state (Fig. 3). Indeed, the DNA breaks appear to be restricted to euchromatic regions and further influence the expression of key genes involved in the differentiation process. As caspase signaling appears to be a conserved component of many cellular differentiation processes, the involvement of caspase-dependent DNA breaks may be a conserved module in inducing rapid genomic reprogramming [27]. A function for CAD and as an extension genomic DNA fragmentation in contributing to the noncell autonomous signaling of apoptotic cells to differentiation has not been examined. DNA fragmentation and cytoplasmic release of histones is observed in the apoptotic myoblasts that accompany differentiation [40]. As such it will be an area of considerable interest to determine if these apoptotic events can further contribute to differentiation.

As previously discussed, loss of function of CAD in murine models does not impact development [20-22]. It is reasonable to suggest this may indicate compensation by additional caspase-directed nucleases to support development in the absence of CAD. Identification of these nucleases may assist in developing a more comprehensive picture of the roles that caspase-directed DNA breaks play in directing differentiation.

Cell fate translation of CAD-induced DNA breaks

How the induction of DNA breaks by CAD can promote nonapoptotic cell fate transitions remains to be clarified. Indeed, through activation of the DDR, in particular p53, CAD would exert potent influence on cell fate decisions including proliferation, senescence, and apoptosis. Furthermore, the formation and repair of DNA breaks can lead to alterations in the surrounding chromatin environment influencing local gene expression [50, 51]. As such both mechanisms could account for transitions in global signal transduction and gene expression that propel cell fate transitions.

The DDR response to CAD-induced DNA breaks in cells treated with either BH3-mimetics or antimitotic agents ultimately leads to a p53 response and a cell fate decision. It is tempting to draw similarities between these two cellular responses as both lead to a p53 response, however, a conserved DDR response to the breaks is not apparent. The phosphorylation of H2AX in BH3-mimetics treated cells can be either ATM- or JNK2-dependent, depending on the cell lines and treatment [12, 29]. In mitotically arrested cells, the phosphorylation of H2AX in response to CAD-induced breaks is DNA-PK-dependent [31]. Of note, the phosphorylation of H2AX that is observed during skeletal muscle differentiation has also recently been suggested to be DNA-PK-dependent [52]. These observations are clearly an early survey into the DDR response to the CAD-induced DNA breaks. More thorough characterization is warranted to fully appreciate how the CAD-induced DDR activation can orchestrate cell fate decisions.

Localization of CAD-induced DNA breaks during these nonapoptotic processes is also of considerable importance. Although apoptotic DNA breaks are largely thought to occur at random across the genome, more recent sequence-based mapping of the breaks has indicated enrichment at specific sites [53]. Indeed, CAD-induced DNA breaks that appear during prolonged mitotic arrest localize to telomeric regions. This telomeric localization likely reflects chromatin accessibility of mitotic cells, which would restrict CAD's activity [31]. The specificity of CAD-induced DNA breaks in an interphase cell is apparent during skeletal muscle differentiation, here the breaks appear restricted to euchromatic regions. Further localized breaks were mapped within the p21 promoter. [11]. These results suggest that specific regions can be targeted by CAD, particularly during differentiation. Genome wide mapping of the CAD-induced breaks both during differentiation and BH3-mimetic-induced senescence will provide important insight into the molecular function of these breaks.

The possibility of site-specific DNA breaks, particularly in promoter elements, could point to a function for CAD in regulating gene expression. How DNA breaks directly impact local gene expression has been an area of extensive study, and a clear picture has not emerged. In many instances, repression of gene expression has been indicated with the induction of specific breaks distal or within gene bodies [54, 55]. Here, DSB generation delays transcription through ATM-dependent limiting of RNA polymerase II decondensation of distal chromatin that is rapidly recovered with DSB repair [54]. Conversely, DNA breaks have also been linked to activating gene expression. Topoisomerase-IIB-induced DNA breaks are involved in gene activation, supporting transcription initiation and elongation, following a diverse range of stimulation [50, 56, 57]. These breaks lead to the activation of ATM, DNA-PK, and PARP-1 that stimulated the repair of the DNA break and also contributed to gene activation. How CAD-induced DNA breaks may regulate gene expression remains to be fully elucidated. CAD-induced DNA breaks within the p21 promoter during skeletal muscle differentiation do correspond to gene activation, whether the breaks directly contribute to the expression of p21 during differentiation remains to be examined.

Regulation of CAD-induced DNA breaks

It is clear that the activation of CAD attacks genome stability both in apoptosis and in nonapoptotic events. The absence of progressive DNA fragmentation in nonapoptotic events supports the limited activation of CAD and a robust DNA damage repair response. Importantly, the nuclease activity of CAD does not directly support continued self-activation. As discussed, several mechanisms can limit the activity of CAD, in particular the extent of cellular ICAD proteolysis. ICAD long and short isoforms, as well as the proteolytic fragments, can disassemble the active CAD dimer [58]. Incomplete proteolysis of total cellular ICAD with continued protein production could effectively limit CAD activity to avoid whole genome degradation. Indeed, rapid removal of whole cellular ICAD leads to unfettered CAD activation sufficient to drive apoptosis [59]. These results achieved by auxin degron tagging and regulation of endogenous ICAD appear to contrast targeted proteolytic cleavage. However, the kinetics of CAD activation accomplished in this model could more closely resemble exposure to an acute genotoxic agent. Here, the extent of DNA damage may accumulate more rapidly than can be repaired activating apoptosis. These results highlight the importance of ICAD in promoting controlled CAD activation.

Transient and partial cleavage of ICAD, at D117, is observed in skeletal muscle differentiation. Cleavage at this site has been shown to be necessary and sufficient for CAD nuclease activation during apoptosis and is required for skeletal muscle differentiation [11, 60]. The significance of this partial cleavage remains to be elucidated. In particular, how it may influence the activity of CAD and if it relates to CAD-induced ssDNA breaks observed during differentiation and apoptosis [15, 39].

These observations position ICAD as a key regulator of CAD activity during nonapoptotic events. Allowing regulated activation to prevent whole genome fragmentation and limiting feedback from the DNA damage to further activate caspase signaling. Post-translation modifications (PTMs) to ICAD, in addition to caspase cleavage, could further contribute to the regulation of CAD. Extensive characterization of ICAD PTM's such as phosphorylation, ubiquitination, or acetylation has not been explored. Notably, phosphorylation has the potential to influence caspase substrate cleavage and can prevent cleavage at specific sites [61]. Understanding if such regulation of ICAD by PTM modifications is in play during subapoptotic caspase signaling could provide further explanation into the apparent regulated activity of CAD.

Oncogenic consequences of CAD-induced DNA breaks

Apoptosis is a central tumor-suppressive mechanism, both in restraining tumor formation and in therapy [16, 28]. Apoptotic signaling, however, has the potential to promote an oncogenic cell fate, particularly as an outcome of cancer treatment. CAD exemplifies this emerging paradoxical function attributed to apoptotic proteins. As discussed, CAD is linked both as a tumor suppressor and a contributing factor in instances of oncogenic apoptotic signaling. Ichim and Tait [41] recently proposed that one contributing consideration for this paradox maybe the stage in tumor development. As such the functional outcome of CAD-induced DNA damage changes as tumorgenesis progresses. Further understanding of this process may support improved therapeutic approaches.

To better define CAD's role in oncogenic transformation, it will be important to appreciate how the DNA damage translates into cell fate decisions and how this damage may progress into oncogenic mutations. The cell fate translation of CAD-induced DNA damage under conditions of subapoptotic caspase signaling implicates p53 signaling [12, 13]. Compromised p53 function may limit both the apoptotic and senescence response to CAD-induced DNA damage. Indeed, the mutagenic potential of CAD-induced DNA damage is most apparent in senescence impaired cells [29].

Appropriate repair of CAD-induced DNA breaks would limit an oncogenic outcome. However in cancer cells, mutations in key DDR genes are frequently observed and may further potentiate the introduction of new oncogenic mutations [62]. Additionally, the accompanying subapoptotic caspase activity could transiently promote loss of function of proteins involved in nonhomologous end joining NPM (NHEJ) and homologous recombination (HR) [63]. In both scenarios, repair of CAD-induced DNA breaks by error prone backup DNA repair pathways could contribute to the oncogenic genomic changes.

Final comment

CAD-induced DNA damage at the terminal of subapoptotic caspase signaling is emerging as potent cell fate determinate. Subapoptotic caspase activity is emerging to regulate a range of biological events. As such it is of interest to gain further understanding of how broadly this activity of CAD may contribute to cell fate adoption. In particular if it may contribute to the DNA damage seen within aging stem cell populations [64, 65]. It is clear that the activity of CAD can serve as a source of genome instability. To this end, utilization of CAD to promote cell fate changes may be subject to erroneous consequences.