DSBs are sensed by the heterotrimeric Mre11/Rad50/NBS1 (MRN) complex which serves as an activation platform for the DNA damage checkpoint kinase ATM [16, 17]. In undamaged cells, quiescent ATM exists as homodimers, which dissociate into active monomers upon activation [18]. Inactive ATM auto-phosphorylates at Ser1981 and subsequently phosphorylates histone variant H2AX (γH2AX) proximal to the DNA break. Phosphorylation of H2AX leads to further enhancement of ATM binding and allows the DNA damage signal to spread along the chromatin [16]. Activated ATM phosphorylates hundreds of proteins including p53, c-Abl, BRCA1, and NBS1 which further propagates the DNA damage signal into numerous cellular pathways and processes [19].

ATM coordinates DNA repair primarily via homologous recombination (HR) which is a relatively error-free method of DSB repair that uses a homologous sister chromatid as a template for repair [20]. ATM is responsible for recruitment of HR proteins such as Mre11 and CtIP (CtBP Interacting Protein) which leads to creation of 3′ single-stranded nucleotide overhangs, followed by replication protein A (RPA) and RAD51 nucleofilament formation [21, 22]. The nucleofilament mediates the homology search, strand invasion, and formation of the D-loop, leading to promotion of DNA synthesis. A full review of ATM involvement in DNA repair is beyond the scope of this work; however, a recent review provides excellent discussion on this topic [23].

An important effector of ATM signaling is checkpoint kinase 2 (Chk2), which is phosphorylated at residue Thr68 by ATM following DSB formation [24]. Once activated, Chk2 is thought to dissociate from sites of damage and disperse as a monomer throughout the nucleus to phosphorylate over 20 substrates involved in apoptosis, gene transcription, and cell cycle progression [25]. In the presence of DSBs, Chk2 arrests the cell cycle by several mechanisms including through phosphorylation and cytoplasmic-translocation of Cdc25C phosphatase [19]. In the cytoplasm, Cdc25C can no longer dephosphorylate and activate the cyclinB1/cyclin-dependent kinase 1 (Cdk1) complex, maintaining Cdk1 in an inert form and preventing progression into mitosis [26]. Chk2 also phosphorylates p53 to promote p21 accumulation and G₂/M arrest. Through similar molecular mechanisms, Chk2 activation can also lead to G₁/S arrest in the presence of DSBs.

In the classical model, ATM and ATR were thought to act on different types of DNA breaks: ATM was activated in response to DSBs, whereas ATR was thought to act in response to single-strand breaks (SSBs). Studies now show that in fact, ATR responds to single-stranded regions of DNA generated at stalled replications forks, but it also appears to be activated in the presence of single-strand DNA (ssDNA) generated by processing of DSBs [27]. The activation of ATR in the presence of DSBs requires both ATM and the nuclease activity of Mre11 to generate ssDNA coated with RPA that is necessary for ATR recruitment [27, 28]. ATR, like ATM, has a large number of molecular substrates. ATR appears to exert its effects on cell cycle arrest primarily through the phosphorylation of checkpoint kinase 1 (Chk1) [29]. Chk1 has an effect on S-phase progression through inhibition of Cdc25 phosphatases and regulation of cyclin-dependent kinases [30]. Chk1 also plays a role in preventing cellular progression into mitosis with unrepaired DNA damage through sequestration of Cdc25C into the cytoplasm and degradation of Cdc25A which maintains Cdk1 in its inactive state resulting in G₂/M arrest [30, 31]. Importantly, it appears that IR-induced G₂/M-phase arrest involves the cooperation of ATR and ATM, since double deletion of ATR and ATM eliminates nearly all IR-induced delay [32]. While the activation of ATM occurs essentially irrespective of the cell cycle phase, ATR is primarily activated in the S and G₂ phase, suggesting that ATR activation is regulated by ATM in a cell cycle-dependent manner [27].

GBM is a highly radioresistant tumor that tends to recur locally despite aggressive surgery and chemo-radiation. One explanation for this nearly eventual recurrence following an initial response to treatment is that GBMs appear to exhibit upregulation of the DNA damage response. This is supported by an analysis of multiple human glioma cell lines by Bartkova and colleagues, in which constitutive activation of the DDR pathway and upregulation of the ATM-Chk2 axis was observed at increased levels in GBM when compared to lower grade gliomas [33].

Other investigators have suggested that high rates of local failure may be attributed to failure to sterilize a subpopulation of radioresistant GBM cancer stem cells (CSCs), which are cells that are capable of repopulating a GBM tumor in vivo. In one study, Bao and colleagues demonstrated that CD133-positive cancer stem cells exhibited radioresistance due to preferential activation of the DDR and an increase in DNA repair capacity when compared to CD133-negative cells [34]. This radioresistance was overcome with an inhibitor of Chk1 and Chk2, suggesting that targeting DDR pathways in CSCs could provide an important therapeutic strategy for malignant brain cancers [34].

Recently, it was demonstrated that primary GBM cells grown in stem cell conditions exhibited increased radioresistance relative to differentiated cell populations originating from the same parental tumor [9]. The stem-like cells exhibited enhanced G₂/M checkpoint activation and DSB repair following IR relative to their differentiated tumor cell counterparts. The observed radioresistance was capable of being overcome by treatment with the ATM inhibitor KU-55933, which produced potent radiosensitization of the GBM CSCs [9]. ATR and Chk1 are also expressed at high levels in GBM stem-like cells under basal conditions, which exhibit rapid Chk1 activation in response to IR [9]. Combined inhibition of PARP and ATR using olaparib and VE-821 resulted in profound radiosensitization of the GBM CSCs, which exceeded the effect of ATM inhibition alone [9].

There have been many other studies demonstrating the efficacy of ATM and ATR inhibitors in malignant glioma models both in vitro and in vivo. One group found that ATM inhibition with KU-55933 led to improved radiosensitization when compared with a DNA-PK inhibitor in glioma stem cells and led to prolonged survival for mice with intracranial xenografts [35]. The ATM inhibitor KU-60019 was also shown to cause radiosensitization of mutant p53 human glioma cells when delivered via intracranial convection-enhanced delivery [36]. ATR inhibition with VE-821 was shown to be effective in inhibiting the growth of cancer cells in 3-D spheroid models of glioblastoma [37]. These studies, among others, provide evidence that inhibition of ATM and ATR may represent an effective therapeutic strategy moving forward for targeting even the most radioresistant subpopulations of GBM.