Abstract
Ataxia-telangiectasia mutated (ATM) is a pivotal protein with versatile kinase activity that responds to DNA damage. While its well-established role as a DNA repair protein is widely recognized, the understanding of its noncanonical functions in ovarian cancer remains limited. Numerous studies have investigated the potential of targeting ATM for ovarian cancer treatment. In addition to its involvement in homologous recombination repair (HRR), an increasing body of research suggests that ATM plays a role in cellular metabolism and adaptive immunity. This review focuses on the current evidence and provides a perspective on how targeting ATM in ovarian cancer can address HRR-deficient genotypes, influence macropinocytosis, and enhance immune checkpoint blockade (ICB) therapy. It underscores the diverse avenues through which targeting ATM is a potential tailored treatment for ovarian cancer.
Ovarian cancer is the deadliest gynecological malignancy worldwide (1). The Food and Drug Administration (FDA) of the USA has approved poly(ADP-ribose) polymerase (PARP) inhibitors for use in ovarian cancer therapy. PARP inhibitors induce synthetic lethality in cancer cells deficient in the homologous recombination (HR) pathway, specifically those harboring BRCA1/2 mutations (2). However, the resistance of tumors to PARP inhibitors limits the efficacy of these drugs (3-5). Therefore, it is crucial to develop new therapeutic strategies for ovarian cancer.
Ataxia-telangiectasia mutated (ATM), located at chromosomal region 11q22, is a tumor suppressor gene that encodes a serine/threonine protein kinase. ATM, a member of the phosphatidylinositol 3′-kinase (PI3K)-related kinase (PIKK) family, is composed of the Tel1/ATM N-terminal motif (TAN), focal adhesion targeting (FAT), and PI3/PI4-kinase domains (6). It also plays a critical role in the maintenance of genome integrity (7). Moreover, targeting ATM with inhibitors, such as KU-60019 for renal tumors and AZD0156 for colorectal cancer (8) (clinicaltrials.gov), has emerged as an essential strategy for cancer treatment.
ATM deficiency or ATM inhibition impairs the ATM-mediated DNA repair pathway and results in an HRR-deficient genotype in ovarian cancer cells. Furthermore, treatment of tumor cells with PARP inhibitors leads to synthetic lethality in ATM-deficient cells and conditional synthetic lethality in ATM-proficient cells (Figure 1A). The loss of ATM induces metabolic reprogramming by activating macropinocytosis. Blocking macropinocytosis causes tumor cell death due to metabolic vulnerability in ATM-deficient cells. In addition, the combination of micropinocytosis inhibitors with ATM inhibitors results in the death of ATM-proficient cells (Figure 1B). Interactions between ATM-mediated double-strand break (DSB) responses and immune checkpoints have been identified. The use of immune checkpoint inhibitors may selectively suppress targeted tumors, which is particularly influenced by ATM status (Figure 1C). In this review, we describe the potential role of ATM in treating ovarian cancer in terms of HRR deficiency, metabolic changes, and tumor immunity, which rely on ATM status and activity (Figure 1).
Targeting ATM Based on the HRD Genotype or by ATM Inhibitors
Homologous recombination repair (HRR) is a highly demanding process in which DNA double-strand breaks are repaired to ensure correct genetic information and maintenance of genome stability. ATM plays a critical role as a regulator of DNA repair and cell cycle progression by directly phosphorylating BRCA1, p53, and Chk2 upon DNA damage (6). Upon DNA double-strand breaks, Mre11, Rad50, and Nbs1 (MRN complex) binds to the damage site (9). Then, the inactive dimeric form of inactive ATM is stimulated by DNA ends, which results in the formation of the monomeric of active ATM through autophosphorylation at serine 1981 (10). The activated ATM further phosphorylates its downstream substrates, including Chk2 at Threonine 68 and p53 at Serine 15, as well as H2AX in adjacent nucleosomes (11, 12). Importantly, ATM mediates the activation and stabilization of the p53 protein in response to DNA damage by phosphorylating both p53 itself and many of its regulators, such as p21 (13). As p21 stands out as a primary target under p53 regulation, the ATM/p53/p21 pathway plays a pivotal role in determining the cell fate choice between apoptosis and senescence in reoxygenated hematopoietic progenitor cells (14) and modulating the cell-cycle checkpoint under chronic γ-irradiation (15) and senescence in human umbilical vein endothelial cells following exposure to oxidative stress (16). ATM regulates cyclin-dependent kinase 1 and 2 (CDK1/2) through the inactivation of CDC25 phosphatases. Furthermore, ATM regulates the formation of the CDK1–cyclin B complex and the recovery from DNA damage-induced G2 arrest (13). ATM is a multifunctional protein that plays a role in cell cycle arrest and DNA repair following DNA damage and is involved in DNA repair via homologous recombination (HR). HR involves utilizing a structurally similar and undamaged chromosome as a template for repairing damaged DNA during the DNA repair process. This mechanism allows for the restoration of damaged genetic material, ensuring the maintenance of genomic integrity (17, 18). Hence, when the ATM gene is mutated, the HR repair mechanism in the cell becomes impaired. The inability to execute proper repairs under these circumstances can lead to cancer. However, ATM enhances resection at these breaks through CtIP phosphorylation, which results in the dissociation of the DNA-ends sensor and NHEJ protein Ku. This suggests an antagonistic role of ATM in the assembly of NHEJ proteins and the synapsis of DNA ends in single-ended DNA double-strand breaks (19).
DNA-damage induced through ATM signaling can lead to cell senescence and synthetic lethality. The combined application of PARPi and senolytics induces synthetic lethality in preclinical models of ovarian and breast cancer (20). Wang et al. demonstrated that continuous administration of olaparib at a low dosage induced senescence in ovarian cancer in a P16- or P53-dependent mechanism (21). More recently, this group reported that combining rosiglitazone with olaparib aids in the management of ovarian cancer by alleviating olaparib-induced senescence and enhancing antitumor effects (22). Chen et al. showed that employing the glycolysis inhibitor fenofibrate (23) in combination with ATM inhibitors (KU-60019 and AZD0156) synergistically induces senescence in ovarian cancer cells (24). These studies indicate that DNA damage induced through ATM signaling can trigger senescence, suggesting a potential role of ATM in synthetic lethality.
Mutations in the BRCA1 and BRCA2 genes and other genes, such as PALB2, ATM, and CHEK2 is the most common cause of ovarian cancer. PARP is a DNA repair enzyme involved in DNA single-strand break repair (2). PARP uses nicotinamide adenine dinucleotide (NAD+) as a substrate to form a repair complex and transfers ADP-ribose to target proteins to promote DNA repair. After DNA repair is completed, PARP and the repair complex dissociate from the DNA (25-27). When PARP activity is inhibited, PARP and the repair complex cannot dissociate from the DNA, resulting in stalled DNA replication. If the stalling effect is not corrected, additional DNA breaks can occur, causing double-stranded DNA breakage (DSB) (2). Therefore, treating tumors harboring mutant ATM with PARP inhibitors may cause severe DSBs in a large number of cells, ultimately leading to apoptosis (Figure 2). As ATM-deficient ovarian cancer cells are more sensitive to PARP inhibitors (28-31), emerging studies have aimed to use ATM inhibitors to inactivate the HRR mechanism in ovarian cancer cells with nonhomologous recombination-related gene mutations. These studies indicated that ATM loss is critical for the efficacy of PARP inhibitors in treating ovarian cancer.
PARP inhibitors, such as olaparib, niraparib, rucaparib, and talazoparib are used for treating ovarian cancer in clinical settings (32) (Table I). The mode of action of PARP inhibitors involves the formation of a PARP-DNA complex that blocks DNA replication and further causes replication fork collapse and DSB (33). Olaparib was the first PARP inhibitor approved by the U.S. FDA for breast cancer treatment (32). In early (phase 1/2) clinical studies, single olaparib treatment effectively inhibited growth of ovarian cancer cells containing BRCA1/2 gene mutations, with a response rate of 41%; however, for ovarian cancer patients without BRCA1/2 gene mutations, the treatment response rate decreased to 24% (32). Niraparib maintenance therapy reduced the risk of disease progression or death by 68% and prolonged progression-free survival (PFS) in patients with platinum-sensitive recurrent ovarian cancer compared with a placebo (32, 34). Rucaparib was developed by Clovis Oncology Pharmaceuticals for treating BRCA-mutated ovarian cancer and was launched on December 19, 2016, through the FDA Priority Review pathway (32, 35). Among the 106 subjects, 96% had BRCA mutations, and approximately 54% had complete or partial tumor shrinkage that lasted an average of 9.2 months (32). Although emerging evidence shows that talazoparib exhibits greater PARP-DNA trapping activity than other PARP inhibitors, the clinical outcome of talazoparib in ovarian cancer treatment is not as promising as that in breast cancer treatment (36, 37). Notably, the primary concern in treating ovarian cancer patients with PARP inhibitors is severe side effects, such as gastrointestinal upset, fatigue, muscle and joint pain, lethargy, neutropenia, and hematotoxicity (anemia) (34, 38, 39). The hematotoxicity caused by PARP2 trapping is considered the primary reason for the side effects of PARP inhibitors (40, 41). To enhance the specificity of PARP inhibitors for better therapeutic effects and reduce side effects, Illuzzi et al. designed and synthesized a next-generation inhibitor, AZD5305. AZD5305 selectively traps PARP1 and suppresses multiple cancer types in vitro and in vivo with decreased hematotoxicity (42).
As ATM is required for homologous recombination upon DNA DSBs by phosphorylating KAP1, stimulating MRE11/CtIP nucleolytic activity, and sustaining RAD51 nucleofilament formation (43, 44), treating ATM-deficient cancers with PARP inhibitors or combining ATM inhibitors with PARP inhibitors have been proposed as potential opportunities for precision oncology (28, 45, 46). Mak et al. investigated the synergism between ATM and PARP1 inhibitors. They found increases in DNA damage and PARylation in response to treatment with ATM inhibitors (KU-60019 or AZD0156) and treatment with PARP1 inhibitors (olaparib or veliparib) combined with ATM inhibitors by abrogating the G2 DNA damage checkpoint (47). Consistent with this result, Aguilar-Quesada et al. reported that ATM-deficient cells are more susceptible to PARP inhibition because of the accumulation of γH2AX foci (48). A similar study was performed by Weston et al. The authors demonstrated that the PARP inhibitor olaparib significantly suppressed ATM-deficient lymphoid tumor cell growth in a cell culture system and a xenograft tumor model (49). These results imply that treatment of ATM-deficient cells with PARP inhibitors leads to synthetic lethality. In contrast, cotreating ATM-proficient cells with an ATM inhibitor and a PARP inhibitor also led to conditional synthetic lethality (Figure 1A and Figure 2). As 53BP1 is known to function upstream of ATM in response to DSB (50), Miyamoto et al. demonstrated that olaparib enhanced anticancer drug cytotoxicity by inducing γH2AX foci through the involvement of 53BP1 (51). Taken together, these findings indicate that targeting ATM is a potential avenue for ovarian cancer treatment.
Inhibiting ATM Leads to Metabolic Reprograming and Drives Macropinocytosis
A series of studies revealed the correlation between macropinocytosis and cancer progression. In recent years, Commisso and his colleagues demonstrated that pancreatic cancer cells harboring KRAS mutations exhibit the unique morphology of macropinocytosis in response to nutrient starvation (52). Numerous studies have demonstrated that the most common drivers of macropinocytosis are Rac1, Pak1, Na+/H+ exchangers, and Cdc42 (53-55). Upon nutrient deprivation, cells unselectively consume substances from the extracellular environment by macropinocytosis. Glucose, glutamine, branched-chain amino acids (BCAAs), and nonessential amino acids have been identified as extra fuel from tissue interstitial or necrotic cells for cancer cell survival via macropinocytosis (55, 56). Therefore, Soraj et al. and Zhang et al. sought to determine whether activating macropinocytosis facilitates internalization of targeted drugs, such as compounds, DNA, or siRNAs carried by proteins or liposomes by the cells (57, 58).
Blocking macropinocytosis may result in cancer cell death through nutrient deprivation. To test this hypothesis, Commisso et al. treated KRAS mutant pancreatic ductal adenocarcinoma (PDAC) cells and MiaPaca 2 cells with the macropinocytosis inhibitor 5-(N-ethyl-N-isopropyl)amiloride (EIPA, a Na+/H+ exchanger NHE inhibitor). They found that MiaPaca 2 cells were more sensitive to EIPA in vitro and that the growth of MiaPaca 2 xenograft tumors was suppressed by EIPA in vivo (52). Kim et al. also demonstrated that blocking macropinocytosis by inhibiting AMPK, RAC1, or PI3K limited PTEN-deficient prostate cancer cells and tumor growth in low-nutrient environments (59).
More recently, Katherine Aird and her colleagues reported that ATM mediates macropinocytosis in ovarian cancer cells in a nutrient-dependent manner. They found that inhibiting ATM results in metabolic reprogramming and demonstrated that the use of the glycolysis inhibitor fenofibrate (23) synergizes with ATM inhibitors (KU-60019 and AZD0156) (24). As glucose and glutamine consumption are enhanced by ATM inhibition, Katherine Aird and her colleagues hypothesized that the inhibition of ATM may induce a nonselective process of extracellular metabolite intake (60). These authors demonstrated that blocking ATM with inhibitors (KU60019 and AZD0156) or shRNA leads to the activation of macropinocytosis and increased BCAA consumption via the mTORC1 pathway (60). As a result, they reported that cotreating ovarian cancer cells with inhibitors of ATM and macropinocytosis suppressed proliferation and induced cell apoptosis in vitro and in vivo in ovarian orthotopic mouse models. These findings reveal a novel mechanism through which ATM mediates macropinocytosis and provide potential therapeutic regimens targeting metabolic vulnerability in patients with ovarian cancer (60) (Figure 1B and Figure 3).
Translating the DNA Damage Response to Adaptive Immunity through ATM
Cancer vaccines. Cancer immunotherapy can be divided into active and passive immunotherapy (61). Active immunotherapy destroys tumors by stimulating or restoring the human immune system, through the use of cytokines and cancer vaccines. The suitability of general cancer vaccines usually depends on the nature of the tumor antigen (TA). The basic requirements of an ideal tumor antigen include limited or no expression in normal tissues. Furthermore, the unusually high frequency of TA expression in tumors contributes to tumor progression, and these tumor antigens exhibit immunogenic properties. However, no TA fully meets these criteria. Cancer-testis antigen (CTA) is the best example for the following reasons: 1) it is mainly expressed in the germ cells of the testis and placental trophoblasts but not in other normal tissues; 2) approximately half of the CTA present originates from the CT-X gene on the X chromosome; 3) non-specific ectopic expression of CTA is found in malignant tumors; and 4) CTA derive from a multigene family that encodes a family of antigens associated with tumors and expressed in human malignancies (62). CTA-rich tumors are found in ovarian cancer, and New York esophageal squamous cell carcinoma 1 (NY-ESO-1) is considered one of the antigens of spontaneous immunogenic tumors (63). NY-ESO-1 ranks among the top ten antigens for developing immunotherapies because of its testicular-restricted expression in normal tissues and immune activation (63). NY-ESO-1 is a protein of 180 amino acids and is exclusively expressed in normal germ cells and placental cells (63). However, NY-ESO-1 is found in 43% of ovarian cancer patients (64). Cellular and humoral immune responses are usually triggered in tumor patients with a high proportion of NY-ESO-1 expression (63). Emerging evidence shows that cancer stem cells may selectively express NY-ESO-1 and additional CTAs (63). Therefore, developing strategies to target ovarian cancer NY-ESO-1 antigens may have potential therapeutic benefits (62). As CTA is associated with genomic instability (65-67), it may be an additional biomarker for HR repair status (68). This novel evidence indicates that the activity of ATM may influence CTA activity or vice versa. Thus, further studies are necessary to determine the circumstances under which the correlation between ATM and CTA can be elucidated, either through cancer vaccine analysis, biomarker detection, or whole-exome sequencing (69) (Figure 4).
Immune checkpoint inhibitors. Immune checkpoints regulate immune responses on T lymphocytes and prevent the immune system from attacking cells indiscriminately (70, 71). In attacking cancer cells, dendritic cells, which are antigen-presenting cells, usually present tumor-associated antigens, even neoantigens, to cytotoxic T cells (71). Nonetheless, ovarian cancer cells often stimulate or inhibit immune responses to avoid attacks by cytotoxic T cells (72). Inhibition signals regulate the T lymphocyte-mediated immune response, and activation signals and checkpoints on T lymphocytes participate in accepting inhibition or activation signals and subsequently transmitting them to downstream signaling pathways (71). Usually, tumor cells tend to inhibit immune responses (71).
Tumor cells suppress naïve T lymphocytes in lymph nodes or deactivate T lymphocytes in tumor peripheral tissue in the late phase to evade immune responses (71). Two immune checkpoints are commonly involved in this process: T lymphocyte-associated antigen 4 (CTLA-4) and programmed death 1 (PD-1). Hu et al. reported that ATM inhibition promotes mitochondrial DNA (mtDNA) leakage and cGAS/STING activation, and further enhances the effectiveness cancer immunotherapy (73). Their study revealed the potential role of ATM in immune checkpoint blockade (ICB) therapy (73).
Immune checkpoints on T lymphocytes: Cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4). An immune response is initiated by two signals that activate T lymphocytes to carry out forward responses. The first signal is triggered by a tumor-associated antigen on the tumor cell surface (74). The major histocompatibility complex on antigen-presenting cells brings tumor-associated antigens to resting T lymphocytes in the lymph nodes near the tumor tissue (74). CD28 initiates the second signal on T lymphocytes, and CD80/86 activates the second signal on antigen-presenting cells. The second signal is a costimulatory molecule, similar to how a gun cannot fire without pulling a trigger (74). T lymphocytes cannot execute a complete immune response without CD28 binding to CD80/86. Some receptors activate T lymphocytes, and these receptors inhibit T lymphocyte activation (74). However, ligands with stronger affinities for receptors on T lymphocytes block the transmission of the second signal. CTLA-4 is an immune checkpoint whose structure is similar to that of CD28, and it has a stronger affinity for CD80/86. It competes with CD28 for binding to CD80/86, thereby blocking the transmission of the second signal (Figure 5A) (75).
CD28, which efficiently binds to CD80/86, can induce T lymphocyte proliferation, increase T lymphocyte survival, and improve cell energy metabolism, thereby fully activating T cells and increasing the production of IL-2 (74). Then, IL-2 impedes the PD-1-mediated inhibition of AKT and induces AKT activation via STAT5 (74). In contrast, CTLA-4 is secreted by exocytosis only when CD28 is bound to CD80/86. Because of the activated loop, CTLA-4 and T lymphocytes inhibit the response and enter a vicious cycle (74). The greater the number of T-cell receptors involved in signaling is, the greater the production of CTLA-4. When CTLA-4 combines with CD80/86, IL-2 production and the cell cycle progression of T lymphocytes are suppressed, thereby impeding the activation of T lymphocytes (75). Currently, a monoclonal antibody against CTLA-4 was developed as an inhibitor of the CTLA-4 checkpoint (74). It blocks the CTLA-4 site, allowing CD28 to later contribute to the activation and proliferation of T cells, enhancing TCR-derived signals to stimulate an immune response (74, 75). Yan et al. reported that CTLA-4 binds to the ATM inhibitor protein phosphatase 2A (PP2A) and sustains ATM activity in response to DNA damage (76).
Immune checkpoints on tumor cells: PD-1 and PD-L1. PD-1 is a receptor that regulates the T lymphocyte immune response and prevents the overexpression of immune response components that damage healthy cells in the human body (77). Programmed cell death ligand 1 (PD-L1) is usually found on the tumor cell surface and binds to PD-1 on activated T lymphocytes, thereby deactivating T lymphocytes (77). Moreover, the interaction between PD-1 and PD-L1 is affected by T-cell exhaustion, which is accompanied by antigen exposure and chronic inflammation (75). Then, the T lymphocytes gradually lose their normal function. This situation commonly occurs in people diagnosed with cancer or chronic inflammation. T lymphocytes lose the ability to fight against invasive cells (77). As a result, the extent of infection and tumor growth cannot be effectively controlled (Figure 5B).
Enhanced PD-L1 expression in tumor cells is due to the loss of PTEN and PI3K signaling and the persistent increase in the IFNγ concentration in the tumor microenvironment (78). PD-1-initiated signal transduction starts at the membrane of T cells and subsequently inhibits the activity of PI3K and the activation of downstream AKT. PI3K and AKT are essential signaling molecules involved in transporting glucose and glycolysis. Therefore, when PD-1 inhibits the activity of PI3K and AKT, the production of cellular energy is also hindered (75). PD-1 and PD-L1 ligation can lead to the phosphorylation of receptor tyrosine kinases in the cytoplasm, causing the phosphorylation of proximal signaling molecules (75). This inhibited signal transmission impairs the PI3K and Akt pathways and reduces T-cell growth and survival, protein synthesis, and IL-2 production (79).
Anti-PD-1 or anti-PD-L1 monoclonal antibodies inhibit the PD-1 checkpoint, block the binding sites of PD-1 or PD-L1, and retain the function of CD28 in T-cell activation and proliferation in ovarian cancer (80). Enhanced signaling from the T-cell receptor stimulates the immune response. As DSBs trigger PD-L1 expression (81), Wang et al. reported that ATM mediates the expression of PD-L1 in docetaxel-treated prostate cancer (82). Their study provides a new avenue for understanding the immune balance that is mediated by ATM-NEMO-NF-B signaling (82). However, Gao et al. reported that ATM inhibition causes irradiation-induced PD-L1 expression in tumor-associated macrophages (83). These data suggest that there is tight crosstalk between the DNA damage response and immunity through ATM during cancer progression and treatment. Furthermore, these findings provide a potential new paradigm for immune therapy in ovarian cancer by targeting ATM.
Conclusion
Several studies have shown that the loss or suppression of ATM in ovarian cancer cells impairs HR repair activity, enhances metabolic activity, and modulates immune checkpoints. These studies indicate that ATM plays a unique and crucial role in response to DNA damage, cellular metabolism, immune reactions, and other factors in ovarian cancer. Notably, PARP inhibitors induce synthetic lethality in ovarian cancer cells with homologous recombination deficiency, including ATM-deficient cells. In contrast, combining a PARP inhibitor with an ATM inhibitor may be a new avenue for treating ATM-proficient ovarian cancer patients. Moreover, recent findings have shown the potential significance of ATM-mediated regulation of key aspects of macropinocytosis. The involvement of ATM in the CTLA-4 immune checkpoint system highlights its potential influence on immune responses in ovarian cancer. The association of ATM with the PD-L1 immune checkpoint system underscores its impact on immune modulation within the ovarian cancer microenvironment. Although further research is warranted to determine the underlying mechanisms of ATM through its canonical or noncanonical roles, ATM is essential for tumorigenesis, immune function, and pharmacological treatment in ovarian cancer.
Acknowledgements
The Authors thank the support from Prof. Han-Jung Lee (Department of Natural Resources and Environmental Studies, College of Environmental Studies, National Dong Hwa University) and the Center for Teaching Excellence at the National Dong Hwa University. This study was supported by grants from the National Science and Technology Council, Taiwan (111-2320-B-259-001 and 111-2320-B-259-002-MY3).
Footnotes
Authors’ Contributions
Conceptualization: C.-C.T, M.-H.K., Y.-C.W., W.-L.H, W.-M.W., C.-H.P., C.-W.C.; writing – original draft: C.-C.T, M.-H.K., Y.-C.W., W.-L.H, W.-M.W., C.-H.P., C.-W.C.; writing – review & editing: C.-W.C.
Conflicts of Interest
The Authors have no financial conflicts regarding the subject of the review or materials discussed in the manuscript. These include employment, consultancies, honoraria, stock ownership or options, expert testimony, grants, or patents received or pending, or royalties.
- Received December 13, 2023.
- Revision received January 16, 2024.
- Accepted January 18, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).