Keywords

14.1 Introduction

Preserving genomic integrity is pivotal for cell survival; consequently, cells rely on a network of complex signaling pathways to facilitate faithful DNA replication and maintain genomic stability [1]. Increased proliferation rates are associated with genomic instability via the accumulation of DNA damage in the form of DNA double strand breaks (DSBs) or DNA single strand breaks (SSBs) caused by a variety of events, including replication stress induced by stalling of replication forks [2,3,4]. These events distort genetic material due to subsequent fusion of DSBs and shortening of telomeres, which can result in translocations, gene amplification, and gene mutations [5,6,7].

When DNA damage or replication stress is sensed, cells are prevented entry into mitosis by activating DNA Damage Response (DDR) pathways at varying phases within the cell cycle [8]. DDR signaling pathways orchestrate tightly regulated kinase cascades to resolve DNA damage and replication stress by pausing the cell cycle and initiating repair [9, 10]. Ataxia telangiectasia and Rad3-related protein (ATR) is a key PI3K-related kinase within the DDR that senses replication stress and regulates checkpoints within the cell cycle’s synthesis (S) and gap 2/mitosis (G2/M) phases to preserve genomic stability [1, 10,11,12].

Notably, replication stress and genomic instability are hallmarks of cancer cells, making them dependent on protective DDR pathways for survival [10, 12,13,14]. As such, targeting ATR in cancer medicine is an attractive therapeutic approach to circumvent cancer cell survival by exploiting their dependence on ATR-driven processes.

14.2 ATR Acts as a Gatekeeper of DNA Damage Repair

Endogenous and exogenous sources of DNA damage lead to a wide variety of adducts including DSBs, SSBs, base damage, bulky adducts and base mismatches [1]. Cells have evolved complex DNA repair mechanisms designed to specifically repair these types of damage and maintain genome stability. Repair of DNA DSBs is of particular importance as it is estimated that a single unrepaired DNA DSB can initiate cell death, highlighting the critical role of DDR pathways in cell survival [16]. DSBs are repaired by a number of mechanisms: the best characterized being the error-free homologous recombination repair (HRR) pathway, the highly efficient -but error prone- nonhomologous end joining (NHEJ) pathway, and the error prone microhomology mediated end-joining pathway (regarded as a backup pathway to HRR) [15].

Repair by HRR is initiated by an upstream activator of DDR, the ataxia telangiectasia-mutated (ATM) PI3K-related kinase, which upon sensing DSBs triggers a cascade of events that include cell cycle arrest, repair and apoptosis [17, 18]. During S or G2-phases, exposure of single-stranded DNA can occur as an intermediate of HRR at areas of resected DNA and also at stressed replication forks. These single-stranded DNA regions quickly become coated with the high-affinity ssDNA binding protein, replication protein A (RPA), which protects against DNA degradation. The coating of single-stranded DNA by RPA recruits ATR/ATR-interacting protein (ATRIP) complexes to sites of damage [1, 16]. Following localization to sites of damage, ATR is activated by either topoisomerase II binding protein (TopBP1) or Ewing tumour-associated antigen 1 (ETAA1) [17]. Moreover, recruitment of TopBP1 is mediated by Rad17, which loads the 9-1-1 (Rad9-Hus1-Rad1) complex onto chromatin, binds to TopBP1, and results in ATR activation [4]. Once activated, ATR proceeds to phosphorylate a series of downstream targets; however, its activation of checkpoint kinase 1 (CHK1) is integral to its regulation of cell cycle checkpoints (Fig. 14.1) [16].

Fig. 14.1
figure 1

ATR activation, inhibition and modulation of antitumor immunity. ATR is activated in response to replication stress, single-stranded DNA, and increased R-loops. ATR activation triggers a kinase cascade of CHK1 and WEE1, resulting in checkpoint activation and cell cycle arrest for DNA repair. ATR inhibition induces inappropriate mitotic entry that culminates in mitotic catastrophe and the release of DNA fragments into the cytosol, which in turn activate the cGAS-STING pathway and a type I IFN response. DDR may also be a means of increasing tumor mutational burden and therefore the generation of neoantigens

Full size image

ATR commands control over the S and G2/M checkpoints by phosphorylating and activating CHK1 [1, 18]. Active CHK1 kinase in turn phosphorylates and inactivates the cell division cycle 25A/25C (CDC25A/CDC25C) phosphatase proteins, leading to their respective degradation [1]. Degradation of CDC25A thereby renders CDK2 and its associated complexes inactive by removing an inhibitory phosphorylation present on CDK2 [19]. Consequently, progression to S phase is interrupted, preventing DNA replication and promoting DNA repair [16]. ATR-mediated activation of CHK1 also interrupts the G2/M checkpoint in a Wee1-like protein kinase (WEE1) dependent manner. Active CHK1 phosphorylates and stabilizes WEE1, enhancing its activity toward CDK1 [19]. Both CDK1 and CDK2 remain in inactive states induced by WEE1’s inhibitory phosphorylation [19]. Inactivation of CDC25C by CHK1 prevents the removal of the inhibitory phosphorylation on CDK1, which halts the G2/M checkpoint to allow time for post-replicative DNA repair and prevent replication of unrepaired DNA [16].

Additionally, ATR also plays a role in regulating replication forks through multiple mechanisms [1]. One mechanism involves ATR-mediated fork remodeling: ATR phosphorylates the helicase SWI/SNF-related, matrix associated, actin dependent regulator of chromatin, subfamily-A-like 1 (SMARCAL1), promoting maintenance of fork stability and fork restart in cooperation with RAD51 and zinc-finger RANBP2-type containing 3 (ZRANB3) [20]. Another known mechanism underscores the importance of the ATR-CHK1 axis in resolving replication stress during the formation of R-loops, which are RNA–DNA hybrid transcription intermediates that induce genome instability. Here, the ATR-CHK1 pathway is activated by R-loop induced reversed replication forks. Upon activation, ATR protects the genome by regulating the activity of the MUS81 endonuclease, preventing excess nucleolytic degradation of reversed forks. Active ATR also suppresses R-loop accumulation and enables replication recovery, while promoting arrest of the cell cycle at the G2/M-phase [21]. Resolution of replication stress triggers ATR to resume HRR activities by promoting fork reversal and restart, a process involving the recruitment of BRCA2 and RAD51 to sites of damage [11, 22].

In addition to its role in HRR, ATR has further roles in DNA repair through its involvement in the inter-strand crosslink repair (ICLR) and nucleotide excision repair (NER) pathways. ICLR removes toxic inter-strand DNA crosslinks (lesions involving both strands of DNA that can result in replication fork-stalling and inhibition of transcription). The presence of ICLs activates ATR and requires ATR-mediated phosphorylation of Fanconi Anemia proteins, which are key players in mediating ICLR [23, 24]. Lastly, NER is a critical mechanism that repairs a wide variety of DNA lesions caused by chemical agents or environmental factors (particularly UV radiation) [25]. In this repair mechanism, ATR phosphorylates and stabilizes Xeroderma pigmentosa group A (XPA), thereby recruiting the protein to sites of damage during the S-phase of the cell-cycle and aiding in the activation of the NER pathway [26]. As a key player in multiple processes, ATR is a master regulator of DNA repair and more broadly in the replication stress response.

14.3 ATR Signaling Fosters Cancer Cell Survival

Activation of DDR in normal cells can either resolve DNA damage and/or replication stress to promote cell survival, or it can trigger programmed cell death when DNA damage cannot be removed, inhibiting tumorigenesis and preventing inheritance of DNA mutations in daughter cells [28,29,30,31]. Since DDR is often dampened in cancer cells, these cells present increased DNA damage that is tolerated due to a simultaneous amelioration of unrepaired DNA damage response. In this way, DDR is used by cancer cells as a decoy mechanism to shield against cell death and allow genomically unstable cells to traverse the cell-cycle unscathed, making cancer cells dependent on DDR for survival. That is, tolerance of replication stress is crucial for tumor viability, and oncogene-induced dysregulation of DNA replication generates high levels of replication stress in cancer cells [27,28,29]. Furthermore, most cancers also feature defective G1 checkpoints, largely due to p53 signaling loss [30], rendering cancel cells dependent on S- and G2-phase checkpoints, which are ATR-regulated processes [30].

Frequent mutations in DDR genes also increase dependency on ATR signaling [28, 31]. In these situations, cancer cells rely on ATR to respond to and resolve replication stress and repair DNA damage in order to bypass cell death [27]. Given that ATR function is conserved in the vast majority of cancers, it is has emerged as a favorable target in cancer medicine [32]. ATR inhibition in normal cycling cells, with intact cell-cycle checkpoints, leads to moderate cytotoxicity due to replication fork stalling and collapse; however, as cancer cells have high replication stress, they are more dependent on ATR for survival and as a result are more sensitive to its inhibition than normal cells [27, 28]. Furthermore, in-vitro data suggests that chronic use of ATR inhibitors impairs the cell’s ability to repair damage by HRR while also impacting the availability of necessary HRR proteins, such as TopBP1, BRCA1, and RAD51 [28]. This targeted approach has culminated in the development of many clinical studies aimed at evaluating the clinical efficacy of ATR inhibition as both a monotherapy in certain DDR defective cancer backgrounds and in combinational approaches [19, 33].

14.4 Early Development of ATR Inhibitors

The development of potent and selective ATR kinase inhibitors has been closely related to (1) the availability of well-characterized assays that permit accurate measurements of selective kinase activity, (2) the availability of structural and functional insights to guide a drug design strategy that maximizes selectivity, and (3) the development of screening tools and biomarkers that can identify suitable patients. Notably, the first generation of small molecule ATR inhibitors struggled to find balance between potency and selectivity to reach clinical usage. For instance, while Schisandrin B, an active ingredient of the magnolia berry (Schisandra chinensis), inhibited ATR kinase activity at high concentrations leading to off-target effects and toxicity, the small molecule NU6087 demonstrated moderate selectivity over ATM homologs, but did not display selectivity over the wider kinase family [34]. With the development of a cell-screening assay that measured ATP-dependent phosphorylation of H2AX as a more accurate quantification of ATR kinase activity in experimental conditions, the small molecule ETP-46464 was selected from a library of compounds based on its increased selectivity over other ATR homologs. Although poor pharmacological properties in mice prevented ETP-46464 from advancing to clinical studies, the discovery of this compound provided proof of concept for a more reliable biomarker of replication stress that accounts for double stand breaks and has become the standard marker for quantifying DNA damage [35]. Based on recent advances in the development of well-characterized assays, as well as new insights in the structure–function relationship of ATR, many pharmaceutical companies have taken on the challenge to design and develop potent and selective ATP competitive inhibitors of ATR with the most advanced targeted therapies described here.

14.5 ATR Inhibitors in the Clinic

14.5.1 Berzosertib (M6620/VX-970/VE-822)

Shortly after the development of the assay that measured ATR-dependent phosphorylation of γH2AX, a high throughput screen that combined structure–activity relationship with homology modeling led to the discovery of VE-821, a selective inhibitor with 600-fold selectivity for ATR over ATM, DNA-PK, mTOR, and PI3K [36]. In addition to increased selectivity, VE-821 also showed strong inhibition of CHK1 phosphorylation in cellular models of ATM and/or p53 deficiency [37]. In-vitro experiments showed that VE-821 sensitized ovarian cancer cells to DNA damaging agents such as cisplatin and gemcitabine, and the effects of gemcitabine were potentiated when combined with VE-821 in pancreatic cancer cells [38]. It was also observed that VE-821 could further sensitize BRCA1-depleted cells to DNA damaging agents [44]. Such synergistic effects appear to be stronger with DNA-damaging agents, such as cisplatin and carboplatin, since DNA crosslinking triggers early activation of ATR and the DDR machinery. Interestingly, p53-deficient cancer cell lines were shown to be more sensitive to the combination of VE-821 and cisplatin than normal cell lines, and significant synergistic activity was observed in ATM-deficient cell lines [43]. These results were further confirmed by treating ATM-proficient cells with the triple combination of VE-821, cisplatin, and a highly selective ATM inhibitor (KU-55933) [37]. Taken together, these results suggested that cancer cells with defective ATM signaling are more reliant on ATR; hence, demonstrating a synthetic lethal interaction between the S-phase specific ATR and the ATM-p53 pathway mediating the G1 checkpoint [39]. However, it must be noted that recent studies have shown that ATM mutations and p53 status are not enough to predict clinical benefit to ATR inhibition, and mutations in other DDR genes—such as PTEN, XRCC1, BRCA1, BRCA2, and ARID1A—may promote synthetic lethality with ATR inhibitors [40].

Based on promising pre-clinical data, VE-822, an optimized analog of VE-821 with increased potency and selectivity for ATR, became the first inhibitor to enter clinical trials labeled as VX-970, later named M6620 and berzosertib [41]. Berzosertib alone was found to sensitize multiple lung cancer cell lines to a wide variety of DNA-damaging chemotherapeutic agents (cisplatin, oxaliplatin, gemcitabine, etoposide, and the active metabolite of irinotecan, SN38), and the combination of berzosertib and cisplatin showed sustained tumor regression in non-small cell lung cancer (NSCLC) patient-derived xenograft models [42]. A recent CRISPR-Cas9 screen suggested that the ATR-CHK1 pathway has the potential for synthetic lethality in small cell lung cancer (SCLC) [43]. In that study, the combination of berzosertib with cisplatin displayed greater synergistic activity in different SCLC cell lines and primary lung fibroblasts when compared to treatment with the combination of cisplatin and etoposide. Interestingly, while SCLC cell-derived xenografts showed that the combination of berzosertib with cisplatin inhibited tumor growth, other studies showed that pediatric solid tumor xenografts treated with berzosertib and cisplatin displayed a larger event free survival relative to those treated with cisplatin monotherapy [44, 45]. Together, these studies were the first to confirm the clinical potential of berzosertib as a chemo-sensitizer of DNA damaging agents in lung cancer patients, as well as in pediatric solid tumors, setting the stage for several other studies that also showed the clinical potential of berzosertib in combination with cisplatin in other cancer types, including colon cancer, triple negative breast cancer (TNBC), and esophageal tumors, amongst others [44,45,46,47].

Berzosertib monotherapy has already advanced to a phase II clinical trial investigating antitumor activity in molecularly selected solid tumors, leiomyosarcoma and osteosarcoma (NCT03718091). Although clinical trials are currently studying the combination of berzosertib with radiotherapy, chemo-radiotherapy agents, PARP inhibitors, VEGF inhibitors, as well as with anti-PD-L1 antibodies, such as avelumab, the most common strategy for berzosertib treatment combinations in registered clinical trials appears to be with DNA damaging agents such as cisplatin, carboplatin, gemcitabine, topotecan, irinotecan, and paclitaxel, amongst others (Table 14.1).

Table 14.1 Trials investigating berzosertib combinational strategies
Full size table

The first-in-human trial of berzosertib in combination with cytotoxic chemotherapy agents in patients with advanced solid tumors started with a lead-in safety phase of berzosertib monotherapy, followed by three dose escalation arms aimed to determine the safety profile and recommended phase 2 dose (RP2D) of the combinations of berzosertib with (1) cisplatin, (2) gemcitabine with and without cisplatin, and (3) irinotecan. This trial also included 3 expansion cohorts to further elucidate preliminary anti-tumor activity for the combination of berzosertib and gemcitabine in NSCLC patients harboring p53 mutations and/or loss of ATM expression, the combination of berzosertib and cisplatin in TNBC patients with germline (g) BRCA wild type status, and the combination of berzosertib and cisplatin or carboplatin in patients with platinum-resistant advanced SCLC.

Recent results from two dose escalation arms of this first-in-human trial demonstrated preliminary antitumor activity for berzosertib when combined with gemcitabine and/or cisplatin [48]. That is, most patients who received berzosertib in combination with cisplatin (73.1%), and those who received berzosertib in combination with gemcitabine (68.7%) or berzosertib in combination with gemcitabine and cisplatin (71.0%) achieved disease control with partial response (PR) or stable disease (SD) as their best response per RECIST v1.1 [48]. Interestingly, all patients who received prior platinum-based chemotherapy, and had experienced disease progression, achieved PR when treated with berzosertib in combination with cisplatin. Since ATR inhibition can disrupt DNA replication fork stability and homologous recombination repair (the two major mechanisms of PARP inhibitor resistance), preliminary results from this trial suggest that berzosertib inhibition may contribute to re-sensitizing solid tumors to cisplatin [49].

The RP2D for the combination of berzosertib and cisplatin was determined as 140 mg/m2 of berzosertib (administered on days 2 and 9), and 75 mg/m2 of cisplatin administered every 3 weeks (Q3W) on day 1. This RP2D was generally well tolerated, and the safety profile of this combination was consistent with that of cisplatin alone. Importantly, while the human equivalent dose required for berzosertib target engagement was estimated from preclinical models to be ~60 mg/m2, results from the first-in-human trial show that dosing berzosertib at 140 mg/m2 induces a reduction in serine 345-phosphorylated CHK1, without evidence of PK interactions in a range of malignancies, including ovarian, breast, thyroid, and pancreatic cancers [50]. In a similar manner, the RP2D combination of berzosertib and gemcitabine, which is currently being evaluated in patients with advanced NSCLC in an expansion arm of this trial, was established as 210 mg/m2 of berzosertib (administered on days 2 and 9), and 1000 mg/ m2 of gemcitabine, administered Q3W on days 1 and 8. Yet, the dose escalation for berzosertib in combination with both gemcitabine and cisplatin was terminated after two patients experienced febrile neutropenia or neutropenia as dose limiting toxicities (DLTs). Taken together, results from the first two arms of the first-in-human trial of berzosertib demonstrate that a tolerable safety toxicity profile is observed when berzosertib is combined with either gemcitabine or cisplatin, but not when combined with both agents [48].

The pharmacokinetic (PK) profile of a berzosertib monotherapy lead-in was determined across the dose range of 18–210 mg/m2 (n = 30), and it was characterized by biphasic decline with a moderate-to-high clearance, a high distribution volume, and an apparent terminal half-life of approximately 17 hours [48]. While the PK characteristics of berzosertib in combination with either gemcitabine or cisplatin were consistent with the corresponding doses of berzosertib monotherapy, the collective PK data from these two arms suggest that pre-administration of cisplatin 24 hours before berzosertib administration does not affect the PK profile of berzosertib.

Another trial that is currently investigating the combination of berzosertib and gemcitabine is a multicenter, randomized, phase II study that recently published preliminary efficacy and safety data suggesting that this combination provides clinical benefit to platinum-resistant high-grade serous ovarian cancer (HGSOC) patients. At the cutoff date of publication, 70 patients had been randomly assigned to either receive treatment with the berzosertib and gemcitabine combination (n = 34) and achieved a median profession-free survival (PFS) of 22.9 weeks (90%, CI 17.9–72.0), or they were assigned to receive treatment with gemcitabine alone (n = 36) and achieved a median PFS of 14.7 weeks (90%, CI 9.7–36.7). Yet, while the combination of berzosertib and gemcitabine displayed a promising PFS with a hazard ratio of 0.57 (90%, CI 0.33–0.98), a higher objective response was observed for the group of patients who received treatment with gemcitabine alone. According to the authors, discrepancies between ORR and PFS are not uncommon in platinum-resistant ovarian cancer patients [51]. A sub-analysis of the patient population based on the length of the platinum-free interval also showed that patients who are treated with the berzosertib and gemcitabine combination, and who have had a platinum-free interval of 3 months or less, have a 30% increase in median PFS (27.7 weeks compared to 18.6 weeks in patients with intervals larger than 3 months). Since the PFS benefit observed for patients with a platinum-free interval of 3 months or less may be related to the enrichment for biomarkers of replicative stress, the authors followed up on this finding with further correlative assays [51]. Interestingly, results from follow-up studies using the same replication stress signature show that the combination of berzosertib and gemcitabine benefited more patients with tumors displaying low replication stress (RS-low) in contrast to patients with high replication stress tumors (who appeared to receive a greater benefit from the increase of replication stress caused by gemcitabine monotherapy [52]. Based on these results, it is suggested that increasing replication stress in RS-loss with gemcitabine concomitant with ATR inhibition by berzosertib is necessary for lethality [52].

Finally, a few clinical trials have published results about the preliminary efficacy and safety profile of the combination of berzosertib with topotecan in patients with lung cancers. A proof-of-concept phase I clinical trial that investigated the combination of berzosertib with topotecan in patients with platinum-resistant small cell lung cancer (SCLC) showed that 60% (3/5) of the patients treated achieved a PR or prolonged SD lasting ≥6 months, and the combination seemed to be well tolerated with no additive toxicity observed [54]. Yet, shortly after the interim results of the DDRiver SCLC 250 phase II trial investigating the combination of berzosertib with topotecan in platinum-resistant SCLC patients reported an objective response rate of 36% (9/25) and a median duration of response of 6.4 months, the trial was discontinued based on a low probability of meeting the primary objective [55, 56]. Further results from ongoing clinical trials are needed to demonstrate whether treating patients with advanced cancers, whose tumors are undergoing high replicative stress, with the combination of berzosertib and DNA-damaging chemotherapeutic agents may potentially help overcome platinum and/or PARP inhibitor resistance.

14.5.2 Ceralasertib (AZD6738)

Ceralasertib is a potent and selective ATR inhibitor with a promising preclinical data package showing efficacy in DDR-deficient settings [57]. Early preclinical studies showed that ceralasertib increases γH2AX phosphorylation, while inhibiting phosphorylation downstream of CHK1 in a variety of ATM-deficient cell lines and inducing accumulation of unrepaired DNA damage and cell death in ATM/p53-deficient leukemia cells [58, 59]. Recently, a growth inhibition assay assessing the sensitivity of 276 cancer cell lines to ceralasertib reported that cell lines harboring CCNE1 amplification or ARD1A, ATRX, and SETD2 mutations were associated with sensitivity. At first sight, cancer cell lines harboring ATM mutations were not associated with sensitivity; yet, upon stratifying the cancer cells based on ATM expression levels, it was shown that complete absence of ATM function is significantly associated with sensitivity to ceralasertib [60]. This finding, along with previous observations of antitumor responses from patients harboring ATM loss-of-function, supports the idea that patient selection for ATR inhibitors should consider biallelic deleterious mutations and ATM-null expression [61].

Preliminary results from the dose escalation and expansion monotherapy arms of the PATRIOT phase I clinical trial reported that ceralasertib was better tolerated when administered at an intermittent schedule of 2-weeks-on/2-weeks-off because only 20% of the patients experienced grade ≥3 treatment related adverse events (TRAEs) compared to 67% of patients when ceralasertib was administered in a continuous schedule (NCT02223923). Although it was previously shown that ceralasertib monotherapy in-vivo only induces significant tumor control/stasis and that the synergistic effects resulting in tumor regression are pronounced when ceralasertib is combined with DNA damaging agents or certain targeted small molecules, preliminary antitumor activity of the ceralasertib monotherapy arms of the PATRIOT trial show that 7% of patients achieved PR and 48% of them achieved SD as best overall response [60, 62, 63].

In-vivo, the combination of ceralasertib and cisplatin induced significant tumor reduction in HER2-positive breast cancer cells, as well as tumor regression in ATM-deficient lung cancer xenograft models and synergistic effects in ATM-deficient NSCLC cell lines [64, 65]. Results from a phase I trial investigating the combination of ceralasertib and carboplatin in advanced solid tumor patients reported that 2 (6%) of patients with low ATM or SLFN11 expression achieved PR as best response by RECIST v1.1, while 53% patients (including two unconfirmed PRs) achieved SD for ≥35 days (NCT02264678) [66]. Although no association between ATM and SLFN11 expression level and antitumor activity was reported, likely due to the sample size, these findings support the notion that further investigations on the interaction between ATR and loss of ATM function are needed. In contrast, a phase I clinical trial investigating the safety and antitumor activity of the ceralasertib and paclitaxel combination in advanced solid tumors (enriched for melanoma patients) reported one patient achieved complete response (CR), while 21% achieved PR and 32% achieved SD. Even though the ORR for the entire population was 22.6% (95% CI, 12.5–35.5), an ORR of 33.3% (95% CI, 10.8–51.8) was reported for the subset of melanoma patients resistant to PD1/L1 treatment [67].

Although both phase I trials studying the combinations of ceralasertib with chemotherapy agents reported that the combinational strategies are safe and well tolerated, thrombocytopenia, neutropenia and anemia were reported as the most common grade ≥3 TRAEs, with schedule limiting consequences observed with the combination of ceralasertib and carboplatin. Since toxicity may be one of the major challenges in the implementation of ATR inhibitor combinations with DNA damaging agents and other targeted small molecules, the success of clinical trials investigating ceralasertib combinations depends on the optimization of the dose scheduling sequences and targeted genetic tumor aberrations. For instance, recent in-vivo studies suggest that to achieve tumor regressions, concurrent dosing for the ceralasertib and irinotecan combination should be extended at least one day, while a few days of ceralasertib dosing should be included after concurrent dosing with carboplatin [60]. In a similar fashion, the ATRiUM phase I clinical trial is investigating the safety and antitumor activity of ceralasertib with either intermittent or continuous gemcitabine dosing in advanced solid tumors, particularly in patients with advanced pancreatic ductal adenocarcinoma with ATM-loss-of-function [68, 69]. In all, results from an ongoing phase II trial investigating ceralasertib monotherapy in advanced solid tumors (enriching for mCRPC with low ATM expression), as well as results from the remaining arms of the PATRIOT phase I clinical trial and the ATRiUM phase I are required to further assess the clinical efficacy of ceralasertib monotherapy and in combination with chemotherapy agents in a molecularly targeted population.

Although synergistic effects in-vivo were observed when combining ceralasertib with either PARP or WEE1 inhibitors, only the PARP inhibitors and ceralasertib combination has successfully reached phase II clinical trials. Out of the six clinical trials that are currently investigating the combination of ceralasertib and olaparib in the advanced cancer setting, one phase I study reported on safety and preliminary antitumor efficacy, as well as established the RP2D of the ceralasertib and olaparib combination in patients with advanced solid tumors, and two Phase II trials have presented contrasting preliminary results based on patient selection (Table 14.2) [63, 70, 71]. Briefly, results from one of the first modular phase I clinical trials to test the combination of ceralasertib and olaparib established a concurrent RP2D of ceralasertib at 160 mg QD on days 1–7 and olaparib at 300 mg twice daily (BID) on days 1–28, with thrombocytopenia and neutropenia defined as dose limiting toxicities [72]. Within the module that tested the dose escalation of ceralasertib and olaparib, antitumor responses were observed in patients with advanced breast, ovarian, prostate, pancreatic, and ampullary cancer. Interestingly, while some of the responding tumors had BRCA1/2 mutations, antitumor responses were independent of ATM status [72]. Such results are in accordance with recent preclinical studies suggesting that the combination of ceralasertib and olaparib in a concurrent schedule induces tumor regression in TNBC BRCA-wild type and BRCA2-mutated xenograft models [60], and the development of a phase II clinical trial currently recruiting patients to investigate the combination of ceralasertib and olaparib in advanced germline BRCA mutated breast cancer (NCT04090567).

Table 14.2 Trials investigating ceralasertib combinational strategies
Full size table

Although differences in study design preclude us from direct comparisons, preliminary antitumor activity from two phase II clinical trials investigating the combination of ceralasertib and olaparib seemed to be influenced by the selection of targeted genetic tumor aberrations. That is, the phase II clinical trial investigating the clinical benefit of this combination in patients with advanced solid tumors with or without ARID1A-deficiency (defined as lack of expression of BAF250a by IHC) reported an ORR of 20% for patients with ARID1A-deficiency, including two patients that achieved sustained CRs, while no objective responses were observed in the cohort of patients with active ARID1A function [73]. In contrast, the phase II clinical trial investigating signals of activity of the ceralasertib and olaparib combination in patients with HGSOC reported no partial or complete responses in a PARP naïve, genetically unselected, platinum-resistant cohort of 12 patients. Nonetheless, 75% of the patients in that trial achieved SD as best overall response by RECIST v1.1 and 27% of the patients achieved ≥ 50% decrease in CA-125, most of them harboring tumors with somatic BRCA1 mutations [74].

As mentioned by investigators of the HGSOC phase II trial, it is likely that more responses may have been achieved by focusing the patient population to ovarian cancer patients harboring tumors with BRCA1/2 mutations and/or CCNE1 copy number amplification [74]. Taking it all together, the ceralasertib and olaparib combination appears to be well tolerated, but it is necessary to continue optimizing patient selection strategies based on the selection of genetic aberrations that induce synthetic lethality in different types of cancer types. Such concept seems to be reflected in recent preliminary results from the HUDSON trial, a phase II multidrug, biomarker selected umbrella study investigating the combination of multiple targeted small molecules with durvalumab, including ceralasertib for NSCLC patients who progressed after anti-PD-1/PD-L1 and platinum therapy (NCT03334617). Although no correlation was found between ATM biomarker status and clinical responses by RECIST 1.1, the HUDSON trial reported an improved ORR (11.1%) and longer PFS (7.43 months) for patients whose tumors harbored ATM mutations or low protein expression -when compared to an ORR of 8.3% and PFS of 4.96 months for NSCLC patients with acquired resistance to prior immunotherapy, regardless of ATM status [75].

By comparing gene expression profiles in paired blood samples from patients with controlled disease and patients whose disease progressed with the ceralasertib monotherapy run-in, the HUDSON trial showed increases in an antigen presentation gene signature and decreases in exhausted T-cell and NK-cell signatures, supporting a model in which ceralasertib also has an active role in the immune activation caused by the combination of ceralasertib with durvalumab [77]. These findings are also in agreement with results from a phase II clinical trial investigating the clinical activity of the ceralasertib and durvalumab combination in advanced gastric cancer patients, which reported (1) significantly longer PFS for patients whose tumors harbored ATM-deficiency and/or HRD-deficiency when compared to patients with active ATM function and HRD-proficient (5.60 months versus 1.65 months, HR 0.13., 95% CI 0.045–0.39, p < 0.001), as well as, (2) upregulation of the innate immune response, (3) activation of intratumoral lymphocytes, and (4) increase of tumor reactive CD8+ T-cells in patients who responded to treatment [76].

Finally, a phase III clinical trial (LATIFY, NCT05450692) will compare the clinical benefit of the ceralasertib and durvalumab combination versus docetaxel monotherapy in NSCLC patients who progressed after anti-PD-1/PD-L1 and platinum therapy. This is based on the finding that NSCLC patients with primary resistance to immunotherapy only responded to the combination of ceralasertib and durvalumab in the HUDSON trial [75].

14.5.3 Elimusertib (BAY1895344)

By evaluating the molecular interactions of available ATR inhibitors within the binding pocket of an ATR homology model created using the crystal structure of a PI3K kinase and performing a high-throughput screen, Bayer identified a lead compound that was further optimized to reduce potential off-target toxicity [78]. BAY1895344, also called elimusertib, is a potent and selective ATR inhibitor shown to increase γH2AX phosphorylation in HT-29 cells and inhibit cell proliferation in a variety of cancer cell lines, including different lymphoma cells and cell lines harboring mutations that affect the ATM pathway, Elimusertib induced stronger antitumor activity than ceralasertib and berzosertib in a lymphoma cell line-derived xenograft (CDX) model, with antitumor activity also observed in ovarian, prostate and colorectal CDX models harboring DDR defects [40]. In addition, elimusertib treatment inhibited neuroblastoma cell growth and induced strong tumor growth inhibition in neuroblastoma xenograft and ALK-driven genetically modified mice models [79]. Interestingly, RNA-seq data from mice who achieved tumor size decrease after elimusertib treatment revealed expression of inflammatory response and immune tumor infiltration, suggesting that ATR inhibition by elimusertib positively impacts the tumoral immune response [79].

Synergistic antitumor efficacy for the combination of elimusertib and DNA-damaging treatments was observed in colorectal cancer cells treated with elimusertib and cisplatin, as well as in colorectal xenograft models treated with elimusertib and radiation therapy [40]. In contrast, antagonistic interactions were observed with the combination of elimusertib and docetaxel [40].

Treatment with elimusertib and olaparib displayed strong antitumor efficacy and a tolerable profile in a TNBC xenograft model and delayed tumor growth in a PARP inhibitor resistant prostate cancer xenograft model [40]. In a similar manner, synergistic antitumor activity was also observed with sequential dosing of anti-PD-1/PD-L1 antibodies and elimusertib in immunocompromised and lymphoma mice models [40]. Taken together, preclinical studies suggest that combining elimusertib with certain DNA damaging agents, as well as with DDR and checkpoint inhibitors, may result in synergistic antitumor activity when compared to the respective singe-agent treatments.

Further studies are currently being conducted to determine the precise combination schedules that are safe and well-tolerated in humans. For instance, results from the dose escalation of the first-in-human trial of elimusertib in patients with advanced solid tumors determined that intermittent dosing of 40 mg BID 3 days on/4 days off is the maximum tolerated dose (MTD) of single-agent elimusertib, with pharmacodynamic data showing on-treatment tumor increases in γH2AX levels (NCT03188965) [80]. The most frequently observed adverse events (AE) in the dose escalation was grade 3 anemia, likely due to limited differentiation and expansion of erythrocyte precursors that are sensitive to replication stress [81]. Based on the safety results from the first elimusertib monotherapy trial, combinational strategies with chemotherapy agents may induce overlapping hematologic toxicity and dose escalations should be approached with caution.

Nonetheless, this trial provided proof-of-concept for the clinical antitumor activity of elimusertib: 4 patients achieved PRs, while 8 achieved SD with a median duration of response of 11.25 months and resulting in 69% disease control rate in patients treated at MTD or above [80]. More importantly, 3 of the 4 patients that achieved PRs had tumors with low ATM expression by IHC, with two of them harboring deleterious ATM mutations. Albeit a small sample size, an ORR of 33.3% was reported for the subgroup of patients with ATM protein loss, and an ORR of 37.5% was calculated for the subgroup of patients harboring ATM deleterious mutations [80]. Within the responders for this trial, the investigators noted one heavily pretreated ovarian cancer patient who had received 9 chemotherapy lines, as well as prior PARP inhibitor and immunotherapy, achieved SD for more than year [80]. The clinical benefit observed for this PARP-resistant ovarian cancer patient harboring a BRCA1 deleterious mutation seem to suggest that PARP inhibitor resistance may be mediated by protection of the DNA replication fork by ATR, opening the possibility of expanding ATR inhibitor treatments to PARP inhibitor-resistant patient population and providing clinical rationale for a phase I clinical trial that investigates the combination of elimusertib and niraparib in patients with advanced ovarian cancer and other solid tumors (Table 14.3, NCT04267939) [80].

Table 14.3 Trials investigating elimusertib combinational strategies
Full size table

14.5.4 Gartisertib (M4344/VX-803)

As an ATP-competitive inhibitor, gartisertib is a selective ATR inhibitor with 100-fold selectivity over a wide range of kinases and strong potency demonstrated by suppression of ATR-driven checkpoint kinase-1 (CHK1) phosphorylation in a prostate cancer cell line, as well as by induction of γH2AX phosphorylation in a small-cell lung cancer cell line [82]. Remarkably, sensitivity assays and gene expression analysis of a variety of cancer lines showed that cancer cells with higher replication stress and high neuroendocrine expression signatures are highly sensitive to gartisertib treatment, suggesting that those genomic signatures may be useful for patient selection and as biomarkers of response [82].

As a single-agent, gartisertib was shown to suppress proliferation in prostate cancer cells at a lower concentration and at a higher rate than berzosertib and ceralasertib, and it was shown to induce tumor stasis and tumor regression in ALT mice models [83, 84]. A variety of preclinical models also demonstrated synergistic effects of different gartisertib combination strategies. For instance, the combination of gartisertib and TOP1 inhibitors showed synergistic antitumor activity in several small-cell lung cancer cell lines and cell-derived mouse xenografts, as well as in prostate cancer patient-derived tumor organoids [85]. In addition, combining gartisertib with DNA damaging agents such as gemcitabine and cisplatin, as well as with PARP inhibitors such as talazoparib, displayed synergy at noncytotoxic concentrations in a small-cell lung cancer cell line.

14.5.5 Camonsertib (RP-3500)

Camonsertib, developed by Repare Therapeutics and recently licensed to Roche, is a highly selective ATR inhibitor that demonstrated potent single-agent efficacy by a dose-dependent inhibition of CHK1 phosphorylation and induction of γH2AX, DNA-PK and KAP1 phosphorylation in-vivo [86]. Camonsertib monotherapy induced significant tumor growth inhibition in an ATM-deficient colorectal xenograft model and also induced complete tumor regression in a gastric xenograft model [86]. Unlike other ATR inhibitors, tumor growth inhibition with minimal hematological adverse effects was observed with intermittent camonsertib treatment in ATM-deficient mouse models [86]. In line with preclinical data suggesting that intermittent camonsertib dosing schedules with dose holidays of at least 4 consecutive days allow for reticulocyte regeneration to avoid hematological toxicities in-vivo, recent preliminary data from the TRESR phase I/IIa clinical trial investigating the safety and preliminary efficacy of camonsertib showed a significant reduction of grade 3 anemia in advanced cancer patients (NCT04497116) [87]. In this study, 14.5% of patients treated with intermittent camonsertib dosing experienced grade 3 anemia, compared with 65.7% of patients who experienced grade 3 anemia after intermittent elimusertib treatment [87, 88]. Preliminary data from the TRESR trial also showed clinical activity across different tumor types, with meaningful clinical benefit in 49% of evaluable patients and an ORR of 25% [87]. Aligned with preclinical data, clinical activity was observed in CRPC patients whose tumors harbored ATM and CDK12 mutations, ovarian cancer with BRCA1 and RAD51C mutations, as well as breast cancer, melanoma, and HNSCC patients with tumors harboring BRCA1 and BRCA2 mutations. Notably, 37 patients whose tumors harbored relevant genomic mutations achieved molecular responses in ctDNA, suggesting that ctDNA responses may predict clinical benefit [87].

Intermittent concomitant rather than sequential administration of camonsertib and PARP inhibitors in different ATM and BRCA1 deficient models led to stronger synergistic antitumor activity without increases in hematological toxicity [86]. Along with the additional modules of the TRESR clinical trial that are currently investigating the combination of camonsertib and talazoparib, the ATTACC phase I/IIa clinical trial investigating the safety and preliminary efficacy of camonsertib in combination with either olaparib or niraparib is currently recruiting patients (NCT04972110).

14.5.6 M1774

Building on learnings from berzosertib, Merck KGaA developed M1774 as a potent and selective ATR inhibitor that has demonstrated antitumor activity in PDX models. The modular DDRiver Solid Tumors 301 clinical trial is currently investigating the safety and tolerability and preliminary efficacy of M1774 in patients with advanced solid tumors harboring selected mutations, including deleterious mutations in ATM, ARID1A, ATRX and/or DAXX (NCT04170153) [89]. Recent results from the dose escalation of this trial suggested an MTD of 180 mg QD continuous dosing and a recommended dose for expansion of 180 mg 2 weeks on/1 week off, with modulation of γH2AX in peripheral blood mononuclear cells achieved in doses starting at 130 mg QD [89]. While the DDRiver 301 trial is currently recruiting patients for two dose expansion modules in biomarker selected cohorts and food effects cohort, it is also recruiting patients in a module investigating the safety and tolerability of the combination of M1174 and niraparib. In addition, a recent clinical trial investigating the safety and tolerability of M1774 in combination with a DDR inhibitor or an immune checkpoint inhibitor has recently started to recruit patients (NCT05396833).

14.5.7 ART0380

ART0380, licensed by Artios Pharma Ltd from The University of Texas MD Anderson Cancer Center and ShangPharma Innovation, demonstrated target engagement by γH2AX and pKAP1 modulation in-vivo and, is currently being investigated as monotherapy and in combination with gemcitabine or irinotecan in a modular phase I/IIa clinical trial for advanced solid tumor patients (NCT04657068). In order to measure target engagement, Artios has developed an assay in normal peripheral blood mononuclear cells (PMBCs) and in circulating tumor cells (CTCs) [90]. Although interim results for this trial have not been presented thus far, Artios recently mentioned in a press release that upon treatment with single-agent ART0380, modulation of γH2AX in patient blood samples is observed at a larger magnitude in CTCs than in PBMCs, and that based on preliminary results from the dose escalation phase of the trial, ART0380 has a safe and tolerable profile. Therefore, the intermittent dose escalation ART0380 has progressed to the dose expansion phase in patients with ATM-deficient tumors.

14.6 ATR and PARP Inhibitor Combination Strategies

Synthetic lethal strategies for cancer treatment, where cell death is induced by targeting proteins or pathways that are redundant in normal cells but not cancer cells, are showing clinical promise. Inhibitors of PARP1 (Poly (ADP)-ribose polymerase 1, a key DDR enzyme) are prime examples of anti-cancer therapeutics capable of harnessing the synthetic lethal mechanism and have revolutionized the field of cancer therapeutics. Seminal work led by multiple teams in the early 2000s identified HR-deficient BRCA1/2-mutated cancers as selective targets for PARP inhibitor-induced lethality [91,92,93]. Today, several PARP inhibitors are FDA-approved for the treatment of BRCA1/2-mutated cancers, including in multiple settings of ovarian cancer, metastatic breast cancer, pancreatic cancer, and advanced castration-resistant prostate cancer (CRPC) [94,95,96,97]. Unfortunately, PARP inhibitor resistance is ubiquitous in the clinic. Acquired PARP inhibitor resistance can occur following prolonged treatment, whereas primary PARP inhibitor resistance is observed in many patients with BRCA1/2-mutated cancers and fail to respond at treatment initiation [98]. One strategy to overcome PARP inhibitor resistance is to develop rational combination treatments to sensitize cells to PARP inhibitors.

Growing evidence suggests that ATR inhibition may help to overcome PARP inhibitor resistance [99, 100]. The ATR gene was identified as a mediator of PARP inhibitor sensitivity in a synthetic lethal siRNA screen [92]. DNA DSBs that are produced following exposure to PARP inhibitors renders cells dependent on ATR for DNA repair [8]. As such, exposure to an ATR inhibitor disables ATR-mediated repair pathways and promotes cell death. Additionally, a known mechanism of PARP inhibitor resistance involves restored replication fork stabilization that may involve ATR, as well as other DDR proteins, such as CHK1 and WEE1 [99]. Preclinical studies have demonstrated that ATR inhibition leads to replication fork collapse that produces irreparable DNA DSBs [101, 102]. Building on this, the rationale for the combination of PARP and ATR inhibitors was demonstrated in another preclinical study in which PARP inhibitor resistant cells exhibited enhanced sensitivity in response to dual ATR and PARP inhibition in ovarian cancer patient-derived xenograft (PDX) models [103]. Furthermore, there are multiple ongoing clinical trials currently evaluating this drug combination, with at least 10 active studies taking place world-wide at the time of publication (Table 14.4).

Table 14.4 Ongoing clinical trials combining ATR and PARP inhibitors
Full size table

With an expansive landscape of trials evaluating ATR and PARP inhibitor combinations, it is important to understand the tolerability and clinical efficacy of this approach. Overlapping toxicities stemming from combined ATR and PARP inhibition may be an issue for this drug combination. One example of such toxicity was reported in a dose-finding phase I trial in which ceralasertib was combined with the olaparib PARP inhibitor, which resulted in dose limiting toxicities (DLTs) in the form of thrombocytopenia and neutropenia that restricted continuous concurrent dosing of these agents [104].

Nonetheless, promising clinical activity produced by this drug combination was reported in a separate phase II trials evaluating a cohort of patients with recurrent ovarian cancer who had progressed on prior PARP inhibitor treatment [105]. In a cohort of thirteen patients, the reported objective response rate (ORR) was 46% across six patients who had achieved radiologic PR [105]. Of these patients, 69% had germline BRCA mutations, 23% had somatic BRCA mutations, and 8% had other homologous recombination deficient mutations [105]. Although no patient discontinued treatment due to toxicity, reported adverse events included thrombocytopenia, anemia, and neutropenia, with dose reductions reported for both ceralasertib and olaparib [105]. Interestingly, this same study reported no objective responses in a cohort of PARP inhibitor naïve patients with platinum-resistant ovarian cancer [106]. Enrichment of therapeutic responses in the cohort of patients with past PARP inhibitor exposure further supports the notion that combined ATR and PARP inhibitor strategies may be key to overcome PARP inhibitor resistance in the clinic [106].

14.7 ATR and Immune-Checkpoint Inhibitor Combination Strategies

An emerging body of preclinical and clinical evidence supports the immunomodulatory role of ATR inhibitors in the tumor microenvironment. For instance, a recent single DNA fiber analysis after ATR inhibition showed induction of chromatin bridge formation and chromosome lagging, which in turn accelerated mitotic entry and further activated the cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) tumor sensing axis [107]. Since genotoxic stress also induces the release of cytosolic DNA fragments that activate the cGAS-STING pathway, the combination of ATR inhibition with chemotherapeutics seems a rational combination to activate the innate immune response [111, 112].

It was also recently shown that treating prostate cancer cell lines with elimusertib induced S-phase DNA damage, activation of cGAS-STING signaling, as well as upregulation of CCL5 (chemokine ligand 5) and CXCL10 (C-X-C motif chemokine ligand 10) expression that culminated in activation of innate immunity [108, 109]. This is further supported by the increase in activated cGAS-STING and TBK1 levels, CD8+ T-cell infiltration, reduction of regulatory T-cell infiltration, and T-cell exhaustion observed in immunocompetent hepatocellular carcinoma mouse models treated with the triple combination of radiation, followed by ceralasertib and PD-L1 inhibition [110]. In addition, shortly after treatment with ceralasertib there was a modest increase in the intratumoral concentration of IFN-γ and proliferating CD8+ T-cells that was accompanied by a reduction of the PD-L1 tumor upregulation induced by radiation. At later time points, the combination of ceralasertib and radiation induced an increase in infiltrating CD8+ T-cells, as well as production of INF-γ and tumor necrosis factor α [110]. Similar results were obtained by studying the combination of ceralasertib and radiation on immunocompetent mouse models of HPV-driven cancer, where a signature of type I and II IFN gene expression and modulation of cytokine gene expression (including CCL3 and CXCL10) were associated with treatment. Interestingly, increased antigen presentation and levels of major histocompatibility complex class I were also observed in vivo with the combination of ceralasertib and radiation [111]. Taken together, results from multiple preclinical studies suggest that the combination of radiation and ATR inhibitors stimulates IFN response and triggers antigen presentation.

ATR inhibition has also been shown to suppress upregulation of the natural killer group 2D (NKG2D) cell surface ligand that binds to activated CD8+ T-cells to trigger pro-inflammatory cytokine production [112]. It has also been suggested that ATM/ATR/CHK1 signaling upregulation leads to transcriptional activation of PD-L1 via the signal transducer and activators of transcription STAT1 and STAT3 and the IFN regulatory factor (IRF1) pathway [113]. In fact, an increase in PD-L1 expression, accompanied by increased infiltrating macrophages and reduced infiltrating CD3+ T-cells, was observed in ATR deficient melanoma models (Fig. 14.1) [114]. Remarkably, such preclinical data is supported by results from the phase I clinical trial investigating elimusertib monotherapy, where paired tumor samples from patients with PD-L1 positive tumors revealed upregulation of PD-L1 [80]. Interestingly, patients with metastatic melanoma that were previously resistant to PD-L1 inhibitors achieved durable responses when treated with the combination of ceralasertib and paclitaxel [115]. In this combinational trial, interlukin-12 fluctuations were also observed in patients that received clinical benefit suggesting activation of the innate immune response [115].

14.8 Candidate Biomarkers of ATR Sensitization

Therapeutic biomarkers are used as indicators of disease prognosis and predictive measures of treatment response [116]. Emerging data from various preclinical and clinical studies that evaluating ATR inhibitors as monotherapy or in combination strategies have identified candidate predictive biomarkers that may indicate sensitivity to ATR inhibition. Here, we summarize key genetic biomarkers and discuss their role in defining target patient populations that may respond best to ATR inhibitors.

ATM is a DDR kinase that senses and repairs dsDNA breaks and whose mutation may confer dependency on the ATR-CHK1 axis, offering an exploitable target for ATR inhibitors [117]. Although ATM is frequently mutated in cancer, the functional impact of many ATM variants is not well established [118]. Furthermore, there is significant overlap between ATM and ATR signaling pathways, as supported by various preclinical and clinical studies evaluating various cancer types including hematological and solid tumors [8, 12, 119]. Clinical responses have been reported from phase I studies of ATR inhibitors specifically in patients with ATM aberrations, including ATM deleterious mutations or protein loss [12, 16, 120]. Although ATM is frequently mutated in cancer, the functional impact of many ATM variants is not well established [118]. However, a large proportion of ATM mutations derive from missense variants, which can lead to a reduction in ATM protein expression levels [31]. This highlights the potential use of immunohistochemistry (IHC) analysis as a clinical tool to probe ATM expression levels and identify those who could benefit from ATR inhibition [31]. Pilie et al., further demonstrated the utility of IHC to understand ATM mutation annotations reported as variants of unknown significance (VUS), in which IHC analysis reported loss of protein in up to 25% of ATM VUS mutations, thus clarifying their functional impact [118]. This study also identified ATM loss of protein in patient tumor samples without identified ATM mutations, which points to the involvement of other mechanisms such as epigenetic or post-translational loss [118].

Another widely evaluated biomarker of ATR inhibitor sensitivity is p53, which plays a prominent role in G1 checkpoint control and whose loss comprises a high proportion of cancer cases [31]. Although there is preclinical data to support p53’s role as a predictive biomarker, the data remains inconsistent. For instance, Toledo et al., showed that cells with defective p53 had augmented replication stress in response to ATR inhibitors compared to cells with wildtype p53 [35]. A similar finding was reported in Kwok et al., in which treatment with the ATR inhibitor AZD6738 resulted in selective toxicity in p53 defective xenografts and cell lines [121]. In contrast, another study showed no increase in sensitivity to single agent ATR inhibition with VE-821 in p53 mutant cell lines compared to matched wildtype p53 cells [122]. Although Dillion et al., reported radio-sensitization by AZD6738 to single radiation fractions in a panel of cell lines, the narrow sensitivity range to AZD6738 was independent of p53 status [123]. Cumulatively, despite strong rationale supporting the use of ATR inhibitors to treat p53 deficient tumors, the conflicting data suggests further studies are necessary to assess its utility as a predictive biomarker of response. Although data is still pending, multiple clinical trials are underway to evaluate ATR inhibitors as monotherapy or in combination strategies in patients with solid tumors harboring TP53 mutations [48, 124].

A link between ATR sensitivity and deficiency of the BAF complex component AT-Rich Interactive Domain-containing protein 1A (ARID1A) was established in a large-scale genetic screen reported in Williamson et al. [125]. In this study, both in-vitro and in-vivo models were used to demonstrate wide-ranged genomic instability and cell death in ARID1A mutant cancer cell lines and tumors in response to ATR inhibition [125]. The clinical significance of this finding is highlighted by the fact that up to 7% of all cancers are associated with ARID1A loss and the frequency of loss is increased in certain cancers, for example, ARID1A loss is reported in up to 50% of clear cell ovarian carcinoma cases [126]. Further support for ATR inhibition in the setting of ARID1A loss was demonstrated in Tsai et al., in which an accumulation of R-loop formation was identified as a driver of replication stress in an ovarian cancer line with ARID1A knockout [126]. Translation of these data to the clinical setting has also produced compelling results. Antitumor activity was observed in patients with ARID1A-deficient solid tumors treated with the single-agent ATR inhibitor ceralasertib, including two patients that achieved RECIST-confirmed complete responses [127]. In addition, treatment with M6620 monotherapy resulted in a RECIST-confirmed complete response after 16 cycles in a patient with metastatic colorectal cancer with ARID1A mutation and IHC confirmed loss of both ARID1A and ATM, with a reported progression free survival of 29 months at their last assessment [128]. The use of cell-free DNA (cfDNA) as an indicator of treatment response was also evaluated in this study, which revealed declining levels of allele frequencies for ARID1A, among other identified mutations, to undetectable levels after 9 cycles of treatment with M6620 compared to baseline [128].

Targeting deficiencies in homologous recombination DNA repair (HRR) also offers a potential opportunity for ATR inhibition [129, 130]. For example, Krajewska et al., demonstrated sensitivity of the breast cancer cell line MCF-7 to ATR inhibitors upon inactivation of RAD51 in the HRR pathway [129]. Other studies have since further elucidated the major role ATR plays in regulating homologous recombination processes. For instance, Kim et al., showed that increased ATR signaling promotes the capacity of HRR in cancer cells by regulating the abundance of homologous recombination factors [28]. In support of this, a phase I trial of the Repare ATR inhibitor RP-3500 monotherapy observed multiple clinical responses in ovarian cancer patients with PARP-inhibitor resistant cancers that harbored actionable BRCA1 and RAD51C mutations [87, 117]. Other responses described in this study included patients with homologous recombination deficiency (HRD), with molecular alterations in ATM, BRCA2 and RAD51B/C [87]. ATR inhibition in HRD-cancers is largely under clinical investigation via multiple trials that are actively recruiting patients with deleterious mutations in HRR genes.

ATR deficiency has also been shown to confer a strong synthetic lethal response with many other DDR genes as well as with inducers of DNA replication stress [29]. For example, ATR inhibition is synthetic lethal in cells with genetic defects in genes such as APOBEC3A and B as well as with overexpression of cyclin E1 (CCNE1) and with c-MYC amplifications [28, 131,132,133]. In addition to those mentioned above, molecular defects in DDR genes such as ERCC1, XRCC1, CHK1, and FANCD2, and even accumulation of R-loops, all have been shown to produce a synthetic lethal effect in response to ATR inhibition [8, 28, 29]. With so many potential synthetic lethal partners possible, results from ongoing preclinical and clinical studies will be instrumental in identifying biomarkers and factors associated with therapeutic response as ATR inhibitors appear poised to enter the clinic in the coming years.

14.9 Concluding Remarks

This chapter provides a rationale for targeting ATR and summarizes the current landscape of ATR inhibitors in clinical evaluation. As a key component of the DDR, ATR is a promising druggable target that is being widely evaluated in phase I, II and III clinical trials as monotherapy and in combinations with other agents, including DNA repair inhibitors, chemo- and radiotherapy, and immunotherapy. Regardless of the approach taken, ongoing clinical studies must address optimization of the therapeutic window for this drug class. A predominantly reported toxicity across ATR inhibitors trials is myelosuppression, which is a mechanism-based toxicity that ultimately limits the therapeutic window in both monotherapy and combination approaches [134]. This carries key implications particularly for combination strategies due to potentiating of overlapping toxicities that may deepen myelosuppression and reduce drug tolerability. Proposed rational combination strategies should limit overlapping toxicity, which may be achieved by coordinating intermittent dosing schedules to facilitate tissue recovery. Another prevalent challenge is refining the target patient population most likely to benefit from ATR inhibition. Molecular technology advances and companion diagnostics have opened the door to precision oncology and the opportunity to offer personalized treatment strategies to patients [135]. Today, many clinical studies are designed on the basis of mutational status, which has led to the approval of several tumor-agnostic drugs [136]. Interestingly, many ongoing ATR inhibitor trials are recruiting patients based on molecular alteration rather than relying solely on tumor-type. A spectrum of molecular alterations have already been identified as potential predictive biomarkers that may sensitize to ATR inhibition; however, to be clinically efficacious, the biomarkers must be sensitive and easy to measure to allow for successful integration into the clinic. In closing, although several ATR inhibitors in development are poised to address a clinically unmet need, no ATR inhibitor has received FDA-approval for cancer indications thus far. We eagerly await the results from ongoing clinical studies as FDA-approval of ATR inhibitors lies close in sight.