Mutual ATRaction: Assessing Synergy of Berzosertib with Sacituzumab Govitecan

In this issue of Clinical Cancer Research, Abel and colleagues report results from the phase I portion of their phase I/II trial combining the antibody–drug conjugate (ADC) sacituzumab govitecan (SG) with the ataxia telangiectasia and Rad3-related (ATR) inhibitor (ATRi) berzosertib (VX-970, M6620) in patients with advanced solid tumors (ref. 1; Fig. 1). DNA damage repair (DDR) pathways are activated in response to genotoxic stress, and defects in the DDR response represent an important mechanism of acquisition of genetic alterations resulting in tumorigenesis, cancer progression, and resistance to treatment. These defects can also lead to therapeutic vulnerabilities such as to inhibitors of components of the DDR response, the most common and successful class targeting PARP, which is involved in DNA single-strand break repair (2). PARP inhibitors (PARPi) as monotherapy have demonstrated clinical activity for the treatment of BRCA-mutated cancers including breast, ovarian, prostate, and pancreatic, and are FDA-approved in these indications. Inhibitors of PARP and other DDR components can synergize with chemotherapeutics, radiotherapy, immune checkpoint inhibitors, and other DDR inhibitors in preclinical models, but there is limited evidence of clinical synergy for these combinations in trials. A major challenge in developing combination regimens including DDR inhibitors is overlapping toxicity, particularly myelosuppression, limiting dosing of the individual agents.

Berzosertib is an inhibitor of ATR, which is a serine/threonine kinase involved in sensing DNA damage and activating the DDR response (3). Unlike the ataxia telangiectasia–mutated (ATM) serine/threonine kinase, which responds primarily to DNA double-strand breaks (DSB), ATR responds to a wide variety of genotoxic stresses, including replication stress. Berzosertib has demonstrated synergistic activity in combination with cytotoxic chemotherapies in preclinical models (4), with modest benefit seen in clinical trials (5, 6), and more promising activity seen with berzosertib plus gemcitabine in ovarian cancer (7). Dose-reduction of berzosertib from the recommended phase II dose of 240 mg/m² as monotherapy to 90 mg/m² is required when combining with platinum chemotherapies (8), which may have contributed to the limited activity observed in combination with cisplatin/gemcitabine in bladder cancer (ref. 9; where the doses of the chemotherapy drugs were also reduced) and with carboplatin in prostate cancer (10). Hence, further exploration of berzosertib combinations would benefit from more targeted delivery of chemotherapeutic agents without dose-limiting myelosuppression.

ADCs utilize antibodies specific to cancer cell surface proteins to deliver cytotoxic payloads resulting in tumor death while minimizing off target effects. SG is an ADC comprised of an antibody targeting trophoblast cell-surface antigen 2 (Trop-2) conjugated through a hydrolyzable linker to the cytotoxic agent SN-38, the active metabolite of the topoisomerase 1 inhibitor irinotecan (11). Trop-2 is highly expressed in multiple cancer types, with overexpression linked to aggressive clinical behavior (12) including acquisition of neuroendocrine phenotype in prostate cancer (13). Therefore, in the study by Abel and colleagues, patients with solid tumors known to have DNA repair mutations or high replication stress, including neuroendocrine cancers, were selected to receive the combined treatment. All patients had previously received cytotoxic chemotherapy. Confirmed partial responses were demonstrated in two of 12 enrolled patients, and one patient achieved a metabolic response on positron emission tomography imaging. No grade 4 treatment-related adverse events other than lymphopenia were reported, and maximally selected dosing of SG 10 mg/m² with berzosertib 210 mg/m² (the dose previously tested in combination with gemcitabine (7) and topotecan; ref. 6) was tolerated. Lower rates of hematologic toxicity were seen compared with a previous report of berzosertib with the conventional TOP1 inhibitor topotecan (6), suggesting delivery of cytotoxics through ADCs as a promising approach to improving patient tolerability and maximizing dosing for antitumor activity.

High Trop-2 IHC expression was seen in pretreatment biopsy specimens from 2 of 3 responding tumors, but was also seen on non-responding tumors, which is concordant with studies demonstrating inconsistent response to SG even with high Trop-2 expression (14). Nonresponse to SG in these cancers is likely related to underlying cellular resistance to SN-38, which could not be overcome by the combination with berzosertib in this study. Furthermore, the lack of response in patients whose tumors harbored deleterious BRCA or ATM mutations even when berzosertib could be safely delivered at full dose in a combination regimen suggests the need for better biomarkers to predict sensitivity and resistance to this agent in this heavily pretreated population.

ATRis have been reported to be selectively toxic in cells with DDR defects, but also in cells reliant on alternative lengthening of telomeres and those with high levels of replication stress induced by oncogenes (e.g., Ras isoforms, Myc, Cyclin E) and hypoxia (3); however, evidence for clinical activity in these molecular contexts is limited. Because ATR and ATM both belong to the phosphatidylinositol 3-kinase-like kinase family and coordinate in activating the DDR response, ATRi is predicted to lead to synthetic lethality in tumors that harbor loss-of-function ATM alterations. Indeed, preclinically, the majority of cell lines sensitive to the ATRi BAY-1895344 (elimusertib) were characterized by mutations in the ATM pathway (15). Elimusertib as monotherapy demonstrated some clinical activity in a patient cohort with ATM loss by immunohistochemistry (16) with favorable disease control rate of 65%, but objective response rate (ORR) was modest at 9%. In the TRESR study of the ATRi camonsertib (RP-3500), ORR was 13% (13/99) in a biomarker-selected population, with responses seen in tumors with ATM, BRCA1, BRCA2, CDK12, RAD51C, and SETD2 genotypes (17). PLANETTE (NCT04564027) is a Phase IIa trial exploring the ATRi ceralasertib (AZD6738) in ATM-mutated castration-resistant prostate cancer, but results are not yet reported.

Thus, even in biomarker-selected populations, novel strategies are needed to increase activity of currently available ATRis. A number of combination strategies are under study (18), including multiple ongoing trials with standard chemotherapeutics. The combination of ceralasertib with the HER2-targeting ADC trastuzumab deruxtecan is being explored in the DASH trial (NCT04704661), and based on this report from Abel and colleagues, other ATRi + ADC combinations warrant further study. Preclinical studies suggest synergistic activity of ATRis with PARPis given their complementary roles in the DDR response, and trials of this strategy are in progress (NCT04267939, NCT04972110, NCT04497116, NCT04090567, NCT03787680). These combinations can lead to dose-limiting myelosuppression, so combination with the less myelosuppressive PARP1-selective inhibitor AZD5305 is attractive and under study (NCT02264678). Other studies in progress include combinations with immune checkpoint inhibitors (NCT04216316, NCT04095273, NCT03682289, NCT05450692), and inhibitors of BTK (NCT03328273) and PKMYT1 (NCT04855656).

Strategies for patient selection as well as drug delivery and dosing should be further optimized. Circulating biomarkers have been proposed to help guide clinical management with ATRis: in the TRESR study, circulating tumor DNA (ctDNA) responses on treatment correlated with clinical partial responses, suggesting that earlier treatment switch or treatment intensification could be considered in patients with sub-optimal ctDNA response. In a Phase I trial of the ATRi ART0380 (19), a blood-based assay found that γH2AX (a marker of DNA DSBs) increased in circulating tumor cells but not in peripheral blood mononuclear cells in patients with drug exposure predicted to be biologically effective, suggesting the possibility of using real-time pharmacodynamic markers to guide dosing. Further advances are needed in drug development and drug targeting incorporating novel technologies (20) to exploit therapeutic vulnerabilities related to the DDR response while minimizing toxicities and improving quality of life for patients.

S.A. Berg reports other support from Exelisis, Eisai, and BMS outside the submitted work. A.D. Choudhury reports grants from Bayer and Pfizer and personal fees from Eli Lilly, AstraZeneca, Astellas, Blue Earth, Janssen, Bayer, Tolmar, Sanofi Aventis, Pfizer, and Lantheus outside the submitted work.