PARP inhibitors, use early in the first-line maintenance therapy setting but with caution

In recent years, the advent of poly-(ADP-ribose) polymerase (PARP) inhibitors (PARPis) has profoundly changed the treatment landscape of ovarian cancer (OC) (1-5). Mechanistically, PARPis act by blocking the catalytic activity of PARP enzymes as well as by trapping PARP enzymes on their DNA substrates, resulting in toxic PARP-DNA complexes, which in turn trigger the collapse of replication forks (6). It is well documented that such emergent DNA double-stranded breaks (DSBs) are mainly resolved by a homologous recombination DSB repair machinery (7). As such, approximately half of high-grade serous OC attains greater benefit from PARPis because of its homologous recombination-deficient (HRd) phenotypes (8). Notably, earlier clinical trials also demonstrated clinical benefit with PARPis in homologous recombination-proficient (HRp) OC (2-3,5) and rucaparib became one of the three U.S. Food and Drug Administration (FDA)-approved PARPis, for the 2^nd or greater line maintenance therapy setting for both HRd and HRp OC based on the findings from ARIEL 3 study (5).

In this ATHENA-MONO trial (9), Monk et al. tested the efficacy of rucaparib maintenance therapy in the first line setting for a broad population. ATHENA/GOG-3020/ENGOT-ov45 is a multicentric, randomized, double-blind, placebo-controlled phase III trial, consisting of two independent cohorts, which evaluate rucaparib alone (ATHENA-MONO) or in combination with anti-PD1, nivolumab (ATHENA-COMBO) in the upfront maintenance setting. While ATHENA-COMBO is currently ongoing, results from ATHENA-MONO met its primary endpoints. Specifically, at median follow-up of approximately 26 months, rucaparib monotherapy significantly improved the median PFS (mPFS) compared with placebo in HRd tumors [28.7 vs. 11.3 months; hazard ratio (HR) 0.47; 95% confidence interval (CI), 0.31–0.72], as well as in the intent-to-treat (ITT) population (20.2 vs. 9.2 months; HR 0.52; 95% CI, 0.40–0.68). Moreover, in preplanned exploratory analysis, HRp population also derived the benefit from rucaparib monotherapy with 12.1 months of mPFS compared to 9.1 months with placebo (HR 0.65; 95% CI, 0.45–0.95) (9).

One of the strengths of ATHENA-MONO is a broad population enrolled, providing more data in non-HRd OC, while other trials investigated the role of PARPis in more specific populations. SOLO-1 was the first trial to demonstrate the clinical benefit of olaparib as a first line maintenance therapy in BRCA mutant OC (1). PRIMA study investigated niraparib in a high-risk OC population e.g., stage IV, or stage III with residual disease, but not limited to BRCA mutation or HRd groups (2). ATHENA-MONO extends the patient population to both high and low risks of relapse [i.e., patients diagnosed with the International Federation of Gynecology and Obstetrics (FIGO) stage III, and patients undergoing primary debulking surgery with no residual disease] to examine potential benefit from rucaparib maintenance in all advanced OC patients, beyond DNA repair dysfunction or risk of relapse status (8).

While exciting, there is recognition that PFS benefit with PARPis may not translate into increased overall survival (OS) in the recurrent setting. Updated survival data of each of the three different PARPis prompted recent revisions to the FDA approvals for PARPis. For instance, no OS benefit was seen in the final analysis of ARIEL3 (10). In the nested primary analysis of ARIEL3, a HR was 0.832 (95% CI, 0.581–1.192) in the BRCA mutant patients, while HR was 1.005 (95% CI, 0.766–1.320) and 0.995 (95% CI, 0.809–1.223) in the HRd and ITT populations, respectively. In subsequent analysis of non-nested exploratory subgroups related to tumor molecular characteristics, no statistically significant differences among treatment arms were found; HRs of 1.280 (95% CI, 0.841–1.948), 1.153 (95% CI, 0.784–1.695), and 0.673 (95% CI, 0.305–1.483) in the BRCA wild-type HRd, BRCA wild-type HRp, and BRCA wild-type homologous recombination unknown subgroups, respectively. Also, in final OS analysis of ARIEL4, a median OS (mOS) favored the standard of care chemotherapy arm over rucaparib monotherapy in relapsed BRCA-mutated OC patients; mOS of 19.4 vs. 25.4 months in rucaparib and chemotherapy groups, respectively (HR 1.313; 95% CI, 0.999–1.725) (11). However, it is noteworthy that there was a high crossover rate (69%) in ARIEL4, with a total of 90% of patients having received rucaparib from randomization or after crossover. This may have provided a bias in evaluating the OS data, as they were unadjusted for subsequent PARPis therapy. Despite these challenges of interpreting the results, the FDA withdrew rucaparib indication as monotherapy for the late line treatment of BRCA mutated OC patients and anticipated the future withdrawal from 2^nd or greater line maintenance therapy for non-BRCA mutant patients (12). Similarly, niraparib indication was already withdrawn for 2^nd or greater line maintenance therapy for non-BRCA mutant patients based on the long-term follow-up data of the NOVA trial, in which OS demonstrated an HR of 1.06 (95% CI, 0.81–1.37) in the non-germline BRCA mutated cohort (12,13). Although it is hard to dissect the exact mechanisms of those contributing to the OS outcomes at this time, it is evident that our approach to the use of PARPis will most likely focus on the earlier maintenance setting based on the updated indications (12).

There are potential explanations to the discrepancies between PFS and OS data. It is possible that OS data may have been influenced by post-progression therapies and cross-over to the investigational treatments, and none of the aforementioned trials were designed to investigate those effects. Secondly, overlapping mechanisms of resistance may have affected those final analyses, thus indicating the importance of tailoring the subsequent treatments. Thus far, several mechanisms of resistance to PARPis have been identified via clinical and preclinical studies. Specifically, mechanisms involved in restoring homologous recombination functionality e.g., BRCA reversion mutations and epigenetic upregulation or RAD51 reversion mutation are well known cross-resistant mechanisms with platinum drugs (14-16). Thus, it is possible that tumors progressed on the PARPis may have had poor response to subsequent platinum-based therapies. Other mechanisms were also studied via various preclinical models e.g., loss of p53-binding protein 1 (53BP1), loss of REV7 gene (17,18), replication fork stabilization and increased drug efflux, all requiring further studies in the clinical settings (16,19-21). Of note, those resistance mechanisms are not mutually exclusive and may cause cross-resistance to other non-platinum agents. Hence, the complexity of PARPis clinical resistance underscores the necessity of tailoring the subsequent treatments in a molecularly selected subgroups after progression on PARPis. As such, molecular analysis of fresh tissue and/or circulating tumor DNA samples will be important to study clinical resistance of PARPis and design the next generation clinical trials for this newly emerging subgroup of patients with unmet medical needs.

In conclusion, given the undisputable benefit of PARPis in the OC treatment landscape, the use of PARPis should not be discouraged, especially in the upfront maintenance therapy setting. It is noteworthy that the 7-year follow-up data from the SOLO-1 trial (22) and the 5-year OS data from the PAOLA-1 study (23) exhibit encouraging OS data (SOLO1: HR for OS of 0.55; 95% CI, 0.40–0.76; PAOLA-1: HR for OS in HRd population of 0.62; 95% CI, 0.45–0.85). As such, long term survival data from ATHENA-MONO are eagerly awaited. At the same time, it is necessary to broaden our understanding on clinically relevant PARPis resistance and conducting hypothesis driven translational studies to identify primary vs. acquired resistance to PARPis as well as our best use of PARPis with caution in the right setting for the right population.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.