BRCA1/2 Reversion Mutations in Japanese Patients with Metastatic Breast Cancer Progressing on Olaparib: OLIVE (WJOG15321B)

Abstract

Background

A key resistance mechanism to PARP inhibitors is the development of BRCA1/2 reversion mutations that restore the protein function.

Methods

Patients with metastatic breast cancer and germline BRCA1/2 mutations were recruited from 14 institutions and treated with olaparib. Targeted next-generation sequencing was performed on circulating tumor DNA (ctDNA) extracted from blood collected at base line and progression on olaparib (Guardant360). The primary objective was to evaluate the frequency of BRCA1/2 reversion mutations in this group.

Results

From November 2021 to October 2023, 60 patients were enrolled. The median age was 50.5 years (32–85). Of these, 12 (20.0%) had germline BRCA1 mutation and 48 (80.0%) had germline BRCA2 mutation. The median number of prior chemotherapies was 2 (0–7). Two patients had received platinum. Among patients with tumor progression on olaparib who had available ctDNA results, BRCA1/2 reversion mutations were identified in 25.0% (2/8; 95% CI 3.2–65.1%) with germline BRCA1 mutation and 23.5% (8/34, 95% CI 10.7–41.2%) with germline BRCA2 mutation. In one patient, BRCA2 reversion mutations were found in ctDNA seven months before radiological confirmation of progression on olaparib and again after progression. Among the ten cases with BRCA1/2 reversion mutations, all but one case involved secondary indels on primary indel sites. Reversion mutations were not observed in the C-terminal regions of BRCA1/2.

Conclusion

BRCA1/2 reversion mutations were detected in ctDNA as a mechanism of resistance to olaparib. Certain genomic regions and mutation types appeared to be particularly prone to reversion mutations.

Clinical trial registration

UMIN000046007.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12282-026-01855-2.

Introduction

Poly (adenosine diphosphate–ribose) polymerase (PARP) inhibitors have revolutionized the treatment of patients with breast cancer harboring germline BRCA1/2 mutation [1, 2]. However, resistance to PARP inhibitors can occur. Mechanisms of resistance to PARP inhibitors are classified into three categories: BRCA-dependent homologous recombination repair (HR) activity restoration (e.g. BRCA1/2 reversion mutation), BRCA-independent HR activity restoration (e.g. reversion mutations of RAD51C, RAD51D and PALB2 and loss of 53BP1) [3, 4], and HR-independent mechanisms (e.g. SLFN11 inactivation and overexpression of ABCC1 and ABCG2) [5, 6]. Secondary mutations that restore HR activity of BRCA1/2, known as reversion mutations, have been documented as a cause of resistance to PARP inhibitors and platinum [7, 8].

However, there are no studies reporting somatic BRCA1/2 reversion mutations in circulating tumor DNA (ctDNA) in Japanese populations with metastatic breast cancer and germline BRCA1/2 mutation. Several reports have suggested population-specific prevalence of BRCA1/2 variants [9, 10]; therefore, it is meaningful to evaluate reversion mutations in the Japanese population. The objective of this study is to evaluate the frequency of BRCA1/2 reversion mutations using ctDNA in Japanese patients with metastatic breast cancer harboring germline BRCA1/2 mutation who received treatment with olaparib. Additionally, we intended to characterize the clinical and molecular features of cases with BRCA1/2 reversion mutation.

Patients and methods

Patients with metastatic breast cancer carrying germline BRCA1/2 mutation were enrolled from 14 institutions and treated with olaparib. The study included patients who had not yet initiated olaparib (cohort 1), those who were receiving olaparib without evidence of disease progression (cohort 2), and those who had already completed olaparib therapy (cohort 3). Blood collection was performed at the baseline and end of treatment with olaparib. Treatment duration was defined as the time from the initiation of olaparib to the date of last dose. Study protocol was approved by the institutional review board (Approval number: 21–029-B) and was done in accordance with the Declaration of Helsinki and Ethical Guidelines for Medical and Biological Research Involving Human Subjects. Patients provided written informed consent before participation.

Circulating tumor DNA analysis

Blood was collected using Streck Cell-Free tubes, while frozen samples of plasma previously collected in EDTA and obtained with patient consent were also permitted for the ctDNA analysis for four samples. Tumor specimens before and/or after olaparib were collected from a limited number of participants; however, the results are not reported here. Deidentified blood samples were sent to Guardant Health (California, US) and plasma were extracted for targeted next-generation sequencing of 74 cancer-related genes (Supplementary Table 1), mostly driver genes, using the Guardant360 [11, 12]. Putative germline mutations were determined by a betabionomial algorithm [13]. A BRCA1/2 reversion mutation was defined as a base substitution changing a deleterious mutation to a non-deleterious mutation, an insertion or deletion (indel) restoring the open reading frame (ORF) disrupted by the deleterious mutation, or an in-frame deletion spanning and removing the deleterious mutation entirely.

Statistical analysis

The median duration of treatment was estimated using the Kaplan–Meier method, and the 95% confidence interval (CI) was calculated accordingly. Statistical analysis was carried out using JMP Student Edition 19.

Results

Patients and blood samples

The enrollment period was from November 2021 to October 2023, during which 60 patients were registered. Among them, 20.0% of patients had germline BRCA1 mutation and 80.0% of patients had germline BRCA2 mutation. Two and 42 patients had prior platinum and anthracycline treatment, respectively (Supplementary Table 2).

Forty-five patients (75%) discontinued olaparib treatment during the observational period, all due to disease progression (Supplementary Fig. 1). Median treatment duration was 11.7 months (95% CI 9.0–13.1). Baseline blood samples were collected from all the patients (N = 60), and post-treatment blood samples were obtained from 42 patients. Tumor tissue analysis data are not included in this report.

Incidence of BRCA1/2 reversion mutations

No BRCA1/2 reversion mutations were detected in any of the 13 pre-olaparib samples. Among those who discontinued olaparib due to tumor progression, reversion mutations occurred in 25.0% (2/8; 95% CI 3.2–65.1%) with germline BRCA1 mutations and 23.5% (8/34, 95% CI 10.7–41.2%) with germline BRCA2 mutations. In total, the incidence was 23.8% (10/42; 95% CI 12.1–39.5%). Multiple reversion mutation clones in BRCA2 were detected in a patient receiving olaparib (NG001), about 7 months before radiological progression. In this case, VAF of c.1403_1436del dramatically increased from 0.25% to 16.55%, much higher than that of the other reversion mutations, suggesting that this revertant clone may represent the dominant resistant population (Supplementary Fig. 5).

Median treatment duration was numerically longer in patients without detectable BRCA1/2 reversion mutation in the post progression sample (N = 32) compared to those with detectable BRCA1/2 reversion mutation (N = 10) (10.1 months [95%CI 8.1–11.8] versus 7.3 months [95% CI 3.6–8.3]) (Supplementary Fig. 2).

The distribution of primary BRCA1/2 mutation that underwent reversion

In cases with BRCA1 reversion mutations, the primary mutations originated in two regions: one case in the Really Interesting New Gene (RING) domain, and the other between the nuclear localization signal (NLS) and the coiled-coil domain (Fig. 1a). For BRCA2, four cases had primary mutations located between the N-terminus and the BRC repeat, two cases within the BRC repeats, and two cases between the BRC repeats and the DNA-binding domain (DBD) (Fig. 1b).

Fig. 1.

Distribution of primary mutation reverted in BRCA1 and BRCA2. a The positions of primary mutations are shown on the domain structure of BRCA1 (n = 2). Coding DNA sequence (CDS) numbers are indicated. RING, Zinc finger domain; NLS, nuclear localization signal; BRCT, BRCA1 C-terminal domain. b Same as in panel a but for BRCA2 (n = 8). According to the report by Pettitt SJ et al., BRCA2 c.750–775 was a frequent site of primary mutations in BRCA2 reversion mutations, which is shown as “Hotspot” [19]. Primary mutation of 2 cases is located on the same position (c.6626). Abbreviations: TA, transcriptional activation domain; H, helical domain; OB Folds, oligonucleotide binding folds; Tower, tower domain; TR2, Rad51 binding domain

Type of mutations that cause reversion mutations

Among the ten cases with BRCA1/2 reversion mutations, all but one cases involved secondary indels on primary indel sites (Table 1, Supplementary Fig. 3, Supplementary Fig. 4) and the one case harbored a secondary single nucleotide variant (SNV) on a primary SNV site, suggesting that the additional indels on the original indels were the major mechanisms of acquired resistance against olaparib through restoration of the homologous recombinant activity (Table 1).

Table 1.

The frequency of different pathogenic mutation types

A. BRCA1
	Primary mutations
	Deletion	Insertion	SNV	Total
Allele numbers	1	0	1	2
%	50.0 (95%CI: 1.3–98.7)	0.0 (95%CI: 0.0–84.2)	50.0 (95%CI: 1.3–98.7)	100
	Secondary mutations
Allele numbers	1	0	2	3
%	33.3 (95%CI: 0.8–90.6)	0.0 (95%CI: 0.0–70.8)	66.7 (95%CI: 9.4–99.2)	100

B. BRCA2
	Primary mutations
	Deletion	Insertion	SNV	Total
Allele numbers	5	3	0	8
%	62.5 (95%CI: 24.5–91.5)	37.5 (95%CI: 8.5–75.5)	0 (95%CI: 0–36.9)	100
	Secondary　mutations
Allele numbers	16	3	0	19
%	84.2 (95%CI: 60.4–96.6)	15.8 (95%CI: 3.4–39.6)	0.0 (95%CI: 0.0–17.6)	100

CI confidence interval

Mutational landscape in post progression on PARP inhibitor

Most post progression samples (41/42, 97.6%) harbored mutation(s) in genes other than BRCA1/2. Gene alterations of TP53 (22/42, 52.4%), TERT (15/42, 35.7%), ERBB2 (11/42, 26.2%), ESR1 (11/42, 26.2%), PIK3CA (9/42, 21.4%) and EGFR (9/42, 21.4%) were most commonly observed in post progression ctDNA samples (Fig. 2).

Fig. 2.

Analysis of ctDNA post progression on PARP inhibitors. Mutational landscape of ctDNA from 42 patients. Post-specific mutations are indicated by circles. Note that post-specific mutations cannot be determined in cohort 3. “Multi-hit” includes either the same or different types of classifications

Among patients whose paired samples were evaluable, frequency of post specific mutations were TP53 (9/22, 40.9%), TERT (0/22, 0%), ERBB2 (0/22,0%), ESR1 (3/22,13.6%), PIK3CA (4/22, 18.2%), and EGFR (3/22, 13.6%). The relevance of these gene alterations is not clear if they contributed to resistance specific to the PARP inhibitor although they have been involved in drug resistance to certain therapies [14–16] and may have implications for combination therapy with olaparib (e.g., oral Selective Estrogen Receptor Degrader (SERD) or phosphoinositide 3-kinase (PI3K) inhibitors) [17, 18].

Discussion

This study examined the frequency of BRCA1/2 reversion mutations occurred in Japanese patients with metastatic breast cancer that had progressed on olaparib. Most participants had not previously received platinum-based treatments, which made it possible to evaluate how frequently reversion mutations appeared upon PARP inhibitor resistance in actual clinical samples of breast cancer.

BRCA1/2 reversion mutations generally arise under selective pressure of homologous recombination deficiency (HRD)-targeted therapy, especially PARP inhibitors or platinum [19]. In this study, blood samples were collected from 40 patients either prior to treatment or during treatment—before clinical progression. Among them, the majority of participants (37; 92.5%) had a history of chemotherapy, with 27 (67.5%) receiving prior anthracycline and 2 (5.0%) receiving prior platinum-based therapy. Notably, no BRCA1/2 reversion mutations were detected in the pretreatment blood samples from those patients. However, substantial patients have developed BRCA1/2 reversion mutations after olaparib exposure, which indicates selective pressure of olaparib at least in part had played a role in inducing BRCA1/2 reversion mutations. The emergence of multiple BRCA1/2 reversion mutations—up to 8 mutations in the present study—indicates strong selective pressure by olaparib on tumor subclones [20]. In one case, BRCA2 reversion mutation emerged during treatment and before clinical progression, and VAF of which increased after continuous olaparib exposure, suggesting strong selective pressure of olaparib induced expansion of resistance clone (Supplementary Fig. 5).

A recent study reported that BRCA1/2 reversion mutations in ctDNA from 60% of patients with metastatic breast cancer at the time of disease progression on HRD-targeted therapy [21]. In contrast, studies including multiple cancer types have consistently shown lower frequencies, generally in the range of approximately 20–40% [22–24]. In our cohort, the frequency of the reversion mutations at disease progression on olaparib was approximately 24%. Several factors may have been attributed to discrepancy between OLIVE and a study by Harvey-Jones E, et al. [21]. First, patients’ backgrounds were different. For example, the prevalence of pathogenic germline BRCA1 mutation was higher in the study by Harvey-Jones E, et al.; 46.8% [21]. Second, the incidence of prior platinum exposure was higher in the study by Harvey-Jones E, et al.; 74.5% received prior (neo)adjuvant chemotherapy and 21.3% prior platinum containing palliative chemotherapy [21]. Third, they utilized GuardantINFINITY as ctDNA assay, which has higher sensitivity than Guardant 360, enabling detection of low-frequency variants [21].

ctDNA analysis has several inherent limitations. First, the limit of detection for mutant allele fraction increased with lower amount of cell free DNA input. Second, large reversion mutations may not be detected. Third, determining cis/trans configuration can be challenging because ctDNA fragments are short (< 200 bp). Therefore, comparisons across studies require caution. On the contrary, ctDNA is a powerful tool in evaluating the evolution of resistant subclones. Further investigation in a larger cohort is necessary.

A previous study showed that the resultant BRCA1 missense variant L63E in the RING domain by reversion mutation has HR activity [25]. Considering the data, the reversion mutation in BRCA1 would have acquired HR activity and became resistant to PARP inhibitor in one case (JF014). The other BRCA1 and BRCA2 alterations by reversion mutation, except for SU006, were located in exon 10 of BRCA1 and exon 10–11 of BRCA2. These regions are considered “coldspots”, that tolerate variation, and pathogenic missense variants are rarely observed [26]. Although functional relevance should be interpreted cautiously, all revertant have the homologous recombination activity and resulted in resistance to PARP inhibitor.

Previous studies identified regions more likely to acquire reversion mutations, as well as regions less prone to reversion events [19, 24]. Consistent with our study, BRCA2 reversion mutations were frequently observed in BRC repeats or the N-terminal region, between the PALB2-interacting domain and the BRC repeats. In contrast, reversion mutations were relatively rare in the C-terminal region, encompassing the DNA-binding domains (DBD), possibly because this region is highly conserved and critical for homologous recombinant function [27, 28]. Particularly, RAD51-binding TR2 motif in the C-terminal region is essential in protection and restart of stalled replication forks [29]. In BRCA1, the BRCA1 C-terminal (BRCT) domain and RING domain were found to be the hotspots [24]. In our cohort, only two cases with germline BRCA1 mutations were identified to be a revertant, which located them outside of the hotspots and limited the comparison with previous reports across these domains due to the small sample size. A subgroup analysis of the phase III PAOLA-1/ENGOT-ov25 trial indicated that maintenance olaparib and bevacizumab was especially effective for advanced high-grade ovarian cancer patients with DBD mutations in BRCA1/2 [30]. These findings suggested that BRCA1 and BRCA2 pathogenic variants in the DBDs were less prone to acquiring reversion mutations, which may have allowed tumor cells to stay homologous recombination deficient leading to durable responses to PARP inhibitors.

Consistent with previous reports, BRCA1/2 frameshift indels were more prone to reversion than SNV [31]. Previous studies have indicated that restoration of the reading frame can be achieved by diverse secondary indels generated through error-prone classical nonhomologous end joining (NHEJ) or microhomology-mediated end-joining (MMEJ), allowing partial recovery of homologous recombination function without requiring exact sequence restoration [31].

This study is subject to several major limitations. Firstly, the sample size is relatively small. The caution should be used for the interpretation of the contents. A future integrated analysis across multiple cohorts would enhance the robustness of the current findings. Secondly, we didn’t evaluate the clinical endpoints (e.g. progression free survival, objective response rate, overall survival) with olaparib. Therefore, we couldn’t correlate BRCA1/2 reversion mutations with those endpoints. A previous study showed that pre-treatment BRCA1/2 reversion mutation inversely correlated with response to rucaparib in patients with ovarian cancer [32]. As increasing number of patients with metastatic breast cancer currently receive platinum-based therapy prior to olaparib, evaluating the correlation of pre-treatment existence of BRCA1/2 reversion mutation and response might be meaningful. Thirdly, we cannot exclude co-existence of other resistant mechanisms to olaparib. Guardant 360, which we utilized in the study, didn’t cover HR related genes including PALB2, RAD51C, or RAD51D and genes implicated in resistance to PARP inhibitors such as TP53BP1, RIF1 and PAXIP1. Fourth, we didn’t collect detailed treatment histories other than prior platinum and anthracycline exposure. Although two patients had received platinum, we don’t have information on whether it was administered in the perioperative or metastatic setting. Finally, all patients received olaparib. PARP inhibitors vary in their ability to trap; olaparib shows relatively moderate trapping activity [33]. It is unclear whether PARP trapping level among the various PARP inhibitors impacts on the emerging mutations in patients with that treatment. Further studies are warranted. Despite the limitations, our findings are of clinical value for further development of precision medicine in patients with BRCA1/2 through monitoring of reversion mutations with ctDNA and devising strategies to overcome resistance.

Conclusions

BRCA1/2 reversion mutations were detected in ctDNA as a mechanism of resistance to olaparib. Certain genomic regions and mutation types appeared to be prone to reversion mutations. In addition, the presence of alternative resistance mechanisms besides BRCA1/2 reversion was yet to be elucidated.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the conceptualization of the study. Data collection was carried out by Hitomi Sakai, Junji Tsurutani, Mari Hosonaga, Kazuki Nozawa, Manabu Futamura, Yuko Tanabe, Toru Mukohara, Kazuhiro Shiraishi, Tatsuya Toyama, Hiroyuki Yasoji, Noriyuki Watanabe, Hideki Sakai, Tsutomu Iwasa, Nobumichi Takeuchi, and Hiroshi Tada. Methodological development was undertaken by Rika Kusumoto-Matsuo, Kazuko Sakai, and Natsuko Chiba. Hitomi Sakai and Rika Kusumoto-Matsuo were responsible for data analysis. The study was supervised by Junji Tsurutani. The initial manuscript draft was prepared by Hitomi Sakai, with all authors providing feedback on earlier versions. All authors reviewed and approved the final manuscript.

Funding

Open Access funding provided by SHOWA Medical University. This work was supported by AstraZeneca K.K.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

HS received honoraria from Daiichi Sankyo, Eli Lilly, Gilead Sciences, AstraZeneca, and Pfizer, and medical writing support from Daiichi Sankyo outside the submitted work. MF received honoraria from Chugai and Daiichi Sankyo. TM received research funding (to the institution) from MSD, AstraZeneca, Eisai, Sanofi, Novartis Pharma, Bayer, ONO Pharmaceutical, Daiichi Sankyo, Chugai Pharmaceutical, Pfizer, Gilead Sciences, Merck Biopharma, Seagen, Astellas Pharma, Sysmex, and Exact Sciences; and honoraria from Terumo, Bayer, MICIN, Kyowa Kirin, Daiichi Sankyo, Eli Lilly, MI-Force, AstraZeneca, Eisai, Taiho Pharmaceutical, Chugai Pharmaceutical, Shionogi, Pfizer, and Ono Pharmaceutical. He is an advisory board member of Pfizer and AstraZeneca. TaT received honoraria from Chugai, Daiichi Sankyo, Pfizer, MSD, Eli Lilly, Novartis, and Taiho. HT received honoraria from Chugai Pharmaceutical, Eli Lilly, Pfizer, and Daiichi Sankyo. KS received honoraria for lectures from Takeda Pharmaceutical Co., Ltd., Nippon Kayaku Co., Ltd., and Qiagen, Inc., and honoraria for manuscript writing from Life Technologies Japan Ltd. NC received research funding from AstraZeneca outside the submitted work. ToT received honoraria from Daiichi Sankyo, Eli Lilly, and Gilead Sciences. JT received research funding from Daiichi Sankyo Co., Ltd., outside the submitted work; honoraria from Daiichi Sankyo Co., Ltd., AstraZeneca K.K., Eli Lilly Japan K.K., Pfizer Japan Inc., Eisai Co., Ltd., Kyowa Kirin Co., Ltd., Taiho Pharmaceutical Co., Ltd., MSD K.K., and Gilead Sciences, Inc.; consulting fees from Daiichi Sankyo Co., Ltd., AstraZeneca K.K., Seagen Inc., Pfizer Japan Inc., and Gilead Sciences, Inc.; and his institution received research funding from Eli Lilly Japan K.K., Kyowa Kirin Co., Ltd., Taiho Pharmaceutical Co., Ltd., Chugai Pharmaceutical Co., Ltd., MSD K.K., and Eisai Co., Ltd., outside the submitted work. All other authors declare no competing financial or non-financial interests.

Ethical approval and consent to participate

Study protocol was approved by the institutional review board (Approval number: 21-029-B). Informed consent was obtained from all participants involved in the study.