Phase II trial of veliparib and temozolomide in metastatic breast cancer patients with and without BRCA1/2 mutations

We evaluated the efficacy and safety of poly-(adenosine diphosphate-ribose) polymerase (PARP) 1 and 2 inhibitor veliparib and temozolomide in metastatic breast cancer patients with and without germline BRCA1/2 mutations. In this single-arm phase II trial, patients with metastatic breast cancer received veliparib 30 to 40 mg twice daily on days 1 to 7 with concurrent temozolomide 150 mg/m2 on days 1 to 5 of a 28-day cycle. The primary cohort was unselected for BRCA mutation status, and an expansion cohort enrolled only BRCA1/2 carriers. The primary endpoint was objective response rate (ORR) in each cohort. Secondary endpoints included progression-free survival (PFS), clinical benefit rate (CBR), and evaluation of safety and tolerability. In the primary cohort of 41 unselected patients, which included 9 BRCA mutation carriers, the ORR was 10% and clinical benefit rate at 4 months (CBR) was 27%. In the expansion cohort of 21 BRCA1/2 carriers, the ORR was 14% and CBR was 43%. Among all 30 BRCA1/2 carriers, the ORR was 23% versus 0% among non-carriers. In the subset of BRCA1/2 carriers, the ORR was 32% among platinum-naïve patients versus 9% among platinum-exposed patients. The median PFS was 3.3 months among BRCA1/2 carriers compared to 1.8 months among non-carriers (HR: 0.48, p = 0.006). A longer median PFS of 6.2 months was observed among BRCA1/2 carriers who had no prior platinum therapy. The most common grade 3 and 4 toxicities were thrombocytopenia (32%) and neutropenia (21%) that generally improved with dose modifications. Veliparib and temozolomide demonstrated clinical activity in platinum-naïve BRCA-associated metastatic breast cancer with manageable toxicity at doses of veliparib well below the single-agent active dose. Although the study did not meet its primary endpoint in unselected nor BRCA-associated breast cancer, this regimen was further evaluated in the BROCADE 2 study. NCT01009788 (ClinicalTrials.gov), November 9, 2009


Introduction
BRCA1/2-deficient breast cancers account for 5-10% of all breast cancers and about 20-25% of hereditary breast cancers [1][2][3][4]. Women with germline BRCA1/2 mutations have a cumulative risk for developing breast cancer ranging from 49 to 57% by age 70 [5]. BRCA1 and BRCA2 are required for double-strand DNA break repair (DSBR) through the homologous recombination (HR) pathway, which is defective in BRCA-mutant cells [6,7]. Poly-(adenosine diphosphateribose) polymerase (PARP) enzymes play important roles in DNA repair. In particular, PARP-1 is critical for singlestrand break repair (SSBR) and also facilitates base excision Jing Xu and Tanya E. Keenan have contributed equally to this work 1 3 repair (BER). Although the mechanism of cell death caused by PARP inhibition is not fully understood, it is generally accepted that PARP inhibition promotes synthetic lethality in BRCA-mutant cells [6,8,9], by preventing SSBR, thus allowing the conversion of single-strand breaks into double-strand breaks, which are not efficiently repaired by cells deficient in HR. Accumulation of double-strand breaks leads to apoptosis. Additionally, PARP inhibitors can trap PARP enzymes at damaged DNA sites by forming DNA-PARP inhibitor complexes, which can cause DNA damage and cell death [10]. In the presence of the BRCA deficiency, PARP inhibition sensitizes tumor cells to DNA-damaging chemotherapies, such as platinum compounds, topoisomerase inhibitors, and alkylating agents [6].
Clinical studies in breast cancer have demonstrated that certain single-agent PARP inhibitors have substantial antitumor activity in patients with BRCA1/2 mutations. This has led to the approval of olaparib and talazoparib for germline BRCA-mutated, HER2-negative metastatic breast cancer [8,11,12]. Combination regimens of PARP inhibitors with chemotherapy agents, targeted therapies, antibody drug conjugates, checkpoint inhibitors, and other therapies are currently being explored. The optimal combination agent for PARP inhibitors in breast cancer remains unknown [13][14][15][16][17]. Early studies reported PARP inhibition potentiates the activity of temozolomide (TMZ), an orally administered alkylating agent, which has the advantage of crossing the blood-brain barrier (BBB) [18][19][20][21][22]. Veliparib, an investigational oral PARP-1 and PARP-2 inhibitor, also efficiently crosses the BBB [23]. Single-agent TMZ demonstrated no significant clinical activity in metastatic breast cancer [24,25]. However, in preclinical studies, veliparib potentiated the activity of TMZ, demonstrating antitumor activity in in vivo models of breast cancer, including tumors resistant to TMZ monotherapy [26]. In clinical settings, this combination has demonstrated activity in relapsed small cell lung cancer, metastatic colorectal cancer, and acute myeloid leukemia and may represent a promising treatment option for metastatic breast cancer [27][28][29].
In this phase II trial, we evaluated the efficacy and safety of veliparib and TMZ in patients with metastatic breast cancer (MBC). The initial cohort of 41 patients included all subtypes of MBC, and an expansion cohort of 21 patients included only patients with germline BRCA1/2-mutated MBC.

Patients
The initial study population included patients with all subtypes of metastatic breast cancer treated with at least one prior line of chemotherapy for metastatic disease. The expansion cohort enrolled patients with metastatic breast cancer and a known deleterious BRCA1/2 mutation, without limitation on prior treatment. Other eligibility criteria included measurable disease by Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 [30], normal organ and marrow function, and Eastern Cooperative Oncology Group (ECOG) performance status of ≤ 2. Both cohorts allowed for previously treated stable brain metastases. Key exclusion criteria included treatment with chemotherapy, biological therapy, targeted therapy, or radiotherapy within 2 weeks, or anti-cancer hormonal therapy within 24 h before starting the study treatment. All patients provided written informed consent prior to study entry.

Study design and treatment
This open-label single-arm phase II study was conducted at Massachusetts General Hospital, Dana-Farber Cancer Institute, and Beth Israel Deaconess Medical Center. The study protocol and informed consent form were reviewed and approved by the Dana-Farber/Harvard Cancer Center (DF/ HCC) institutional review board. This study was performed according to the Declaration of Helsinki and International Conference on Harmonization (ICH) Good Clinical Practice (GCP) guidelines.
The study treatment consisted of veliparib on days 1 through 7 with concurrent TMZ on days 1 through 5 of a 28-day cycle. The study initially dosed veliparib at 40 mg orally twice daily recommended in Phase I study and later reduced veliparib to 30 mg twice daily for all patients after one patient experienced grade 4 thrombocytopenia during the first cycle of treatment. The expansion cohort dosed veliparib at 30 mg twice daily for all patients. All patients in both cohorts received TMZ 150 mg/m 2 oral once daily, which was increased to 200 mg/m 2 in cycle 2 as tolerated.
Patients received study treatment until disease progression, unacceptable toxicity, death, withdrawal of consent, or loss to follow-up. Treatment could be interrupted for grade 3 or 4 treatment-related toxicities for up to 21 days, after which the study drugs were stopped if the toxicity was still not resolved. Only one dose reduction was allowed for veliparib before discontinuation of study treatment, and the TMZ dose could not be reduced below 75 mg/m 2 daily.

Study endpoints and assessment
The primary endpoint of the study was objective response rate (ORR) based on tumor assessment every 8 weeks according to RECIST 1.1. Secondary endpoints included clinical benefit rate (CBR) at 4 months, progression-free survival (PFS), and adverse event rate. CBR at 4 months was defined as the percentage of patients who had CR, PR, or SD for greater than 16 weeks of follow-up. Adverse events were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE).

Statistical analysis
A target ORR of 20% was considered to be of clinical benefit for this combination regimen in the primary cohort based on the fact that TMZ showed no single-agent activity in breast cancer at that time. A sample size of 41 was estimated to have 93% power to detect a response rate of 20% compared to a null response rate of 5% at a one-sided significance level of 0.05. If at least 8 of 41 patients (20%) had response (CR/ PR), this combination therapy would warrant further investigation. Based on the initial finding that 50% of BRCA1/2 carriers achieved a partial response in the primary cohort, a target ORR response rate of 45% was selected for the expansion cohort, with a null response rate of 15% deemed to be of little clinical interest. The enrollment of 20 patients provided 87% power to identify the combination as worthy of further investigation if the true response rate was 45% at a one-sided significance level of 0.025. If at least 7 of 20 patients had response (CR/PR) at the end, this combination therapy would warrant further investigation.
Efficacy analyses were assessed in all patients, who received at least one cycle of study drugs. ORR and CBR at 4 months were reported as point estimates with 90% confidence intervals. PFS, defined as the time from enrollment until disease progression or death, was estimated by the Kaplan-Meier method. In subgroup analyses stratified by BRCA1/2 mutation status and prior platinum therapy, ORR and median PFS (mPFS) were compared by the Fisher's exact test and log rank test, respectively. Safety data were evaluated for all patients who received at least one dose of either study drug.

Patient characteristics
Of 63 patients with metastatic breast cancer were enrolled in this study, 62 patients received study drugs and were included in the analysis. The primary cohort included 41 patients unselected for BRCA1/2 mutations or breast cancer subtype, and the expansion cohort included 21 patients who all had BRCA1/2 deleterious mutations. Table 1 shows the baseline characteristics of patients in these two cohorts. Most patients had an ECOG score of less than 2 and one patient in the primary cohort was male. The primary cohort included TNBC (54%), HR+ (37%), and HER2+ (10%) patients, while the expansion cohort included predominantly HR+ patients (57%). Nine of 41 (22%) patients in the initial cohort had BRCA1/2 mutations compared to all subjects in the expansion cohort. The most common metastatic sites were lymph nodes in the primary cohort and bone lesions in the expansion cohort. The median number of prior chemotherapeutic, hormonal, or HER2 directed regimens was 3 (range, 1-9) in the primary cohort and 2 (range, 0-9) in the expansion cohort. A total of 14 patients (34%) in the primary cohort and eight patients (38%) in the expansion cohort had received prior platinum therapy for metastatic disease.

Response to treatment
Of 62 patients who received treatment, 10 patients had no follow-up imaging due to rapid clinical progression (7 in the primary cohort and 2 in the expansion cohort) or early death (one subject in the expansion cohort). 34 patients in the primary cohort and 18 patients in the expansion cohort were evaluable for response ( Table 2).
In the combined overall population of 62 patients, the ORR (CR + PR) was 12% (7/62) and the CBR at 4 months was 32% (20/62); the primary cohort had a 10% (4/41) ORR and 27% (11/41) CBR at 4 months, while the BRCA1/2 expansion cohort had a 14% (3/21) ORR and 43% (9/21) CBR at 4 months. The waterfall plots demonstrated the best overall response of evaluable patients in each cohort (Supplemental Figure S1). In the primary cohort (n = 34 evaluable, Supplemental Figure S1A), all four patients who achieved CR or PR had BRCA1/2 mutations. Thirteen patients in this cohort, including one who had an unconfirmed partial response, demonstrated stable disease according to RECIST 1.1 criteria. Seventeen patients had PD, including eight with stable target lesions but with non-target progression or new lesions elsewhere. In the BRCA1/2-positive expansion cohort (n = 18 evaluable, Supplemental Figure S1B), three patients achieved PR, an additional 10 patients had SD, and five patients had PD, including one patient who achieved SD in target lesions but had new lesions. Combining both cohorts, all seven patients who had achieved a response (CR or PR) were BRCA1/2 carriers, including five patients with BRCA2 mutations and two patients with BRCA1 mutations (Fig. 1). There were no responses in patients without BRCA1/2 mutations (Figs. 1 and 2).

Survival analysis
The median PFS was 2.1 months (90% CI 1.8 to 3.0 month) in the overall population (Fig. 3a), and 1.8 months and 3 months in the primary and expansion cohorts, respectively (Supplemental Figure S2). The median PFS among BRCA1/2 carriers was 3.3 months (90% CI 2.3 to 6.2 month) compared to 1.8 months (90% CI 1.6 to 2.3 month) among non-carriers (HR: 0.48, p = 0.006; Fig. 3b). Among patients who had not received prior platinum therapy, the median PFS was 2.7 months (90% CI 1.9 to 4.0 month) compared to 1.9 months (90% CI 1.3 to Table 2 Best objective response to treatment This table shows the number of patients with each response (%) in he primary cohort, the expansion cohort and the total study population; the denominators used in all calculations consisted of the total numbers of patients in each group; the subgroup analyses by status of BRCA1/2 mutation and the prior platinum treatment are presented in Table 3 CR complete response, PR partial response, SD stable disease, PD progressive disease, NE non-evaluable, ORR objective response rate; CBR at 4 month Clinical benefit rate at 4 months a 7 patients had no follow-up imaging due to rapid clinical progression; one of these patients had a BRCA2 mutation b Patients had no follow-up imaging due to rapid clinical progression (n = 2) or early death (n = 1)  Fig. 3d). One patient, a BRCA2 carrier who had no prior platinum therapy, had a durable complete remission for at least 5 years as of last contact. No significant difference in PFS was observed by breast cancer subtype (Supplemental Figure S3). Of the 62 patients, one patient received only one dose of veliparib and TMZ and experienced no treatment-related adverse events. The most common all-grade AEs were thrombocytopenia, nausea, fatigue, anemia, and leukopenia (Table 4), and the most frequent grade 3 or higher AEs were thrombocytopenia, neutropenia, leukopenia, nausea, and vomiting. The most common AEs that caused dose delays or modifications were thrombocytopenia and neutropenia. One patient discontinued study treatment due to prolonged thrombocytopenia. Another patient died of sepsis after receiving 4 cycles of treatment.

Discussion
Although neither cohort in this study met the prespecified target ORR, the combination of veliparib and TMZ demonstrated antitumor activity among metastatic breast cancer patients with BRCA1/2 mutations, particularly  those who had not received prior platinum chemotherapy. The objective response rate was significantly greater and the median PFS was significantly longer among BRCA1/2 carriers than non-carriers, and among patients who had not received prior platinum chemotherapy than those who had. A higher ORR (32%) and CBR at 4 months (58%), and a longer PFS (6.2 months) were observed in the subset of BRCA1/2-positive patients who had not received prior platinum therapy. Although the study was initially designed to test the hypothesis that the combination of TMZ and veliparib was active in metastatic breast cancer regardless of BRCA mutation status, there was no clinical activity observed in non-carriers despite encouraging preclinical data. This was the first study to evaluate the efficacy of veliparib and TMZ in breast cancer. A previous trial of singleagent veliparib (400 mg BID) in platinum-naïve BRCAmutant breast cancer showed response rates of 14% (3/22) for BRCA1+ patients and 36% (8/22) BRCA2+ patients [31]. These response rates were comparable to the ORR of 32% observed in this study among platinum-naïve BRCA+ patients (Table 3). Notably, patients in the present study received veliparib twice daily at 30-40 mg, about one tenth of the dose used in the previous trial, suggesting that the observed activity in our study was unlikely due to the single-agent activity of veliparib. Rather, our data support that TMZ enhanced the efficacy of veliparib in BRCAdeficient patients. The observation that this combination led to responses only in patients with BRCA1/2 deficiency suggests that responses were dependent on the mechanism of synthetic lethality, in which inhibition of PARP in the presence of BRCA deficiency made cancer cells more vulnerable to TMZ-induced DNA damage. This may explain why the addition of TMZ to veliparib exhibited antitumor efficacy at a lower dose of veliparib, whereas single-agent veliparib required a higher dose for clinical activity in this population [31].
In this study, the combination of veliparib and TMZ had no activity in patients who had progressed on prior platinum therapy. This result is consistent with prior work showing that PARP inhibitors and platinum agents share common mechanisms of resistance [10,32,33], at least partly explained by secondary somatic mutations that restore the function of BRCA1/2 proteins [32,34,35]. Other overlapping resistance mechanisms to both PARP inhibitors and platinum therapy include microRNA-mediated restoration of HR, replication fork stabilization, PARP-1 mutations, and drug efflux pumps [10,36,37].
Previous studies showed PARP inhibition potentiates the activity of TMZ and exacerbates TMZ hematological toxicity [18][19][20][21][22]. Accordingly, thrombocytopenia was the most common hematological toxicity in this study. The severity of hematologic toxicities in this study was comparable to a subsequent study with the same combination regimen [17] and was more prominent than in patients receiving veliparib monotherapy [31]. One patient in this study discontinued treatment due to thrombocytopenia, and another patient died due to severe infection possibly related to the study treatment. The observed hematological toxicity was effectively managed in most patients with dose reductions of TMZ.
This proof-of-concept phase II study had several limitations. The single-arm design and lack of a comparison group prevent drawing conclusions about the efficacy of this combination therapy compared to single-agent veliparib. Furthermore, although TMZ has no demonstrated activity in metastatic breast cancer, it has not been evaluated in breast cancer patients with BRCA1/2 mutations. Additionally, the small sample size precluded analyses with adequate power to detect enhanced efficacy in subgroups. Despite these limitations, this study suggested that veliparib with TMZ had clinical activity in platinum-naïve BRCA-deficient breast cancer patients and led to a multicenter randomized phase II BROCADE study of veliparib and TMZ compared to carboplatin and paclitaxel with or without veliparib among patients with BRCA-associated metastatic breast cancer [17,38]. The BROCADE trial found that TMZ with veliparib yielded inferior ORR and PFS compared to carboplatin and paclitaxel with or without veliparib [17]. Thus, veliparib combined with platinum-based regimens had better efficacy than veliparib combined with TMZ. The subsequent phase III BROCADE3 trial evaluated the efficacy of carboplatin and paclitaxel with or without veliparib in BRCA1/2-deficient MBC and confirmed improved PFS with the addition of veliparib (HR 0.71) [39].

Conclusion
In this phase II study, veliparib and TMZ showed efficacy in patients with BRCA-deficient metastatic breast cancer with no prior platinum treatment that was comparable to single-agent veliparib at a higher dose. However, the study did not meet its primary response rate endpoint in patients with MBC unselected for BRCA mutations, nor in patients with known BRCA mutation. The regimen was predominantly associated with hematologic adverse events manageable with dose modifications.