Sequence-Specific Pharmacokinetic and Pharmacodynamic Phase I/Ib Study of Olaparib Tablets and Carboplatin in Women's Cancer

DNA damage repair pathways are active therapeutic targets in gynecologic and breast cancers (1). Homologous recombination repair (HRR) is an error-free DNA double-strand break repair mechanism (2). Base excision repair (BER), a DNA single-strand break repair pathway, is activated in response to deficient HRR, requiring PARP activity (3, 4). The absence of single-stranded DNA damage repair causes DNA helix strain at transcription forks and leads to double-strand breaks requiring HRR (3, 5). Therefore, HRR dysfunction sensitizes cells to PARP inhibition, leading to further chromosomal instability, cell-cycle arrest, and cell death (4).

PARP inhibitors (PARPi) have demonstrated clinical potential in women's cancers (6–10). Olaparib, an oral PARP1 and PARP2 inhibitor, is the first FDA-approved PARPi, allowed for germline BRCA mutation–associated ovarian cancers in fourth or greater recurrence (11). Olaparib capsules are approved at the single-agent continuous dose of 400 mg twice daily. Olaparib, now in investigational tablet formulation, delivers the therapeutic range of 400 mg olaparib capsules at 300 mg, with a higher average C_max and AUC_0–12 and little change in terminal half-life (12, 13). The optimal application of olaparib tablets with carboplatin is unknown.

It has been hypothesized that PARP inhibition should sensitize tumor cells to radiotherapy or cytotoxic agents that induce DNA damage (14, 15). Preclinical testing of PARPi demonstrated increased antitumor activity when combined with platinum drugs (16, 17). PARPi increased cytotoxic activity and apoptosis in cisplatin-resistant ovarian cancer cell lines when given with cisplatin (18). Earlier clinical studies demonstrated PARPi was a chemotherapy or radiation sensitizer (19). There are limited data on the optimal sequence of administration of PARPi with chemotherapy. We found that the administration of olaparib prior to carboplatin attenuated carboplatin cytotoxicity in ovarian and breast cancer cell lines in vitro, suggesting the order of administration may be important (Lee and colleagues, in preparation). Thus, examination of the role of drug sequence may improve upon current clinical benefit by optimizing drug administration.

We conducted a phase I/Ib pharmacokinetic and pharmacodynamic study to prospectively evaluate pharmacokinetic/pharmacodynamic endpoints as a function of order of administration of olaparib and carboplatin. The primary objectives of this study were to investigate olaparib pharmacokinetics and pharmacodynamics on the different schedules. Platinum–DNA adducts in peripheral blood mononuclear cells (PBMC) were shown to be related to survival in advanced ovarian cancer patients who were treated with platinum-based therapy (20, 21). We thus studied quantities of PBMC platinum–DNA adducts as a primary pharmacodynamic objective. Secondary objectives included measures of DNA damage and poly(ADP-ribose) (PAR) concentrations and safety and preliminary activity of olaparib tablets with carboplatin in women's cancer.

Eligible patients had recurrent or refractory gynecologic cancers or metastatic or unresectable breast cancer for which no curative therapies exist. Additional eligibility criteria included measurable or evaluable disease, ECOG performance status 0 to 2, normal end organ function except grade 1 anemia, neutropenia, leukopenia, and AST/ALT, no tumor-related therapy for 4 weeks, no platinum therapy for the immediate prior 6 months and no history of NCI Common Terminology Criteria (CTCAE v4.0) grade 4 platinum allergy, no antibiotics within 7 days, and no brain metastases diagnosed or active within the past year. Germline BRCA mutation status was requested at entry. All patients provided written informed consent before enrollment. The trial conformed to the Declaration of Helsinki, Good Clinical Practice guidelines, and was approved by the Institutional Review Board of the Center for Cancer Research, National Cancer Institute (Bethesda, MD; ClinicalTrials.gov identifier: NCT01237067).

An open label 3 + 3 dose–escalation study examined the combination of olaparib 100 to 200 mg tablet formulation every 12 hours on days 1 to 7, with carboplatin AUC4 or 5 on days 1 or 2 in 21-day cycles (Fig. 1 and Supplementary Table S1). Dose-limiting toxicity (DLT) was determined during the first two cycles of therapy. No more than eight cycles of combined therapy was given, after which patients received continuous daily maintenance therapy with full dose olaparib, 300 mg tablets every 12 hours in 4-week cycles. Granisetron (days 1–7) and dexamethasone (days 1–4) were given prophylactically during combination therapy only. Clinical response was assessed every two cycles by imaging using RECISTv1.1 criteria. Study treatment was discontinued for progression of disease, intercurrent illness, adverse events not recovering to ≤grade 1 within 3 weeks, and/or patient preference.

Patients were evaluated for toxicity per CTCAEv4. DLTs included treatment-emergent grade 3 or 4 nonhematologic and grade 4 hematologic adverse events with the following exceptions: grade 3 diarrhea, nausea, or vomiting unresponsive to optimal medical management, and asymptomatic grade 3 hypomagnesemia, hyponatremia, hypophosphatemia, or hypocalcemia rapidly reversible with medical management. Grade 3 thrombocytopenia for ≥7 days or requiring transfusion and grade 4 neutropenia for ≥7 days or with fever were dose-limiting. Complete blood counts and serum chemistries were monitored weekly during the DLT period.

Pegfilgrastim was indicated if the day 1 absolute neutrophil count (ANC) was less than 1,500/mL, necessitating a treatment delay, and used in subsequent cycles if the day 1 ANC was 1,500 to 1,800/mL. Once initiated, pegfilgrastim was continued during all combination treatment cycles. It was not allowed during the first two cycles of the dose–escalation phase or during olaparib maintenance therapy.

Patients were randomly assigned to one of the two schedules administered at the maximum tolerated dose (MTD). Cohort A received olaparib followed by carboplatin, and cohort B received carboplatin followed by olaparib during cycle 1, and schedules were reversed for cycle 2 (Fig. 1B). All patients received concurrent administration of olaparib and carboplatin starting with cycle 3, for up to 8 combination cycles (7), then followed by olaparib tablet maintenance therapy. Pharmacokinetic/pharmacodynamic primary endpoints were examined, comparing cycles 1 and 2 within and between the two cohorts.

The pharmacokinetic/pharmacodynamic schema, including blood collection time points, is shown in Fig. 1B. PBMCs and plasma were collected, separated within 4 hours, and stored in aliquots at −80°C until use.

PBMC DNA was isolated as described previously (20). Platinum–DNA adducts were measured by inductively coupled plasma mass spectrometry (ICP-MS: Supplementary Methods S1).

The alkaline comet assay was used to evaluate DNA double-strand breaks (22, 23) according to the manufacturer's instructions (Trevigen). A minimum of 20 comets was acquired for each data point. Percentage of DNA in the tail moment was measured with CometScore Pro software (TriTek Corp.).

Separate aliquots of PBMCs were used to measure PAR incorporation into DNA as described previously (Trevigen; ref. 24).

Blood samples were collected at 0.5, 1, 2, 3, 4, and 8 hours after the first dose of olaparib and immediately prior to the second dose of olaparib (approximately 12 hours after the first dose) on both cycles 1 and 2. Plasma was separated and stored at −80°C until measurement using a validated assay with a lower limit of quantitation of 0.5 ng/mL (25). Pharmacokinetic parameters for olaparib were calculated using noncompartmental methods (Phoenix WinNonlin v6.4, Pharsight Corp), expressed in geometric mean ± standard deviation, and compared in both an unpaired and a pair-wise manner against the absence or presence of carboplatin to assess drug–drug interactions.

Preclinical examination of the olaparib and carboplatin (Selleck Chemicals) interaction used the same study and dose design. Anonymized healthy volunteer's whole blood was obtained from NIH Blood Bank, spiked with carboplatin to 10 μmol/L concentration, or water (vehicle control), and incubated at 37°C for 24 hours. Avian DT40 PARP-1 knockout cells (PARP-1 KO; generously provided by Dr. Y Pommier, NCI, Bethesda, MD) were incubated with carboplatin (10 μmol/L) or water at 37°C for 24 hours. Isolated PBMCs or PARP-1 KO cells were washed and incubated with 10 μmol/L olaparib at 37°C for 1 hour before cells were washed, lysed, and the intracellular olaparib concentrations measured by LC/MS-MS.

The primary endpoints of this trial examined whether the order of drug administration affected the quantity of platinum–DNA adducts measured in patient PBMCs and olaparib pharmacokinetics. A minimum of 21 patients with paired PBMC samples per cohort was needed to provide 80% power to detect a difference equal to one standard deviation between pre- and on-treatment values between the different schedules at P = 0.05. Secondary endpoints, DNA damage, and PAR concentrations were analyzed using two-way ANOVA (GraphPad Prism v.6). Paired t tests were performed in a parametric manner. Mann–Whitney test was used for ex vivo and in vitro pharmacokinetic measures. P < 0.05 was considered statistically significant, and all statistical tests were two-sided.

Patient distribution is shown in the consort diagram (Fig. 1A; N = 77), and characteristics are detailed in Table 1. Eighteen patients were on the dose–escalation cohort, and 59 patients on the pharmacokinetic/pharmacodynamic cohort received the MTD. Patients were heavily pretreated with a median of 5 prior regimens. Three fourths of patients had ovarian cancer. One third of ovarian cancer patients had platinum-sensitive recurrent disease, and most of those carried a germline BRCA mutation. Fifty-six percent of patients had deleterious germline BRCA mutation.

We previously reported the safety and activity of the 7 of 21-day olaparib capsule schedule with carboplatin (7). We now examine the safety of the intermittent schedule using olaparib tablets. DLT was seen in 2 of 6 patients at dose level 4, olaparib 200 mg every 12 hours for 7 days with carboplatin AUC5. One patient had grade 4 thrombocytopenia, requiring platelet transfusion on cycle 1; the second patient had grade 4 neutropenia lasting ≥1 week, requiring growth factor support. The MTD is defined as olaparib tablets 200 mg every 12 hours on days 1 to 7 with carboplatin AUC4.

All patients had at least one treatment-emergent adverse event (Table 2). Hematologic toxicity was the most common adverse event. Grade 3/4 neutropenia was observed in 19 of 76 patients (25%) with no febrile neutropenia; 29 patients received pegfilgrastim or filgrastim, starting as early as cycle 2 for expansion cohort patients, used to prevent recurrent treatment delays. Grade 3/4 thrombocytopenia occurred in 10 of 76 patients (13%) and 5 received platelet transfusion. Grade 3/4 anemia was observed in 7 patients (9%). Twenty-nine patients (38%) received red blood cell transfusion during the combination therapy (26) or during olaparib maintenance (2), and 8 (11%) received darbepoetin to avoid recurrent treatment delays. Two patients receiving olaparib maintenance had olaparib dose reductions for recurrent grade 2 anemia. Carboplatin allergic reactions occurred in 14 patients; 12 had dose reduction or early carboplatin discontinuation for delayed bone marrow recovery (10) or allergic reaction (2); they then continued on daily olaparib maintenance therapy.

Fifty-eight patients had olaparib plasma concentration–time data for noncompartmental pharmacokinetic analysis. All but 5 patients had data for both cycles available; these 5 patients had cycle 1 data only (cohort A or B). There was a 49% increase in the calculated apparent oral clearance (CL/F) of olaparib when administered after carboplatin compared with before carboplatin (P = 0.012; cohort A, Fig. 2, with increases observed in 21 of 24 patients with paired data (P < 0.0001). This 49% faster clearance of olaparib resulted in a 28% shorter t_1/2 (P = 0.001) and a 25% lower AUC_last (P = 0.046) for olaparib after carboplatin. Cohort B, receiving olaparib in the reverse order, had a +28% change in clearance (Fig. 2B). The magnitude of change in AUC_last (25%) and C_max (∼20%) was almost identical between cohorts A and B, and before/after carboplatin.

The effect of carboplatin pretreatment on olaparib biodistribution was assessed in healthy volunteer PBMCs. A 24-hour pretreatment with carboplatin significantly increased intracellular olaparib concentration by more than 30% (P = 0.013, Fig. 2C). Intracellular olaparib concentrations did not statistically differ between DT-40 PARP^+/+ and ^−/− cells (Fig. 2D).

Platinum–DNA adducts (fg Pt/ng DNA) were measured 24 hours posttreatment with carboplatin on cycles 1 and 2, in cohorts A and B, respectively. No significant difference in platinum–DNA adduct quantities were observed 24 hours post-carboplatin compared with their relative baselines and between the two cycles for both cohorts, with a trend to lower adduct concentrations when administered after olaparib (Fig. 3A).

A significant increase in comet tail percentage DNA was measured with the combination compared with their respective baselines (Fig. 3B).

A significant decrease in PAR concentrations was found 24 hours post-olaparib compared with their respective baselines. This expected finding is a positive control for the presence of olaparib. PAR concentrations after combination treatment were not significantly different to single-agent olaparib exposure (Fig. 3C).

Seventy-one of 77 patients were evaluable for response (Fig. 4). Six patients discontinued treatment during cycle 1 or 2, due to drug toxicity (1), intercurrent illness (1), cardiac arrest related to nonneutropenic pneumonia (1), unsuspected early brain metastasis (1), or withdrawal of consent (2). The objective response rate (ORR) is 46% (33/71) with a disease control rate of 75% [DCR; complete response (CR) + partial response (PR) + stable disease (SD) ≥4 months; 53/71]; 2 patients with BRCA1 mutation [1 each, ovarian cancer and triple-negative breast cancer (TNBC)] achieved CR (11; 28 months) and 31 had PR (median 10 months, range, 5–30+). Clinical benefit was greater in germline BRCA mutation carriers with 68% ORR and 88% DCR (35/40), compared with 19% ORR (6/31) and 58% DCR (18/31) in those without BRCA mutation. One platinum-resistant recurrent ovarian cancer patient with BRCA1 mutation had a CR. One ovarian cancer patient with high BRCAPro score (93% risk) had PR of 5-month durable response (27).

This is the first prospective clinical trial examining the sequence effect of olaparib in combination with carboplatin in women's cancer. We investigated whether a different order of drug administration affected olaparib pharmacokinetics and PBMC platinum–DNA adducts as pharmacodynamic measures. Our findings showed increased olaparib clearance, in patients and in in vitro, caused by carboplatin preexposure and no effect on platinum–DNA adduct quantities. These results suggest drug sequence may improve upon clinical benefit by optimizing drug administration.

The interaction of olaparib tablets and carboplatin is likely influenced by several factors. Carboplatin exposure decreased olaparib C_max and AUC_last 1.2-fold and 1.3-fold, respectively, due to increased olaparib clearance. This plasma clearance effect was due to intracellular sequestration rather than total drug loss from the environment. Our data trend similarly to AstraZeneca's comparison data, although their findings did not reach statistical significance (unpublished data). We also found no change in plasma protein binding (data not shown). Further pharmacokinetic parameters from our study and in vitro experiments will be reported in detail. We posited that the intracellular uptake could be a result of olaparib–PARP1 protein binding. The lack of intracellular olaparib concentration variance between PARP1-expressing and -KO cells argues against the olaparib/PARP-1 complex as an intracellular olaparib sink. This may explain the increased hematologic toxicity observed at equipotent olaparib dosing, requiring a dose of 200 mg twice daily tablets with a lower AUC of carboplatin compared with our prior olaparib capsule combination MTD (7).

PBMC platinum–DNA adduct quantity was shown to be a reliable surrogate for tumor tissue platinum–DNA adduct levels and was correlated with survival in advanced ovarian cancer patients (20, 21, 26). The presence and degree of platinum–DNA adducts could represent the net balance between adduct formation and repair via nucleotide excision repair (28). Preclinical studies suggest PARPi may reduce DNA repair of platinum adducts by inhibiting binding of PARP to single-strand break sites (29–31). The quantities of PBMC platinum–DNA adducts were not significantly influenced by olaparib in our study. It is possible that no difference was found in part due to limited PBMC numbers, the sampling timing, or the different method of platinum–DNA adduct measurements used.

This study established the MTD of olaparib tablets 200 mg every 12 hours for 7 days and carboplatin AUC4 every 21 days. These women's cancer patients were no more heavily pretreated than our prior germline BRCA mutation patients (7) in whom we reported an MTD of carboplatin AUC5/olaparib capsules 400 mg every 12 hours for 7-day schedule; a 400 mg twice daily olaparib capsule dose corresponds to a 250 to 300 mg twice daily tablet dose. Current patients were thus less tolerant of the combination yielding a carboplatin AUC4 MTD, combined with the lower dose of olaparib tablets. The majority of adverse events, primarily hematologic toxicities, were manageable and reversible with supportive care, and no unexpected new events were uncovered. Consistent with our findings, Balmana and colleagues reported a phase I study in breast and ovarian cancer patients of olaparib capsules and cisplatin, requiring intermittent and lower dose olaparib due to hematologic toxicity (olaparib 50–100 mg capsules days 1–5 or 1–10/cisplatin 60–75 mg/m² every 21 days; ref. 32). These observations suggest that an intermittent PARPi schedule can be used as a treatment strategy without attenuating clinical benefit.

We observed promising antitumor activity of the combination in heavily pretreated patients with an ORR of 46%. The ORR of 68% in BRCA mutation carriers was higher than reported in previous monotherapy phase II studies of olaparib capsules in BRCA-mutated breast or ovarian cancer, although direct comparisons cannot be made (10, 33, 34). Clinical activity was also seen in subsets of BRCA wild-type TNBC and platinum-resistant ovarian cancer patients. Gelmon and colleagues demonstrated olaparib monotherapy capsule activity in 24% of BRCA wild-type ovarian cancer patients (10). Our findings suggest this combination may be an opportunity for women's cancer patients without germline BRCA mutation.

In conclusion, we demonstrated olaparib tablets and carboplatin is a safe and active combination in women's cancer. Our pharmacokinetic data suggest a preferable schedule of administering carboplatin prior to olaparib. These findings may further improve the overall clinical utility of this combination. Further clinical exploration of this combination with optimal sequence is warranted in patients with women's cancers.

No potential conflicts of interest were disclosed.

Conception and design: J.-M. Lee, W.D. Figg, E.C. Kohn

Development of methodology: C.J. Peer, M. Yu, A.K.L. Goey, T.M. Sissung, W.D. Figg, E.C. Kohn

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.-M. Lee, L. Amable, N. Gordon, C.M. Annunziata, A.K.L. Goey, T.M. Sissung, B. Parker, L. Minasian, V.L. Chiou, R.F. Murphy, W.D. Figg, E.C. Kohn

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.-M. Lee, C.J. Peer, M. Yu, N. Gordon, A.K.L. Goey, T.M. Sissung, V.L. Chiou, B.C. Widemann, W.D. Figg, E.C. Kohn

Writing, review, and/or revision of the manuscript: J.-M. Lee, C.J. Peer, L. Amable, C.M. Annunziata, A.K.L. Goey, T.M. Sissung, L. Minasian, V.L. Chiou, B.C. Widemann, W.D. Figg, E.C. Kohn

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.-M. Lee, M. Yu, N. Gordon, N. Houston, E.C. Kohn

Study supervision: J.-M. Lee, L. Minasian, E.C. Kohn

Olaparib was supplied to the Center for Cancer Research, NCI, under a Cooperative Research and Development Agreement between the CCR/NCI and AstraZeneca. We thank Dr. Y. Pommier for PARP1-KO cells, D.A. Botesteanu for graphical support, and Drs. J. Hays, A. Noonan, and I, Ekwede RN for their contributions in clinic. We also thank Dr. S.M. Steinberg for statistical support and Dr. J. Murai for experimental advice.

This work was funded in part by the 2011 Jane C. Wright MD ASCO Young Investigator Award from Conquer Cancer Foundation (to J.-M. Lee) and by the Intramural Program of the Center for Cancer Research, NCI and NIMHD, NIH.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.