Neoadjuvant PARP inhibitor scheduling in BRCA1 and BRCA2 related breast cancer: PARTNER, a randomized phase II/III trial

Abstract

Poly (ADP-ribose) polymerase inhibitors (PARPi) exploit DNA repair deficiency in germline BRCA1 and BRCA2 pathogenic variant (gBRCAm) cancers. Haematological toxicity limits chemotherapy-PARPi treatment combinations. In preclinical models we identified a schedule combining olaparib and carboplatin that avoids enhanced toxicity but maintains anti-tumour activity. We investigated this schedule in a neoadjuvant, phase II-III, randomised controlled trial for gBRCAm breast cancers (ClinicalTrials.gov ID:NCT03150576; PARTNER). The research arm included carboplatin (Area Under the Curve 5, 3-weekly); paclitaxel (80 mg/m², weekly) day 1, plus olaparib (150 mg twice daily) day 3-14 (4 cycles), followed by anthracycline-containing chemotherapy (3 cycles); control arm gave chemotherapy alone. The primary endpoint, pathological complete response rate, showed no statistical difference between research 64.1% (25/39); control 69.8% (30/43) (p = 0.59). However, estimated survival outcomes at 36-months demonstrated improved event-free survival: research 96.4%, control 80.1% (p = 0.04); overall survival: research 100%, control 88.2% (p = 0.04) and breast cancer specific survival: research 100%, control 88.2% (p = 0.04). There were no statistical differences in relapse-free survival and distant disease-free survival, both were: research 96.4%, control 87.9% (p = 0.20). Similarly, local recurrence-free survival and time to second cancer were both: research 96.4%, control 87.8% (p = 0.20). The PARTNER trial identified a safe, tolerable schedule combining neoadjuvant chemotherapy with olaparib. This combination demonstrated schedule-dependent overall survival benefit in early-stage gBRCAm breast cancer. This result needs confirmation in larger trials.

Fig. 1. Comparison of DNA damage repair kinetics in bone marrow versus tumour cells.

A Examples of γH2AX immunohistochemistry (IHC) staining (brown colour) of sections of rat bone marrow (top panel) with vehicle control treatment or carboplatin (50 mg/kg) at 6 h and 72 h post treatment. The lower panel is a representative image of PDX tumour γH2AX staining. Quantification of γH2AX from both rat bone marrow (B) and PDX tumour (C) is shown using the pathologists’ scoring system of 0–3 from three independent biological replicates (or four independent biological replicates in the case of the PDX 48 h and 72 h data). Individual scores for vehicle control (circles) or platinum treatment (triangles) for the biological replicates are also shown, as is the group mean (orange bars with SD error bars). Source data is provided as a Source Data file. γH2AX staining is indicative of DNA damage induction and repair over time and shows that carboplatin treatment is resolved in rat bone marrow after 48 h, while in the PDX tumour, DNA damage is still increasing at 72 h post treatment.

Fig. 2. Introducing a 48 h gap between carboplatin and olaparib treatments ameliorates the combination bone marrow toxicity effect while maintaining combination anti-tumour efficacy.

A Flow cytometry quantification of CD90+/Lineage− multipotent progenitor stem cells indicates that a 24 h gap between the carboplatin and olaparib combination is not sufficient to reduce bone marrow toxicity compared to concurrent treatment. Introduction of a 48, 72 or 96 h gap does reduce the combination toxicity effect to that seen for carboplatin alone. ANOVA statistical analysis of CD90+ cell levels from 6 biological replicates used a two-sided Student’s t test. P-values for the statistical significance of comparisons between the % CD90+/Lineage− multipotent progenitor stem cells treated with carboplatin alone and those involving concurrent or gap scheduling combinations with olaparib are indicated. B The use of a 48 h gap schedule for the carboplatin/olaparib combination still maintains greater anti-tumour efficacy in the TNBC gBRCAm PDX model HBCx-17 than the effects of either olaparib or carboplatin alone. Treatment represented by the black line is 28 d of vehicle control, the green line is 28 days (28D) daily 100 mg/kg olaparib starting on Day 1 (D1), the yellow line is a single D1 dose of 50 mg/kg carboplatin and the red line the single D1 50 mg/kg carboplatin and 28D daily 100 mg/kg olaparib treatment starting on D3. Mean tumour volumes are plotted along with error bars shown with ±SEM. Statistical significance was evaluated using a one-tailed t-test from 10 independent biological replicates for the vehicle control and 9 biological replicates for the other three treatment arms. Source data is provided as a Source Data file.

Fig. 3. PARTNER Trial gBRCAm cohort schema and CONSORT diagram.

A gBRCAm cohort trial flow chart. B Trial Consort Diagram. The first main reason for treatment discontinuation was reported.

Fig. 4. Efficacy endpoints of the gBRCAm cohort in modified intention-to-treat (mITT) patients.

A Pathological complete response (pCR) rate by treatment arm in gBRCAm patients. B pCR by cancer subtype in the entire trial population, excluding the dropped arm. The point estimation is the pCR response rate, and the error bar represents the 96% confidence interval of the proportion based on the Clopper–Pearson method. The statistical test was based on the two-sided chi-squared test. Source data is provided as a Source Data file.

Fig. 5. Kaplan–Meier curves of time to event outcomes.

A Event-free survival by treatment arm (control vs research). B Overall survival by treatment arm (control vs research). C Event-free survival by treatment arm (including the dropped arm). D Overall survival by treatment arm (including the dropped arm). E Event-free survival by pathological complete response (excluding the dropped arm). F Overall survival by pathological complete response (excluding the dropped arm) in the mITT population. No adjustments were made for multiple comparisons. Source data is provided as a Source Data file.

Fig. 6. Analysis of different olaparib and carboplatin schedules in gBRCAm SUM149PT cells.

A Schematic of single-agent and combination schedules of olaparib and carboplatin in TNBC gBRCAm SUM149PT cells. Squares represent treatment days with purple for olaparib and green, carboplatin treatment. B Cell cycle distribution analysis of SUM149PT. Day 4 (D4) analysis following olaparib (1 µM) and/or carboplatin (10 µM) treatment (mean values ± SD from three independent biological replicates). C DNA damage assessment at D4 of SUM149PT cells using γH2AX. Following olaparib (1 μM) and/or carboplatin (10 μM) treatment, γH2AX foci were visualised by immunofluorescence staining (mean values ± SD from three independent biological replicates, a.u. arbitrary units). D Cell viability of SUM149PT cells at day six after olaparib (0.3 μM) or carboplatin (1 μM) single agent treatments, concurrent combination or carboplatin first or olaparib first combination schedules. Viability was assessed using a cell titre glow assay (see methods). Shown are the % viability for each treatment (mean values ± SD, from three independent biological replicates). Also provided are the p-values from one-way ANOVA multiple comparisons. Source data is provided as a Source Data file.