The CHK1 inhibitor prexasertib in BRCA wild-type platinum-resistant recurrent high-grade serous ovarian carcinoma: a phase 2 trial

Abstract

The multi-cohort phase 2 trial NCT02203513 was designed to evaluate the clinical activity of the CHK1 inhibitor (CHK1i) prexasertib in patients with breast or ovarian cancer. Here we report the activity of CHK1i in platinum-resistant high-grade serous ovarian carcinoma (HGSOC) with measurable and biopsiable disease (cohort 5), or without biopsiable disease (cohort 6). The primary endpoint was objective response rate (ORR). Secondary outcomes were safety and progression-free survival (PFS). 49 heavily pretreated patients were enrolled (24 in cohort 5, 25 in cohort 6). Among the 39 RECISTv1.1-evaluable patients, ORR was 33.3% in cohort 5 and 28.6% in cohort 6. Primary endpoint was not evaluable due to early stop of the trial. The median PFS was 4 months in cohort 5 and 6 months in cohort 6. Toxicity was manageable. Translational research was an exploratory endpoint. Potential biomarkers were investigated using pre-treatment fresh biopsies and serial blood samples. Transcriptomic analysis revealed high levels of DNA replication-related genes (POLA1, POLE, GINS3) associated with lack of clinical benefit [defined post-hoc as PFS < 6 months]. Subsequent preclinical experiments demonstrated significant cytotoxicity of POLA1 silencing in combination with CHK1i in platinum-resistant HGSOC cell line models. Therefore, POLA1 expression may be predictive for CHK1i resistance, and the concurrent POLA1 inhibition may improve the efficacy of CHK1i monotherapy in this hard-to-treat population, deserving further investigation.

Fig. 1. Correlative study endpoints analyses workflow.

a Workflow to detect molecular correlates to CHK1i treatment. Pre-treatment fresh core biopsies were collected from patients with safely biopsiable diseases enrolled in the clinical trial (NCT02203513). Genomic (BROCA GOv.1 panel and WES) and transcriptomic analyses were performed to identify the molecular characteristics between patients with clinical benefit (PFS ≥ 6 months) and no clinical benefit (PFS < 6 months). b Workflow to detect pharmacodynamic biomarkers reflecting CHK1i therapy. Paired blood samples were collected at baseline and at C1D15 to evaluate dynamic changes in EpCAM+ CTCs and immune cell subsets in patients with and without clinical benefit. The figure was created with BioRender.com. Abbreviations: BRCAwt BRCA wild-type, C1D15 Cycle 1 Day 15, CHK1i CHK1 inhibitor, CTCs circulating tumor cells, DNAseq DNA sequencing, EpCAM epithelial cell adhesion molecules, GSEA gene set enrichment analysis, PBMC peripheral blood mononuclear cells, PR-HGSOC platinum-resistant high-grade serous ovarian cancer, WES whole exome sequencing.

Fig. 2. Clinical trial design and antitumor activity.

a Clinical trial with correlative study endpoints: prior to prexasertib administration, blood and tumor samples were collected for correlative study endpoints, and blood samples were further obtained at C1D15. AEs (CTCAE v4) were evaluated at each study drug administration at cycle 1, and every 4 weeks for subsequent cycles. CT scans were performed every 2 cycles (RECIST v1.1 evaluation). b The CONSORT flow diagram. Overall, 49 patients were enrolled in the study including 24 patients in the biopsy cohort 5 and 25 patients in the non-biopsy cohort 6. 39 patients were evaluable for tumor response per RECIST v1.1 criteria. c Duration of treatment: swimmer plot showing the duration of treatment (time in months) with prexasertib monotherapy for each individual RECIST-evaluable patient (n = 39). d PFS of each patient is shown. PFS was estimated using the Kaplan–Meier method with 95% CI. e Waterfall plot showing the best responses in 39 RECIST-evaluable patients. The horizontal dotted line indicates the threshold for partial response (30% reduction in tumor size from baseline). Sequencing was conducted and analyzed on fresh pre-treatment tissue samples with optimal quality. POLA1 and POLE mRNA levels were measured by RNAseq (Supplementary Data 5). Upregulation is defined as expression ≥median, and downregulation as expression <median for each gene (cohort 5, n = 15). DNAseq exhibited the genetic alteration of genes related to DDR (cohort 5, n = 15). Abbreviations: AEs adverse events, CTCAE Common Terminology Criteria for Adverse Events, CCNE1 cyclin E1, CDK12 cyclin-dependent kinase 12, CI confidence interval, DDR DNA damage repair, DNAseq DNA sequencing, IV intravenously, KRAS Kirsten rat sarcoma virus, MYC Myc proto-oncogene, NF1 neurofibromatosis type 1, PFS progression-free survival, POLA1 DNA polymerase alpha 1, POLE DNA polymerase epsilon, RAD51C RAD51 paralog C, RECIST Response evaluation criteria in solid tumors, RNAseq RNA sequencing.

Fig. 3. DNA replication machinery is associated with CHK1i resistance.

a Analysis of bulk RNAseq indicated no individual gene differentially expressed between CB (PFS ≥ 6 months, n = 6) and NCB (PFS < 6 months, n = 9) after adjustment of multiple testing (FDR q = 0.87–1, Supplementary Data 7), possibly due to the small sample size. We also conducted GSEA of RNAseq data to identify pathways that might contribute to CHK1i resistance and response. The bar plot shows the top five KEGG gene sets associated with CB (n = 6) or NCB (n = 9). NES scores, nominal p-values, and FDR q-values shown in the figure were calculated by GSEA software (Supplementary Data 6). b The DNA replication pathway was enriched in patients with NCB (n = 9) versus CB (n = 6). Genes in the core enrichment of the DNA replication pathway were shown (right). c The mRNA levels of genes were analyzed from RNAseq in patients with CB (n = 6) versus NCB (n = 9). High mRNA expression of DNA replication pathway-related genes, POLA1, POLE, GINS3 are significantly associated with no clinical benefit. A regular t-test (two-sided) was used for the raw P-value (Supplementary Data 7). The boxes extend from min to max values, with the median depicted by a horizontal line. Source data are provided as a Source Data file. Abbreviations: CB clinical benefit, CHK1i CHK1 inhibitor, FDR false-discovery rate, GSEA gene set enrichment analysis, GINS3 GINS Complex Subunit 3, MCM7 minichromosome maintenance complex component 7, NCB no clinical benefit, NES normalized enrichment score, RNAseq RNA sequencing, RSEM RNAseq by expectation-maximization, PFS progression-free survival, POLA1 DNA polymerase alpha 1, POLE DNA polymerase epsilon.

Fig. 4. Targeting B-family polymerases induces significant cytotoxicity with CHK1i in BRCAwt platinum-resistant HGSOC cells.

a XTT assays were performed with a gradient of aphidicolin with or without CHK1i prexasertib (0.8 nM) and survival plotted relative to untreated (n = 3). b, c XTT assays were performed with siRNAs that target either POLE (b, n = 3) or POLA1 (c, n = 3). d XTT assays were done similarly with increasing concentrations of POLA1 specific inhibitor ST1926 (0–500 nM) and survival plotted with ST1926 alone or in combination with sublethal concentration of CHK1i (1.5 nM) (n = 3). e–g Relative inhibition of PrexR cell lines with siRNA specific to POLE (e, n = 3), to POLA1 (f, n = 3), or to ST1926 (g, n = 3) with or without CHK1i (6.25 nM). h Western blot analysis of protein extracts from cells treated with either control or those treated with specific siRNAs targeting either POLE or POLA1 for 48 h prior to be treated with CHK1i overnight at the concentrations mentioned (n = 3). Densitometric analysis was performed using ImageStudio software. Phosphorylated proteins were quantified relative to total proteins, normalized to β-actin, and then expressed relative to loading control β-actin. Data from a–g were analyzed using a standard Student’s t-test (two-sided). Data are shown as mean ± SD. Source data are provided as a Source Data file. Abbreviations: CHK1i CHK1 inhibitor, DDR DNA damage repair, HGSOC high-grade serous ovarian cancer, PrexR prexasertib-resistant, POLA1 DNA polymerase alpha 1, POLE DNA polymerase epsilon.