CHK1 inhibitor SRA737 is active in PARP inhibitor resistant and CCNE1 amplified ovarian cancer

Abstract

High-grade serous ovarian cancers (HGSOCs) with homologous recombination deficiency (HRD) are initially responsive to poly (ADP-ribose) polymerase inhibitors (PARPi), but resistance ultimately emerges. HGSOC with CCNE1 amplification (CCNE1 ^amp) are associated with resistance to PARPi and platinum treatments. High replication stress in HRD and CCNE1 ^amp HGSOC leads to increased reliance on checkpoint kinase 1 (CHK1), a key regulator of cell cycle progression and the replication stress response. Here, we investigated the anti-tumor activity of the potent, highly selective, orally bioavailable CHK1 inhibitor (CHK1i), SRA737, in both acquired PARPi-resistant BRCA1/2 mutant and CCNE1 ^amp HGSOC models. We demonstrated that SRA737 increased replication stress and induced subsequent cell death in vitro. SRA737 monotherapy in vivo prolonged survival in CCNE1 ^amp models, suggesting a potential biomarker for CHK1i therapy. Combination SRA737 and PARPi therapy increased tumor regression in both PARPi-resistant and CCNE1 ^amp patient-derived xenograft models, warranting further study in these HGSOC subgroups.

Keywords: Cancer; Cell biology; Molecular biology.

Graphical abstract

Figure 1

SRA737 is active as monotherapy in a CCNE1^amp HGSOC xenograft, resulting in replication stress and dsDNA breaks (A) OVCAR3 cells (CCNE1^amp) were treated with vehicle or SRA737 (0.1 or 0.2 μM) for 12 days. Crystal violet staining was quantified using ImageJ and percent of control was calculated. Data are representative of three independent experiments. Mean ± SD (n = 3), one-way ANOVA with Dunnett’s multiple comparisons post-hoc test compared to the control ∗∗∗∗, p < 0.0001. (B) OVCAR3 and PARPi resistant PEO1 (PEO1-PR) cells were treated with vehicle or SRA737 (0.2 μM) for 24 h prior to cell lysis and Western blot analysis for indicated markers. (C) FACS analysis for DNA bound p-RPA32 and γ-H2AX expression in OVCAR3 cells following 24 h of treatment with vehicle or SRA737 (1 μM). (D) Mice bearing subcutaneous OVCAR3 tumors were treated with SRA737 (25, 50, 100 or 150 mg/kg QD, PO) or vehicle for three cycles of 5 days on/2 days off (n = 10). Tumor volume was calculated from twice weekly caliper measurements, and data are represented as mean ± SEM. Data were analyzed by one-way ANOVA comparison of the average area under the curve. Statistical significance is shown compared to the vehicle control, ∗, p < 0.05 and ∗∗∗∗, p < 0.0001. (E and F) OVCAR3 tumor bearing mice were treated with a single dose of SRA737 (25, 50, 100 and 150 mg/kg) or vehicle and the tumors were collected 12 h later. (E) Representative immunohistochemistry for p-CHK1 (S345) and γ-H2AX in vehicle and SRA737. Quantification of percent nuclear p-CHK1 (S345) or γ-H2AX staining intensity relative to the vehicle mean. Statistical significance is shown compared to the vehicle control (n = 3, Mean ± SD; one-way ANOVA; ∗∗, p < 0.01, ∗∗∗, p < 0.001, ∗∗∗∗, p < 0.0001). (F) Tumor lysates prepared from three animals per treatment group were analyzed by Western blot for indicated markers. See also Figure S1.

Figure 2

SRA737 is active as a monotherapy in ovarian cancer CCNE1^amp PDX models (A–D) PDX models were tested for response to standard of care carboplatin (WO-19) or cisplatin (#29, #111, #1177), PARPi, and SRA737 in vivo. Tumor volume growth curve (left and middle) and survival analysis (right) after randomization. Tumor growth shown is mean ± SEM. Longitudinal analysis by Linear Mixed-Effects modeling with type II ANOVA and pairwise comparisons across groups. Survival is shown by Kaplan–Meier curve using the Mantel-Cox log rank test. Statistical significance is shown comparing the highest dose of SRA737 tested for each model and PARPi or vehicle control as indicated (∗, p < 0.5, ∗∗, p < 0.01, ∗∗∗, p < 0.001). A, Platinum-resistant HGSOC WO-19 tumors were orthotopically implanted of recipient mice. Once tumors reached ∼80 mm³ by ultrasound measurements, mice were assigned to vehicle (n = 4), SRA737 (50 mg/kg; PO, 5 days on/2 days off; n = 8) or PARPi (50 mg/kg; PO; n = 5). B-D, PARPi-resistant and platinum-refractory #111 and #29 HGSOC and #1177 carcinosarcoma tumors were subcutaneously implanted on the left flank of recipient NSG mice. Once tumors reached 180-250 mm³, mice were assigned to vehicle, SRA737 (100 mg/kg or 50 mg/kg or 25 mg/kg; PO, 21 days continuous dosing) or PARPi (100 mg/kg; IP, 5 days on/2 days off for 3 weeks) (#111, n = 10–11; #29, n = 6; #1177, n = 3–17, per group). See also Figure S1. (E) PDX tumors were analyzed for cyclin E1 protein levels by immunohistochemistry and histopathologic review confirmed high (H-Score >150) cyclin E1 expression (see also Figure 3). Scale bars indicate 100 μm.

Figure 3

Drug response heatmap and oncoprint of PDX tumor characteristics Drug response data, cyclin E1 copy number and protein expression for all PDX models are summarized using a heatmap. Gene alterations present in the PDX models used in this study are presented as an oncoprint (see also Table S1). ND, no data; SD, stable disease; PD, progressive disease; VUS, variant of uncertain significance; SNV, single nucleotide variant.

Figure 4

Dual inhibition of CHK1 and PARP induce replication stress and DNA damage in PARPi-resistant cells (A) Colony formation assay on cell lines treated with SRA737 (0.1 μM in Kuramochi; 0.2 μM in JHOS4, UWB B1+, PEO1-PR, PEO4, OVCAR3; 0.5 μM in PEO1, JHOS4 PR, FUOV1), Olaparib (0.1 μM in JHOS4; 0.2 μM in PEO1; 0.5 μM in Kuramochi, OVCAR3; 1 μM in JHOS4 PR, UWB B1+, PEO1 PR, PEO4, FUOV1), or combination for 10 days. (B) Colonies were quantified using ImageJ. Data shown as mean ± SD (n = 3) of a single representative experiment. One-way ANOVA analysis with Tukey’s multiple-comparisons test were performed and statistical significance is shown comparing single agent to combination treatment (∗, p < 0.05, ∗∗∗∗, p < 0.0001). (C) Mean value (%) of colony formation in (B) was used to calculate the coefficient of drug interaction (CDI). CDI <1 indicated synergism, CDI <0.7 significant synergism, CDI = 1 additivity, CDI >1 antagonism. (D) Western blot analysis for cyclin E1 protein expression in cell lines. (E) Western blot analysis for pCHK1(S345 and S296) 24 h after cells were treated with 1 μM PARPi (except PEO1 0.2 μM), SRA737 at indicated dosages (PEO1, PEO1-PR: SRA737 0.2 μM; JHOS4, JHOS4-PR: SRA737 1 μM, OVCAR3 and FUOV1: SRA737 0.5 μM). (F–H) Cells were treated with optimal doses as indicated in E. After 24 h, FACS analysis was performed pRPA32(S33) (F) and S-phase specific $\gamma$-H2AX (G). After 72 h, FACS analyses was performed to detect Annexin/PI + cells (H). Data are representative of 3 individual experiments, and statistical significance is shown comparing SRA737 and combination treatment (∗, p < 0.05, ∗∗∗∗, p < 0.0001, one-way ANOVA). (I) Western blot analysis for cleaved caspase-3 after cells were treated as in D for 72 h. See also Figure S3.

Figure 5

SRA737 is active in combination with PARPi in HGSOC PDX (A–D) Tumor volume growth curve (left), survival analysis (middle), and body weight (right) after randomization. Tumor growth shown is mean ± SEM. Longitudinal analysis by Linear Mixed-Effects modeling with type II ANOVA and pairwise comparisons across groups. Survival is shown by Kaplan–Meier curve using the Mantel-Cox log rank test. Statistical significance is shown comparing PARPi and PARPi-SRA737 combination (∗, p < 0.5, ∗∗, p < 0.01, ∗∗∗, p < 0.001). Body weight change (%) from baseline +/− SEM are plotted. Representative images of IHC for cyclin E1 staining are shown. Scale bars indicate 100 μm. (A and B) PARPi-resistant and platinum-refractory #111 (n = 7–8) and #206 (n = 6–7) HGSOC tumors were subcutaneously implanted on the left flank of recipient NSG mice. Once tumors reached 180–250 mm³, mice were assigned to vehicle, SRA737 (75 mg/kg; PO, 21 days continuous dosing), PARPi (75 mg/kg; IP, 5 days on/2 days off for 3 weeks) or SRA737+PARPi. CDI = 1.1 for #111 and 1.0 for #206. (C) Orthotopically implanted HGSOC tumors collected from a platinum resistant patient were treated with PARPi until PARPi-resistance developed (24 weeks) and were then harvested. After a second implantation, tumors reaching 50 mm³ were grown in the presence of PARPi (50–100 mg/kg; most 75 mg/kg; QD) until tumors grew to 100 mm³. At that time, mice either remained on PARPi (75 mg/kg; QD) or were switched to SRA737 (75 mg/kg; 5 days on/2 days off) alone or in combination with PARPi. CDI = 1.2. (D) PDX #111 tumors harvested after two weeks of treatment were stained for pCHK1 (Ser345) and γH2AX. See also Figure S4 and Table S2. Representative IHC images for each treatment and antibody are shown. STARDIST analysis was performed to determine % positive cells. Each point represents the mean % positive cells for each tumor from 9 fields of view per tumor, two independent tumors for each treatment. Scale bars indicate 50 μm.

Figure 6

Pathway enrichment analysis for CHK1i monotherapy responder vs. non-responder PDX models Differentially expressed genes were tested separately to identify terms that were over-represented. GO terms and KEGG biological pathways were interrogated. KEGG pathways significantly upregulated (A) or downregulated (B) in responsive PDX compared to non-responder PDX are shown. The FDR (false-discovery rate) is calculated from p-values generated by a modified Fisher’s exact test adopted to measure the gene enrichment in annotation terms. The gene count represents how many genes from the input list of significant genes are present in that pathway.