Dual Src and MEK Inhibition Decreases Ovarian Cancer Growth and Targets Tumor Initiating Stem-Like Cells

Abstract

Purpose: Rational targeted therapies are needed for treatment of ovarian cancers. Signaling kinases Src and MAPK are activated in high-grade serous ovarian cancer (HGSOC). Here, we tested the frequency of activation of both kinases in HGSOC and the therapeutic potential of dual kinase inhibition.Experimental Design: MEK and Src activation was assayed in primary HGSOC from The Cancer Genome Atlas (TGGA). Effects of dual kinase inhibition were assayed on cell-cycle, apoptosis, gene, and proteomic analysis; cancer stem cells; and xenografts.Results: Both Src and MAPK are coactivated in 31% of HGSOC, and this associates with worse overall survival on multivariate analysis. Frequent dual kinase activation in HGSOC led us to assay the efficacy of combined Src and MEK inhibition. Treatment of established lines and primary ovarian cancer cultures with Src and MEK inhibitors saracatinib and selumetinib, respectively, showed target kinase inhibition and synergistic induction of apoptosis and cell-cycle arrest in vitro, and tumor inhibition in xenografts. Gene expression and proteomic analysis confirmed cell-cycle inhibition and autophagy. Dual therapy also potently inhibited tumor-initiating cells. Src and MAPK were both activated in tumor-initiating populations. Combination treatment followed by drug washout decreased sphere formation and ALDH1⁺ cells. In vivo, tumors dissociated after dual therapy showed a marked decrease in ALDH1 staining, sphere formation, and loss of tumor-initiating cells upon serial xenografting.Conclusions: Selumetinib added to saracatinib overcomes EGFR/HER2/ERBB2-mediated bypass activation of MEK/MAPK observed with saracatinib alone and targets tumor-initiating ovarian cancer populations, supporting further evaluation of combined Src-MEK inhibition in clinical trials. Clin Cancer Res; 24(19); 4874-86. ©2018 AACR.

Fig. 1.. Effects of pMAPK and pSrc on HGSOC survival.

(A) Kaplan Meier curve shows worse HGSOC survival with both intratumor pMAPK and pSrc above median on RPPA (n=405; logrank p=0.004). (B) Multivariate analysis shows elevation of both pMAPK and pSrc above median, greater age, lack of macroscopic disease and any residual disease are independent prognostic factors.

Fig. 2.. Effects of dual Src and MEK inhibition on signaling and cell cycle.

A-E PEO1R controls, C, were treated with saracatinib (1μM, Src inhibitor, SI), selumetinib (200nM, MEK inhibitor, MI) or both for 48 hours. The same lysates were used in B-E. Graphs show mean +/− SEM; significant differences between multiple comparisons vs control were calculated by ANOVA with post-hoc comparisons between treated groups and control. (A) Cell cycle distribution ^‡p =0.01 for both drugs vs control, CDI=0.38 (See also Supplementary Fig. S1) (B, C) Westerns show indicated proteins (densitometry in Fig S1C). p27 densitometry in C, right panel: *p=0.0008, **p=0.0004, ^‡p=0.00002. (D) Cyclin E IP/Western blotted for Cyclin E, p27 and Cdk2. Cyclin E-bound p27 densitometry: *p=0.0005, **p=0.0007, ^‡p=0.00001. (E) Cyclin E-Cdk2 activity graphed as % max, *p=0.03, **p=0.006, ^‡p=0.000004, CDI=0.77. (F-I) For OCI-C5x and OCI-P5x, controls, C, were treated as in “A”. (F, H) Cell cycle distribution. Differences in %S for OCI-C5x: *p=0.01, **p=0.03, ^‡p=0.0002, CDI=0.60; for OCI-P5x: *p=0.02, **p=0.1, ^‡p=0.0067, CDI=0.52. (G,I) Westerns blots for OCI-C5x (G) and OCI-P5x (I). Densitometry in Supplementary Fig. S2.

Fig. 3.. Effects of Src and MEK inhibition on apoptosis and autophagy in OVCA cells.

A-F Untreated (controls, C) of the indicated lines were or treated with paclitaxel 100nM, saracatinib (1μM, SI), selumetinib (200nM, MI) or both SI and MI (both) for 48 hours. Graphs show mean +/− SEM; significant differences between multiple comparisons vs control were calculated by ANOVA with post-hoc comparisons between treated groups and control. (A-C) Westerns show PARP and cleaved PARP +/− drug. See Supplementary Fig. S4A for densitometry. (D-F) Quantitation of apoptotic cells by flow cytometry for Annexin V. For PEO1R: *p=0.00008, **p=0.003, ^‡p=0.002, #p=0.00002; for OCI-C5x: *p=0.0009, **p=0.0006, #p=0.00001; for OCI-P5x: *p=0.00055, **p=0.023, #p=0.0026. (G-I) Western for LC3B-1 and LC3B-II after treatment with either SI, MI or both. For densitometry see Supplementary Fig. S4B (J-L) Autophagic lysosomes detected as in Methods. Differences for PEO1R *p=0.002, **p=0.0003, ^‡p=0.00006; for OCI-C5x: *p=0.016, ^‡p=0.0029; for OCI-P5x: ^‡p=0.0052.

Fig. 4.. Gene expression and RPPA analysis of control and drug-treated PEO1R cells.

A-C PE01R was treated with saracatinib (SI, 1μM), selumetinib (MI, 200nM) or both for 48 hours, then gene and protein expression were evaluated by Illumina microarray and RPPA. (A) GSEA plots of cell cycle, DNA replication, aurora and autophagy after combination therapy versus control with normalized enrichment score (NES). See also Supplementary Fig. S5. (B) Clustering analysis of normalized RPPA data show the top 20 proteins down- or upregulated by treatment. LogFC represents the log2-fold change between treatment versus controls. (C) Heat map representation of selected triplicate repeat RPPA data from controls or drug treated cells shows inter-sample variability.

Fig. 5.

Stem cell effects of Src and MEK inhibition (A-B) Sphere *p=0.0005 (A) and soft agar colony formation *p=0.0029 (B), in ALDH1+ and ALDH1- PEO1R cells. Differences by Student’s T Test. (C-E) Westerns ALDH1+ and ALDH1- populations. Densitometry in Supplementary Fig. S6A-C. (F-H) Mean % ALDH1+ cells +/−SEM in indicated lines after 48 hr of saracatinib (1μM, SI), selumetinib (200nM, MI) or both. For PEO1R: *p=0.045, **p=0.106, ^‡p=0.0048 (F). For OCI-C5x: *p=0.0002, **p=0.0003, ^‡p=0.000001 (G). For OCI-P5x: *p=0.003, **p=0.0029, ^‡p=0.00002 (H). See also Supplementary Fig. S6D for OCI-U1a. (I-K) PEO1R, OCI-C5x and OCI-P5x were treated with SI, MI or both or mock treated for 48hrs, followed by 48 hrs drug washout and then 2,000, 1,000, or 500 cells seeded in biologic triplicate sphere assays. Graphs show mean +/− SEM. (I) PEO1R 2,000 cells: *p=0.0006, **p=0.0001, ^‡p=0.00002; 1,000 cells: *p=0.01, **p=0.00005, ^‡p=0.00002; 500 cells: *p=0.3, **p=0.003, ^‡p=0.0005 (J) OCI-C5× 2,000 cells: *p=0.0004, **p=0.0008, ‡p=0.00005; 1,000 cells: *p=0.0009, **p=0.0003, ^‡p=0.00007; 500 cells: *p=0.0006, **p=0.001, ‡p=0.0003 (K) OCI-P5× 2,000 cells: * p=0.0003, ** p=0.0006, ^‡p=0.00008, 1,000 cells: *p=0.007, **p=0.008, ^‡p=0.002, 500 cells: *p=0.009, **p=0.001, ^‡p=0.00043. For OCI-U1a sphere data, see Supplemental Fig. S6E. See also Supplemental Fig. S6F-I Fig for cell cycle profiles at the time of seeding.

Fig. 6.. Synergistic effects of saracatinib and selumetinib on tumor volumes and tumor initiating cells in vivo

(A) Mean volume ±SEM is graphed for PEO1R xenograft controls, C, and groups treated with saracatinib (SI), selumetinib (MI) and both. (B) Mean fold tumor volume change +/−SEM between week 1.5 and harvest. *p=0.0006, **p=0.007, ^‡p=0.00006; for indicated groups versus control; CDI=0.42. Graphs show mean +/− SEM; significant differences between multiple comparisons vs control were calculated by ANOVA with post-hoc comparisons between treated groups and control. (D-F) Tumors from “A” above were assayed by IHC or IF and mean % positive cells graphed as mean +/− SEM; significant differences between multiple comparisons vs control were calculated by ANOVA with post-hoc comparisons between treated groups and control. IHC Ki67 *p=0.99, **p=0.05, ^‡p<0.0001 (D), Caspase 3 IF *p=0.99, **p=0.54, ^‡p=0.004 (E), and ALDH1 IHC *p=0.32, **p=0.01, ^‡p<0.0001 (F). (photomicrographs in Fig. S7). (G) PEO1R xenografts from “A” above were dissociated and 5,000 pooled tumor cells were seeded in triplicated sphere assays. *p=0.004, **p=0.1734, ‡p=0.0002; CDI=0.71. (H, J) Dissociated tumor cells from “A” above were pooled for each group and implanted into NOD SCID mice in a limiting dilution tumor initiating assay as in Methods. Mean tumor volume/time is graphed in (H). Mean % tumor-free mice/time is graphed for the 5000 cell groups in (I); tumor initiating stem cell (TISC) frequency from limiting dilution cell injections, p=0.018 for dual therapy vs control. (J)