Fig 2: GPRC5A is co‐expressed with EphA2 and associates with worse survival of OC patients AKaplan–Meier survival curve illustrates the overall survival (OS) of patients with high or low (top 40% vs. bottom 40%) GPRC5A mRNA expression. Mean OS was 46 months for GPRC5Ahigh vs. 56 months for GPRC5Alow.BScatter plot illustrates the correlation of EphA2 and GPRC5A mRNA in OC (using TCGA by cBioPortal for Cancer Genomics).C–FKaplan–Meier survival curves illustrate the progression‐free survival (PFS) of patients with high or low (top 25% vs. bottom 25%) RSK1 (RPS6KA1; C), RSK2 (RPS6KA3; D), RSK3 (RPS6KA2; E), and RSK4 (RPS6KA4; F) mRNA expression. No significant differences in PFS were observed in these survival analyses.Data information: In (A, C–F), logrank test was used. In (B), Spearman's rank test was used.

Fig 3: RSK‐EphA2‐pS897 inhibition specifically sensitizes GPRC5Ahigh HGSC cells to platinum chemotherapy ARepresentative confocal micrographs show EphA2 (red), RSK (green: top), and GPRC5A (green: bottom) in frozen sections of HGSC patient tumors. S indicates the stroma. Scale bars: 50 μm.BIndicated proteins and EphA2 phosphorylation were assessed by immunoblotting in patient‐derived HGSC cells and CAFs.CGPRC5A in earlier (top) and later passage (bottom) HGSC cells. See normalized GPRC5A relative to OCKI_p01 below the immunoblot images.DCharts illustrate EphA2‐pS897 inhibition (circles) and EphA2‐pY588 increase (squares) by 10 µM BI‐D1870 alone (unbroken line) or in combination with 0–10 µM cisplatin (dotted line) in the GPRC5Ahigh HGSC cells. See Appendix Fig S7B for immunoblots.ECell viability of the GPRC5Ahigh OCKI_p02 and OCKI_p06 after treatment with a combination of 0–20 µM cisplatin and 10 µM BI‐D1870 for 72 h. N = 3.F, GCell viability upon 72‐h combinatorial treatment with 0–20 µM cisplatin and 25 µM LJH685 in patient‐derived CAFs (F; N = 3) and 10 µM BI‐D1870 in mesothelial cells (G; N = 3).HChart illustrates the changes in EphA2‐pS897 and EphA2‐pY588 by BI‐D1870 alone and with cisplatin in mesothelial cells. Upon combination with 10 µM cisplatin for 72 h, 10 µM BI‐D1870 decreased EphA2‐pS897 and EphA2‐pY588. See Appendix Fig S7J for immunoblot.IConfocal micrographs of EphA2 (red) and GPRC5A (green) in primary OCKI_p06 patient cells cultured in 3D collagen for 4 days before 72‐h treatment with or without 10 µM cisplatin. Scale bars: 20 μm.Data information: In (E–G), data are presented as mean (SD). *P < 0.05. Exact P‐values are provided in Appendix Table S10, Student's t‐test.Source data are available online for this figure.

Fig 4: RSK interacts with REPS1 and phosphorylates p-REPS1 S709. (A) The domain structure of REPS1 and its mutants. (B) 293T cells transfected with Flag-REPS1 and HA-RSK1 plasmids were subjected to co-immunoprecipitation with anti-Flag magnetic beads and the western blot were completed with anti-HA and anti-Flag M2 antibodies. (C) HEK239E cells stably expressing RSK1/2/3 shRNA were lysed and the lysates were subjected to immunoblots with the indicated antibodies. (D) In vitro kinase assay was carried out as described in Materials and Methods. The data presented here was a representative of two independent experiments.

Fig 5: MEK-RSK mediated the phosphorylation of REPS1 at S709 in response to EGF and amino acid stimulation. After 293E cells were serum-starved overnight, these cells were treated with the inhibitors for 30 min, and then 10 ng/ml human EGF for 1 hr (A, B). Or the serum-starved cells were cultured in D-PBS for 1 hr and treated with inhibitors for 30 min and changed to DMEM including 25 mM HEPES, 1X Glutamax for 1 hr (C, D). All of the lysates were subject to immunoblots with the indicated antibodies. The intensity of the p-REPS1 bands was quantified by ImageJ program and divided by total REPS1. All of the data were obtained from two independent experiments. *P < 0.05, **P < 0.01, ***P < 0.005.