Fig 2: Flowchart of the study. First, BEV-sensitive and BEV-resistant ovarian cancer mouse models were constructed, and RNA sequencing was performed on tumor tissue. Then, a BEV-related prognostic signature was established using machine learning, and S100B was identified as the most important molecule regulating BEV sensitivity in ovarian cancer. Its function and mechanism were analyzed in vitro. Finally, BEV efficacy when combined with an S100B inhibitor was verified in vivo. BEV: Bevacizumab; DEGs: Differentially expressed genes.

Fig 3: Co-culture of S100B overexpressing ovarian cancer cells promotes HUEVC angiogenesis and migration. (A) Representative Western blot images and their densitometric quantification showing comparative S100B protein expression profiles among the three ovarian cancer cell lines (A2780, HEY, and OVCAR3) (B) Western blot analysis and statistical analysis of OVCAR3 and HEY cells overexpressing S100B. RT-qPCR analysis of alternative angiogenic factor mRNA expression levels in (C) OVCAR3 and (D) HEY ovarian cancer cell lines following S100B overexpression. (E) The level of S100B in the supernatant of OVCAR3 and HEY cells overexpressing S100B significantly increased, as determined by ELISA. (F) There was no significant change in VEGFA levels in the supernatant of OVCAR3 and HEY cells overexpressing S100B. (G, H) HUVEC tube formation significantly increased after co-culture with ovarian cancer cell lines overexpressing S100B, with or without exogenous BEV. (I, J) Co-culture with ovarian cancer cells overexpressing S100B significantly promoted HUVEC sprouting with or without exogenous BEV. (K, L) Co-culture of ovarian cancer cells overexpressing S100B significantly promoted migration of HUVECs in a Transwell system with or without exogenous BEV. (M, N) Co-culture of ovarian cancer cell lines overexpressing S100B significantly promoted healing of HUVECs with or without exogenous BEV in a scratch assay. BEV: Bevacizumab; NC: negative control; OE: overexpression; *: P < 0.05; **: P < 0.01; ***: P < 0.001; ****: P < 0.0001. Scale: All panels are 50 μm.

Fig 4: S100B enters endothelial cells through clathrin-dependent endocytosis to promote angiogenesis and migration. (A) Western blot detection showed that there was no significant change in RAGE receptor expression in endothelial cells after exogenous S100B treatment. (B) Western blot detection showed that protein levels of S100B in endothelial cells significantly increased after co-culture with ovarian cancer cells overexpressing S100B. (C) S100B total protein and nuclear protein in endothelial cells significantly increased after treatment with exogenous S100B, as determined by western blot analysis. (D) Western blot analysis of protein levels of S100B in endothelial cells after pretreatment with the clathrin-mediated endocytosis inhibitor Pitstop-2 or caveolae/caveolin-1-mediated endocytosis inhibitor nystatin. Changes in the (F, G) tubular phenotype, (H, I) sprouting phenotype, (J, K) migration phenotype (Transwell assay), and (L–M) wound healing phenotype (scratch migration assay) of endothelial cells after treatment with the RAGE receptor inhibitor FPS-ZM1, the clathrin-mediated endocytosis inhibitor Pitstop-2, and the caveolae/caveolin-mediated endocytosis inhibitor nystatin. BEV: Bevacizumab; *: P < 0.05; **: P < 0.01; ***: P < 0.001; ****: P < 0.0001. Scale: All panels are 50 μm.

Fig 5: S100B promotes tube formation and migration of endothelial cells through the FOXO1/β-catenin signaling pathway. (A) DIA proteomics thermogram of differentially expressed proteins in HUVECs with or without exogenous S100B. (B) Function and pathway enrichment analysis by Metascape database of differentially expressed proteins. (C–E) Western blot analysis showed that levels of S100B in ovarian cancer tissue of BEV-resistant mice significantly increased, while FOXO1 levels decreased. (F) Histogram of an RT-qPCR analysis of downstream transcriptional target genes of β-catenin. (G–J) Western blot analysis showed that FOXO1 total protein and nuclear protein levels in endothelial cells significantly decreased after treatment with exogenous S100B, while levels of β-catenin and MMP7 significantly increased. (K–M) Western blot analysis of FOXO1 and β-catenin levels in S100B-treated endothelial cells overexpressing FOXO1 compared with the control group. Endothelial cells transfected with a control virus or FOXO1-overexpressing lentivirus and treated with S100B were evaluated for their (N) tubular phenotype by tube formation assay, (O) sprouting phenotype by fibrin bead sprouting assay, and (P) migration phenotype by Tranwell assay. (Q) Schematic diagram of the mechanism of exogenous S100B on endothelial cell angiogenesis. *: P < 0.05; **: P < 0.01;^: P < 0.05 vs. S100B−/oeFOXO1−; #: P < 0.05 vs. S100B−/oeFOXO1+; &: P < 0.05 vs. S100B+/oeFOXO1 − . Scale: All panels are 50 μm.