Fig 2: FBXO2-FBN1 promote EC cells proliferation by regulating the cell cycle and autophagy signaling pathways. (A) Heatmap and volcano plot of differential genes enrichment based on the RNA sequencing of the RL95-2-NC and RL95-2-shFBXO2 groups. (B) Go functional enrichment analysis. (C) KEGG pathway enrichment analysis. (D,E) GSEA (autophagy signaling pathway) enrichment analysis and the corresponding heatmap. (F,G) GSEA (the cell cycle) enrichment analysis and the corresponding heatmap.

Fig 3: (A–L) Verification of differential genes associated with the cell cycle and autophagy signaling pathways. (A) mRNA levels of MCM7, CDK4, CCND1, SMAC1A, CCND2, CHEK1, CDC14B, and CCNA1 in RL95-2-NC and RL95-2-shFBXO2 groups. CDK4, CCND1, CCND2, and CCNA1 were significantly down-regulated in the RL95-2-shFBXO2 group compared with RL95-2-NC group. Data are presented as mean ± SD, **P < 0.01, t-test. (B) mRNA levels of PPP2CA, ATG4A, ATG4B, ATG4C, and ATG4D in the RL95-2-NC and RL95-2-shFBXO2 groups. ATG4A and ATG4D were significantly up-regulated in the RL95-2-shFBXO2 group compared with the RL95-2-NC group. Data are presented as mean ± SD, **P < 0.01, ***P < 0.001, t-test. (C) Protein levels of CDK4, CyclinD1, CyclinD2, and CyclinA1 in RL95-2-NC, RL95-2-shFBXO2, RL95-2-shFBN1, RL95-2-shFBXO2/shFBN1, and RL95-2-ovFBN1 cells. Knockdown of FBXO2 and overexpression of FBN1 revealed the same tendency in decreasing protein levels of CDK4, CyclinD1, CyclinD2, and CyclinA1 compared with the RL95-2-NC group. Knockdown of both FBXO2 and FBN1 showed increased protein levels of CDK4, CyclinD1, CyclinD2, and CyclinA1 compared with the RL95-2-shFBXO2 group. a-Tublin was used as a loading control. (D) The autophagy-specific inhibitor 3-MA (2 mM) reversed the effects of FBXO2 on the RL95-2 cells proliferation. Data are presented as mean ± SD, *P < 0.05, **P < 0.01, one-way ANOVA. (E) The CDK4/6 inhibitor palbociclib (2 µM) reversed the effects of FBXO2 on the Ishikawa cells proliferation. Data are presented as mean ± SD, **P < 0.01, one-way ANOVA. (F–L), FBXO2 silencing inhibited EC carcinogenicity in vivo via FBN1. (F) RL95-2-NC, RL95-2-shFBXO2, RL95-2-shFBN1, and RL95-2-shFBXO2/shFBN1 cells were injected into BALB/c nude mice subcutaneously (0.2 ml, 5 × 106 cells) and harvested at day 27. (G) Mice weight were monitored every 3 days and recorded (P > 0.05). (H,I) Tumor weight was recorded in a time-dependent manner (H) and at harvest day (I) among the RL95-2-NC, RL95-2-shFBXO2, RL95-2-shFBN1, and RL95-2-shFBXO2/shFBN1 groups. Data are presented as mean ± SD, *P < 0.05, **P < 0.01, one-way ANOVA. (J) IHC staining for FBXO2, FBN1, and Ki67 in histologic sections of transplanted tumors. (K) Inverse correlation between FBXO2 and FBN1 staining, Pearson r2 = –0.77, P < 0.001. (L) Positive correlation between FBXO2 and Ki67 staining, Pearson r2 = 0.857, P < 0.0001.

Fig 4: FBXO2 degrades FBN1 protein through its ubiquitination. (A,B) Overexpression or knockdown of FBXO2 shortened or prolonged FBN1 protein half-life in Ishikawa or RL95-2 cells, respectively. Ishikawa-NC, Ishikawa-ovFBXO2, RL95-2-NC, and RL95-2-shFBXO2 cells were treated with 100 µg/ml of the protein synthesis inhibitor cycloheximide (CHX), and harvested at indicated time points for evaluation of FBN1 protein levels. *P < 0.05, **P < 0.01. (C) FBXO2 enhanced FBN1 polyubiquitination in a dose-dependent manner. HEK-293T cells were transfected with Flag-FBXO2, Myc-FBN1, and HA-ub for 48 h and then treated with 20 µM MG132 for 8 h. Cell lysates were prepared for co-IP with anti-Myc beads and western blotting analysis was used to detect the HA-ub (FBN1 ubiquitination levels). (D) Schematic representation of the FBXO2 protein. FL, full long; N, N terminus, including F-box domain; N, N terminus, including FBA domain; Tyr278A, FBA domain was replaced with non-functional Tyr278A mutant; Trp279A, FBA domain was replaced with non-functional Trp279A mutant. (E) The FBXO2 C terminus down-regulated FBN1 protein levels in a dose-dependent manner while the FBXO2 N terminus did not have the function. HEK-293T cells were transfected with 1.5 µg of Myc-FBN1 and Flag-FBXO2 FL/N/C terminus plasmids at doses of 0, 0.75, 1.5, or 3.0 µg for 48 h before harvesting for western blotting analysis. a-Tublin was used as a loading control. (F) The FBXO2 Tyr278A and Trp279A mutants did not down-regulate the FBN1 protein levels. HEK-293T cells were transfected with 1.5 µg Myc-FBN1 and Flag-FBXO2 FL/Tyr278A/Trp279A plasmids at doses of 0, 0.75, 1.5, or 3.0 µg for 48 h before harvesting for western blotting analysis. a-Tublin was used as a loading control. (G) The FBXO2 C terminus (FBA domain) enhanced FBN1 polyubiquitination. HEK-293T cells were transfected with Flag-FL/N/C/Tyr278A/Trp279A, Myc-FBN1, and HA-ub for 48 h and then treated with 20 µM MG132 for 8 h. Cell lysates were prepared for co-IP with anti-Myc beads and western blotting analysis was used to detect the HA-ub (FBN1 ubiquitination levels).

Fig 5: FBN1 identified as a novel substrate for FBXO2. (A,B) Cluster analysis and volcano plot of the differential proteins regulated by FBXO2 ubiquitination based on the LC MS/MS data. (C) FBN1 protein was directly bound to FBXO2 protein. (D) The intersection of IP-MS, LC MS/MS, and our RNA sequencing data. (E) FBN1 mRNA expression measured by RT-PCR in RL95-2-shFBXO2 and Ishikawa-ovFBXO2 cells compared with their parental cells (P > 0.05). ***P < 0.001, t-test. (F) Endogenous FBXO2 and FBN1 protein levels in RL95-2 and Ishikawa cells. (G) FBN1 protein levels as measured by western blot analysis in shFBXO2-RL95-2 cells and Ishikawa-ovFBXO2 cells compared with their parental cells. (H) Immunohistochemical staining of FBXO2 and FBN1 in 90 EC samples and paired para-carcinoma samples. Left four photographs: FBXO2 was highly expressed in carcinoma tissues (upper panel), while FBN1 showed low expression in the same patient (lower panel). Right four photographs: FBXO2 showed low expression in carcinoma tissues (upper) while FBN1 was highly expressed in the same patient (lower). Original magnification, ×60 and ×400; scale bar: 200 and 50 µm, respectively. (I) Inverse correlation between FBXO2 and FBN1 staining, Pearson r = –0.305, P = 0.004. (J) Exogenous FBXO2 and FBN1 proteins interacted with each other in HEK-293T cells. HEK-293T cells were transfected with Flag-FBXO2, Myc-FBN1, and co-transfected with Flag-FBXO2 and Myc-FBN1 for 48 h, respectively. After treatment with 20 µM MG132 for 8 h, cell lysates were prepared for co-IP with anti-Flag or anti-Myc beads and western blot analysis. (K) Endogenous FBXO2 and FBN1 proteins interacted with each other in endometrial cancer cell lines. RL95-2 and Ishikawa cell lysates were prepared for co-IP with anti-FBXO2 or anti-FBN1 and western blot analysis. (L) FBXO2 and FBN1 co-localized in RL95-2 and Ishikawa cells cytoplasm and membrane. EC cells were immunostained with anti-FBXO2 (red) and anti-FBN1 (green) antibodies and visualized with confocal microscopy. DAPI (blue) was used to indicate cell nuclei. Scale bar, 25 µM.