Fig 2: Schematic diagram. DNMT3B inhibits miR-451a transcription by methylation modification of the miR-451a promoter, thus activating EPHA2 and the following PI3K/AKT signaling, leading to BCa progression

Fig 3: Overexpression of miR-451a or knockdown of DNMT3B inhibits the aggressiveness of BCa cells in vitro. BCa cells were transfected with shDNMT3B alone or with miR-451a inhibitor. a, miR-451a expression in BCa cell lines determined by RT-qPCR; b, the number of T24 and 5637 cells at the S phase determined by EdU staining where red represents the replicating DNA stained by EdU and blue represents the nucleus stained by DAPI; c, apoptosis in T24 and 5637 cells detected by flow cytometric analysis; d, the formation of apoptotic bodies in T24 cells and 5637 cells determined by TUNEL staining where green represents apoptotic bodies stained by TUNEL and blue represents nuclei stained by DAPI; e, number of migratory T24 and 5637 cells was detected by Transwell assays; f, number of invasive T24 and 5637 cells was detected by Transwell assays. Data are represented as the means ± SD from three independent assays. Statistically significant differences were calculated using one-way or two-way ANOVA, followed by Tukey’s multiple comparison test. **p < 0.01 vs sh-NC or sh-DNMT3B + inhibitor NC treatment

Fig 4: DNMT3B promotes miR-451a promoter methylation. a, DNMT3B expression in BCa and matched adjacent tissues in the Oncomine database; b, DNMT3B expression in normal and tumor tissues in TCGA databases predicted by the GEPIA website; c, the association between DNMT3B expression and the histological subtype of BC patients; d, the association between DNMT3B expression and the clinical stage of BC patients; e, the association between DNMT3B expression and the gender of BCa patients; f, DNMT3B mRNA expression in BCa cell lines T24, 5637, TCCSUP, UM-UC3 and SV-HUC-1 cells determined by RT-qPCR; g, DNMT3B protein expression in BCa cell lines T24, 5637, TCCSUP, UM-UC3 and SV-HUC-1 cells determined by western blot (full-length blots/gels are presented in Supplementary Figure 1A, B); h, the binding relationship between DNMT3B and miR-451a promoter in T24 and 5637 cells tested by ChIP-qPCR; i, promoter methylation in T24 cells in response to shDNMT3B (the top 3 haplotypes of high frequency were presented); j, promoter methylation in 5637 cells in response to shDNMT3B (the top 3 haplotypes of high frequency were presented); k, miR-451a expression in BCa cell lines in response to shDNMT3B determined by RT-qPCR. Data are represented as the means ± SD from three independent assays. Statistically significant differences were calculated using unpaired t-test (panel b), one-way (panel c and d) or two-way ANOVA (panel f-k), followed by Tukey’s multiple comparison test. *p < 0.05 vs tumor tissues, **p < 0.01 vs sh-NC or IgG treatment, ***p < 0.001 vs SV-HUC-1 cells

Fig 5: DNMT3B/miR-451a/EPHA2/PI3K/AKT axis modulates BCa cell growth in vivo. T24 cells (2 × 106 cells/mL) with stable poor expression of DNMT3B or miR-451a were injected into the right armpit of NOD/SCID mice. a, from the fifth day, the growth rate of T24 cells was evaluated by measuring the volume every 5 days in vivo; b, tumor weight; c, KI67 staining intensity in tumors evaluated by immunohistochemistry; d, EPHA2 staining intensity in tumors evaluated by immunohistochemistry; e, phos-PI3K staining intensity in tumors evaluated by immunohistochemistry; f, phos-AKT1 staining intensity in tumors evaluated by immunohistochemistry. Data are represented as the means ± SD from three independent assays. Statistically significant differences were calculated using one-way or two-way ANOVA, followed by Tukey’s multiple comparison test. **p < 0.01 vs mice injected with cells transfected with sh-NC or sh-DNMT3B + inhibitor NC