Fig 2: BACH1 promoted in vivo cell growth and angiogenesis in ESCC. (A‐C) KYSE170‐shBACH1 and KYSE170‐shCtrl cells were subcutaneously injected into the flanks of nude mice (n = 6). The xenograft tumors were obtained after 4 weeks, and (B) the tumor volume and (C) weight were measured. (D‐F) KYSE150‐BACH1 and empty vector control cells were subcutaneously injected into the flanks of nude mice (n = 6). The xenograft tumors were obtained after 4 weeks, and (E) the tumor volume and (F) weight were measured. (G‐H) Immunohistochemical staining for N‐cadherin, vimentin, and Slug in xenograft tumors derived from KYSE170‐shBACH1 and KYSE170‐shCtrl cells. Scale bar (40×), 50 μm. (I, J) Immunohistochemical staining for (I) CD31 and (J) VEGFC in xenograft tumors derived from KYSE170‐shBACH1 and KYSE170‐shCtrl cells. Scale bars (20×), 200 μm, scale bars (40×), 50 μm. (K) HE and immunofluorescence staining for CD31 in xenograft tumors derived from KYSE150‐BACH1 and KYSE150‐vector cells. Scale bars (20×), 100 μm, scale bars (40×), 50 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student's t test.

Fig 3: BACH1 transcriptionally activated CDH2, SNAI2, and VEGFC to promote the EMT and angiogenesis. (A) Upper panel, schematic diagram showing the BACH1/antioxidant response element (ARE) consensus motif and predicted BACH1‐binding sequence in the CDH2 promoter region ~3400 bp upstream of the transcriptional start sites (TSSs). Lower panel, schematic diagram showing 350 bp promoter sequences containing the BACH1‐binding sequence or deletion subcloned into the pGL3 luciferase reporter vector. (B, C) ChIP‐qPCR analysis of BACH1 binding to the promoter regions of HMOX1, SNAI2, VIM, and CDH2 in (B) KYSE170 and (C) KYSE150 cells (comparison vs. control IgG binding; HMOX1 was used as a positive control). (D) Luciferase assays of the CDH2 promoter and the mutant promoter in KYSE170 cells with exogenous overexpression of BACH1 or cells transfected with an empty vector. (E) The mRNA levels of VEGFC in BACH1‐knockdown (KYSE170‐siBACH1‐1/‐2 and KYSE30‐siBACH1‐1/‐2) and BACH1‐overexpressing (KYSE150‐BACH1) cells were quantified by qPCR. (F) Analysis of VEGFC mRNA levels in ESCC tissues and paracancerous tissues based on the GSE23400 dataset. (G) The mRNA expression of VEGFC was positively correlated with BACH1 expression according to the analysis of the data of the TCGA database. (H) ChIP‐qPCR analysis of BACH1 binding to the promoter regions of VEGFC compared with that for the control IgG binding. **P < 0.01, ***P < 0.001 by Student's t test.

Fig 4: The expression of BACH1 in ESCC cells. (A) Western blot analysis and (B) qPCR analysis of the levels of BACH1 in various ESCC cell lines. (C) Immunofluorescence analysis of BACH1 (green) and DAPI (blue, cell nuclei) in ESCC cell lines. Scale bars, 30 μm. (D) Western blot analysis and (E) qPCR analysis of the expression of BACH1 in the transient or stable knockdown of BACH1 in KYSE30 and KYSE170 cells and the corresponding control cells and the cells with stable overexpression of BACH1 using KYSE150‐BACH1 and the corresponding KYSE150‐vector control cells. The protein levels of BACH1 were normalized to those of β‐actin, and fold changes are shown relative to the control group. *P < 0.05, **P < 0.01 and ***P < 0.001 by Student's t test.

Fig 5: BACH1 is expressed at a high level in ESCC. (A) Representative images of BACH1 expression in paired esophageal epithelial tissues (upper row) and ESCC tissues (lower row) obtained by IHC staining of ESCC tissue arrays. (B) Comparison of the IHC scores of BACH1 in paired esophageal epithelial tissues and ESCC tissues by Mann‐Whitney U test. (C) Comparison of BACH1 mRNA levels in ESCC tissues and noncancerous tissues in the TCGA database by the Mann‐Whitney U test. (D, E) Kaplan‐Meier plot of the TCGA data showing overall survival (OS) and disease‐free survival (DFS) of ESCC patients with high and low levels of BACH1 assigned based on the cut‐off values obtained by the ROC curve analysis.