Fig 2: LSD1 acts cooperatively with Slug to inactivate ESR1 promoters by demethylating H3K4me2. (a, b) Western blot (a) and RT–PCR (b) analyze effects of LSD1 and Slug on ERa expression via Slug overexpression and LSD1 knockdown in MCF-7 cells. (c, d) Western blot (c) and RT–PCR (d) analysis of effects of LSD1 and Slug on ERa expression via Slug and LSD1 knockdown in MDA-MB-231 cells. (e) MDA-MB-231 cells expressing pHAGE-CMV-Flag-HA-LSD1 and pcDNA3.1-Slug-Flag or pcDNA3.1-?N-Slug-Flag was analyzed by IP. Cell lysates were immunoprecipitated with anti-HA antibody. Immunocomplexes were then immunoblotted using antibodies against Flag and HA. (f) A dual-luciferase reporter assay is used to investigate ESR1 promoter 2 activity in MDA-MB-231 cells by overexpressing wild-type or mutant Slug. (g) ChIP-qPCR analysis of LSD1 recruitment on ESR1 promoter in MDA-MB-231shSlug cells and their control cells. (h) ChIP-qPCR analysis of H3K4me2 recruitment on ESR1 promoter in MDA-MB-231shLSD1 cells and their control cells. All the ChIP-qPCR results represent % of input chromatin. (i) A dual-luciferase reporter assay is used to investigate ESR1 promoter activity in MDA-MB-231 cells after LSD1 knockdown. Statistical analysis is performed using GraphPad Prism version 7.0. Means±s.d. is calculated by at least three independent experiments. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 (Student’s t-test) as compared to the control group.

Fig 3: ERa expression is directly regulated by Slug. (a) A schematic representation of the ESR1 promoter region in which the promoter is divided into three subregions including eight E-box-binding sites. (b) Different concentration of Slug siRNA (20, 40 or 80 nm, respectively) was co-transfected with the ESR1 promoter-luciferase plasmid in MDA-MB-231 cells. Variation in transfection efficiency was normalized by Renilla luciferase activity. (c) Different concentration of Slug (0.1, 0.3 or 0.9 µg, respectively) was co-expressed with the ESR1 promoter-luciferase construct in MCF-7 cells. After 24 h, Renilla luciferase activity was used to normalize variation in transfection efficiency. (d) Normal IgG or anti-Slug antibodies were used in a ChIP assay to determine which E-box-binding site Slug binds in the ESR1 promoter in MDA-MB-231 cells. Seven primer pairs covering different binding sites within promoter 1 and 2 are designed. (e, f) After ChIP, qPCR was used to analyse Slug recruitment on CDH1 (e) and ESR1 (f) promoter in MDA-MB-231shSlug cells and their control cells. The results represent % of input chromatin. Statistical analysis is performed using GraphPad Prism version 7.0. Means±s.d. is calculated by at least three independent experiments. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 (Student’s t-test) as compared to the control group.

Fig 4: SOX9 knockdown drives MET and impairs metastasis. (A) Representative Western blot of SOX9, E-CAD, SNAIL, SLUG and BMI1 protein levels in Panc-1 and RWP-1 cell lines lentivirally infected with virus harboring a specific short hairpin RNA against SOX9 (shSOX9) or the corresponding empty vector (pLKO) (n = 4). (B) Images show filamentous actin (F-actin) labeled with phalloidin in RWP-1 pLKO and shSOX9 cells. Nuclei were marked with DAPI, and the merged images are presented in right panel. (C) Representative images of E-Cadherin and N-Cadherin expression determined by immunofluorescence in Panc-1 and RWP-1 pLKO and shSOX9 cells (n = 3). (D) Relative migration determined by transwell assays based in Boyden chamber (n = 4) in indicated cells and genotypes. (E) Frame at the time of wound closure (from Supplementary Video S2: down, middle panel) obtained in a scratch assay performed with a mix of IMIMPC-2 cells transduced with GFP or shSOX9. (F) Relative invasion of Panc-1 and RWP-1 cell lines with shSOX9 with respect to pLKO cells. Invasive cells analyzed in vitro using collagen transwell assays based in Boyden chamber. (G) Tumor volume at the indicated time points of subcutaneous tumors formed by control pLKO and shSOX9 RWP-1 cells (n = 12). (H) Representative images of SOX9, E-Cadherin and ß-Catenin expression determined by immunohistochemistry in subcutaneous tumors represented in 3G. (I) Percentage of nude mice that developed metastatic foci derived from Panc-1 and RWP-1 pLKO and shSOX9 cells in tail vein injection assays (n = 5). Asterisks (*, **, and ***) indicate statistical significance (p < 0.05, p < 0.01, and p < 0.001, respectively).

Fig 5: Negative relationship between Slug and ERa in human breast cancer cells. (a, b) Analysis of Slug and ERa expression levels in MCF-7, T47D, SKBR3, MDA-MB-231 and BT549 by western blot and RT–PCR. (c, d) Western blot (c) and RT–PCR (d) analysis of protein and mRNA levels as indicated in MCF-7 cells transfected with increasing concentrations (2, 4 and 8 µg, respectively) of Slug. (e, f) Western blot (e) and RT–PCR (f) analysis of protein and mRNA levels in MDA-MB-231 cells transfected with two different interference sequences of Slug (40 nm, separately). (g) Western blot analysis of Slug knockdown level in stable cell lines of MDA-MB231shSlug. (h) Immunofluorescence micrographs of Slug (red) and ERa (green) expression in MDA-MB-231shNC cells (upper) and MDA-MB-231shSlug (lower). DAPI is for nuclei (blue). Magnification, × 400. Protein and mRNA expression was normalized to ß-actin. In figure 2d, NC means transfecting with control pcDNA3.1. In figure 2e and f, siNC is scrambled control siRNA used for transient transfection. ShNC without shRNA is used as a control by infecting pLKO.1 lentivirus in MDA-MB-231 cell line. Statistical analysis is performed using GraphPad Prism version 7.0. Means±s.d. is calculated by at least three independent experiments. *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 (Student’s t-test) as compared to the control group. RT–PCR, reverse transcription–PCR.