Fig 2: S100A10 and PAI-1 are regulated by Smad4-dependent TGFβ1-mediated signaling and FOXC2-mediated PI3K signaling. The model proposes that the treatment of epithelial cells with TGFβ1 increases S100A10 and PAI-1 mRNA and protein levels through canonical Smad-dependent TGFβ1 signaling. S100A10 and PAI-1 are also affected by the pro-growth PI3K pathway. Serum starvation, PI3K inhibition or mTOR inhibition upregulate S100A10 expression suggesting an inhibitory effect through this pathway. The transcription factor FOXC2, which is downstream of PI3K, mediates the repression of S100A10 expression and induction of PAI-1.

Fig 3: Stabilization of vimentin-S56 phosphorylation in CSC-enriched populations increases multinucleation and inhibits SSCDs(A) Schematics showing monolayer culture (top) and suspension culture, which is CSC enriched (bottom) before cells were disassociated, plated, and treated with 2 μg/mL doxycycline or 0.5 μM FiVe1. Image created with BioRender.com.(B) Relative percentages of multinucleated 4T1 VIM-S56E cells treated with doxycycline following monolayer or suspension growth.(C) Relative percentages of multinucleated 4T1, SUM159, and MDA-MB-231 cells treated with FiVe1 following monolayer or suspension growth.(D) Representative fluorescence-activated cell sorting analysis of HMLER and HCC1806 cells gated for CD44high/CD24low and CD44low/CD24high populations.(E) Relative percentages of multinucleation of CD44high/CD24low and CD44low/CD24high HMLER (left) and CD44high/CD24low and CD44low/CD24high HCC1806 (right) cells treated with FiVe1.(F) Left: Representative images of FOXC2 and Numb immunofluorescence staining of 4T1 cells showing staining patterns of cells undergoing symmetrical differentiated cell division (SDCD) and SSCD. Right: Percentages of cells undergoing SDCD, SSCD, and asymmetrical stem cell division (ASCD) in 4T1 cultures treated with 0.1% DMSO or 1 μM FiVe1.(G) Left: Representative images of FOXC2 and Numb immunofluorescence staining of SUM159 cells showing staining patterns of cells undergoing SDCD and SSCD. Right: Percentages of cells undergoing SDCD, SSCD, and ASCD in SUM159 cultures treated with 0.1% DMSO or 1.2 μM FiVe1. An unpaired t test was used to calculate significance. *p < 0.05; **p < 0.01; ***p < 0.001; and ****p < 0.0001; NS indicates not significant. Error bars are ± SEM.

Fig 4: β-Catenin mediates FOXC2-associated EMT and acquisition of stem cell-like properties. (A) Phase contrast images of HMLE-Vector (left) and HMLE-Snail (right) cells treated with a control shRNA (top row, shControl) or a shRNA targeting CTNNB1 (bottom row, shβ-catenin). (B) qRT-PCR quantification of expression levels of CTNNB1 and FOXC2 in HMLE-Snail cells expressing indicated shRNA. (C) Western blot analysis of HMLE-Vector (left) and HMLE-Snail (right) cell lines for mesenchymal markers after transducing with shControl or shβ-catenin. (D) FACS analysis of CD44 and CD24 expression in HMLE-Vector (top) and HMLE-Snail (bottom) after transducing with shControl or shβ-catenin. (E) FACS quantification of CD24high cells in HMLE-Snail cells with or without β-catenin knockdown. (F) Quantification of mammospheres per 1000 cells in HMLE-Vector and HMLE-Snail cells with or without β-catenin knockdown. Error bars represent s.d. from triplicate experiments. Statistical comparisons were performed by Student’s t-test. *** p < 0.0005; ** p < 0.005.

Fig 5: β-catenin and FOXC2 are correlated in clinical samples [38]. (A) TCGA data of breast cancer subtypes vs. expression of RNA, Protein/RPPA data for CTNNB1 (β-catenin), ESR1, and FOXC2. FOXC2 and β-catenin are highly expressed in basal breast cancer subtypes as compared with those in Her2 or luminal breast cancer subtypes. (B) EMTome generated data for correlation of Wnt/β-catenin signaling and FOXC2 expression during different stages in breast invasive carcinoma.