Fig 1: EGR1 and ELF3 are mutually exclusive markers for undifferentiated and differentiated HMEC cells, respectively. (A) Top five differentially expressed genes in GO term: 0045893 (positive regulation of transcription, DNA-templated). Microarray data sets were compared between cells before and after YDAC removal for 2 days (YDAC (+) and YDAC (-), respectively). (B, C) Immunostaining showing EGR1 expression in nearly half of the YDAC-maintained undifferentiated basal cell population. EGR1 expression was rare after differentiation through YDAC removal. In contrast, ELF3 expression was rare in YDAC-maintained cells, but observed in nearly half of all cells 3 days after YDAC removal. Note that cells with a decrease in EGR1 and an increase in ELF3 almost coincided with those expressing Claudin-4. (D) Quantifications of stained cells and (E) immunoblotting with samples in B and C supports the data. Hundred cells were counted in each group (n=5). (F) Left, Double staining for EGR1 and ELF3 during the transient status (day 2). Note the mutually exclusive expression. Right, the quantification. (G) Tissue expression data from the Human Protein Atlas (HPA) website, where anti-ELF3 stains differentiated luminal cells, and anti-EGR1 partially stains basal cells, including undifferentiated cells. The specificity of antibodies is validated by the presence of single bands in endogenous immunoblotting (see also Materials and Methods section). (H) Our new method for mammary cell culture in an in vivo-like bilayer. The bilayer was created by culturing cells in YDAC medium for approximately 3 days after they reached confluence (see also Supplementary Figure 2 G). Immunostaining and confocal microscopy analysis revealed ZO-1 and ELF3 expressions in the upper layer and EGR1 expression in the lower layer. Note their relevance with the natural organization of mammary glands in G and mammary organoids in Supplementary Figure 2 F. Scale bars, 10 µm. All cell data were obtained from donor 1. See also Supplementary Figure 2 (donor 2, independent donor as a biological replicate).
Fig 2: Endogenous EGR1 suppresses Claudin expression in HMECs.(A) Left, immunostaining of YDAC-maintained undifferentiated basal cell population treated with siEGR1 for 3 days, which caused the disappearance of EGR1 expression and an increase in the population of Claudin-4- and ELF3-stained cells. 537 and 538 are the designations of siRNAs. Right, quantification of cells stained with each marker. Hundred cells were counted in each group (n = 5). (B) Immunoblotting with samples in A supports the data. (C) Ectopic induction of EGR1 in differentiated cells with Dox-containing YDAC-free F medium. Left, immunostaining indicating disappearance of Claudin-4 at cell-cell borders (top), impaired formation of actin circumferential ring (middle), disappearance of ELF3 (bottom), in ectopically EGR1-expressing cells. Right, quantification of cells stained with each marker. Hundred cells were counted in each group (n = 5). (D) Immunoblotting with samples in C. Mock, empty vector. siControl corresponds to a negative control. Scale bars, 10 μm.
Fig 3: Endogenous ELF3 localizes Claudin at cell-cell border in HMECs.(A) Left, immunostaining of YDAC-maintained basal cell population treated with siELF3, whereafter YDAC was withdrawn for 3 days to induce differentiation. ELF3 knockdown cells showed a disturbance in Claudin-4 localization at cell-cell borders. Note that, 623 and 625 are the designations of siRNAs. Right, quantification of cells stained with each marker. Hundred cells were counted in each group (n = 5). (B) However, immunoblotting did not confirm ether decrease in Claudin-4, or increase in EGR1, TWIST1, and SNAI2 but confirmed a decrease in actin organizer GRHL3, implying the importance of ELF3 not for EMT but for cell polarity in the differentiated state. (C) Left, another immunostaining with the same sample in A, indicating impaired formation of actin circumferential ring in ELF3 knockdown cells, with similarity to ectopic EGR1 expression. Right, quantification of cells. Hundred cells were counted in each group (n = 5). (D) Ectopic induction of ELF3 in undifferentiated basal cells with Dox-and YDAC-containing F medium. Left, immunostaining indicating no increase in Claudin-4 staining in ectopically ELF3-expressing cells. Right, quantification of cells stained with each marker. One-hundred cells were counted in each group (n = 5). (E) Immunoblotting with samples in D. (F) Graphical abstract of our cellular model and findings. See also: main text. Mock, empty vector. siControl corresponds to a negative control. Scale bars, 10 μm.
Supplier Page from DNASU for EGR1 (Homo sapiens) in pDONR221 (Gateway donor/master vector)