Fig. 1

High-content screening of normal human mammary epithelial cells (hMECs). a Schematic representation of a normal human mammary duct in vivo showing the major cell types: basal cells (red) and luminal cells (blue). b Light micrographs of immunohistochemical staining of K19 and K14 on FFPE sections of normal human breast delineating luminal and basal cells respectively. c Representative image of immunofluorescence staining of primary hMECs acquired using the IN Cell Analyser 2000 (GE Healthcare, Silverwater, NSW, Australia) at 20× magnification. Cells were stained for K14 (red) and K19 (blue) and counterstained with DAPI (white). Areas of the four possible cell phenotypes are indicated (arrows): double negative or K14⁻/K19⁻; double positive or K14⁺/K19⁺; K14⁺ (K14⁺/K19⁻); and K19⁺ (K14⁻/K19⁺). d Identification of cell phenotype on a single cell basis using digital image analysis. The image analysis protocol defines cells based on nuclear segmentation (first column), then subsequently measures the average intensity of K14 (second column) and K19 (third column) within each cell. The four possible outcomes for K14 and K19 are shown by row. e Intensity distributions for K14 (left panels) and K19 (right panels) in primary hMEC culture of negative secondary only control (black) and stained samples plotting their respective cytokeratin distribution (K14, red; K19, blue). f Subpopulation frequency of the four cell phenotypes in the first passage of hMECs for five donors. Mean subpopulation frequencies across (n = 5–6) technical replicates were calculated as the average percentage of total cell number within each well for each of the five individual donors (± standard error of the mean (SEM)). g Subpopulation frequency with passage and enrichment using FACS (Fluorescence-activated cell sorting). Cells from reduction mammoplasties were dissociated and cultured in vitro (P1). After 5–7 days, cells were harvested and sorted based on expression of CD10 and MUC1 using FACS. CD10⁺, MUC1⁺ and unsorted (US) cells were cultured for 5 days (P2) before fixation and analysis of population heterogeneity. Mean subpopulation frequencies were calculated as the average percentage of total cell number within each well for each sample (± SEM). DAPI 4′,6-diamidino-2-phenylindole, FACS fluorescence-activated cell sorting, K14 cytokeratin 14, K19 cytokeratin 19, P1 passage 1, P2 passage 2