Fig. 1: Developing a physiologically relevant organotypic model of ER+ breast cancer.

a MCF-7 and T47D breast tumour cells were cultured in 3D in collagen type I matrices ranging in concentration from 0.7–2.0 mg/mL for 7 days. H&E stained 4 µm sections of Formalin-fixed Paraffin-embedded (FFPE) organotypic cultures are shown. Each brightfield image is obtained from a single organotypic model at 40x magnification, scale bar: 20 µm. Arrows indicate morphologies of interest; either mass (blue) or mesenchymal (green). b Number of cell clusters in different concentrations of collagen matrix. Cell clusters counted in ImageJ from three images per treatment, across three experiments (n = 9). c Percentage positive staining of pan-cytokeratin (filled squares), α-SMA (filled triangles) and CD68 (filled squares) from images of multiplex IHC slides from full face sections of ER+ and TNBC breast cancer samples from the BeGIN study (Breast Cancer: Correlating Genetic, Immunological and Nutritional Predictors patient cohort; not published), colour coded according to patient ID. d Breast tumour cells (MCF-7 or T47D) co-cultured with myoepithelial cells (MECs) in a type I collagen matrix for 7 days. Tri-cultures contained monocyte-derived macrophages (MDM) or HFFF2 fibroblasts (FIBs), and quadra-cultures contained all four cell populations. (i) Arrows highlight morphologies of interest; mass clusters (blue and yellow) and tumour islands at the collagen edge (green). Images taken at 10x and 40x magnification under brightfield. Scale bars: 50 and 20 µm, respectively. (ii) Circumference of cell clusters in each culture condition. (iii) Number of cell clusters with 4 or more nuclei in each field of view. Quantification performed on three fields of view from n = 3 organotypic samples. Data represent mean ± SEM with c one-way ANOVA with Friedman test or d two-way ANOVA used to test for statistical significance; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and non-significant (ns) P > 0.05.