Probing the consequences of cell-to-cell variability during epithelial morphogenesis: a programmed assembly approach

The formation, maintenance, and function of epithelial tissues are guided by dynamic interaction between epithelial cells and their surrounding microenvironment. Three-dimensional (3D) epithelial cell culture models are valuable ex vivo tools for understanding the global effects of soluble factors, extracellular matrix components, and genetic perturbations on multicellular processes such as morphogenesis and tissue polarization. Probing the consequences of cell-to-cell variability in the activation of specific pathways remains challenging even in 3D culture models, however, because controlling the patterns of variability within each multicellular structure has been difficult under the culture conditions necessary to support morphogenesis. To better model cell-to-cell variability within heterogeneous epithelia, we are using a DNA-templated programmed assembly strategy to build aggregates of MCF10A human mammary epithelial cells with defined cell number, composition, and initial cell-cell connectivity. We do so by labeling cell surfaces with synthetic, single-stranded DNA and directing the assembly of complementarily labeled cell populations. When grown in 3D culture, assembled wild-type (WT) aggregates condense into polarized microtissues that undergo morphogenesis. MCF10A aggregates expressing low levels of oncogenic H-Ras (MCF10ARas) similarly polarize and undergo morphogenesis. Surprisingly, heterogeneous aggregates where single MCF10ARas cell are assembled with WT cells display emergent phenotypes such as cell extrusions and multicellular protrusions. These results add to accumulating evidence that emergent behaviors may arise from heterogeneity within local populations of epithelial cells and highlight the utility of our bottom-up method for constructing mosaic epithelial microtissues for studying such interactions in defined microenvironments. Understanding the mechanisms and conditions required for such emergent behaviors may inform upon similar behaviors observed in early stages of cancer development and progression