Rho activation drives luminal collapse and eversion in epithelial acini.

Epithelial cells lining a gland and cells grown in a soft extracellular matrix polarize with apical proteins exposed to the lumen and basal proteins in contact with the extracellular matrix. Alterations to polarity, including an apical-out polarity, occur in human cancers. Although some aberrant polarity states may result from altered protein trafficking, recent observations of an extraordinary tissue-level inside-out unfolding suggest an alternative pathway for altered polarity. Because mechanical alterations are common in human cancer, including an upregulation of RhoA-mediated actomyosin tension in acinar epithelia, we explored whether perturbing mechanical homeostasis could cause apical-out eversion. Acinar eversion was robustly induced by direct activation of RhoA in normal and tumor epithelial acini, or indirect activation of RhoA through blockage of ß1-integrins, disruption of the LINC complex, oncogenic Ras activation, or Rac1 inhibition. Furthermore, laser ablation of a portion of the untreated acinus was sufficient to induce eversion. Analyses of acini revealed high curvature and low phosphorylated myosin in the apical cell surfaces relative to the basal surfaces. A vertex-based mathematical model that balances tension at cell-cell interfaces revealed a fivefold greater basal cell surface tension relative to the apical cell surface tension. The model suggests that the difference in surface energy between the apical and basal surfaces is the driving force for acinar eversion. Our findings raise the possibility that a loss of mechanical homeostasis may cause apical-out polarity states in human cancers.

Reverse Plasticity Underlies Rapid Evolution by Clonal Selection within Populations of Fibroblasts Propagated on a Novel Soft Substrate.

Mechanical properties such as substrate stiffness are a ubiquitous feature of a cell's environment. Many types of animal cells exhibit canonical phenotypic plasticity when grown on substrates of differing stiffness, in vitro and in vivo. Whether such plasticity is a multivariate optimum due to hundreds of millions of years of animal evolution, or instead is a compromise between conflicting selective demands, is unknown. We addressed these questions by means of experimental evolution of populations of mouse fibroblasts propagated for approximately 90 cell generations on soft or stiff substrates. The ancestral cells grow twice as fast on stiff substrate as on soft substrate and exhibit the canonical phenotypic plasticity. Soft-selected lines derived from a genetically diverse ancestral population increased growth rate on soft substrate to the ancestral level on stiff substrate and evolved the same multivariate phenotype. The pattern of plasticity in the soft-selected lines was opposite of the ancestral pattern, suggesting that reverse plasticity underlies the observed rapid evolution. Conversely, growth rate and phenotypes did not change in selected lines derived from clonal cells. Overall, our results suggest that the changes were the result of genetic evolution and not phenotypic plasticity per se. Whole-transcriptome analysis revealed consistent differentiation between ancestral and soft-selected populations, and that both emergent phenotypes and gene expression tended to revert in the soft-selected lines. However, the selected populations appear to have achieved the same phenotypic outcome by means of at least two distinct transcriptional architectures related to mechanotransduction and proliferation.