Mechanical imbalance between normal and cancer cells drives epithelial defense against cancer.

Cell competition enables normal wildtype cells of epithelial tissue to eliminate mutant cells expressing activated oncoproteins such as HRasV12. However, the driving force behind this fundamental epithelial defense against cancer remains enigmatic. Here, we employ tissue stress microscopy and theoretical modeling and invent a new collective compressibility measurement technique called gel compression microscopy to unveil the mechanism governing cell competition. Stress microscopy reveals unique compressive stress experienced by the mutant cells, contrasting with predominantly tensile stress experienced by normal cells. A cell-based computer simulation then predicts that this compressive stress arises out of a mechanical imbalance between two competing populations due to a difference in their collective compressibility and rigidity. Gel compression microscopy empirically confirms the prediction and elucidates a three-fold higher compressibility of the mutant population than the normal population. Mechanistically, this difference stems from the reduced abundance and coupling of junctional E-cadherin molecules in the mutant cells, which weakens cell-cell adhesions and renders the mutant population more compressible. Taken together, our study elucidates both the physical principle and the underlying molecular mechanism driving cell competition in epithelial defense against cancer and opens new directions for mechanomedicine in cancer.

Splicing of branchpoint-distant exons is promoted by Cactin, Tls1 and the ubiquitin-fold-activated Sde2.

Intron diversity facilitates regulated gene expression and alternative splicing. Spliceosomes excise introns after recognizing their splicing signals: the 5'-splice site (5'ss), branchpoint (BP) and 3'-splice site (3'ss). The latter two signals are recognized by U2 small nuclear ribonucleoprotein (snRNP) and its accessory factors (U2AFs), but longer spacings between them result in weaker splicing. Here, we show that excision of introns with a BP-distant 3'ss (e.g. rap1 intron 2) requires the ubiquitin-fold-activated splicing regulator Sde2 in Schizosaccharomyces pombe. By monitoring splicing-specific ura4 reporters in a collection of S. pombe mutants, Cay1 and Tls1 were identified as additional regulators of this process. The role of Sde2, Cay1 and Tls1 was further confirmed by increasing BP-3'ss spacings in a canonical tho5 intron. We also examined BP-distant exons spliced independently of these factors and observed that RNA secondary structures possibly bridged the gap between the two signals. These proteins may guide the 3'ss towards the spliceosome's catalytic centre by folding the RNA between the BP and 3'ss. Orthologues of Sde2, Cay1 and Tls1, although missing in the intron-poor Saccharomyces cerevisiae, are present in intron-rich eukaryotes, including humans. This type of intron-specific pre-mRNA splicing appears to have evolved for regulated gene expression and alternative splicing of key heterochromatin factors.

Matrix mechanics regulates epithelial defence against cancer by tuning dynamic localization of filamin.

In epithelia, normal cells recognize and extrude out newly emerged transformed cells by competition. This process is the most fundamental epithelial defence against cancer, whose occasional failure promotes oncogenesis. However, little is known about what factors determine the success or failure of this defence. Here we report that mechanical stiffening of extracellular matrix attenuates the epithelial defence against HRasV12-transformed cells. Using photoconversion labelling, protein tracking, and loss-of-function mutations, we attribute this attenuation to stiffening-induced perinuclear sequestration of a cytoskeletal protein, filamin. On soft matrix mimicking healthy epithelium, filamin exists as a dynamically single population, which moves to the normal cell-transformed cell interface to initiate the extrusion of transformed cells. However, on stiff matrix mimicking fibrotic epithelium, filamin redistributes into two dynamically distinct populations, including a new perinuclear pool that cannot move to the cell-cell interface. A matrix stiffness-dependent differential between filamin-Cdc42 and filamin-perinuclear cytoskeleton interaction controls this distinctive filamin localization and hence, determines the success or failure of epithelial defence on soft versus stiff matrix. Together, our study reveals how pathological matrix stiffening leads to a failed epithelial defence at the initial stage of oncogenesis.