Quantification of topological features in cell meshes to explore E-cadherin dysfunction

In cancer, defective E-cadherin leads to cell detachment, migration and metastization. Further, alterations mediated by E-cadherin dysfunction affect cell topology and tissue organization. Herein, we propose a novel quantitative approach, based on microscopy images, to analyse abnormal cellular distribution patterns. We generated undirected graphs composed by sets of triangles which accurately reproduce cell positioning and structural organization within each image. Network analysis was developed by exploring triangle geometric features, namely area, edges length and formed angles, as well as their variance, when compared with the respective equilateral triangles. We generated synthetic networks, mimicking the diversity of cell-cell interaction patterns, and evaluated the applicability of the selected metrics to study topological features. Cells expressing wild-type E-cadherin and cancer-related mutants were used to validate our strategy. Specifically, A634V, R749W and P799R cancer-causing mutants present more disorganized spatial distribution when compared with wild-type cells. Moreover, P799R exhibited higher length and angle distortions and abnormal cytoskeletal organization, suggesting the formation of very dynamic and plastic cellular interactions. Hence, topological analysis of cell network diagrams is an effective tool to quantify changes in cell-cell interactions and, importantly, it can be applied to a myriad of processes, namely tissue morphogenesis and cancer.