Unraveling the cytoskeletal architecture of cancer cells: development of a novel computational approach

Diogo João Fróis Vieira

Key information

Authors:

Diogo João Fróis Vieira (Diogo Fróis Vieira)

Supervisors:

João Miguel Raposo Sanches (João Miguel Raposo Sanches); Maria Sofia Fernandes

Published in

11/21/2022

Abstract

Cancer remains a major health problem and one of the leading causes of mortality worldwide. This is mostly due to the development of metastatic disease and the lack of powerful tools to detect invasive cancer cells at an early stage. Therefore, there is an urgent need to unravel new strategies to identify invasive cancer cells. It is well established that the structure and organization of the cytoskeleton are dynamically orchestrated during many cellular processes, including cancer invasion. Thus, in this study, a novel computational approach was developed to characterize the cytoskeletal architecture of cancer cells. Immunofluorescence images were used to devise a pipeline to characterize the structural pattern of α-tubulin, a cytoskeleton component, in cells with E-cadherin mutations leading to loss of cell-cell adhesion and a potential invasive phenotype. Microtubule organization features were evaluated, including morphology, orientation, compactness, and radiality. The strategy involved a preprocessing step for image enhancement followed by feature extraction (processing) of manually segmented microtubules based on three methods: analysis of grayscale pixel intensity distribution, 2D spatial rearrangement of automatically-detected line segments/microtubules and graph theory features. Results of the proposed method applied to cells with E-cadherin mutations have shown that the microtubules of cells with disrupted E-cadherin are shorter, have uniform length patterns and are more compactly distributed as compared with cells with wild type E-cadherin. This study reveals that cytoskeletal features could provide an efficient strategy to identify cells with invasive potential, ultimately impacting cancer diagnosis and prognosis.