Abstract
The identification of cancer cells with invasive and metastatic potential remains challenging. In recent years, it became evident that the organization of the cytoskeleton is dynamically orchestrated during cell transformation, but the impact of its remodelling is still largely unknown. In this study, we have developed a computational pipeline to characterize the cytoskeletal architecture of cancer cells and investigate fine-tuned cytoskeletal alterations. Our results have shown that the proposed computational framework was able to dissect unique cytoskeletal cues associated with invasive capacity. These include quantity, orientation, compactness, radiality, and morphology of microtubules. Validating our approach, we verified that microtubules of cells with disrupted E-cadherin and increased invasive rates are shorter, have disperse orientations and are more compactly distributed. Ultimately, this work provides a comprehensive portrait of the cytoskeleton reorganization that could be used as a proxy for automated analysis of cellular behaviour.