No abstract

No abstract

Background: Cellular heterogeneity in tumor cells is a well-established phenomenon. Genetic and phenotypic cell-to-cell variability have been observed in numerous studies both within the same type of cancer cells and across different types of cancers. Another known fact for metastatic tumor cells is that they tend to be softer than their normal or non-metastatic counterparts. However, the heterogeneity of mechanical properties in tumor cells are not widely studied. Results: Here we analyzed single-cell optical stretcher data with machine learning algorithms on three different breast tumor cell lines and show that similar heterogeneity can also be seen in mechanical properties of cells both within and between breast tumor cell lines. We identified two clusters within MDA-MB-231 cells, with cells in one cluster being softer than in the other. In addition, we show that MDA-MB-231 cells and MDA-MB-436 cells which are both epithelial breast cancer cell lines with a mesenchymal-like phenotype derived from metastatic cancers are mechanically more different from each other than from non-malignant epithelial MCF-10A cells. Conclusion: Since stiffness of tumor cells can be an indicator of metastatic potential, this result suggests that metastatic abilities could vary within the same monoclonal tumor cell line.

Wide tumour excision is currently the standard approach to surgical treatment of solid cancers including carcinomas of the lower genital tract. This strategy is based on the premise that tumours exhibit isotropic growth potential. We reviewed and analysed local tumour spreading patterns in 518 patients with cancer of the uterine cervix who underwent surgical tumour resection. Based on data obtained from pathological examination of the surgical specimen, we applied computational modelling techniques to simulate local tumour spread in order to identify parameters influencing preferred infiltration patterns and used area-proportional Euler diagrams to detect and confirm ordered patterns of tumour spread. Some anatomical structures, e.g. tissues of the urinary bladder, were significantly more likely to be infiltrated than other structures, e.g. the ureter and the rectum. Computational models assuming isotropic growth could not explain these infiltration patterns. Introducing ontogenetic distance of a tissue relative to the uterine cervix as a parameter led to accurate predictions of the clinically observed infiltration likelihoods. The clinical data indicates that successive infiltration likelihoods of ontogenetically distant tissues are nearly perfect subsets of ontogenetically closer tissues. The prevailing assumption of isotropic tumour extension has significant shortcomings in the case of cervical cancer. Rather, cervical cancer spread seems to follow ontogenetically defined trajectories.

No abstract