Skip to main content

System level modelling of metastasis from signalling to cell motility

Final Report Summary - METASTASIS SYS MOD (System level modelling of metastasis from signalling to cell motility)

Migrating cancer cells exhibit various motility modes, which differ in cell shape dynamics as well as cell-environment and cell-cell interactions. Cellular blebbing is involved in amoeboid motility, an important motility mode which is in the focus of this research. Blebs are large, unstable membrane structures that are driven by hydrostatic pressure and occur when the cell cortex is ruptured or disconnects from the plasma membrane. Blebbing occurs in different cell types during apoptosis, development and migration, and are related to specific signalling in the cell, although the exact mechanism is still barely understood.
This project focused on cancer cell motility from intercellular mechanics and signalling to cell-environment interactions. In the intercellular level, we studied bleb initiation and retraction mechanisms using a mathematical model and computer simulations as well advanced experimental data analysis. We used the model to compare different signalling variations in the cells and identify the putative role that each signalling molecule is playing in the cell shape dynamics.
Initially, bleb characteristics were measured in HEK293T cells, wild-type as well as with various mutations. We found that overexpression of Met increases blebbing rate, and dominant-negative Met inhibits blebbing significantly. Overexpression of Gab1 resulted in increased blebbing rate, while overexpression of Grb2 induced stable podia in addition to blebbing. These measurements were preformed manually, which raised the need for an automated method for bleb characterization.
In order to process a large amount of experimental data we have developed a novel, state-of-the-art algorithm for automated blebbing analysis. This new hybrid algorithm combines image analysis with physical considerations and can be applied to a variety of microscopy images, from bacterial colonies to intercellular compartments. This tool, named ABC (Automated Bleb Characterization) will be publically available as soon as the corresponding paper is published, and we expect it to become useful for a variety of cellular systems.
Current biological knowledge is insufficient to explain how the signalling changes result in the observed changes in the cell shape. To try and bridge over this gap, we have developed a mathematical model for cellular blebbing. The model includes a representation of the plasma membrane as well as the cellular cortex, and defines a set of forces on each of them and between them. The force on each of the membrane nodes is given by:
where P is the internal hydrostatic pressure (or the pressure difference), is the membrane arc element, is the force between the membrane and the cortex (R represents the connection), is the surface tension, is the membrane curvature and represents the constant volume constraint that leads to retraction of the rest of the cell.
The cortex and the membrane are assumed to be attached via a spring-like force, presumably by Myosin-tethering molecules such as Ezrin. A spontaneous rupture, or break down, of the attachment force between the membrane and the cortex, leads to a cytoplasmic burst by the hydrostatic pressure and to the onset of a bleb. Reconstruction of the connection between the membrane and the cortex leads to bleb retraction.
Using the simulation, we have tested two mechanisms for cortex-membrane reattachment: gradual build-up from the sides of the bleb, and spontaneous attachment at points along the bleb. Our model shows that while bleb retraction is obtained in both scenarios, only with central build-up membrane wrinkles are formed. Comparing this result to the experimental data showing substantial membrane wrinkling during bleb retraction leads to the conclusion that retraction probably occurs by membrane-cortex spontaneous attachment along the bleb. Our model therefore leads to an experimentally-testable prediction and points on the molecule that should be the focus of the experiment.
After establishing the correct mechanisms for bleb onset and retraction, we moved on and included signalling dynamics. Instead of randomly choosing a point for cortex-membrane separation, we included a set of stochastic FitzHu-Nagumo equations of an activator B and an inhibitor R that control the signalling. According to the experimental data we assume that the level of Met is governing the dynamics of bleb onset, and integrate it in the stochastic equations via a parameter that controls the system’s excitability. Another signalling component that was introduced is Grb2, which was found to induce stable podia. We hypothesize that this phenomenon results from inhibition of the ERM (Ezrin-Radixin-Moesin) complex, and introduce another parameter into the stochastic signalling equations that represents the value of Grb2. When Grb2 is high, the system shifts to a different fixed point, so that the inhibitor remains high enough to prevent bleb retraction. Importantly, the simulation results were analysed using the same automatic tool that was used for the experiments, and the model parameters were calibrated and tuned to produce statistically similar distributions.
In the intra-cellular level we studied cell-cell and cell-environment interactions in a group of migrating cells with either mesenchymal or amoeboid attributes. We showed that effective cooperation emerges even without direct interaction, as mesenchymal cells may enhance the invasion of amoeboids. We therefore predict that cell-cell interaction should be regulated as a stress-response mechanism, which improves cellular migration under strong metabolic or mechanical stress. Our results are in good agreement with experimental data of migrating cancer cells.
A somewhat different scenario of group motility was experimentally studied by wound healing assays. We used DA3 (murine breast cancer) cells with and without HGF, and developed a new algorithm (MultiCellSeg) to separate between multi-cellular and background regions for bright field images, which is based on classification of local patches within an image: a cascade of Support Vector Machines (SVMs) is applied using basic image features. Our algorithm was shown to provide better segmentation compared to other available software, and was used to study the wound closure rate with and without supplement of HGF. More experiments with different metabolites (e.g. glucose and glutamate) and specific markers for mitochondrial activity have been performed and are now being analysed.
The roles of metabolism and cell proliferation were studied using a mathematical model and simulations of different cell phenotypes with metabolic-like energy dynamics. Using experimental data-based modelling, we focused on the fundamentals of metastatic invasion: motility, invasion, proliferation and metabolism, and study how they may be combined to maximize the cancer’s ability to metastasize. The cells’ performance is measured by the number of cells that succeed in migration in a maze, which mimics the extracellular environment. We show that co-existence of different cell clones in the tumour, as found in experiments, optimizes the metastatic ability in a frequently-changing environment. We studied the role of metabolism and stimulation by a growth factor, and showed that metabolism plays a crucial role in the metastatic process and should therefore be targeted for successful treatment.