Systems microscopy analysis of cell migration

Abstract

Single cell migration is heterogeneous and a complicated process. It arises from a hugely complex network of multi-scale interactions between molecular and macromolecular entities. Though the full, spatiotemporally resolved molecular complexity of the cell migration system is currently inaccessible, two macromolecular entities, namely cell-matrix adhesion complexes (CMACs) and F-actin, provide a means to abstract this complexity to a level that is tractable with imaging approaches, while also enhancing the significance of information captured from the cell migration system. Based on this rationale, we combined quantitative imaging and acquisition of multi-scale quantitative data, describing simultaneously both cell behavior (migration) and organization (e.g. CMAC and F-actin status) on a per cell basis, thereby leveraging a natural cellular (spatial and temporal) heterogeneity. This was then further combined with multivariate statistics and mathematical modelling, resulting in an approach referred to as systems microscopy. Subsequently, we employed this approach to interrogate several biological aspects.

In the first study, we used a systems microscopy approach, including use of the Granger causality concept, to map pairwise causal (directional) relationships between organizational and behavioral features of the cell migration system, advancing on the commonly used correlative (non-directional) relationships. This way, we were able to leveraging the natural cellular heterogeneity to better understand the cell migration system. We found that organizational features such as adhesion stability and adhesion F-actin content causally determined the cell migration speed. Contrary to previous findings, we observed that cell speed also acted upstream of organizational features, including cell shape and adhesion complex location. A comparison between unperturbed and modulated cells provided evidence that Granger causal interaction patterns are in fact plastic and context dependent rather than stable and generalizable.

In the second study, we employed a systems microscopy approach to separate the regulatory associations underlying either cell migration or its membrane dynamics. We introduced a new measure of relative membrane dynamics, corrected membrane dynamics (CMD), which is independent of cell speed. We found that F-actin features (e.g. F-actin concentrations at CMACs and F-actin concentrations per cell) were strongly associated with membrane dynamics while cell migration was more strongly correlated with adhesion-complex features (e.g. variance in CMAC age and CMAC shape). Moreover, these correlative linkages were often non-linear and context-dependent, changing dramatically with spontaneous heterogeneity in cellular behavior.

In the third study, cellular plasticity was studied, using the Nuclear-Golgi positioning as a model system addressing the coordination between cell migration and cellular asymmetry. We systematically analyzed these processes over a two-dimensional experimental array wherein intracellular tension and matrix ligand density were progressively co-varied. We found plastic responses of cellular behaviours, e.g. for the cell motion angle, cell polarity angle and the polarity and motion alignment. Moreover, polarity and motion alignment and cell motion angle dynamics displayed non-linear and non-monotonic relationships to cell speed and the correlative relationship between them were context-dependent. Some of these relationships were susceptible to decoupling with a reduction in tension or attachment strength. Moreover, we found that the forward polarity of the Golgi is an ordered cellular state, in contrast to backward polarity. More broadly, we found that in the majority of cases, motion and asymmetry were coordinated and that the different types of coordination coincide with specific cellular behaviors.

In the fourth study, we employed the systems microscopy approach to demonstrate the existence of two divergent modalities of mesenchymal cell migration, spontaneously emerging in parallel under a uniform environmental condition. The discontinuous migration acquires faster and less persistent migration and is characterized by a dramatic cell rearretraction events that are temporally decoupled from protrusion. Quantification of cell-matrix adhesion, F-actin and cell morphological features in each mode revealed that the cell speed within each mode is controlled by the unique assemblage of organizational features, suggesting the differential mechanism of regulating cell speed within each mode. We also demonstrated that the cell adaptive response is mediated by an adaptive switching rather than a progressive adaptive stretching, rendering adaptive switching as a dominant mechanism. We also provided evidence of important molecular regulators involved in adaptive switching, involving the sub-cellular systems of actomyosin contractility and cell-ECM interactions in this regulation.