Collective movement of vascular endothelial cells on microgrooved substrates viewed as the flow of an active fluid

Constraining the movement of vascular cells by culturing them on engineered microgrooved substrates generates a new pattern of collective motion that can be predicted using the physical framework of active fluids.
Collective movement of vascular endothelial cells on microgrooved substrates viewed as the flow of an active fluid

Movement is at the core of life, and collective movement, i.e. coordinated displacement of large groups of individuals, can be observed at very different scales. Nature provides many striking examples of collective movement including swarms of insects, schools of fish, flocks of birds or sheep herds. Physics, and in particular the field of fluid mechanics, has long taken interest in studying and predicting this collective behavior.

If we go down in scale, collective movement can also be found within our own bodies, beginning at the very moment we are conceived. Collective cellular migration is present at every step of an organism’s life, shaping the embryo and the organs, healing our wounds in response to injury, and fueling our diseases. Since the first observations of cell movement in the 17th century, careful investigations of this process by biologists have unraveled many of its key mechanisms. Historically, cellular processes have been principally viewed from a biochemical perspective; however, more recently, the interest of the physics community in biological systems has opened new avenues in our understanding of these processes, and in particular of collective cell migration. Indeed, by using tools and theoretical frameworks from their field, physicists have highlighted many similarities between cellular dynamics and the motion of active fluids. For instance, groups of cells moving collectively on 2D surfaces can exhibit swirling-like motions akin to turbulent fluid flow. Interestingly, constraining the adhesive area and therefore the shape of cell assemblies can organize the chaotic motions observed in 2D into particular patterns of movement such as global rotation on circular patterns or bidirectional movements on adhesive stripes.

When starting this study, we reasoned that global adhesive confinement of cell assemblies in 2D was not the most physiological setting. In our lab, we use microstructured topographic culture substrates which we believe do a better job. Indeed, cells in vivo usually reside on extracellular matrices that present a topographical organization. More specifically, we used a culture substrate composed of parallel arrays of micrometric grooves that mimic the anisotropy often found in native extracellular matrices and provide a constraint on each cell’s basal surface. Various cell types have been cultured on microgrooved substrates in the past, and most of them show cellular alignment and elongation in the direction of the grooves, a process called contact guidance. However, most of these studies focused on single cells, and collective migration on these structures remained largely unknown. We worked in this study with our favorite cells, vascular endothelial cells, a specialized epithelium that lines the inner surfaces of our blood vessels and is crucial for vascular homeostasis.

Left: artery without endothelial cells showing the topography of the surface on which the cell reside. Right: our microgrooved culture substrate
Left: artery without endothelial cells showing the topography of the surface on which the cell reside. 
Right: our microgrooved culture substrate

We started by imaging over a period of 24 hours the movement of confluent vascular endothelial cell monolayers cultured on microgrooved surfaces (grooves of identical width, depth, and spacing of 5 µm) coated with fibronectin. By extracting cell trajectories from these recordings, we observed that the direction of migration was constrained along the groove axis, coherent with the aligned and elongated cell shapes observed by immunostaining on fixed samples. More surprisingly, however, when we plotted the cumulative cell trajectories during the entire recording period, we observed a more striking pattern in the form of periodic and alternating corridors of cells moving from left to right or right to left (the groove axis being horizontal). We termed this pattern “antiparallel cell streams”. The characteristic size of these streams (hundreds of microns), which is much larger than either the groove or cell size, pointed to a collective nature of this pattern and the existence of long-range interaction among cells. In line with this idea, the cell streams were only visible when cells began to establish contacts, and inhibition of the junctions in between cells highly perturbed the emergence of this pattern. In addition, all cells appeared to be equal participants in the establishment of the streams since we observed no differences in morphology, contractility, or polarization among cells in the monolayer.

Left: recording of endothelial cells (nucleus stained in blue) movement on microgrooves (horizontal). Right: map of the cell speed along the groove axis showing the emergence of the streams (alternance of red and blue corridors)

We then wondered what can explain the emergence of the antiparallel stream pattern. We turned to physics for help and teamed up with a theoretician who proposed a simplified view in a model where the monolayer is considered as an active fluid and cells as nematic particles. By considering the effect of the microgrooves as a constraint on cell orientation and an initial state of immobile cells aligned with the grooves, the model was able to predict the emergence of an instability in cell orientation and velocity, giving rise to a sinusoidal velocity profile along the groove direction, i.e. the formation of the antiparallel cell streams observed experimentally. Interestingly, the model also predicted a linear relationship between the streams’ dimensions (width and length) and the mean cell orientation. This prediction was also verified in experiments in which changing the groove dimensions (and particularly increasing groove depth) led to more aligned cells as well as thinner and longer streams.

When considering the terms describing the effective free energy of the nematic particles representing the cells, an equation at the core of the model, we realized that the main physical element explaining the emergence of the streams was not the grooves themselves but rather the constraint they exert on cell orientation. Consequently, we would expect this pattern to appear in cell monolayers subjected to any kind of aligning external factor. To test this hypothesis, we subjected endothelial cell monolayers on flat surfaces (i.e. no grooves) to an apical flow which, similar to blood flow in vivo, also aligns and elongates the cells. Consistent with our hypothesis, a similar pattern of antiparallel cell streams was also visible in that situation.

When looking in the literature for similar patterns of movement in other systems, the few examples we were able to find always referred to situations where individual elements of a system (cells or reconstituted cytoskeleton, for instance) were also externally constrained. We therefore propose that the mechanism described in this study can be generalized and may constitute a common framework to explain the emergence of such patterns of collective behavior in very different systems. 

A pending question in this study, and more generally in the field of biological active matter, is the physiological relevance of the observed patterns. The cell streams described here could in theory be present in vivo (although difficult to verify experimentally), but many factors not taken into account in our system (blood flow, vessel curvature, etc.) may also influence the collective motion of vascular endothelial cells within our blood vessels. Other studies have identified links between active fluid-like collective cellular dynamics and physiological processes such as cell division or cell extrusion. In our system, although it remains to be demonstrated, collective motion in the endothelium is likely to influence vascular permeability.  From a more applications-oriented perspective, we believe that the results of this study can also prove useful in the field of implantable endovascular devices where surface topographic motifs constitute a potentially promising strategy for improving device efficacy.

To conclude, this study illustrates once again the potential of bringing together different disciplines to open new perspectives and enrich our vision of biological phenomena. To end on a more personal note, this project was for the biologist that I am a great adventure into foreign territories and a heartwarming confirmation that as different as our fields can be, we learn from one another to produce better science. It was also a particularly rich human experience: as I look proudly at the list of co-authors of this study, I see great collaborators in physics and mathematics but also friends and even family, reminding us that science is a story of human beings above all.