Coordination between cells requires efficient cell–cell communication, with mechanical signalling implicated in collective cancer invasion. In particular, traction force from one cell will impact the local ECM and thereby affect the force generation of nearby cells. To investigate this idea of coordinated force generation, US researchers developed a model based on the DIGME (diskoid in geometrically micropatterned ECM) tumour system. They used their model to examine how the cooperativity of cell contraction affects breast cancer cell migration in the DIGME system (Phys. Biol. 14 045005).

"Collective cell migration is crucial to many biophysical processes. And we still don't understand it," said Yang Jiao from Arizona State University and Bo Sun from Oregon State University. "The problem is that there are too many details – an ever-growing list of molecules all seeming to play important roles in multicellular dynamics. We were thus motivated to develop the phenomenological model, which allows us to quantify the degree of collectiveness during collective cancer cell migration without explicitly considering the detailed mechanisms."

Instead, the model characterizes cell–cell communication using a single parameter, a, which quantifies the correlation length of cellular migration cycles. Specifically, each tumour cell undergoes a series of periodic contraction cycles (for simplicity, all cycles possess the same periodicity) and is also characterized by a unique contraction phase. A high a value indicates synchronized contraction between cells, and zero indicates completely random cell contractions.

Jiao, Sun and colleagues used the model to compute the deformation fields resulting from collective cell contraction with different a values. They compared these with experimentally measured ECM deformation fields in the DIGME system, in order to identify the correlation length for such a system.

"Knowing the correlation length will guide us to investigate exactly what is the major communication pathway between the cells," explained Jiao and Sun. "We can also use the same approach to measure the correlation length in other types of tumours, or cells directly from patients. Cancer is a complicated disease, the more metrics we have to characterize the tumour, the more precisely we can apply treatments."

Experimental measurements

The researchers used DIGME to create circular diskoids of human breast cancer cells, labelled with green fluorescent protein, confined in a 3D collagen matrix. They embedded red fluorescent particles in the collagen to enable simultaneous monitoring of the cells and the matrix deformation. To probe the effects of mechanical forces generated by the cells, they imaged the diskoids continuously for three days at 5 min. intervals.

Three distinct phases were seen in the collective force generation. Immediate following seeding, the tumour diskoid pushed the matrix outward, due to initial spreading of the closely packed cells. This pushing phase was followed by a transition period of almost zero velocity, then the diskoid started to generate a contractive force that pulled the matrix radially inward.


Next, Jiao, Sun and colleagues used computational modelling to examine the extent to which the cells coordinately apply force to the extracellular matrix. To do this, they reconstructed the distributions of contraction phase in tumour cell aggregates with different phase a values. As a increased, regions containing cells with similar contraction phase values grew in size.

To investigate the ECM deformation due to collective cell contraction, the team mechanically coupled the cells to a continuum ECM model. They found that as a increased, the magnitude of displacement at a fixed distance from the tumour diskoid also increased, due to the larger pulling forces generated by more collective cell contraction.

The correlation length also affected the temporal evolution of the ECM strain profile. For the totally uncorrelated case (small a values), the large fluctuations in individual cell contraction phases smoothed out the temporal variation in the ECM deformation field. For large a values (synchronized cell contraction), a significant temporal variation of the overall deformation field was seen.

For intermediate correlations, as a increased, the temporal variation in the volume strain profile was more significant. The researchers used this intrinsic dependence of the time-dependent ECM strain profile on a to identify the correlation length of the tumour diskoids from the experimental DIGME data. Quantitative comparison of simulated and experimentally measured deformation fields revealed a correlation length of about 25 µm, corresponding to the size of about two to three cells. They note that one possible mechanism for this intermediate correlation length is fibre-mediated stress propagation in the ECM network in DIGME.

The researchers are now developing a model that considers the fibrous nature of ECM and mechanical properties of cells, as well as the mechanical coupling between cells and ECM. "We hope this model will guide us to engineer ECM in order to program collective behaviour of cells," said Jiao.

Related stories

• Research briefs: biophysics
• Cell stiffening could provide cancer clues
• Advances in modelling tumour growth
• Tumour metastasis: environment matters