Viscoelastic relaxation of collagen networks provides a self-generated directional cue during collective migration

Andrew G. Clark, Ananyo Maitra, Cécile Jacques, Anthony Simon, Carlos Pérez-González, Xavier Trepat, Raphaël Voituriez, Danijela Matic Vignjevic

Preprint posted on July 12, 2020

Beyond chemotaxis: understanding collagen networks as highways allowing cell migration

Selected by Mariana De Niz


Collective cell migration is an essential process for various physiological phenomena, including development, tissue homeostasis, and cancer metastasis. Besides conventionally studied migratory cues such as chemotaxis, there is evidence that the physical properties of the cellular environment, such as collagen fiber thickness and alignment, can act as directional cues during cell migration (1-3). Previous work has shown that large cell aggregates embedded in collagen networks in vitro physically pull on collagen fibers, leading to a re-organization that facilitates invasion of cells into the surrounding matrix. Yet, it is not well understood how active physical remodeling of the network by cells affects their migration dynamics. In their work, Clark et al used long-term imaging, traction force microscopy, and theoretical modeling to investigate how the viscoelastic properties of collagen networks affect network organization during collective cell migration and how these local changes in network topology affect cell migration dynamics (4).


Figure 1. Theoretical model of persistent migration on a viscoelastic substrate (from Ref 4).

Key findings and developments

The authors first investigated how collective cell migration dynamics differed within deformable collagen networks and soft elastic substrates. For this purpose, they generated clusters of A431 cells. Clusters appeared to migrate faster, more persistently, along straighter tracks, and would explore a larger area when migrating on collagen networks, as opposed to collagen-coated elastic gels.

Front-back polarization is a common mechanism by which cells migrate persistently. Therefore, next the authors investigated whether persistent collective migration on collagen networks could be the result of front-back polarity mechanisms at a cluster scale. For this purpose, they began by staining cells for Rac1 – a marker associated with the protrusive leading edge during cell migration. Their findings suggest that cell clusters behave as large super cells. The authors then generated reporter cells expressing myosin-2 light chain fused to GFP, and imaged clusters by live imaging. Their findings suggest that clusters do not display typical front-back polarity mechanisms during collective migration on collagen networks.

Next, the authors aimed to determine whether cell clusters display asymmetric traction forces during migration. They found that cell clusters generate radial inward-facing tractions around the perimeter of the cluster, and that the peak in traction magnitude was highest at the cluster rear. The traction peak at the rear of the cluster was 10% higher compared to the leading edge of the cluster. This altogether suggests that cell clusters induce asymmetric traction force profiles on collagen networks during migration.

The next step was to investigate whether asymmetric traction would correlate with asymmetric deformation of the collagen network during cell migration.  Using live cell imaging, the authors observed that cell clusters reorganized collagen networks during migration, generating radial arrays of collagen fibers around the cluster. Collagen density seemed asymmetrically patterned, with a region of high collagen density observed between the center of mass and the trailing edge of the cell cluster. Further analysis suggested that cells tend to move away from regions of high collagen density. Altogether, the authors suggest that observing parameters such as cell position and underlying collagen density enable predicting a most probable direction of cluster migration, and that clusters generate inverse patterns of collagen density and nematic order during cluster migration, with a peak that is offset toward the trailing edge of the cluster.

The authors then investigated viscoelastic behaviour in collagen networks in response to traction forces generated by cell clusters. Using particle image velocimetry (PIV) on collagen images, and 3D displacement microscopy, they suggest that collagen networks behave in a viscoelastic manner in response to stresses generated by cell clusters during migration. They found that the relaxation time of the displacements was significantly smaller than the relaxation time of collagen density or collagen movements measured by PIV.

The authors then developed a theoretical model to identify the main requirements for persistent polarized migration of a cell cluster without any intrinsic polarity or shape asymmetry. In the model, persistent migration arises from the interaction of the moving cluster with the viscoelastic substrate. The activity of the cluster induces a perturbation in the substrate, analogous to traction forces causing local changes in filament density/orientation. The model presents a picture whereby motion results from a spontaneous symmetry-breaking mechanism similar to a previous model proposed for autophoretic colloids. Moreover, the model allows two additional predictions: that migration persistence decreases for substrates with lower relaxation times and that migration persistence is lower for small clusters.

The authors then tested whether reducing substrate relaxation time leads to reduced persistence, by comparing control collagen networks, with collagen networks treated with threose, which crosslinks collagen networks. Crosslinking with threose led to an increase in network stiffness and faster relaxation. Altogether, findings on crosslinked gels support the idea that reducing substrate relaxation time leads to symmetric substrate deformations, and that clusters migrate significantly less persistently than in control collagen.

Based on predictions from the model, the authors compared cluster migration to single cell migration to test the stress both setups generate on collagen networks. They found that single cells exert less total displacements and less stress on the substrate. Moreover, live imaging suggested that single cells were too small to sense collagen gradients during migration. They migrated with lower speed and lower persistence compared to clusters.


What I like about this preprint

I like the interdisciplinary nature of the work, combining biophysics and known biological principles. I liked also the approach the authors took towards the questions. It’s a very interesting topic with great relevance for various phenomena involving migrating cells. It is also a very understandable manuscript with very clear messages.



  1. Riching, K. M. et al. 3D collagen alignment limits protrusions to enhance breast cancer cell persistence. J. 107, 2546–2558 (2014).
  2. Fraley, S. I. et al. Three-dimensional matrix fiber alignment modulates cell migration and MT1-MMP utility by spatially and temporally directing protrusions. Rep. 5, 14580 (2015).
  3. Sapudom, J. et al. The phenotype of cancer cell invasion controlled by fibril diameter and pore size of 3D collagen networks. Biomaterials 52, 367–375 (2015).
  4. Clark AG, Maitra A, et al. Viscoelastic relaxation of collagen networks provides a self-generated directional cue during collective migration. bioRxiv (2020).


Posted on: 26th August 2020


Read preprint (No Ratings Yet)

Author's response

Clark Andrew and Ananyo Maitra shared

Open questions 

1.This is a very interesting work. My first question is what is the minimal cluster size capable of inducing the changes in the collagen network necessary for migration?

Thank you. That is indeed a great question and something we are currently looking into. It is not clear whether there is truly a threshold size for producing collagen gradients, but we do see a correlation between cell/cluster size and gradient strength (which we can approximate by taking the ratio between the maximum and minimum collagen densities across the cell). From our initial analysis, it seems that the gradient strength seems to be very sensitive to changes in cell/cluster size near the size of a single cell. For larger clusters, the gradient strength is less sensitive. This could hint at a sharp transition between formation of a gradient or no gradient that looks similar to a size threshold near the size of a single cell.

We find a very similar scaling when we analyze how cluster size scales with migration dynamics. Migration persistence is very sensitive to changes in size for smaller clusters and single cells, but not for larger clusters. That means that the transition from non-persistent to persistent migration occurs for very small clusters, just above the size of a single cell. This corresponds with the threshold-like behavior we observe for the formation of collagen gradients.

The model does not tell us exactly how cluster size should scale with persistence, though it does allow us to predict a size dependence. Being able to measure this scaling experimentally, it will be informative to incorporate this into the model in the future.


2.You discuss your findings in the context of tumour migration. Do you expect this might differ from collective migration in other contexts, such as infectious pathogens migrating collectively within tissues inside a host, or immune cells migrating in response to infection?

I think our study deals with a particular form of collective migration, where the cells migrate together as a single cohesive unit. This would be most similar to border cell migration in Drosophila and lateral line formation in Zebrafish. In these systems, I would also expect that viscoelastic relaxation of the local environment could also play a role in collective migration.

In the case of collective migration of pathogens, like bacterial swarming, or directed migration of immune cells, I could also imagine similar effects of local ECM remodeling. These systems are inherently more complex because individual cells can interact with each other, and perturbations of the local environment by one cell could affect its neighbors. Here I would draw an analogy with self-generated chemotactic gradients, in which groups of cells migrate directionally in the presence of uniform chemokine concentration (Tweedy et al, 2016 PLoS Biol.). There, a wave of cells migrates up a local chemokine gradient as the cells consume the chemokine (which further creates the gradient). I could imagine a similar mechanism of substrate deformation: a wave of migrating cells at the leading edge create a wave of high substrate perturbation just behind the leading edge, which further directs the cells to move forward. In the case of swarming bacteria, this could be a hydrodynamic effect in the surrounding medium, while in immune cells or perhaps epithelial cells, this could be a wave of high ECM density, similar to what we observe in our system. However, this may differ for cell migration under confinement or for adhesion-free migration modes, where migration can occur in the absence of interactions with ECM and without exerting high stresses on the substrate. It will be interesting in the future to better understand how viscoelastic properties of confined environments can influence migration.


3.You discuss chemotaxis as a separate phenomenon throughout your work. In a physiological situation, how would both phenomena combine (ie chemotaxis and collagen remodelling) to result in cell migration?

This is a really great question that our group and others are starting to look into. Since we have only looked at migration in random conditions, we do not yet know how substrate remodeling and chemotaxis function simultaneously. We could well imagine that the sort of dynamic ECM remodeling that we see could further enhance chemotaxis because it promotes persistent migration. We would also expect that ECM remodeling could permit chemotaxis with a “pulsed” chemokine gradient. If a gradient were formed transiently, that would be enough to break symmetry and start cell migration toward the gradient; if the chemokine gradient were then removed, we would expect the cells to continue migrating in the same direction. In this sense, the ECM remodeling mechanism could reduce the burden of making a chemokine gradient since just a pulse of chemokine, and not a sustained gradient, may be enough to stimulate cells to migrate in the right direction. For similar reasons, it may also make it more difficult for a group of cells to “turn” if the gradient changes direction, since the substrate deformation would push the cells to continue moving forward. We would expect this to result in a time lag or delay when changing directions.


4.Different tissues have a different composition in their parenchyma. In a hypothetical situation considering the same cell type (individual or cluster) migrating, do you expect there are organs where collagen network remodelling is more efficient/easier than others, and this influences cell behaviour (including metastasis)?

Another excellent question. Our results suggest that if collagen is too stiff or recovers too quickly, migration persistence is reduced. With less persistent migration, groups of cells tend to explore a smaller area since they spend their time migrating back and forth in small circles. From this, it would follow that tissues with more deformable stromal networks would be more permissive for migration and potentially metastasis. However, high stiffness in the pre-metastatic niche could also encourage cells to stop migrating and start growing, which would favor metastasis formation. The requirement for migration at a secondary site or the ability of metastases to grow depends on a number of factors and is likely very tissue specific. This gets back to the old “seed and soil” hypothesis, which states that in order for cancer cells (the “seeds”) to mestasize (or “grow”), they must be placed in the right microenvironment (the “soil”; Paget 1889 Lancet). It is becoming increasingly clear that the mechanics of the microenvironment must be considered alongside biochemistry and signaling, and I think our study offers some insight into the importance of substrate mechanics in cell migration and potentially metastasis.


5.You discuss in the model and experimentally, relaxation times. Considering a physiological situation of a combination of various cell clusters and individual cells migrating collectively, how would collagen network remodelling by one cluster influence migration of individual cells following this cluster, and within which time-scale?

I would expect this sort of mixed population to display different behaviors depending on the relative positions of the clusters and single cells. Although single cells cannot generate strong enough gradients to migrate persistently on their own, they are still able to respond to imposed gradients. Based on preliminary analysis using similar theoretical models of autophoretic motion, we would predict that single cells in front of large clusters (where front and back are defined with respect to the direction of motion of the cluster) would be persistently pushed forward by the cluster. In contrast, any single cell behind a large cluster would be kicked backward, as it would experience a collagen gradient in the opposite direction, and the cell would not be predicted to migrate persistently. These behaviors would all occur on the timescale of collagen relaxation, ~10 min. Numerical models and simulations are really handy for exploring these questions, as the behavior would be expected to be very different for different cellular arrangements, even though the basic interaction rules are the same.

Considering long-timescale interactions, we would expect to see a gradual stiffening of the collagen after many migration events. We observe in our cell removal assays that collagen density does not completely relax to levels of the background density, suggesting that some of the collagen reorganization is permanent. Over time, we would expect collagen density to accumulate and result in stiffening. This would mean that the first clusters to migrate across a region of collagen would experience a softer, more deformable network, while clusters that migrate later may experience a less deformable network and, as a result, migrate less persistently.

Have your say

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up to customise the site to your preferences and to receive alerts

Register here

Also in the cell biology category:

FENS 2020

A collection of preprints presented during the virtual meeting of the Federation of European Neuroscience Societies (FENS) in 2020


List by Ana Dorrego-Rivas

Planar Cell Polarity – PCP

This preList contains preprints about the latest findings on Planar Cell Polarity (PCP) in various model organisms at the molecular, cellular and tissue levels.


List by Ana Dorrego-Rivas

BioMalPar XVI: Biology and Pathology of the Malaria Parasite

[under construction] Preprints presented at the (fully virtual) EMBL BioMalPar XVI, 17-18 May 2020 #emblmalaria


List by Gautam Dey, Samantha Seah


Cell Polarity

Recent research from the field of cell polarity is summarized in this list of preprints. It comprises of studies focusing on various forms of cell polarity ranging from epithelial polarity, planar cell polarity to front-to-rear polarity.


List by Yamini Ravichandran

TAGC 2020

Preprints recently presented at the virtual Allied Genetics Conference, April 22-26, 2020. #TAGC20


List by Maiko Kitaoka, Madhuja Samaddar, Miguel V. Almeida, Sejal Davla, Jennifer Ann Black, Gautam Dey

3D Gastruloids

A curated list of preprints related to Gastruloids (in vitro models of early development obtained by 3D aggregation of embryonic cells). Preprint missing? Don't hesitate to let us know.


List by Paul Gerald L. Sanchez and Stefano Vianello

ECFG15 – Fungal biology

Preprints presented at 15th European Conference on Fungal Genetics 17-20 February 2020 Rome


List by Hiral Shah

ASCB EMBO Annual Meeting 2019

A collection of preprints presented at the 2019 ASCB EMBO Meeting in Washington, DC (December 7-11)


List by Madhuja Samaddar, Ramona Jühlen, Amanda Haage, Laura McCormick, Maiko Kitaoka

EMBL Seeing is Believing – Imaging the Molecular Processes of Life

Preprints discussed at the 2019 edition of Seeing is Believing, at EMBL Heidelberg from the 9th-12th October 2019


List by Gautam Dey


Preprints on autophagy and lysosomal degradation and its role in neurodegeneration and disease. Includes molecular mechanisms, upstream signalling and regulation as well as studies on pharmaceutical interventions to upregulate the process.


List by Sandra Malmgren Hill

Lung Disease and Regeneration

This preprint list compiles highlights from the field of lung biology.


List by Rob Hynds

Cellular metabolism

A curated list of preprints related to cellular metabolism at Biorxiv by Pablo Ranea Robles from the Prelights community. Special interest on lipid metabolism, peroxisomes and mitochondria.


List by Pablo Ranea Robles

BSCB/BSDB Annual Meeting 2019

Preprints presented at the BSCB/BSDB Annual Meeting 2019


List by Gautam Dey


This list of preprints is focused on work expanding our knowledge on mitochondria in any organism, tissue or cell type, from the normal biology to the pathology.


List by Sandra Franco Iborra

Biophysical Society Annual Meeting 2019

Few of the preprints that were discussed in the recent BPS annual meeting at Baltimore, USA


List by Joseph Jose Thottacherry

ASCB/EMBO Annual Meeting 2018

This list relates to preprints that were discussed at the recent ASCB conference.


List by Gautam Dey, Amanda Haage