Tissue size controls patterns of cell proliferation and migration in freely-expanding epithelia

Matthew A. Heinrich, Julienne M. LaChance, Tom J. Zajdel, Ricard Alert, Andrej Košmrlj, Daniel J. Cohen

Posted on: 27 March 2020

Preprint posted on 28 February 2020

Article now published in eLife at http://dx.doi.org/10.7554/eLife.58945

Understanding collective cell behaviour at multiple scales.

Selected by Mariana De Niz

Categories: bioengineering, biophysics, cell biology, developmental biology, physiology

Background

Key advances in our understanding of collective cell migration and tissue growth have been made by studying the expansion of epithelial monolayers in vivo. Since 1859, the concept of the cell state was first used to describe tissues as a ‘society of cells, in a tiny well-ordered state’. This concept has continued into present work, which has led multiple labs to study how cell-cell interactions are important in complex biological phenomena such as locomotion, proliferation, and collective cell migration. These phenomena are pivotal in biological processes as varied as wound healing, and tumour invasion. In the past decades, research on cell-cell interactions and cell dynamics has continuously evolved to become an interdisciplinary field joining perspectives from biology, engineering, and physics to understand how local cellular interactions can give rise to globally coordinated motions. While interdisciplinary research has been a huge asset, until now most studies have focused on sub-millimetric scales. Therefore, cell proliferation and migration at larger scales (reminiscent of those in living organisms) remains poorly understood. To fill this gap, in their work, Heinrich et al study various cellular phenomena occurring in millimetre-scale, freely expanding monolayers to understand how tissue size and shape affect long-term epithelial growth.

Key findings and developments

Overall findings

Different initial sizes exhibit different spatiotemporal patterns of cell proliferation and collective cell migration in their internal regions.
Within several cell cycles, the core of large tissues becomes almost quiescent, and ceases cell-cycle progression. Conversely, the core of smaller tissues develops a local minimum of cell density as well as a tissue-spanning vortex.
The edge zone of both large and small tissues displays rapid cell-cycle progression and radially-oriented migration with a steady velocity independent of tissue size.
The overall conclusion is that cell proliferation and migration are regulated in a collective manner that decouples the internal and edge regions of the tissue, leading to size- and history-dependent internal patterns in expanding epithelia.
These data are the first comprehensive study of macro-scale, long-term epithelial expansion.
Findings summarised in Figure 1 (adapted from Heirich et al 2020 (1)).

Figure 1. Top: Footprint of cell expansion dynamics over 46h in small and large circular tissues. Middle: Vortex formation in expanding tissues- 10h traces of cell trajectories. Bottom: Coordinated spatiotemporal cell cycle dynamics observed by FUCCI marker showing G1 (magenta) and G2 (green) phases (Adapted from ref 1).

Specific developments

The team generated arrays of precisely patterned, unconfined epithelia at the macro-scale, and performed long-term time-lapse imaging.

They then investigated the role of tissue size and shape on boundary motion and tracked every cell in the tissues to relate the overall expansion kinetics to cell migration speed, cell density, and cell-cycle dynamics.

Specific findings

Effects of size and shape on expansion of millimetre-scale epithelia: phenomena at the tissue edge.

To address how tissue size influences long-term growth, the authors measured the expansion of confluent circular tissues with the same cell density but different initial diameters. MDCK epithelial cells, expressing a 2-colour FUCCI cell cycle marker were cultured in circular silicone stencils of controlled size for 18h, after which the stencils were removed, and the tissues allowed to freely expand for another 46h.
Small tissues experienced a transient density decrease while large tissues exhibited a monotonic increase in cell density.
The average radial velocity of the tissue boundary was independent of tissue size and cell density. The authors hypothesized that the overshoot was due to the formation of fast multicellular finger-like protrusions that emerge at the tissue edge early during expansion, and later diminish.
The authors then analysed tissue expansion in various tissue shapes with the same area and tissue density, but different perimeter. They showed that increasing the perimeter-to-area ratio of a tissue by increasing its aspect ratio, increases the areal expansion rate.

Spatiotemporal dynamics of migration speed and radial velocity: phenomena at the tissue bulk

The authors investigated tissue areal expansion rate relative to internal collective cell migration dynamics. They used Particle-Image-Velocimetry (PIV) to obtain flow fields describing cell migration within freely expanding epithelia, and constructed kymographs to describe the spatiotemporal flow patterns of the tissue.
The study revealed differences in small and large tissues: a motile interior region was observed, which was separated from the outer region by a transition zone of lower speed. In small tissues, this region starts to become quiescent almost twice the time after this occurs in large tissues.
In contrast to overall speed, radial velocity profiles for both large and small tissues match almost identically.

Emergence of large-scale vortices

The propagation of low radial velocity out from the centre of small tissues coincided with the formation and expansion of a vortex. While smaller and more transient vortices are observed in large tissues, small tissues exhibit tissue-spanning, fast, and persistent vortices.
In addition to vortices near the tissue centroids, the study showed smaller off-centre vortices and co-rotating vortex pairs. Altogether, cell tracking showed that in small tissues, cell trajectories are mainly radial in the boundary zone, but mainly tangential in the central zone.
Quantification of the radial and tangential displacements revealed a spiralling vortical flow that combines tangential shear with radial expansion.
Vortices in small tissues had higher intensity than those in large tissues, and were also larger relative to tissue size. However, vortices became maximally strong several hours later in small tissues than they did in large tissues. The authors hypothesized that this difference is due to large tissues featuring a faster density increase than small tissues.

Spatiotemporal dynamics of cell density

The authors further explored the spatiotemporal evolution of cell density, and showed that the vortex region in the centre of small tissues is accompanied by an unexpected local density minimum. However, in large tissues, vortices are often off-centred, and the low-density region does not appear where expected in the kymograph.
The density evolution of the center and boundary zones was characterized across tissues with different starting densities and sizes. The cell density at the centre of large tissues increased throughout the experiment. In contrast the cell density at the centre of small tissues initially increased, then underwent a transient density decrease, and finally increased again.
The authors aimed to model cell expansion and cell density. The model showed a discrepancy in cell density prediction, and did not quantitatively reproduce the cell density profiles at the edge regions. This suggests that more complex cell proliferation behaviour occurs, and further understanding is required to understand the full complexity of cell proliferation.

Spatiotemporal dynamics of cell cycle

In cells stably expressing the FUCCI markers, cells in the G0-G1-S phase of the cell cycle fluoresce in red, while cells in the S-G2-M phase of the cell cycle fluoresce in green. The outer zone of both large and small tissues was primarily populated by G2 cells and post-mitotic cells (which do not fluoresce). Conversely, the inner zone of large tissues was almost entirely comprised of G1 cells while in small tissues it was a mixture of cell-cycle states.
The authors showed that the contact-inhibited state that dominates the centre of large tissues does not originate in, or propagate out from, the dense core of the tissue. Instead, a front of contact inhibition of proliferation (CIP) seems to initiate between the central and edge zones and then propagates inwards.
In both small and large tissues, the boundary region was populated by rapidly-cycling cells so that after about 12h of tissue expansion, the boundary zone predominantly contains cells that have recently divided or are likely to divide soon.
In contrast, the center zone of large tissues undergoes cell-cycle arrest at about 30h.
Correlations between local measurements of cell cycle, cell speed, and cell density suggested that in large tissues, the cell-cycle state transitions from G1-dominated to G2-dominated when cell density increases above a certain threshold, and cell speed falls. Small tissues, by contrast, lack the GI-dominated, high-density and slow cell population found in the center of large tissues.

What I like about this preprint

I enjoyed this preprint very much. Particularly in the meticulous exploration of cell dynamics in multiple biophysical aspects. I like interdisciplinary science, and this is a good example of it. Further, the authors ask multiple relevant questions and explore them at scales that begin to be relevant for biological processes, beyond the conventional in vitro system which, as they show, might be governed by different dynamics.

Open questions

My first question is how consistent do you think this is among different cell populations? Furthermore, the usual cell lines used in many labs in vitro are tumour-derived cell lines. Does this affect their behaviour in vitro and at different scales?
Your work is very exciting for the potential it has to study various biological phenomena. In your work, you mentioned wound healing, or tumour invasion. With organoids and organs-on-chips becoming more available across scientific labs, how do you envisage that your findings will impact how we study tissue dynamics in vitro? Perhaps so far, the assumption for organoids and organs-on-chip, has been dictated from small scale tissues, not resembling scales of the organs they aim to mimic?
Also from a biophysical point of view, across labs, specific material is used for cell culture including plastics, glass, etc. Each of these influences how cells behave, including their proliferation and expansion. What do you think are suitable surfaces to reproduce and study tissue expansion, in a way that is relevant to in vivo aspects of cell biology?
You mention in your discussion that the vortices you found were a striking an unexpected example of internal dynamics. I found so as well while reading your work. Can you expand further on the significance of this from both a biophysics point of view, and what its relevance could be for a biological process?
You started in your introduction describing a society of cells. While you focused here on very key biophysical aspects, from a cell biology point of view, your work is also extremely interesting. Do you envisage studying the cell biology of the different cells of the tissue, and cell-cell communication? For instance, nutrient acquisition or gas exchange differing in the centre versus the edges? or actin dynamics, or organelle organization and function that might give us a more holistic view of how tissues control expansion?

References

Matthew A. Heinrich, Julienne M. LaChance, Tom J. Zajdel, Ricard Alert, Andrej Košmrlj, Daniel J. Cohen, Tissue size controls patterns of cell proliferation and migration in freely-expanding epithelia, bioRxiv, 2020, doi:10.1101/2020.02.28.970418

Acknowledgement

Thank you to Daniel Cohen and Matthew Heinrich for their engaging input.

doi: https://doi.org/10.1242/prelights.17676

Read preprint

(No Ratings Yet)

Author's response

Daniel Cohen and Matthew Heinrich shared

Open questions

1.My first question is how consistent do you think this is among different cell populations? Furthermore, the usual cell lines used in many labs in vitro are tumour-derived cell lines. Does this affect their behaviour in vitro and at different scales?

We expect our results to be representative across non-cancerous epithelial cell lines, specifically the results pointing to the decoupling of edge and centre zones and the relationship of cell cycle progression with cell speed and density. Non-epithelial cell lines that are less cohesive and therefore less collective in their motion may exhibit similar behaviours over smaller length scales or different behaviours altogether. However, we would expect even tumour-derived cell lines to exhibit some of the same principles that we observed, with the differences perhaps shedding more light on how cancer affects tissue ‘social dynamics’.

2.Your work is very exciting for the potential it has to study various biological phenomena. In your work, you mentioned wound healing, or tumour invasion. With organoids and organs-on-chips becoming more available across scientific labs, how do you envisage that your findings will impact how we study tissue dynamics in vitro? Perhaps so far, the assumption for organoids and organs-on-chip, has been dictated from small scale tissues, not resembling scales of the organs they aim to mimic?

We certainly hope that our study can inspire researchers to push the limits of the length scales of their experiments! Very small-scale studies of tens or hundreds of cells can be invaluable for connecting the onset of multicellularity to single-cell behaviour, but the field of collective dynamics in general is showing that distinct behaviours that we would not expect, like our vortex, can emerge in large populations. To perform an experiment at larger length scales with more physiological relevance, one may be forced to give up single-cell-resolution or the ability to image the entire tissue, but advances in automated microscopy and image processing can help limit these consequences. For organoids and organs-on-a-chip, we would love to see experiments performed across all relevant length scales (and bridging 2D-3D!) eventually up to actual organ size.

3.Also from a biophysical point of view, across labs, specific material is used for cell culture including plastics, glass, etc. Each of these influences how cells behave, including their proliferation and expansion. What do you think are suitable surfaces to reproduce and study tissue expansion, in a way that is relevant to in vivo aspects of cell biology?

Differences between plastic and glass substrates are probably not critical as long as the surface is coated with an ECM protein relevant to the cell type’s in vivo context; for example, we used Collagen-coated plastic substrates. More thought may be required as to whether an experiment on a 2D substrate is appropriate to draw conclusions about the 3D context that cells experience in vivo. In the case of epithelial tissues, which often live on a 2D manifold in a single layer or several stacked layers, we think that the 2D experiment is appropriate and quite instructive.

4.You mention in your discussion that the vortices you found were a striking and unexpected example of internal dynamics. I found so as well while reading your work. Can you expand further on the significance of this from both a biophysics point of view, and what its relevance could be for a biological process?

From a biophysical point of view, the presence of intense, large scale vortices point to potential length scales of collective behaviour that dwarf the expected limit of the velocity-velocity correlation length of the agents that participate in it; in particular, the vortices included thousands of cells despite the fact that the velocity-velocity correlation length for this cell type is usually only 5-10 cells! For biological processes, the long-time evolution of the vortices over several cell cycles emphasizes the ability for biological processes synchronize cell divisions with cell migration over longer timescales, and the decoupling of the centre vortex region from the boundary growth region demonstrates that a single, cohesive tissue may undergo multiple biological processes independently.

5. You started in your introduction describing a society of cells. While you focused here on very key biophysical aspects, from a cell biology point of view, your work is also extremely interesting. Do you envisage studying the cell biology of the different cells of the tissue, and cell-cell communication? For instance, nutrient acquisition or gas exchange differing in the centre versus the edges? or actin dynamics, or organelle organization and function that might give us a more holistic view of how tissues control expansion?

These are very exciting things to think about! We scratched the surface by tracking cell-cycle using the FUCCI marker, but other cell biological aspects of the different domains of the tissue surely promise to offer further insights into the regulation of tissue growth. We’re very interested in relating these large-scale dynamics to various aspects of homeostasis. The systems biology of these large-scale cellular ensembles seems quite exciting.