Spatial heterogeneity of cell-matrix adhesive forces predicts human glioblastoma migration

Rasha Rezk, Bill Zong Jia, Astrid Wendler, Ivan Dimov, Colin Watts, Athina E Markaki, Kristian Franze, Alexandre J Kabla

Preprint posted on 8 May 2020

Article now published in Neuro-Oncology Advances at

Cell matrix adhesive forces and glioblastoma migration

Selected by Mariana De Niz

Categories: biophysics, cancer biology


Glioblastoma (GBM) is a highly aggressive and currently incurable brain tumor. The main cause of mortality in GBM patients is the invasive rim of cells migrating away from the main tumor mass, and invading healthy parts of the brain. This extensive and infiltrative growth pattern makes surgical resection difficult, and limits the efficacy of radiation therapy.

While it is known that cell migration and invasion are driven by mechanical forces, our current understanding of the physical factors involved in glioma infiltration remains limited. It is known that GBM display a mesenchymal mode of migration, using focal adhesion proteins as molecular clutches to transmit force to their environments. Although attractive, so far, therapeutics designed to target adhesion receptors or proteases have failed in clinical trials. It is thought that this failure might be due to the heterogeneity in expression of the adhesion proteins both, between patients, and between tumour regions in the same patient. In their work, Rezk and colleagues for the first time investigated adhesive and migratory properties of different GBM subpopulations from different patients, and different regions within the tumours.

Figure 1. Vinculin and F-actin localises at the edge of cells derived from strong fluorescent lines (tumour core) compared to cells derived from weak- and non-fluorescent lines (tumour rim or margin). (Figure reproduced from Figure 2, Ref1).

Key findings and developments

In this study, Rezk and colleagues investigated the adhesion properties within and between patients’ tumors on a cellular level and tested whether these properties correlate with cell migration. For this purpose, the authors used 5-aminolevulinic acid (5-ALA) fluorescence guided therapy in various patients to separate different tumour sections for further study. These sections included the tumour core (highest fluorescence intensity), the tumour rim (medium fluorescence intensity), and tumour margins (not fluorescent). The authors then adapted a microfluidic device to detach adherent cells through shear stress, and cultured them in PDMS with a low modulus consistent with the stiffness of the environment that GBM infiltrate.

Cell spreading area is known to be an indicator of how cells mechanically interact with their extracellular environment. The authors detected morphological differences in cells derived from different patients, and demonstrated that morphological heterogeneity within and between tumours was related to 5-ALA fluorescence intensity (i.e. tumour region). To investigate whether cell morphology relates to the way cells adhere to their environment, the authors went on to explore specific patterns of key cytoskeletal proteins. For this, they imaged the localization and structure of actin filaments and the actin-binding protein vinculin in cells from different tumour regions, and demonstrated that shape and distribution of actin and vinculin differed between cells derived from each region.

The authors propose that the organization of actin filaments and vinculin suggests spatial differences in focal adhesion assembly and enlargement between cells of different tumour regions. The authors went on to build a microfluidic device to quantitate cell matrix adhesion strength. Cells were injected into channels and subjected to a controlled flow rate to create a constant shear force on the cell. The output of this setup is quantification of the fraction of detached cells over time. For the purpose of the work, the authors took measurements of cell detachment after 5 minutes of flow, for quantifications of adhesion strength. They found that cells derived from areas corresponding to the tumour rim and margins had significantly lower cell matrix adhesion and smaller spreading areas, than cells derived from the tumour core.

They then quantified the migratory behaviour of cells derived from the different tumour regions. For this, the trajectory of the different cell populations on compliant PDMS substrate was recorded. This showed differences in migratory capacity between cells derived from different patients, and importantly, it showed that cells derived from the tumour core were significantly slower than cells derived from the tumour rim or margins.

Altogether, the authors conclude that cell-adhesion strength could be an accurate predictor of tumour cell migration, unlike molecular classification which have so far failed to provide accurate predictions. They further suggest that preclinical tests aimed at developing anti-invasive drugs or adhesion inhibitors against GBM, would be more accurate if using cell lines derived from the tumour periphery, rather than the tumour core.

What I like about this preprint

I like this preprint because it investigates a new aspect of GBM, and aims to determine the source of the heterogeneity that so far has hindered treatment, and the focus of clinical trials. I like inter-disciplinary research and I like approaches that are out-of-the-box to address scientific questions. I think this work beautifully does this. Also, I think there are pieces of work in multiple fields showing more and more the importance of biophysics and biomechanics. For me this is a new era of research, in which we no longer only focus on the molecular and OMICS, but have broadened our approach and questions.



  1. Rezk R, et al, Spatial heterogeneity of cell-matrix adhesive forces predicts human glioblastoma migration, bioRxiv, 2020
  2. Heinrich MA, et al, Tissue size controls patterns of cell proliferation and migration in freely expanding epithelia, bioRxiv, 2020.
  3. Plodinec, Marija, et al. The nanomechanical signature of breast cancer, Nature nanotechnology, 11, 2012.


Posted on: 17 June 2020


Read preprint (No Ratings Yet)

Author's response

Alexandre Kabla and Rasha Rezk shared

Open questions

1.I liked your work a lot. One of the first questions I have is how does your work, focusing on GBM, apply to other types of cancer? Does one observe the same phenotype in other invasive forms or does this differ vastly among tissues?

Tumours are known to be genetically heterogeneous, and there is evidence in other cancers that the material properties of the tissues may depend on location in the tumour. For instance, the stiffness of malignant tissue varied from core to periphery in human breast cancer biopsies (3). The methodology we used, consisting in sampling different regions and testing derived lines in standardised conditions, will help establish in other cancers how much of the phenotypic variability is due to the cell lines themselves, and how much is influenced by the cell’s environment itself.

2.Another concern in cancer in general, is the time of diagnostics. Is it possible, with your setup, to investigate when is it likely in GBM, to provide a timely diagnosis? Is there a time when the tumour does not yet have these 3 very specific regions that behave so differently in terms of migration capacity, adhesion, etc? Specifically, when does the tumour rim and tumour margins become a cause of concern?

 The short answer is that we don’t know at this point when and how the patterning of migratory and adhesive properties appears. It is difficult to monitor the evolution of a malignant glioma tumour from an early stage; sadly diagnosis tends to be late in the disease progression, most often grade IV by then. Our study investigated tissue derived from newly diagnosed GBM patients who underwent their first surgical resection. In vitro or animal models, assuming they can be developed, may however enable us to answer this question in future.


3.What defines the tumour core? In a preprint I recently covered by the Cohen lab, they studied cell motility and vortex formation in cultures of different sizes (2). How do you envision you could keep the entire tumour behaving as a tumour core? Is this mostly through avoidance of its growth?

Core and marginal lines are defined based on the location in the tumour of the cells from which lines are derived. The morphological and biophysical measurements are not made in-situ, but performed on cells derived, placed in similar conditions. This ensures that we are characterising the cell lines and not their microenvironments. The approach presented in Heinrich et al. is different since in their case the cells are all the same, but their physical location influences their emerging collective behaviour. In our case, what triggers the cellular heterogeneity remains unclear. This brings us back to your second question, and we would love to be able to answer it.

4.Why is metastasis beyond the central nervous system unusual in this disease, if the cells in the tumour rim and margin have large migratory capacity?

This is a great question. This may be partially due to the presence of the blood–brain barrier and overall low median survival. GBM is the only solid tumour defined as Grade IV without any evidence of metastasis, and we are interested in finding why GBM only seems to thrive in the brain microenvironment (recurrence is local).

5.It is known for instance in some infectious diseases, that immune cells unfortunately in an attempt to control infection, rather contribute to its spread. Is it known whether angiogenesis or immune-recognition play a role in tumour spread for GBM? If so, how? Is immunotheraphy against cells in the tumour rim and margins something that has been considered?

Brain tumours, especially GBM, show the highest level of vascular proliferation, compared to other solid tumours. Vessel co-option and angiogenesis both exist in GBM, increasing the complexity of heterogeneity of the tumour microenvironment. GBM are highly heterogeneous tumours that grow in a highly immunosuppressive microenvironment protected by the blood–brain barrier. So far, phase III immunotherapy clinical trials have not been successful. Identifying a treatment that is accessible to a large patient population will always be a challenge. In order to mitigate tumour immune escape, and overcome the intrinsic resistance imparted by GBM’s intra- and intertumoral  heterogeneity, we need to understand the genetic, functional and phenotypic heterogeneity, as well as the interactions between tumours and the tightly regulated immune system in the central nervous system.

Have your say

Your email address will not be published.

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