Heterotypic interfacial tension between oncogenic and wild-type populations forms the mechanical basis of tissue-specific oncogenesis in epithelia
- Department of Bioengineering, Indian Institute of Science, India
- Department of Physics, Indian Institute of Science, India
- Department of Physics, Syracuse University, United States
Figures
Figure 1 with 2 supplements
Tissue-specific outcomes of HRasV12 oncogenic mutants in epithelial monolayers.
(a) Schematic representation of the experimental setup and outcome. (b) Representative images showing extrusion of single oncogenic cells (singlets) in mammary epithelium (upper panel), but protrusive behavior of the same oncogenic singlets in bronchial epithelium (bottom panel). (c) Quantification of shape indices of HRasV12 singlets showing a reduction in shape indices, with increasing wild-type cell density in mammary epithelium (upper panel), but the opposite trend in bronchial epithelium (bottom panel), where HRasV12 cells continue to spread. Shaded regions indicate the results of simple linear regression analysis with 95% confidence intervals. (d) Extrusion rates of HRasV12 singlets are significantly higher in mammary epithelium, while (e) the percentage of HRasV12 cells forming protrusions is markedly higher in bronchial epithelium. Representative data are plotted from one of three independent experiments, with the median shown as a bold, dashed line and the first and third quartiles are shown as thin dashed lines. Statistical significance was calculated using an unpaired t-test with Welch’s correction. (f) Behavior of HRasV12 oncogenic clusters in the two tissues. In mammary epithelium (upper panel), clusters become spatially confined with a smooth, circular interface with the wild-type population, while in bronchial epithelium (bottom panel), HRasV12 clusters expand, forming long protrusions over time. (g) Immunofluorescence images of HRasV12 clusters in mammary and bronchial epithelia, highlighting differences in spreading patterns in the two epithelia. (h) Quantification of cluster segregation: mammary epithelium exhibits a higher segregation index and reduced eccentricity compared to bronchial epithelium, indicating greater spatial confinement of oncogenic clusters. The segregation index (SI), defined as the average ratio of homotypic and all cell neighbors, quantifies the degree of demixing and is averaged over all cells inside an oncogenic cluster. Data are mean ± sem and plotted from different clusters from one representative experiment (scale bars = 50 μm).
Figure 1—video 1
An HRasV12 singlet is extruded from MCF10A monolayer but persists in BEAS2B monolayer, multiplies and forms long protrusions.
Scale bar = 20 μm.
Figure 1—video 2
A group of HRasV12-transfected cells gradually gets confined into a compact cluster in MCF10A monolayer and forms a smooth interface with the wild-type population, while a group of HRasV12-transfected cells gradually grows in size in BEAS2B monolayer, forming a jagged interface with the wild-type population along with protrusions.
Scale bar = 50 μm.
Figure 2 with 3 supplements
Distinct mechanical signatures of HRasV12 clusters in mammary and bronchial epithelia.
(a) Representative images showing regions of interest (ROIs) selected for analysis in mammary (upper panel) and bronchial epithelia (lower panel). (b) Immunofluorescence images of HRasV12 clusters stained for F-actin in mammary (upper panel) and bronchial epithelia (lower panel). (c) Oncogenic cluster-wild-type interface showing distinct actin belt in mammary marked by white arrows (upper panel) and absence of it in bronchial epithelia (lower panel). (d) ROI-based F-actin-stained regions of wild-type and oncogenic cells in mammary (upper panel) and bronchial epithelia (lower panel), revealing distinct shape differences in different regions – in mammary epithelia, wild-type cells at the interface show elongated shapes (ROI2), and oncogenic cells inside the cluster show jammed shapes (ROI4), in comparison to wild-type cells at a random location (ROI1), while in bronchial epithelia, oncogenic cells (ROI3 and ROI4) show more elongated shapes compared to wild-type BEAS2B cells (e). (f) Quantifications of F-actin intensities and shape indices in mammary epithelia (e) and bronchial epithelia (f) across the four ROIs confirming jammed oncogenic clusters, with unjamming at the interface, in mammary epithelia and unjammed clusters in bronchial epithelia. (g) Schematic illustrating the Bayesian force inference pipeline used to estimate relative intracellular pressure and cell-cell edge tension. (h) Heatmaps depicting relative cell pressures and cell-cell edge tensions in mammary epithelium, revealing localized mechanical heterogeneity. (i) Heatmaps of relative cell pressures and edge tensions in bronchial epithelium, showing a distinct mechanical landscape compared to mammary epithelium. (j) Quantification of relative intracellular pressures and cell-cell edge tensions across the four ROIs in mammary (upper panel) and bronchial epithelia (lower panel), highlighting tissue-specific differences. All data represented as mean ± sem in box and whisker plots and are plotted from one out of three independent experiments with whiskers extending from the minimum to the maximum values, showing the full range of the dataset including outliers. Line represents the median. Statistical significance was calculated using an unpaired t-test with Welch’s correction. Scale bars = 20 μm.
Figure 2—figure supplement 1
Myosin belt and per-cell myosin intensities.
(a) Images of HRasV12 clusters in mammary epithelia stained for phosphomyosin (left panel) and phosphomyosin belt around the cluster marked with white arrowheads (right panel). (b) Plot showing lower phosphomyosin levels inside oncogenic clusters in MCF10A. Representative data is plotted from one experiment with the median shown as a bold line. Statistical significance was calculated using an unpaired t-test with Welch’s correction. Scale bars = 50 μm.
Figure 2—figure supplement 2
HRasV12 clusters show invasive phenotype in MDCK monolayer.
(a) Representative images of wild-type MDCK monolayers stained for F-actin (left) and E-cadherin (right), showing intact cell-cell junctions. (b) HRasV12 clusters expand, forming long protrusions over time. Scale bars = 50 μm.
Figure 2—figure supplement 3
Actomyosin enrichment at the interface of HRasV12 clusters and wild-type MCF10A.
(a) Images of HRasV12 clusters in mammary epithelia stained for F-actin, showing higher F-actin deposition at heterotypic interfaces. (b) Representative images of HRasV12 clusters in mammary epithelia stained for phosphomyosin, showing higher myosin deposition at the heterotypic interface. Scale bars = 10 μm.
Figure 3 with 5 supplements
Particle image velocimetry (PIV) reveals distinct cellular movement patterns in HRasV12 clusters in the two monolayers.
(a), (b) Representative snapshots of HRasV12 clusters, and the corresponding PIV maps, in mammary (a) and bronchial epithelium (b), as the density of wild-type cells increases, showing wild-type cells along the cluster-wild-type interface exhibit a higher velocity attributed to tangential motion in mammary epithelium (a), but undergo a more kinetically arrested state in bronchial epithelium (b). (c) Oncogenic cluster in MCF10A monolayer with concentric circles drawn from the center of the cluster, radially outward till its boundary (upper panel) with the red (highlighted) circle representing the one along which tangential velocity was the highest and plot showing tangential motion (expressed as a function of the circles), revealing highest tangential motion (data point highlighted with red) along the interface of oncogenic cluster with the wild-type cells in mammary epithelia (lower panel). No significant tangential motion in random wild-type region. (c) Representative snapshots of HRasV12 clusters in bronchial epithelium at the same two time points and the corresponding PIV velocity maps. (d) Quantification of cellular velocities in local regions as cell densities reach confluency showing oncogenic cells clustered in mammary epithelium (top panel) jam and show a significant reduction in velocities. In bronchial epithelium (bottom panel), while wild-type cells show reduced movement as cell density increased, transfected cells continued to show higher motion, signifying that jamming of wild-type cells has no effect on mutants. All data are represented as mean ± sem plotted from different clusters from one representative experiment. Scale bars = 50 μm.
Figure 3—figure supplement 1
Particle image velocimetry (PIV) maps of wild-type mammary and bronchial epithelial monolayers with increasing cell density.
(a) Representative time-lapse images of wild-type (control) mammary epithelium at two different time points, with corresponding PIV velocity maps showing a reduction in overall velocity as cell density increases. (b) Representative time-lapse images of wild-type (control) bronchial epithelium at two different time points, with corresponding PIV velocity maps demonstrating a similar reduction in velocity with increasing cell density. (c) Quantification of velocity reduction in wild-type mammary and bronchial epithelia as a function of cell density, highlighting tissue-specific differences in collective cell dynamics. All data are represented as mean ± sem plotted from different clusters from one representative experiment. Scale bars = 50 μm.
Figure 3—video 1
Particle image velocimetry (PIV) video during formation of a representative HRasV12 cluster in MCF10A monolayer.
Scale bar = 50 μm.
Figure 3—video 2
Particle image velocimetry (PIV) video showing the dynamics of a representative wild-type MCF10A (control) monolayer without any oncogenic mutants.
Scale bar = 50 μm.
Figure 3—video 3
Particle image velocimetry (PIV) video showing the dynamics of a representative HRasV12 cluster in BEAS2B monolayer.
Scale bar = 50 μm.
Figure 3—video 4
Particle image velocimetry (PIV) video showing the dynamics of a representative wild-type BEAS2B (control) monolayer without any oncogenic mutants.
Scale bar = 50 μm.
Figure 4 with 5 supplements
Bi-disperse vertex model reveals interfacial tension-driven segregation of oncogenic clusters.
(a) Schematic of the bi-disperse vertex model, where interfacial line tension (red) is applied at the boundary between mutant (green) and wild-type (gray). (b) The value of Λ determines whether the mutant cluster remains compact or spreads into the wild-type population, where the red line gives the outline of the morphology of the mutant cluster interface. (c) Difference in shape index between interfacial and bulk cells as a function of Λ (interfacial tension), shown for different stress thresholds. Experimental average (red) overlaid for comparison is the difference between the shape indices for ROI2 (interface), and ROI1, ROI3, and ROI4 (away from interface) from Figure 2e. (d) Eccentricity of mutant clusters (given by the red lines in (b)) over time for different values of Λ (with Π0=0.3), indicating changes in cluster morphology. (e) Segregation index at t=300 as a function of Λ for Π0=0, quantifying the extent of cluster segregation. (f) Time evolution of the segregation index for different values of Λ (with Π0=0.3 with Π0=0.3), demonstrating interfacial tension-driven segregation. (g) Hydrostatic pressure maps of mutant and wild-type cells at t=300, with mutant cells labeled by green directors and wild-type cells by black directors. (h) Velocity field maps of epithelial monolayers for different values of Λ, revealing flow patterns around oncogenic clusters. (i) Tangential velocity loop integral for circles of varying radius centered at the mutant cluster (Region 1) and centered away from the cluster (Region 2), showing tissue-specific velocity variations. (j) Extrusion probability of a single mutant cell as a function of Λ, demonstrating how interfacial tension influences mutant cell fate.
Figure 4—figure supplement 1
Normalized tangential velocity loop integral plotted vs. circle number (increasing radius).
The integral has been normalized by the perimeter of the circle. The same trend as in the unnormalized analysis is seen. Average has been taken over eight different wild-type and mutant cluster positions.
Figure 4—figure supplement 2
Normalized tangential velocity loop integral plotted for circle number 22 (cells at the interface of the mutant clusters) vs. time.
The integral has been normalized by the perimeter of the circle. We see tangential velocity persists for over 24 hr, till the end of the time-lapse video, starting when the mutants have formed a mature, compact, rounded, well-defined cluster.
Figure 4—figure supplement 3
Line tensions from simulation for Λ = −0.2.
Figure 4—figure supplement 4
Line tensions from simulation for Λ=1.6.
Figure 4—video 1
Video showing output of the bi-disperse vertex model with cluster dynamics varying with interfacial tension (Λ).
Figure 5
Actomyosin belt disruption prevents mutant cluster compaction in mammary epithelia.
(a) Representative actin staining image upon blebbistatin treatment, done after oncogenic induction in mammary epithelial monolayer, showing an absence of an actin belt. (b) Quantification of F-actin intensities (left) and shape indices (right), showing differences between oncogenic clusters and wild-type cells in these parameters, is getting small – although not completely vanished because of low concentration of blebbistatin used for the experiment. (c) Representative actin and E-cadherin staining images of the wild-type monolayers showing shorter stress fibers, higher E-cadherin intensities in MCF10A wild-type monolayer, and the opposite in BEAS2B. (d) Plots comparing F-actin stress fiber length (left panel) and shape indices (right panel) in wild-type MCF10A and BEAS2B monolayers. Stress fiber lengths were calculated manually in ImageJ. (e) Plots comparing F-actin and E-cadherin levels in the two wild-type tissues. Representative data is plotted from one experiment with the median shown as a bold line. Statistical significance was calculated using an unpaired t-test with Welch’s correction. Scale bars = 50 μm.
Appendix 1—figure 1
Illustration of shape-tension coupling at the mutant (green)-wild-type (gray) interface (red), where angles θ and θ’ represent the orientation of boundary cells relative to their shared edge.
Appendix 1—figure 2
Shape indices vs. the interfacial line tension for no shape-tension coupling.
Appendix 1—figure 3
Interface dot product for different interfacial line tension with and without shape-tension coupling.
Appendix 1—figure 4
For positive differential interfacial tension (Λ >0), the mutant cluster has lower bulk actin level compared to the wild-type cells surrounding it.
Appendix 1—figure 5
For negative differential interfacial tension (Λ<0), the mutant cluster does not have a very different actin level compared to the wild-type cells surrounding it.
Appendix 1—figure 6
Difference in shape index between interfacial and bulk cells as a function of Λ (interfacialt ension) for the shear stress-activated model.
Tables
Key resources table
| Reagent type (species) or resource | Designation | Source or reference | Identifiers | Additional information |
|---|---|---|---|---|
| Cell line (Homo sapiens) | MCF10A | ATCC | CRL-10317 | Immortalized, non-tumorigenic human mammary epithelial cell line |
| Cell line (Homo sapiens) | BEAS2B | Gift from Arjun Guha, inSTEM, Bangalore | Immortalized, non-tumorigenic human bronchial epithelial cell line | |
| Cell line (Canis lupus familiaris, female) | MDCK | Gift from Tamal Das, TIFR Hyderabad | Immortalized, non-tumorigenic, canine kidney epithelial cell line; widely used as a model for epithelial polarity and mechanics | |
| Transfected construct (H. sapiens (human)) | HRasV IRES GFP plasmid | Addgene; Ohoka, 2015 | Plasmid #18780 | Retroviral plasmid constitutively expressing along with HRasV12 and GFP |
| Antibody | E-cadherin 24E10 (Rabbit monoclonal) | Cell Signaling Technology | Cat# 3195 | IF(1:1000) |
| Antibody | Anti-rabbit IgG (H+L), F(ab')2 Fragment (Alexa Fluor 594 Conjugate) (Rabbit polyclonal) | Cell Signaling Technology | Cat# 8889 | IF(1:500) |
| Antibody | Phospho-Myosin Light Chain 2 (Ser19) (Rabbit polyclonal) | Cell Signaling Technology | Cat# 3671 | 1:50 |
| Chemical compound, drug | Lipofectamine 3000 | Thermo Fisher | Cat# L3000015 | Reagent used for transfecting HRasV12-GFP plasmid |
| Peptide, recombinant protein | Fibronectin | Sigma-Aldrich | Cat# F1141 | Surface coating on glass-bottom dishes |
| Chemical compound, drug | Blebbistatin | Sigma-Aldrich | B0560 | Myosin II inhibitor (final concentration: 5 µM) |
| Software, algorithm | SPSS | SPSS | RRID:SCR_002865 | |
| Other | DAPI stain | Cell Signaling Technology | Cat# 4083 | 1:1000 |
| Other | Phalloidin stain | Cell Signaling Technology | Cat# 12877 | 1:300 |
| Software, algorithm | ImageJ/Fiji | NIH | RRID:SCR_002285 | Image analysis |
| Software, algorithm | MATLAB | MathWorks | RRID:SCR_001622 | PIV (PIVlab) and Bayesian force inference analysis |
| Software, algorithm | Cellpose | Stringer et al., 2021. | RRID:SCR_021716 | Cell segmentation |
| Software, algorithm | Tissue Analyser | Aigouy et al., 2016 | Cell segmentation for force inference | |
| Software, algorithm | PIVlab | Thielicke and Sonntag, 2021 | Particle image velocimetry | |
| Software, algorithm | GraphPad Prism 10 | GraphPad Software | RRID:SCR_002798 | Statistical analysis |
| Software, algorithm | Python | https://www.python.org/ | Vertex model code was written in Python |
Additional files
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Heterotypic interfacial tension between oncogenic and wild-type populations forms the mechanical basis of tissue-specific oncogenesis in epithelia
eLife 14:RP106893.
https://doi.org/10.7554/eLife.106893.4
