[Paper Review] Cancer Metastasis: Collective Invasion in Heterogeneous Multicellular Systems

Adrien Hallou, Joel Jennings+1 more|arXiv (Cornell University)|Dec 31, 2014

Nanoplatforms for cancer theranostics3 citations

TL;DR

This study uses in silico modeling with a Cellular Potts framework to investigate how motility heterogeneity in multicellular tumour spheroids drives collective invasion. It demonstrates that mechanical interactions and cell motility variation—rather than average mechanical properties—dominate invasion morphology and efficiency, revealing that even a small fraction of highly motile cells can initiate and sustain collective invasion patterns consistent with in vitro and in vivo observations.

ABSTRACT

Heterogeneity within tumour cell populations is associated with an increase in malignancy and appears to play an important role during cancer metastasis. Using in silico experiments, we study the interplay between collective behaviours and cell motility heterogeneities in a model system. Working with tumour spheroids that contain two non-proliferating cell populations of different motile properties, we explore the conditions required for maximal invasion into surrounding tissues. We show emerging spatial patterns of cellular organisation and invasion which are consistent with in vitro and in vivo observations. This demonstrates that mechanical interactions at the cellular level are sufficient to account for many of the observed morphologies of invasion and that heterogeneity in cell motility can be more important than average mechanical properties in controlling the fate of large cell populations.

Motivation & Objective

• To understand how phenotypic heterogeneity in tumour cell populations influences collective invasion dynamics.
• To determine whether cell motility variation, rather than average mechanical properties, governs invasion morphology and efficiency.
• To model and simulate collective invasion in heterogeneous multicellular systems using a 2D Cellular Potts framework with active motility.
• To compare invasion outcomes in homogeneous versus heterogeneous tumour spheroids under controlled mechanical and motility parameters.
• To identify conditions under which minimal populations of highly motile cells trigger maximal invasion in multicellular systems.

Proposed method

• A 2D Cellular Potts model simulates tumour spheroids composed of two non-proliferating cell populations with distinct motility levels.
• Cellular motility is modeled via a self-propelled force term, with motility magnitude μi defining individual cell activity.
• Inter-cellular adhesion and cortical tension are captured by a single interfacial energy parameter J.
• Cell volume is constrained to a target value V0 using a bulk modulus κ.
• Invasion is quantified by measuring the gradient of invasion length Λb over time, with Γb representing the invasion rate.
• Simulations are run over short timescales to exclude proliferation, focusing on mechanical and motility-driven dynamics.

Experimental results

Research questions

• RQ1 What role does motility heterogeneity play in initiating and sustaining collective invasion in multicellular tumour spheroids?
• RQ2 How do mechanical interactions between cells with varying motility levels give rise to spatial patterns of invasion?
• RQ3 Can emergent collective invasion morphologies arise solely from cell-cell mechanical interactions and motility differences, without subcellular signaling?
• RQ4 What proportion of highly motile cells is sufficient to drive maximal invasion in a heterogeneous population?
• RQ5 How does the invasion rate Γb depend on the motility parameter μb in both homogeneous and heterogeneous systems?

Key findings

• Heterogeneous tumour spheroids with a small fraction of highly motile cells exhibit significantly enhanced invasion rates compared to homogeneous populations.
• The invasion rate Γb increases with μb, and the maximum invasion is achieved when motility heterogeneity is introduced, even at low concentrations of motile cells.
• Spatial patterns of collective invasion—such as strands and clusters—emerge naturally from mechanical interactions and motility differences, matching in vitro and in vivo observations.
• The study shows that motility heterogeneity is more critical than average mechanical properties in determining the fate of large cell populations during invasion.
• Invasion length Λb increases non-linearly with μb, and the gradient Γb reaches a peak at intermediate motility levels, indicating an optimal invasion window.
• The model successfully reproduces experimentally observed collective invasion morphologies without requiring complex subcellular signaling pathways.