Vascular networks due to dynamically arrested crystalline ordering of elongated cells

Recent experimental and theoretical studies suggest that crystallization and glass-like solidification are useful analogies for understanding cell ordering in confluent biological tissues. It remains unexplored how cellular ordering contributes to pattern formation during morphogenesis. With a computational model we show that a system of elongated, cohering biological cells can get dynamically arrested in a network pattern. Our model provides a new explanation for the formation of cellular networks in culture systems that exclude intercellular interaction via chemotaxis or mechanical traction.

💡 Research Summary

The paper investigates how elongated, cohering cells can spontaneously generate vascular‑like network patterns without relying on chemotactic cues or long‑range mechanical traction. The authors build a minimal two‑dimensional computational model in which each cell is represented as an anisotropic (elliptical) particle. Inter‑cellular interactions consist of a direction‑dependent adhesion potential (analogous to a Lennard‑Jones term with angular dependence) and a soft‑core volume‑exclusion term. Thermal‑like noise is introduced to mimic stochastic cellular motility. Simulations start from a random orientation and uniform density, allowing the system to evolve under overdamped dynamics.

Three distinct phases emerge during the simulation. In the early stage, cells collide and locally align along their long axes, forming small ordered clusters. As time progresses, these clusters grow by recruiting neighboring cells, but the system never reaches a global crystalline state. Instead, at an intermediate stage a “dynamic arrest” occurs: cells become trapped in a jammed configuration where further rearrangement is kinetically prohibited. The resulting structure is a percolating network composed of thin linear segments intersecting at nodes, reminiscent of capillary‑like meshes. This arrested state is analogous to glass formation rather than classical nucleation‑driven crystallization, because no single nucleus dominates and the whole ensemble freezes simultaneously.

Parameter sweeps reveal the critical roles of aspect ratio (L/D), adhesion strength (ε), and noise amplitude (T). When L/D ≤ 2 (near‑spherical cells), the system remains disordered and no network forms. For L/D between 3 and 5, cells exhibit a strong tendency to align, and a moderate ε combined with an optimal T yields the most pronounced network. Excessively high ε drives the system toward a dense, ordered crystal lattice, while too low ε leads to fragile connections that break apart. Similarly, high T (strong noise) disrupts the network, whereas very low T accelerates crystallization, bypassing the arrested mesh altogether. These findings map directly onto experimental variables: substrate stiffness modulates effective adhesion, cell density influences the onset of arrest, and external fields can tune the effective noise level.

The authors compare their simulation outcomes with published in‑vitro experiments where endothelial or fibroblast cells form reticular structures in the absence of chemotactic gradients or externally applied forces. The morphological characteristics—branch thickness, node spacing, and overall connectivity—match the model’s predictions for the dynamically arrested regime. This concordance supports the hypothesis that purely physical interactions, specifically shape anisotropy and short‑range adhesion, are sufficient to drive network morphogenesis.

In conclusion, the study provides a mechanistic framework linking cellular geometry and local adhesion to emergent tissue‑scale patterns. By demonstrating that vascular‑like networks can arise from a glass‑like dynamic arrest of elongated cells, the work challenges the prevailing view that long‑range biochemical signaling is a prerequisite for such architectures. The model offers a quantitative tool for tissue engineering: adjusting cell aspect ratio, adhesion molecule expression, or substrate mechanical properties could be used to program desired network topologies without the need for exogenous morphogens. This insight opens new avenues for designing self‑organizing scaffolds and for understanding how physical constraints shape morphogenesis in vivo.