An experimental-computational framework to characterize immune cell migration in the glioblastoma microenvironment

One of the main physiological drivers of glioblastoma progression is the interaction of the immune system with the tumor microenvironment, where secreted chemoattractants establish biochemical gradients that guide immune cell migration and infiltration. Understanding how these signals modulate immune cell motility is essential for characterizing the mechanisms underlying immune activation and suppression. Furthermore, this chemotactic response is significantly moderated by the mechanical properties of the extracellular matrix.

This work presents an integrated pipeline combining in vitro microfluidic experiments with agent-based modeling to characterize peripheral blood mononuclear cell migration dynamics under two collagen concentrations: 2 and 4 mg/mL. The pipeline includes a parametric model that is calibrated from experimental data using specific quantities of interest. The calibration procedure itself is first verified against in silico experiments, in which synthetic data are generated with known parameter values that the pipeline is required to recover.

Results show that higher collagen concentrations are associated with a significant reduction in directed migration. Although this outcome could be attributed to a physical effect, a denser extracellular matrix mechanically restricts cell displacement, the results are more readily compatible with a chemosensory mechanism, in which increased matrix density attenuates chemokine diffusion, thereby impairing the ability of immune cells to orient along chemical gradients. Additionally, a temporal decay in the motility metrics and their associated parameters was observed, reaching statistical significance at the lower collagen concentration, which could reflect progressive immune exhaustion, a phenomenon consistent with previous reports.

Together, these findings suggest that extracellular matrix density and temporal dynamics modulate immune cell migration in this glioblastoma model. The proposed framework enables a mechanistic interpretation of immune cell migration that complements traditional statistical analysis, providing a versatile platform adaptable to other cell populations and tumor contexts.