Matrix–Matrix Interfaces Orchestrate Early Mechanosensitive Transition from Attractor To Track

Hierarchical extracellular matrix (ECM) cues spanning mechanics, architecture, and matrix–matrix interfaces (MMIs) regulate the directionality and efficiency of tumor-cell migration and invasion. Despite their relevance, the contributions of interfacial structures within 3D ECMs remain under-resolved, particularly whether discrete boundaries serve as “attractor-and-track” drivers. Here, we engineered a polydimethylsiloxane (PDMS) microfluidic platform to create controlled MMIs that emulate the tumor microenvironment (TME)’s mechanical heterogeneity, achieved by sequential collagen gelation to create both planar and curved boundaries. With the system, we quantified how MDA-MB-231 (invasive) and MCF-7 (non-invasive) breast cancer spheroids migrate in uniform matrices of graded stiffness and when encountering soft–stiff boundaries. MDA-MB-231 spheroids demonstrated significantly greater migration in soft matrices and exhibited distinct invasive outgrowth at interfacial boundaries, with soft-top pairings (o–o, t–o) gating early detachment, followed by convergence of outgrowth across stiffness pairings at later times. Spheroids positioned above or below planar or curved MMIs showed directional approach toward the boundary and alignment of trajectories within the interfacial plane, consistent with interface-seeking and interface-parallel migration. In contrast, MCF-7 spheroids displayed minimal migration under all tested conditions, underscoring a phenotype-dependent responsiveness to ECM cues. Overall, our findings highlight the critical role of interfacial structures, in addition to bulk stiffness and architecture, in shaping cancer invasion, supporting a two-phase model in which local bulk mechanics license early outward dissemination, whereas interfacial stiffness increasingly sustains expansion. The proposed microfluidic platform offers a tunable and physiologically relevant model for dissecting 3D cell migration mechanisms within complex ECM environments, with optical access, curvature control, and validated passive gradients enabling future chemotaxis studies.