Engineering Tissue-Relevant Collagen Bundle Architectures via Macromolecular Crowding.

Collagen type I forms thick, cell-scaled bundles in native tissues, but standard in vitro collagen gels are composed of disordered thin nanofibrillar networks lacking this architecture. We introduce a simple macromolecular crowding (MMC) strategy, distinct from thermo- or pH-driven gelation, that rapidly assembles collagen into continuous, thick, microscale bundles with tunable dimensions matching healthy and diseased tissue states. These bundles recreate fibrotic and cancer-associated matrix features, enabling direct investigation of how the geometry and topography of collagen fiber networks regulate cell-state transitions and tumor invasion. The method is compatible with collagen from multiple species and collagen-rich decellularized ECM. These single-cell-sized bundles also support robust endothelial sprouting and the formation of aligned microvascular networks in matrices that normally restrict angiogenesis. Bundle suspensions are readily extrudable for bioprinting and injectable delivery applications. Compared to conventional collagen gels, MMC-synthesized bundles are highly tunable to trigger diverse cell behaviors and more accurately replicate native tissue microenvironments, providing a broadly applicable platform for disease modeling and regenerative engineering.