Regenerative engineering harnesses materials science and stem cell biology to develop strategies to repair damaged and diseased tissue. Despite advances in designer materials, few techniques effectively provide auto-regulated feedback mechanisms that govern how cells sense and respond to discrete microenvironmental changes. Here, we demonstrate that the artificial, juxtacrine-like receptor synthetic Notch (synNotch) can be activated by endogenous multimeric cytokines in solution, without immobilizing materials, revealing a previously unreported activation modality and yielding up to 24-fold dynamic range. To broaden synNotch sensing to monomeric cytokines, we developed nMATRIX, a co-engineered material-cell platform that detects endogenous, soluble ligands and routes them to programmed gene circuits with spatially confined effects. nMATRIX can be tuned to recognize the interleukins IL-1beta and IL-6 using synNotch receptors plus cognate biomaterials, yielding up to 68-fold dynamic range and converting these inflammatory inputs into orthogonal outputs that reprogram nearby cell phenotypes. nMATRIX functions across multiple cell types and can incorporate the synNotch-related SNIPR synthetic receptor platform. nMATRIX repurposed inflammatory signals and converted them into anti-inflammatory cues to polarize macrophages (increased CD163, CD206; decreased CD86). Thus, nMATRIX couples native soluble cues to customized cellular responses with tunable sensitivity, offering a flexible materials-based approach for self-regulating regenerative therapies.
The Hidden Power of Normalization: Exponential Capacity Control in Deep Neural Networks
arXiv:2511.00958v1 Announce Type: cross Abstract: Normalization methods are fundamental components of modern deep neural networks (DNNs). Empirically, they are known to stabilize optimization dynamics and