This is a summary of González-Gallego et al. A fully iPS-cell-derived 3D model of the human blood–brain barrier for exploring neurovascular disease mechanisms and therapeutic interventions. Published in Nature Neuroscience (2025). DOI: 10.1038/s41593-025-02123-w
The challenge
BBB failure is implicated in neurovascular and neurodegenerative disorders, yet many mechanistic insights and drug candidates arise from rodent or simplified in vitro systems that do not fully translate to humans. Currently used co-cultures in transwells lack the tubular, perfused architecture and multi-cell interactions that shape endothelial gene programs, junction integrity, and transport behavior in vivo. This gap is especially problematic for studying human genetic risk factors—such as FOXF2, linked to cerebral small vessel disease—where cell-type-specific effects and clinically relevant phenotypes may be more faithfully modeled in a human context. Moreover, the field needs human BBB platforms that can simultaneously reveal disease mechanisms and serve as practical testbeds for therapeutic delivery strategies, including targeted delivery to brain endothelial cells.
Our approach
We differentiated iPSCs into all major BBB-forming cell types (endothelial cells, mural cells, astrocytes), benchmarked identity and function, and assembled them in microfluidic chips to form perfusable vessel-like tubes. Using CRISPR–Cas9, we generated FOXF2 knockout iPSCs and focused on endothelial FOXF2 loss to model BBB dysfunction. Finally, we tested translational rescue by perfusing LNPs carrying Foxf2 mRNA to restore barrier properties.
Our findings
The 3D BBB model expressed canonical BBB markers, formed lumenized vessel-like networks, supported perfusion (including with human blood), and exhibited functional transporter activity and barrier readouts. Endothelial FOXF2 loss triggered hallmark BBB defects: disrupted junction integrity (notably reduced TJP1), increased caveolae-related transport (elevated CAV1), altered endothelial ultrastructure, reduced barrier resistance, and broader proteomic shifts affecting endocytosis and adhesion pathways. Importantly, LNP-mediated Foxf2 mRNA delivery substantially rescued BBB impairment, demonstrating therapeutic-test capability within the human model.
The implications
Our work constitutes the first fully human iPSC model of the BBB used to study neurovascular diseases. It provides a scalable, fully human BBB platform to connect human genetics with BBB mechanisms and to evaluate therapeutic delivery—supporting more predictive drug discovery for neurovascular and neurodegenerative disease.
Creating SyNergies
By integrating stem-cell differentiation, microfluidic tissue engineering, genome editing, and multi-omics validation, the study exemplifies how interdisciplinary neurovascular research accelerates translation. It connects disease genetics to measurable BBB phenotypes and tests a realistic therapeutic strategy (mRNA-LNP rescue) in a human system. Key contributing SyNergy researchers include Martina Schifferer, Ali Ertürk, Thomas Misgeld, Stefan Lichtenthaler, Martin Dichgans, and Dominik Paquet, whose combined expertise bridges mechanism, modeling, and intervention.
