Towards Novel Biomimetic In Vitro Models of the Blood-Brain Barrier for Drug Permeability Evaluation

The blood-brain barrier (BBB) is a highly selective interface formed by the neurovascular unit (NVU) that protects brain homeostasis but severely limits central nervous system drug delivery. Traditional animal models and static in vitro systems fail to recapitulate key human BBB features (cellular composition, tight junctions, transporters, shear stress, and geometry), contributing to poor clinical translation of CNS therapeutics. This review addresses the need for biomimetic, human-relevant BBB models for reliable drug permeability prediction, highlighting organ-on-chip (OOC) microfluidic platforms that reproduce BBB architecture, fluid dynamics, and multicellular interactions.

The paper surveys BBB biology and transport mechanisms (paracellular and transcellular routes, tight junctions, transporters, efflux pumps, glycocalyx) and outlines challenges for CNS delivery and strategies to enhance penetration. It critiques interspecies differences that limit animal model translatability and reviews progression from 2D/static and Transwell co-culture systems to dynamic OOC devices. Key comparisons include impacts of shear stress, co-culture (endothelial cells with pericytes and astrocytes), and basal lamina mimicry. The review contrasts porous membrane-based devices (polycarbonate, PET, polyester, silicon nitride) versus ECM-based hydrogels (collagen, fibrin, collagen-hyaluronic acid) that better emulate the basal lamina and 3D geometry. It discusses TEER as a barrier metric alongside its limitations and the integration of biosensors for real-time monitoring. Application-focused sections summarize BBB-on-chip use in nanoparticle/drug permeability testing, neuroinflammation, glioblastoma and Alzheimer’s disease models, viral infection (VEEV, SARS-CoV-2), ischemic stroke, and multi-organ configurations (e.g., liver–BBB coupling) for ADME-aware permeability assessment.

Narrative review of recent BBB-on-chip advances, compiling and comparing device architectures (membrane vs hydrogel), materials, cell sources and co-cultures, flow conditions, sensing modalities, and application domains for permeability and disease modeling. The article synthesizes reported outcomes (e.g., TEER, permeability, gene expression) across multiple studies without a formal systematic protocol or meta-analysis.

- Organ-on-chip BBB models overcome key limitations of static 2D/Transwell systems by introducing flow-derived shear stress, 3D architectures, and multicellular NVU co-cultures. - Minimum biomimetic requirements proposed: (1) tri-culture of human brain microvascular endothelial cells, pericytes, and astrocytes; (2) replacement of inert semipermeable membranes with ECM-based hydrogels or biological membranes mimicking basal lamina; and (3) inclusion of flow to provide physiological shear stress. - TEER alone is insufficient to validate BBB integrity; it is influenced by biological, physical, and measurement setup factors. Complementary assays (permeability to reference markers, expression/localization of tight junction proteins, transporter function) are necessary. - Flow and co-culture conditions enhance BBB-like properties: fluid flow upregulates glycocalyx-related genes and increases negative surface charge; pericyte/astrocyte co-culture improves barrier function and reduces permeability compared to endothelial monocultures. - Hydrogel-based, membrane-free or hybrid devices better approximate in vivo geometry and cell–ECM interactions, often yielding more in vivo-like permeability than Transwells. - BBB-on-chip platforms enable drug and nanoparticle screening, mapping transport mechanisms, and disease modeling (glioblastoma, Alzheimer’s disease, viral infections, ischemic stroke), with emerging high-throughput formats (e.g., OrganoPlate) and integrated biosensors for real-time monitoring. - Material choice impacts drug studies: PDMS absorption may confound small-molecule assays; thermoplastics (polystyrene, PMMA, polycarbonate, cyclic olefin copolymer) are preferred for low absorption.

By defining minimal design criteria and benchmarking features that enhance physiological relevance (human tri-culture NVU, ECM-mimetic scaffolds, flow), the review provides a framework to develop BBB-on-chip models capable of predicting drug permeability more reliably than animal or static in vitro systems. Incorporating shear stress, 3D microarchitecture, and cell–cell/ECM interactions aligns in vitro behavior with in vivo endothelial function, including tight junction integrity and transporter activity. Application examples illustrate how these platforms can de-risk CNS drug discovery, enable mechanistic insights (e.g., nanoparticle receptor-mediated transcytosis, hormone effects on transport), and evaluate neurotoxicity by coupling BBB with neuronal compartments. Integration of biosensors and multi-organ chips further addresses pharmacokinetics and disease-relevant biomarkers, improving translational relevance.

The review consolidates current progress in BBB-on-chip technologies and proposes practical minimum requirements for biomimicry: human endothelial–pericyte–astrocyte tri-culture, ECM/hydrogel basal lamina mimics, and physiologically relevant flow. It emphasizes comprehensive validation beyond TEER and highlights materials considerations for drug screening. Future directions include: mimicking in vivo vascular geometry, inclusion of neurons to assess neurotoxicity, standardized benchmarking of permeability and protein expression, integration of multiplexed biosensors, adoption of thermoplastics over PDMS for small-molecule studies, and development of multi-organ systems to capture absorption, metabolism, and excretion prior to BBB exposure. These advances position BBB-on-chip as a robust alternative to animal models in preclinical CNS drug development.

Reported limitations of existing models include: reliance on non-human cell sources leading to species-specific discrepancies (e.g., P-gp activity); incomplete NVU composition (omission of pericytes, astrocytes, or neurons); use of inert porous membranes that fail to recapitulate basal lamina properties; variable and setup-dependent TEER measurements that cannot alone guarantee barrier fidelity; material absorption of hydrophobic drugs by PDMS; and hydrogel compositions (e.g., fibrin) that may not fully mimic in vivo extracellular matrix. Many studies lack standardized validation panels and consensus benchmarking, and the review does not report a formal systematic search or quantitative synthesis.