Evaluation of Antiviral Drugs’ Central Nervous System Distribution Using In Vitro Blood–Brain and Blood-CSF Barrier Models: The Impact of Serum Protein Concentration

The blood–brain barrier (BBB) and blood–cerebrospinal fluid barrier (BCSFB) are dynamic neurovascular interfaces that maintain central nervous system (CNS) homeostasis by tightly regulating paracellular and transcellular molecular transport. Intercellular communication between adjacent cells is critical for maintaining barrier function. Also, serum protein binding is a key factor influencing central nervous system (CNS) penetration. Accurate in vitro BBB and BCSFB models that recapitulate these physiological interactions offer valuable alternatives to animal-based systems for screening CNS targeted drugs. Building upon our published human primary cell–based BBB model, we employed complementary BBB and BCSFB systems to evaluate the permeability of nine antiviral agents/drugs (AVs) under low (L) and high (H) serum conditions. Barrier integrity was confirmed by transepithelial electrical resistance (TEER) and FITC-dextran permeability assays. In H serum, AVs like dolutegravir and efavirenz with extensive protein binding demonstrated significantly reduced permeability across both barriers. These in vitro permeability profiles were fairly consistent with reported higher clinical cerebrospinal fluid concentrations for drugs that exhibit low protein binding. Furthermore, astrocytes exposed to dolutegravir and emtricitabine indicated upregulation of ATP-binding cassette transporter multidrug resistance protein (MRP1) and P-glycoprotein, while lacking breast cancer resistant protein and MRP4 transporters, additional factors for low CNS concentration of AVs. Our validated in vitro BBB and BCSFB models reproduce clinically relevant patterns of CNS antiviral penetration. These systems, with further optimization, can provide reliable, human-relevant platforms for preclinical CNS drug screening and align with current efforts to reduce animal use in biomedical research.