Development of an in vitro cerebral ischemia model of the blood-brain barrier based on human induced pluripotent stem cells: The pivotal role of microenvironmental cells

Background: During cerebral insults, such as traumatic brain injury (TBI) and ischemic stroke, flow cessation leads to cerebral ischemia and this in turn to the breakdown of the blood-brain barrier (BBB) which worsens patient survival rate and recovery outcome. Several in vitro models exist simulating cerebral ischemia applying oxygen/glucose deprivation (OGD). A model system that meets in vivo conditions to a high degree and offers the possibility to study genetic heterogenicity in patients are human induced pluripotent stem cell (hiPSC) - differentiated brain capillary endothelial-like cells (BCELCs). The aim of the study was to develop and investigate a hiPSC-based BBB in vitro model of cerebral ischemia. Methods: Transwell ^® models with hiPSC-differentiated BCELCs in co-culture with the rat glioma C6 cells, in triple-culture with human primary astrocytes and pericytes (hAP) or in mono-culture were exposed to OGD for 6h, 7h, 9h or 17h or to OGD for 6h followed by an 18h recovery phase (OGR). BBB breakdown after OGD and OGR was assessed by changes in transendothelial electrical resistance (TEER) and permeability studies with the paracellular marker fluorescein. Transcriptomic changes of BBB-relevant targets were determined by high-throughput qPCR, proteomic changes were assessed with an antibody-based quantitative protein OLINK assay. These in vitro generated data were compared to transcriptomic and proteomic changes in isolated brain capillaries from autopsy samples of patients suffering from TBI with differing survival times. Results: A faster and progressive BBB breakdown with decreased TEER and increased permeability in a time-dependent manner was determined for BCELCs in co-culture and triple-culture. Hierarchical clustering of transcriptomic changes revealed that 85 % of targets grouped together in co- and triple-cultures. BBB integrity was regained after the 18h recovery phase. Several similarities in transcriptomic and proteomic regulations were revealed between in vitro data and clinical samples affirming the quality of the developed cerebral ischemia BBB in vitro model. Conclusion: A physiologically relevant in vitro disease model based on BCELCs in triple-culture with hAP mimicking the BBB microenvironment was established and optimized. It could be applied for studies with regard to functional and molecular changes after OGD treatment as well as after BBB restoration.