Homeostatic maintenance is essential for proper cerebral function. The blood vessel- and non-vascular cells (neurons and glial cells) in the brain form the neurovascular unit (NVU) . The NVU plays an important role in brain maintenance via cellular interactions between microvessels and parenchyma. Under physiological conditions, the NVU regulates nutrient supply, vascular growth, hemodynamics, toxin elimination and brain protection. Adherens junctions (AJs) and tight junctions (TJs) reduce the paracellular flux across brain endothelium, whereas specific transporters and receptors carry glucose, amino-acids, nucleosides, organic anions, large amino-acids, transferrin, lipo-proteins and drugs into the brain. Conversely, pathological stimuli that increase blood–brain barrier (BBB) permeability perturb brain homeostasis. The leakage of ions, water, and serum proteins into the parenchyma modifies oncotic pressure and ionic concentrations, while leukocyte extravasation triggers immune and inflammatory responses. This imbalance leads to abnormal neuronal activity or toxicity. In excitable brain structures such as the hippocampus and cerebral cortex these features induce seizures. In several CNS structures, increased BBB permeability participates in or worsens neurological disorders like Alzheimer’s disease, multiple sclerosis or chronic epilepsy [2–5].
Modeling the NVU in vitro has furthered the understanding of selective mechanisms which regulate permeability, toxin elimination, nutrient supply, brain protection and homeostasis regulation. Several in vitro cell-based BBB models have previously been developed but were unable to fully recapitulate all known features of the BBB [6, 7]. Despite the conservation of endothelial cell properties ex vivo their isolation from multicellular blood vessels is methodologically difficult . The endothelial cell monolayer is one of the most commonly used in vitro models; however, it only represents a simplified view of the BBB. This simplification reduces the interactions with other cell types, which are essential for BBB maintenance [9, 10]. The co-culture of astrocytes and endothelial cells is the most validated cell-based BBB model. This model contains TJs, transporters, ionic channels and high transendothelial electrical resistance (TEER) necessary for a suitable model. However the absence of other cell types such as pericytes is a limitation in dynamic studies of the NVU, including vasomodulation . To counteract the lack of pericytes, the tri-culture has been developed using endothelial cells, pericyte and astrocytes cell lines. In this system, all cell types are necessary for the adequate localization of TJs and transporter functions . This model can be modified depending on the research objectives, using leukocytes or neurons as the third cell type [13, 14]. The tri-culture is currently one of the most representative in vitro models to study BBB regulation in humans .
Clearly, BBB models should contain most or all cellular and molecular players of the NVU and take into account the different environmental factors. Thirty years ago, Gähwiler et al. described an integrative model to study interactions between cell types in brain slices maintained in culture . This model was simplified by growing organotypic brain slices on a membrane surface . These slices maintain all cell types and their interactions for 2 weeks and were mainly used to study the activity of neural cells under diverse physiologic and pathologic conditions [18, 19].
In 2003, it was shown for the first time that, despite the absence of blood flow in organotypic cortical slices, microvessels were present and able to respond to angiogenic stimuli like acidosis or hyperthermia . Furthermore, microvessels preserved within organotypic slices respond to experimental seizures. We have used this in vitro model to study the effects of seizure-like activity on the NVU. We chose slices of rat hippocampus, since the corresponding structure in the human brain is involved in temporal lobe epilepsy (TLE). We found that kainate-induced epileptiform activities induced vascular changes in organotypic slices including angiogenesis and BBB alteration, similar to those reported in human intractable TLE and in vivo models [21, 22].