Development of a Direct-Contact Triculture Model of the Human Blood-Brain Barrier and Potential Applications as a Preclinical Drug Screening Tool

Aimable Ngendahimana, Purdue University


Given the high cost and emerging ethical and public concerns associated with the use of animal models, the use of in vitro techniques is gaining considerable attention in research programs related to CNS drug discovery and development. More specifically, various in vitro cell culture models for studying and predicting drug transport across the Blood-Brain Barrier (BBB) were developed as preclinical research tools for neurological drug discovery and development. The main goal of the present studies has been to improve conventional in vitro methods used in mimicking the BBB physiology and functions by allowing physiologically complex interactions of the human BBB cells (astrocytes, pericytes, and Brain Microvessel Endothelial cells or the BMECs) to occur in vitro. This goal was achieved by carrying out a systematic seeding of layers corresponding to all the principal cellular components of a human BBB, such that a confluent and continuous layer of BMECs would form on a basement of human brain astrocytes and pericytes. Therefore, it was hoped that if complex interactions among these BBB cells were allowed to occur in vitro, then this would lead to an improved multicellular culture model of the BBB system, showing better paracellular tightening in comparison to the conventional BBB monoculture models. Results presented in this dissertation indicate that a confluent and continuous layer of hCMEC/D3 cells (an immortalized BBB cell line) was formed directly on a basal coculture made of astrocytes and pericytes layers to make a physiological configuration resembling that of the human BBB in vivo. The resulting direct-contact triculture model was more restrictive to the permeation of various paracellular markers (mannitol, sucrose, PEG-4000 and inulin) than its corresponding monoculture of hCMEC/D3 cells. In order to compare the transcellular absorption between mono- and tri-culture models, digoxin (an FDA recommended probe substrate for efflux transporters) and elacridar (an FDA recommended potent inhibitor of efflux transporters) were used to compare efflux activity between the mono- and tri-culture models. While, a statistically significant effect of elacridar on the absorption of digoxin was observed in the hCMEC/D3 monocultures, there was no statistically significant effect of elacridar on the absorption of digoxin in the triculture model. The observed difficulty to increase the absorption of digoxin using a potent Pgp inhibitor (elacridar) in the triculture model is in agreement with previously published research showing that it is difficult to improve brain drug exposure even when efflux transporters at the human BBB are inhibited. In addition, the triculture model showed a valid ranking order of transcellular BBB permeants; L-histidine (a non-Pgp substrate) showed higher apparent permeability, followed by thiamine (a Pgp-substrate, but essential for brain function), followed by propranolol (a weak Pgp substrate), then paclitaxel, and the lowest permeability corresponded to verapamil which is a strong Pgp substrate. The final chapter of this dissertation provides possible strategies on how the triculture methodology can be further improved and applied in an industrial setting as a screening tool for studying the effect of pharmaceutical drugs and drug products on the BBB physiology and functions. Ultimately, it is hoped that the triculture will be useful in the creation of IVIVC models to predict clinical pharmacokinetic and pharmacodynamic parameters.




Knipp, Purdue University.

Subject Area

Bioengineering|Neurosciences|Pharmaceutical sciences

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server