(84a) A Retinoic Acid-Enhanced Human Blood-Brain Barrier Co-Culture Model Constructed from Scalable Cell Sources

Authors: 
Lippmann, E. S., University of Wisconsin-Madison
Al-Ahmad, A., University of Wisconsin
Azarin, S. M., Northwestern University
Palecek, S. P., University of Wisconsin
Shusta, E. V., University of Wisconsin-Madison

The blood-brain barrier (BBB), which consists of the endothelial cells that line brain capillaries, is a highly impermeable interface between the bloodstream and the brain parenchyma. While the physiological purpose of the BBB is to maintain brain health and homeostasis, its restrictive properties also present a challenge for drug delivery. Over the last several decades, researchers have developed many permutations of in vitro BBB models to create a suitable system that mimics the in vivoBBB for drug screening applications. These models are most often comprised of primary brain microvascular endothelial cells (BMECs) and support cells of the neurovascular unit such as pericytes, astrocytes, and neurons. However, due to yield and availability constraints, the cells in these models are typically derived from various animal species which may not appropriately represent the properties of the human BBB. In an effort to overcome these limitations, we previously introduced a method for deriving BBB endothelium from human pluripotent stem cells (hPSCs). Here, we present improvements made to this system, including the identification of retinoic acid (RA) as a substantial contributor to BBB properties and the incorporation of primary human brain pericytes and human astrocytes and neurons derived from neural progenitor cells (NPCs).

The addition of RA to differentiating hPSC cultures consisting of neural cells and immature BMECs induced VE-cadherin expression in the PECAM-1+/GLUT-1+ BMECs, indicating their maturation along the endothelial lineage. RA did not increase BMEC differentiation efficiency but did promote BMEC proliferation to increase yield by approximately 2-fold. Upon purification, RA-treated BMECs exhibited greater organization of tight junction strands, possessed transendothelial electrical resistance (TEER) that was substantially greater than untreated controls, and demonstrated increased multidrug resistance protein (MRP) efflux activity. When co-cultured with primary human brain pericytes and NPC-derived astrocyte/neuron mixtures, the hPSC-derived BMECs exhibited TEER reproducibly exceeding 5000 Ωxcm2. This model possesses barrier properties approaching in vivo measurements in rats (5900 Ωxcm2) and the theoretical maximum of BBB resistance (8000 Ωxcm2), whereas no previous human model has exceeded 500 Ωxcm2 and most animal models fall below 2000 Ωxcm2.

This fully human BBB model should have significant utility for drug screening applications given its substantial barrier properties. Because the human cells in this model are obtained from renewable sources, the overall system can be readily scaled to handle high throughput workloads. Moreover, in conjunction with advancements in cellular reprogramming and genome editing, this system could be highly attractive for disease modeling applications and assist in understanding how an individual’s genetic profile may dictate drug distribution in the central nervous system.