Designing and testing BBB-permeable drugs represents a huge burden for CNS drug development. The vast majority of compounds—approximately 100% of large molecules and more than 98% of small molecules—are excluded from the CNS by the BBB through the physical barrier or by efflux pumps expressed by BMECs [
68]. Due to the species-specific differences in transporter and efflux pump expression, human iPSC-based BBB models are an attractive platform to test drug permeability. These models more accurately predict human BBB permeability compared to non-human BBB models [
69] and hold great promise in providing a high-throughput platform for predicting human CNS drug permeabilities and circumventing the need for animal-based testing [
70]. Early iPSC-derived BBB models highlighted the ability of iBMECs to correlate well with in vivo drug permeability using transwell systems [
12,
71] and subsequent studies have expanded permeability testing to microfluidic platforms under fluid flow that more closely mimic in vivo conditions [
19,
31,
64]. Importantly, iBMECs express many of the necessary efflux pumps and transporters [
18,
31,
39] and have successfully been used to investigate general drug transport as well as specific transporter–drug interactions such as LAT1 with gabapentin [
72]. Furthermore, iBMECs can be co-cultured with other cells of the NVU that can potentially alter drug permeabilities through changes in barrier properties or transporter expression [
24], and hence should be considered when designing drug screening platforms. However, it is worth noting that permeability for candidate large and small molecules does not change above TEER thresholds of 500 and 900 Ω cm
2, respectively [
73], suggesting that complex co-culture models may not be necessary for accurately modeling permeability. Despite these advances in drug permeability testing using iPSC-derived BBB models, limited in vivo human permeability data is available to benchmark in vitro BBB models. Recent work has begun to address this issue by measuring in vitro permeability of positron emission tomography (PET) radioligands, for which in vivo human BBB permeability values are known from clinical PET imaging [
74]. Remarkably, iPSC-derived BBB models show highly significant correlation to in vivo values for the 8 radioligands tested. Interestingly, when Le Roux and colleagues [
74] tested a suite of other drugs in their radioligand-validated model, they generated relative permeabilities that could not have been predicted based on the physicochemical properties of the drugs alone. In addition to permeability testing of small molecules, iPSC-derived BBB models are also being used to test permeability of new classes of CNS drugs such as peptides and antibodies. For example, attaching an Angiopep-2 peptide to fluorescent nanoparticles can increase their BBB permeability by 3.5-fold [
19] and a comparable strategy could be used to increase CNS delivery of larger molecule therapeutics. Similarly, iPSC-derived BBB models are being used to evaluate receptor mediated transcytosis-targeting antibodies to enhance drug delivery [
30], also known as molecular Trojan horses [
75]. The species-specific differences in transporter expression highlighted earlier [
2‐
6] underscores the importance of using human-based models for testing these types of novel delivery mechanisms. Lastly, other alternative drug delivery strategies being explored using iPSC-derived BBB models are polymer nanoparticles [
65] and perfusion of hyperosmolar agents like mannitol to temporarily open the BBB and permit the diffusion of non-permeable therapeutics into the CNS [
19,
61,
67]. The development of iPSC-derived BBB models has significantly enhanced the ability to perform human-relevant in vitro drug screens and will likely continue to aid in the discovery and development of new therapeutics and CNS drug delivery methods.