Protocols to differentiate iPSCs to BMECs for in vitro BBB modeling have not necessarily achieved widespread adoption, potentially due in part to the time required to produce purified BMECs and the cost associated with such differentiation and purification. The purpose of this study was to decrease both the time and cost required for such differentiation while still achieving BMECs of comparable performance to established differentiation methods. These advancements will potentially allow iPSC-derived BMECs to be more readily accessible to researchers, thereby providing high-fidelity human in vitro BBB models for a wide range of applications.
In this study, existing BMEC differentiation protocols were modified to exclude the iPSC expansion phase prior to initiation of differentiation [
24,
25,
37]. By controlling initial iPSC seeding density [
37], maximum TEER values of purified BMECs from such differentiations remained consistently in excess of 2500 Ω × cm
2, and barrier fidelity, as indicated by TEER above 1000 Ω × cm
2, was maintained for a minimum of 8 days, as was similarly achieved in a recent microfluidics-based model using UM-derived BMECs [
40]. E8 medium was also used for iPSC maintenance in place of mTeSR, as described by others [
30]. Given that we routinely use E6 medium (a derivative of E8 medium that lacks growth factors promoting pluripotency) for neural differentiations [
28], we further explored its use for differentiating iPSCs to BMECs. Upon differentiation in E6 medium, immunocytochemical analysis unexpectedly showed PECAM-1
+ cells at 4 days of differentiation rather than 6 days. We note that changes in culture methods and differentiation medium have previously been shown to alter differentiation times to lineages such as neuroectoderm [
28,
41,
42] and midbrain dopaminergic neurons [
41,
43], with all methods ultimately resulting in cells expressing the same characteristic markers. Therefore, it is unsurprising that changes made in differentiation medium resulted in altered differentiation timelines. After establishing this accelerated differentiation timeline from iPSCs to BMECs using E6 medium, BMECs were evaluated for BBB phenotype by TEER measurement and efflux transporter activity. BMECs differentiated in E6 medium maintained a stable barrier above 1000 Ω × cm
2 for 8 days, longer than previously published reports using similar Transwell-based methods [
25,
37], while achieving similar maximum TEER values to UM-derived BMECs. BMECs differentiated using E6 medium and UM also had no statistical difference in efflux transporter activity for P-glycoprotein and MRP family members. BMEC differentiation using E6 medium was further validated in iPSC lines CD12, CC3, and SM14. CD12- and CC3-derived BMECs demonstrated equivalent maximum TEER and long-term stability as IMR90-4-derived BMECs differentiated in both E6 medium and UM. Notably, CD12-derived BMECs achieved TEER greater than 4000 Ω × cm
2 on days 9 and 10 of subculture, a value equivalent to the maximum TEER achieved after initial induction of BMEC phenotype. Finally, SM14 iPSCs, which harbor biallelic loss-of-function
PARK2 mutations associated with familial early onset PD, yielded BMECs with maximum TEER and long term stability equivalent to BMECs derived from control lines, similar fluorescein permeability, and equivalent MRP family efflux transporter activity. Interestingly, SM14-derived BMECs did not show active P-glycoprotein in the apical-to-basolateral transport assays. Studies on advanced PD patients have demonstrated increased brain uptake of P-glycoprotein substrates [
44]. Our data may indicate that patients with familial PD mutations are predisposed to loss of P-glycoprotein function. However, our results are very preliminary and would need to be rigorously confirmed across multiple iPSC lines from different patients harboring the same mutation. As this manuscript is centered on the utility of E6 medium for in vitro differentiation, we have not pursued these studies herein. Even so, this exciting result indicates researchers can probe mechanisms of BBB regulation in the context of genetic disease or evaluate molecular transport and toxicology over extended experimental time points. Our methods can ostensibly be extended to other iPSC disease lines of interest, provided that the genetic mutation does not impact barrier stability.
To further explore the utility of the model, we investigated the effect of co-culturing BMECs with iPSC-derived astrocytes and primary human brain pericytes. Astrocytes and pericytes are known to aid in induction of BBB phenotype in the developing neurovascular environment [
3]. Though co-culture of BMECs with astrocytes and co-culture of BMECs with pericytes individually were both found to increase barrier function as indicated by TEER, co-culture with astrocytes and pericytes concurrently enhanced the TEER above that achieved through co-culture with either cell type alone. This effect has been noted in other in vitro BBB models, though the reported maximum TEER value from co-culture in this study exceeds previously published in vitro TEER values by more than 700 Ω × cm
2 [
25]. Excitingly, this achievement approaches in vivo TEER predictions of 8000 Ω × cm
2 put forth by Smith and Rapoport [
45]. In addition, medium was not changed following barrier induction to minimize external influences on barrier stability. Due to prospective increased metabolic burden in the co-culture system as evidenced by a qualitative decrease in number of co-cultured cells between the start and end of the experiment, we suspect that gradual media changes following barrier induction may further improve the stability of the model by supporting neurovascular health. Owing to the extended barrier phenotype observed, researchers may now conduct long-term experiments without concern that barrier degradation may confound results.