(B-60) A meta-analysis of engineered blood brain barriers in vitro

Introduction:: The blood brain barrier (BBB) is the most impermeable blood vessel in the human body¹. While the barrier is necessary for preventing harmful materials and organisms from breaching the brain, it also poses a tremendous hurdle for drug delivery into the brain. To expedite drug screening for treating brain diseases, in vitro models have been engineered as an alternative to animal models which are costly and time-consuming^2,3. Over the last 3 decades, two major types of BBB models, transwells and microfluidic models, have been investigated to recapitulate the BBB, including its permeability and trans-endothelial electrical resistance (TEER)⁴. The types of BBB cells included in BBB models have been shown to greatly affect resulting permeabilities of the models⁵. While others have investigated the effect of including different combinations of BBB cell types in their models, a statistical meta-analysis of the true effect of model-type, cell combination, and materials used in BBB models has never been performed to evaluate the recapitulation of the human BBB with the models and the particular combination of those parameters which provides the most physiologically-relevant BBB model with current technology. The results provided here can shed insight into developing improved BBB models in the near future.


Materials and Methods:: To obtain permeability data of BBB models, Google scholar (www.scholar.google.com) was used and keywords “blood brain barrier model in vitro permeability” were input into the search bar to find publications. The keywords “blood brain barrier model in vitro TEER” were input into the search bar to find publications on BBB models with TEER data. Publications were limited to the last ~20 years (~1999-2022), and each result required a thorough reading of the data to find permeability and TEER measurements with co-culture variations of the model. Normalized permeability and TEER values were calculated by normalizing reported permeability or TEER values in co-culture models to the data reported for the endothelial mono-culture model in that particular study, to study the impact of the co-culture. For comparison between two groups, statistical analyses were performed using unpaired, two-tailed t-tests, assuming equal variance. For comparison between more than 2 groups with 2 dependent variables, statistical analyses were performed using two-way ANOVA with post hoc Tukey’s multiple comparisons tests.


Results, Conclusions, and Discussions:: In particular, there is no statistical difference in normalized permeability between transwell and microfluidic models of any co-culture combination (Fig. 1A). The microfluidic endothelial cell, pericyte, and astrocyte tri-culture model shows the lowest normalized permeability of all the analyzed combinations and is significantly lower in permeability than the microfluidic endothelial cell and astrocyte bi-culture model (Fig. 1B). There is also no statistical difference in normalized TEER values between the two models (Fig. 1C), however there is a 1.4-fold greater TEER in transwell models than microfluidic models. Surprisingly, the microfluidic endothelial cell, pericyte, and astrocyte tri-culture model shows the lowest TEER of all the analyzed combinations (Fig. 1D), which is contrary to Fig. 1B that demonstrates it is the least permeable combination. Moreover, we show that the choice of ECM used in models has a tremendous effect on resulting BBB properties according to the meta-analysis in this work. Matrigel-based models show greater permeabilities than non-Matrigel-based models (≥ ~3-fold), however there is a striking decrease in permeability when Matrigel is used in microfluidic BBB models compared to transwell models (>3-fold), showing model-dependency for Matrigel-based models, but not in other material-based models. Overall, cell types and ECM are two pivotal design considerations that should be refined for recapitulating the human BBB.

The meta-analysis of transwell and microfluidic models from over the last ~20 years suggests cell combination is a more crucial design parameter for transwell and microfluidic models than the model type. The first ever meta-analysis of BBB model types, cell combinations, and materials provided here gives necessary retrospect and insight into the last several decades of BBB-modeling endeavors, and can guide future advancements in recapitulating the human BBB in vitro.


Acknowledgements (Optional): :

References (Optional): : ¹Stamatovic et al. (2008). Curr Neuropharmacol 6, 3. ²Rouleau et al. (2023). Nat Rev Bioeng 1, 252. ³Hajal et al. (2022). Nat Protoc 17, 95. ⁴Jiang et al. (2019). Micromachines (Basel) 10, 6. ⁵He et al. (2014). Stroke 45, 8.