Sources of cells for modeling the human BBB — 4 considerations

Jan 26, 2023 | Trends

Welcome to Part 3 of our blog series covering the blood brain barrier (BBB). In Part 1, we looked briefly at how the BBB was discovered and presented what we currently know about its composition and functions. Part 2 focused on the need for reliable BBB models in drug development and summarized the current in vitro models. In Part 3, we will look at considerations surrounding the cell types and sources used in in vitro BBB models. 

In vitro models that accurately represent the human BBB are a critical part of the developmental process for central nervous system (CNS) drugs, as well as to understand diseases that involve degradation or dysfunction of the BBB. More than a century of research has greatly improved our understanding of the human BBB as outlined in Part 1, yet our picture of its composition and mechanisms remain incomplete (at best). This is largely down to a lack of reliable models by which to study the BBB, and the remaining gaps in our understanding further impede the development of the ‘perfect’ human BBB model. This dilemma is a major contributing factor to the lack of progress in treating brain diseases, with an over reliance on non-human models that leads to the development of drugs with poor BBB penetration efficacy.

In Part 2, we presented the most commonly used in vitro BBB models, and briefly mentioned some of the cell types used in those models. In this post, we will take a look at the cell types that have been used to build in vitro BBB models throughout the years, and highlight the advantages and limitations associated with each. We will finish off by looking at a more recently explored source of cells for human BBB modeling that addresses the major limitations of the other cell types used. 

Specifically, we will focus on:

  1. The use of Caco-2 and Madin-Darby canine kidney (MDCK) immortalized cell lines
  2. The use of primary human cells from live donors or via autopsy of deceased individuals
  3. Poor translatability using immortalized cell lines and primary cells calls for superior in vitro BBB models 
  4. Human induced pluripotent stem cell-derived models (hiPSC-derived cell types) as an alternative to primary and immortalized cell lines

1. BBB modeling with Caco-2 and MDCK cell lines

Caco-2 cells and Madin-Darby canine kidney (MDCK) cells have been widely used as sources of immortalized cells for in vitro BBB studies. Both cell types are heterogeneous in nature, and in studies of BBB transport through TEER and permeability measurements, they are typically used to form a confluent monolayer of cells in a single-cell transwell setup. Despite their widespread use in BBB modeling, it is important to note that both cell types originate from non-brain and non-human tissue and are thus best described as surrogate models for the human BBB.

The Caco-2 cell line is an immortalized cell line that originates from cells isolated from a human epithelial colorectal adenocarcinoma in the 1970s (1). Caco-2 cells have been widely used as a model of the intestinal epithelial barrier, owing to their ability to spontaneously differentiate in culture into a monolayer of cells with properties typical of the cells normally found in the small intestine. When cultured in a transwell apparatus, Caco-2 cells differentiate to form a polarized epithelial cell monolayer that provides a physical and biochemical barrier to the passage of ions and small molecules. 

Native Caco-2 cells or Caco-2 cells treated with the chemotherapeutic vinblastine (VB-Caco-2), which leads to elevated expression of the human drug efflux transporter P-glycoprotein, are routinely used to predict BBB transport of CNS drugs and other substances. However, in vivo BBB transport studies have revealed that such in vitro predictions are not always reliable (2). In industry, Caco-2 cells are commonly engineered to overexpress ion channels or transporters to assay for the mechanism of a drug-channel or drug-transporter interaction. 

The MDCK cell line was isolated in 1958 from a kidney of a normal cocker spaniel, and is presumed to have since undergone spontaneous immortalization during in vitro culture (3). This cell line comprises numerous clones with different characteristics in terms of molecular transport. Today, MDCK-MDR1 cells, which are engineered to express the human MDR1 gene that encodes the drug efflux transporter P-glycoprotein, are the most widely used version of MDCK cells within BBB studies. MDCK cells exhibit alkaline phosphatase and γ-glutamyl-transpeptidase activity, enzymes that are also expressed at the BBB. MDCK cells are widely used in permeability screens for diffusional transport owing to the ease at which they can be cultured and their propensity to achieve tight and reproducible barrier properties (4). Despite these attractive features and their widespread use, the morphology, gene expression signatures and cell-cell junctions of MDCK is distinct from human BMECs — which are critically important endothelial cells at the human blood-brain-barrier — raising concerns about their robustness and relevance as a model for the human BBB. 

  1. BBB models based on primary human cells 

To avoid the limitations associated with species differences when using non-human cells, a number of human BBB models have been established by culturing primary human BMECs isolated from fresh surgical specimens of human brain tissue derived from tumors or epilepsy patients, or more rarely from autopsy tissue (5). More recently, researchers in the UK described a novel transwell BBB model using four primary human cell sources — astrocytes, pericytes, human BMECs, and neurons — to reconstitute the major cell types that comprise the neuromuscular unit (6). 

Ethical issues and challenges with BMEC yield and fidelity (and donor matching; or mixing donors’ cells in a single experiment) have limited the feasibility of using primary human cells for BBB studies. To address these obstacles, numerous attempts have been made to create human BBB models based on normal human BMECs that were immortalized through viral transduction with human telomerase or the oncogenic SV40 T antigen (6, 7). However, while these cell lines retained some of the morphological features of native human BMECs, their barrier properties have been described as poor, with low baseline TEER and discontinuous tight junction protein expression (8). 

  1. Poor translatability calls for superior in vitro BBB models 

Building an in vitro BBB model calls for many considerations about the cell type(s) to be used, including but not limited to, cell quality and ease of supply, the source of the cells, purity of donor material (i.e. do the cell originate from a single donor), how well they can recapitulate BBB biology, and whether it is most appropriate to work with healthy or disease-relevant cells. Other considerations involve batch-to-batch consistency, e.g., when working with immortalized cell lines, and donor-to-donor variability when using donor cells or pools of donor cells. 

No perfect in vitro BBB model exists, and in practice the choice of cells or model setup (see Part 2 for an overview) is often a trade-off between assay throughput and speed, cost, and how closely a model needs to reflect the in vivo situation (7). However, while in vitro studies will always be a simplification of what is happening in an in vivo situation, cellular BBB models remain a critical tool in drug development and research. 

Despite their continued use in BBB modeling, all of the cell sources described so far in this post suffer major limitations as presented in Table 1. 


Single-cell monolayer modeling of the BBB using Caco-2 and MDCK cells is particularly problematic because these cells are neither from human nor brain tissue, thus they can never recapitulate the biology of human brain microvascular endothelial cells (BMECs). Their non-human origin also precludes the possibility to perform comparative studies in patient cells vs. patient-matched controls. 

Use of a single cell type such as Caco-2, MDCK or primary human BMECs limits the fidelity of these models, since we know that BBB function is dependent not only upon the BMECs but also their critical interactions with pericytes, immune cells, glial cells, and neural cells. 

Although not discussed in detail here, in vitro BBB models based on immortalized non-human brain endothelial cell lines (of porcine, bovine and rodent origin) are also widely used to study BBB function. These provide a stable source of cells with high yield and homogeneity, and exhibit certain features that may allow recapitulation of some native BBB properties; for example, they possess tight junctions and some BBB enzymatic activities, as well as express some BBB-specific cell surface markers. However, these cells are also limited by inevitable species differences compared to human cells, and studies have also demonstrated poor TEER for these cell lines, which reflects loose intercellular junctions and makes it difficult to rely on drug permeability measurements. 

Additional issues arise concerning the genetic purity of immortalized cell lines due to over engineering in laboratories, which can impact genomic stability. Additionally, numerous genetically diverse clones in circulation throughout labs all over the world inhibit interlaboratory data validation and comparisons. A similar limitation applies to primary human cells for which the source and genetic profile (e.g. genotypes) may be uncertain or spurious, which confounds assay results and data interpretation. In addition, limited cell numbers per donor sample may result in multi-donor pooling, resulting in mixed genetic materials as well as cells.

  1. hiPSCs – a practical alternative to primary and immortalized cell lines that is genetically well-defined.

Advances in human induced pluripotent stem cell (hiPSC) technologies have led to the emergence of novel hiPSC-derived BBB models in the last decade. The use of hiPSCs makes it possible to generate reproducible and scalable human-relevant cell resources. Importantly for human BBB studies, hiPSC-derived cell types  provide a uniform baseline that is of human origin, and the technology of reprogramming allows the cell types to be upscaled (10^9 cells per experiment). These are critical for preclinical drug development in the biotech-pharma industry. Furthermore, this allows for meaningful comparisons between researchers and for paired evaluations of “healthy vs. patient” cells during drug development.

The first report of a hiPSC-derived endothelial cell (hiEC) model of the BBB came from the University of Wisconsin–Madison (US) in 2012, when a group led by Eric V Shusta demonstrated that hiPSC-derived ECs (hiECs) acquired BBB properties when co-differentiated with neural cells that provide relevant cues, including those involved in the Wnt/β-catenin signaling pathway, which is critical for BBB development (4). The resulting iECs from the Shusta lab possessed many relevant brain attributes in terms of biomarker expressions, barrier function, and permeability assessment. 

Major challenges remained, however, for deriving the hiECs from a wide variety of human donor samples and scaling up the manufacturing to >10^9 cells per donor line per project. Scaling up the manufacturing of these cells is a critical step for scientists in the pharmaceutical development industry. They need billions of cells per batch (in a batch-to-batch, and donor-to-donor consistent manner) in order to use the cells in preclinical drug development and safety assessment studies. 

Recently (as of October 2022), Tempo Bioscience (“Tempo”) announced that they officially launched products for human blood-brain-barrier models: human iPSC-derived brain microvascular endothelial cells and human iPSC-derived pericytes. 

While it was no surprise to scientists worldwide that iBMECs and iPericytes could be derived from human iPSCs, it was a surprise that the Tempo-iBMECs and Tempo-iPericytes can be manufactured at scale and still retain their functional and genetic/biomarker characteristics appropriate for modeling the human BBB, including tight junction proteins, TEER measurements, co-culture adaptations between iBMECs/iPeri with iAstrocytes and other iPSC-derived neuronal/glial cell types. 

Besides modeling human BBB functions such as barrier and permeability, understanding the transport of any novel drug candidate – whether it is a small molecule, biologic, AAV, or CAR-T cell therapy – with or without any drug delivery vehicle is important for scientists in the pharmaceutical preclinical and clinical settings. 

In our next post in this series, we will discuss how assessing drug delivery across the BBB is critical for measuring toxicity and efficacy of all new CNS-targeted drugs. Did you know that ~99% of CNS drugs are plasma-bound and do NOT penetrate the human BBB? Stay tuned!


  1. Sonia, T.A., & Sharma, C.P. Experimental techniques involved in the development of oral insulin carriers.
  2. Lundquist S, Renftel M, Brillault J, Fenart L, Cecchelli R, Dehouck MP. Prediction of drug transport through the blood-brain barrier in vivo: a comparison between two in vitro cell models. Pharm Res. 2002;19(7):976-981. doi:10.1023/a:1016462205267.
  3. Gaush CR, Hard WL, Smith TF. Characterization of an established line of canine kidney cells (MDCK). Proc Soc Exp Biol Med. 1966;122(3):931-935. doi:10.3181/00379727-122-31293.
  4. Cecchelli R, Berezowski V, Lundquist S, et al. Modelling of the blood-brain barrier in drug discovery and development. Nat Rev Drug Discov. 2007;6(8):650-661. doi:10.1038/nrd2368.
  5. Lippmann ES, Azarin SM, Kay JE, et al. Derivation of blood-brain barrier endothelial cells from human pluripotent stem cells. Nat Biotechnol. 2012;30(8):783-791. doi:10.1038/nbt.2247.
  6. Stone NL, England TJ, O’Sullivan SE. A Novel Transwell Blood Brain Barrier Model Using Primary Human Cells. Front Cell Neurosci. 2019;13:230. Published 2019 Jun 6. doi:10.3389/fncel.2019.00230.
  7. Muruganandam A, Herx LM, Monette R, Durkin JP, Stanimirovic DB. Development of immortalized human cerebromicrovascular endothelial cell line as an in vitro model of the human blood-brain barrier. FASEB J. 1997;11(13):1187-1197. doi:10.1096/fasebj.11.13.9367354.
  8. Weksler BB, Subileau EA, Perrière N, et al. Blood-brain barrier-specific properties of a human adult brain endothelial cell line. FASEB J. 2005;19(13):1872-1874. doi:10.1096/fj.04-3458fje.

Karen O’Hanlon Cohrt is an independent Science Writer with a PhD in biotechnology from Maynooth University, Ireland (2011). After her PhD, Karen relocated to Denmark where she held postdoctoral positions in mycology and later in human cell cycle regulation, before moving to the world of drug discovery. Karen has been a full-time science writer since 2017, and has since then held numerous contract roles in science communication and editing spanning diverse topics including diagnostics, molecular biology, and gene therapy. Her broad research background provides the technical know-how to support scientists in diverse areas, and this in combination with her passion for learning helps her to keep abreast of exciting research developments as they unfold. Karen is currently based in Ireland, and you can follow her on Linkedin here.