Skip to main content
. 2021 Mar 15;8(11):2002085. doi: 10.1002/advs.202002085

Table 6.

Comparative features of in vitro models different BBB models

Preclinical development of CNS drug targets in human diseases involves complementary in vitro modeling methods. Summary advantages and disadvantages of variants of these models of the BBB are herein reviewed.
Advantages Disadvantages
In vivo
  • +

    Animal models are considered the gold standard of preclinical predictive tools, but ≈80% of drug candidates identified in this way later fail in clinical trial

  • Animal models are analogous in many of the cell, molecular, organ system functions and connections found in man, but may not accurately reflect the human condition

  • Relatively low throughput, expensive, and limited by ethical considerations

In vitro: static (transwell and organoid)
  • +

    Includes the barrier phenotypes induced by cellular contacts of EC with a basement membrane, each other, and if included in coculture various other cells of the NVU, notably pericytes/vascular smooth muscle cells and astrocytes

  • Omits the barrier phenotypes that are induced and maintained by the shear forces imparted by circulating plasma on ECs

Transwell A suspended semi‐permeable platform on which endothelial monolayer grown; with or without cocultured cells on abluminal surface or on base of well into which insert placed
Monolayer (transwell)
  • +

    Commonly used (body of comparative work is available)

  • +

    Convenience of treatment and sampling of luminal and abluminal compartments; ease of tissue sourcing

  • +

    Robust barrier properties of animal (porcine or bovine) brain EC monolayers are an impressive option

  • Human monolayers have poor barrier properties by comparison

  • Omit the cellular contacts made by EC with other cells of the NVU, notably pericytes/vascular smooth muscle cells and astrocytes that determine barrier phenotype

  • Omits the barrier phenotypes that are induced and maintained by the shear forces imparted by circulating plasma on ECs

CO‐/multi‐culture (transwell)
  • +

    Can achieve good correlation between in vitro and in vivo drug permeability, validating predictive value, but cellular composition critical: Presence of cocultured glial cells in the abluminal compartment better mimicked the effect of brain tissue binding that occurs in vivo

  • Omits direct cellular contacts that induce barrier properties (that are present in 3D spheroid organoids or microfluidic organoids); also, does not mimic shear forces of fluidic systems

Multi‐culture variants (primary, immortalized, and renewable stem cells)

Features relevant to all in vitro platforms—transwell, organoid, and organ‐on‐a‐chip

  • Primary cells

  • +

    Differentiated phenotype; clinically relevant if human

  • Difficult to source human cells.

  • Brief use (limited growth)

  • Immortalized (cell lines)

  • +

    Convenience of renewable cells; of which commercial banks available

  • Lack of polarization–differentiation state might not reflect normal

  • Generally, well characterized, but susceptible to genetic changes over multiple passages

  • Renewable stem cells

  • +

    Renewable, sustainably available (cells can be frozen/grown up)

  • +

    Human cells are “induced,” not requiring placental material

  • +

    Can avoid immunogenic mismatch if all cells derived from same progenitor

  • +

    Self‐aggregating, thought to recapitulate cues for development of barrier phenotype

  • Requires technical expertise; complexity; delay in induction of phenotypes

Spheroid/organoid

Multicellular, self‐assemblage of brain cells enveloped in endothelia

  • +

    Relatively new model, but now established and becoming validated

  • +

    Available in variants of complexity, using primary, immortalized or “renewable” stem (isolated or induced) human cells

  • +

    Includes the barrier phenotypes induced by direct cellular contacts

  • Omits the barrier phenotypes induced and maintained by shear forces

Dynamic in vitro models (“organ‐on‐a‐chip”)
  • +

    The fullest representation of the complex cellular and shear force interactions that establish and sustain barrier properties

  • Relatively low throughput

In silico
  • +/‐

    Models involve computation based on physiochemical and physiological parameters derived from/requiring validation by in vivo/in vitro studies

  • +/‐

    Provide leads; need validation