Evaluation of the MDR-MDCK cell line as a permeability screen for the blood–brain barrier

https://doi.org/10.1016/j.ijpharm.2004.10.007Get rights and content

Abstract

The objectives of this study were to (1) characterize MDR-MDCK monolayers as an in vitro model to predict brain uptake potential; (2) examine the ability of MDR-MDCK monolayers to identify the brain uptake potential of compounds that interact with P-glycoprotein (P-gp). The study measured the bi-directional transport of 28 compounds across MDR-MDCK monolayers. The brain uptake of a subset of the compounds was determined in the rat brain perfusion model. Drug concentrations were analyzed by LC–MS–MS. CNS-positive drugs exhibited absorptive permeability coefficients (Papp, A–B) values ranging from 3.4 × 10−6 to 20.2 × 10−6 cm/s; whereas CNS-negative drugs showed Papp (A–B) ranging from 0.03 × 10−6 to 0.83 × 10−6 cm/s. Inhibition of P-gp by cyclosporin A (CsA) significantly reduced secretory flux of compounds known to be P-pg substrates, but only enhanced the absorptive flux of compounds with high efflux ratio (>100). In vitro results were confirmed by brain perfusion studies on selected compounds. MDR-MDCK monolayers can be used to classify compounds into CNS-positive or CNS-negative based on the permeability coefficients (Papp, A–B). Under our experimental conditions, compounds with Papp (A–B) >3 × 10−6 cm/s have high brain uptake potential; compounds with Papp (A–B) < 1 × 10−6 cm/s are unable to penetrate the blood–brain barrier (BBB); the brain uptake of compounds with Papp (A–B) < 1 × 10−6 cm/s and a P-gp-mediated efflux ratio of >100 may be enhanced by inhibiting P-gp.

Introduction

The primary interfaces between the central nervous system (CNS) and the systemic circulation are the blood–brain barrier (BBB), formed by the cerebral capillary endothelium, and the blood–cerebrospinal fluid (CSF) barrier, formed by the choroid plexus epithelium (Bendayan et al., 2002). The barrier at each site consists of a continuous single layer of cells joined by tight junctions that provide a highly regulated environment for the brain to function normally. In addition, expression of efflux proteins (i.e., P-glycoprotein (P-gp)) predominately localized in the luminal membranes of endothelial cells further restricts the entry of lipophilic compounds into the CNS (Lee et al., 2001). From a drug delivery point of view, the most desirable outcome is to enhance the BBB penetration by CNS drugs for maximal intended pharmacological effects, while reducing the BBB penetration by non-CNS drugs for minimal adverse neurological effects. Therefore, it is important to select drug candidates possessing desirable brain uptake potential.

Progress has been made in both in vivo and in vitro methodologies to study drug transport across the BBB. In vivo BBB experiments can provide valuable insight on drug permeation across the BBB, such as regional drug distribution, but these studies are laborious and require complicated analytical methods to measure plasma or brain drug concentrations. Thus, various in vitro systems have been used for studying BBB permeability. Brain microvessel endothelial cells, either primary cultures or cell lines, have been investigated from various mammalian species (Gumbleton and Audus, 2001).

There are several drawbacks associated with the use of primary cell culture systems for BBB permeability screening, including (a) time and labor associated with cell isolation and (b) batch-to-batch variability (Gumbleton and Audus, 2001). Brain endothelial cell lines provide a stable source with high yield and homogeneity. Thus, immortalized cell lines from bovine (Otis et al., 2001), rat (Yang and Aschner, 2003) and porcine (Franke et al., 2000) origin have been established. These cell lines form intercellular junctions, retain BBB enzymatic activities, and express BBB-specific cell surface markers. However, an important limitation of these cell lines is their low transendothelial electrical resistance (TEER). For example the measured TEER of bovine brain microvascular endothelial cells (BBMEC) has been reported in the range of 160–200 Ω cm2 (Raub et al., 1992). In contrast, the TEER of tight junctions in vivo is 2000 Ω cm2 or above (Crone and Olesen, 1982). Low TEER reflects loose intercellular junctions and is associated with high, passive paracellular diffusion. The permeability coefficients for sucrose were 10 × 10−6 cm/s in BBMEC (Raub et al., 1992) and 214 × 10−6 cm/s in the RBE4 cell line, derived from rat brain microvascular endothelial cells (Rist et al., 1997). Both systems lack sufficient paracellular restriction to be representative models of the BBB. Caco-2 cells are widely employed to predict oral absorption potential (Hidalgo, 2001); however, Caco-2 cell permeability did not predict in vivo BBB transport (Lundquist et al., 2002, Faassen et al., 2003).

In addition, Di et al. (2003) put forward parallel artificial membranes coated with porcine brain lipids (PAMPA-BBB) as a high throughput permeability assay for BBB. Also, computational models to predict BBB permeability of compounds were developed (Luco, 1999, Liu et al., 2004). However, lack of ability to integrate the effects of protein carriers in these processes greatly limits their use as reliable in vitro models for BBB permeability. This is particularly true if we consider that a high percentage of compounds interact with membrane transporters (Dresser et al., 2001).

Although the utility of MDR-MDCK, MDCK-II transfected with the human MDR1 gene, as a BBB permeability model has been proposed (Gumbleton and Audus, 2001), its potential utility cannot be assumed but needs to be demonstrated experimentally. First, MDR-MDCK cells are not endothelial but epithelial cells. Second, they were derived from dog and not human. Third, they are from kidney and not from brain capillaries. Mahar Doan et al. (2002) carried out pioneering work to investigate the utility of MDR-MDCK as a BBB model: passive permeability and efflux ratio of 93 drugs. A pitfall of this study was that they used the classification of drugs into CNS-indicated or non-CNS-indicated as the reference to judge the accuracy of the MDR-MDCK model to predict CNS uptake. In other words, CNS-indicated drugs were assumed to penetrate the BBB and non-CNS-indicated drugs were assumed not to penetrate the brain. However, these assumptions might be dangerous. Non-CNS-indicated drugs might have substantial brain uptake potential. CNS-indicated drugs might need only minimal brain uptake. CNS uptake potential may be better categorized based on in vivo brain uptake rather than therapeutic indications. Thus, in the present study a set of compounds were selected based on data obtained from in vivo or in situ rat or mice brain uptake experiments (Table 1).

The objective of this study was to assess MDR-MDCK cells as a potential in vitro brain-uptake screening model. The study was carried out in two parts. In the first part, the bi-directional permeability coefficients of marketed drugs that are known to penetrate the BBB were measured to determine whether there was a qualitative correlation between MDR-MDCK cell permeability and in vivo CNS activities. In the second part, the role of P-gp as a determinant of the BBB penetration potential classification of drugs that interact with P-gp was examined.

Section snippets

Materials

All tested compounds were purchased from Sigma (St. Louis, MO, USA). The selection of compounds was based on available clinical indications, i.e., information that supports brain penetration, such as clinical site of effect and in vivo animal experiments or known absence of brain penetration (Table 1). MDR-MDCK cells were obtained from NIH (Bethesda, MD, USA) and maintained in Minimum essential Eagle's medium containing 2 mM l-glutamine, 20 mM sodium bicarbonate, 0.1 mM non-essential amino acids,

Results

The compounds used in this study were divided into two categories, i.e., CNS-positive and CNS-negative. CNS-positive compounds include drugs whose site of pharmacological effects is located in the CNS and compounds known to cross the BBB (Table 1). CNS-negative compounds are those whose pharmacological effects are not in the CNS, or are known not to cross the BBB. The first observation was that CNS-positive compounds in Table 2 exhibited much higher Papp (A–B) values ranging from 3.4 × 10−6 to

Discussion

Finding a suitable in vitro model to study brain drug delivery has remained one of the most formidable challenges in the pharmaceutical industry. It is not easy to simulate in vivo physiological characteristics of the BBB, such as high expression of P-gp and other efflux proteins, which might be responsible for low drug permeation into the brain. MDCK cells transfected with the P-gp gene exhibited high transepithelial electrical resistance (in the range of 1800–2200 Ω cm2). There has been no

Conclusion

Among the drugs tested in the study, the CNS-positive drugs showed high absorptive transport across the MDR-MDCK cells, whereas CNS-negative drugs showed low absorptive transport. The distinction between these two classes of drugs is obvious and there is no overlap: the passive Papp (A–B) of CNS-positive compounds is greater than 3 × 10−6 cm/s and Papp (A–B) of CNS-negative compounds is less than 1 × 10−6 cm/s. The data suggest that the MDR-MDCK cell line could be used as a quick BBB model to aid

Acknowledgements

The authors would like to thank Amy Brooks, Michael Furlong and Karen Elder-Doucette for their technical assistance. No source of fund is provided for the present study.

Reference (43)

  • P.R. Twentyman

    Cyclosporins as drug resistance modifiers

    Biochem. Pharmacol.

    (1992)
  • M. Uhr et al.

    Penetration of amitriptyline, but not of fluoxetine, into brain is enhanced in mice with blood-brain barrier deficiency due to mdr1a P-glycoprotein gene disruption

    Neuropsychopharmacology

    (2000)
  • G. Watt et al.

    Reversal of drug-resistant falciparum malaria by calcium antagonists: potential for host cell toxicity

    Trans. R. Soc. Trop. Med. Hyg.

    (1990)
  • J. Yang et al.

    Developmental aspects of blood–brain barrier (BBB) and rat brain endothelial (RBE4) cells as in vitro model for studies on chlorpyrifos transport

    Neurotoxicology

    (2003)
  • N.D. Barber et al.

    Comparison of the actions of centrally and peripherally administered clonidine and guanfacine in the rabbit: investigation of the differences

    Br. J. Pharmacol.

    (1982)
  • R. Bendayan et al.

    Functional expression and localization of P-glycoprotein at the blood brain barrier

    Microsc. Res. Tech.

    (2002)
  • J. Canena et al.

    Distribution of bismuth in the rat after oral dosing with ranitidine bismuth citrate and bismuth subcitrate

    J. Pharm. Pharmacol.

    (1998)
  • D.J. Castro et al.

    Biodistribution of rhodamine-123 in nude mice heterotransplanted with human squamous cell carcinomas

    Laryngoscope

    (1992)
  • R.G. del Moral et al.

    P-glycoprotein: a new mechanism to control drug-induced nephrotoxicity

    Exp. Nephrol.

    (1998)
  • I.J. Hidalgo

    Assessing the absorption of new pharmaceuticals

    Curr. Top. Med. Chem.

    (2001)
  • D.M. Killian et al.

    Modulating blood–brain barrier interactions of amino acid-based anticancer agents

    Drug Delivery

    (2000)
  • Cited by (0)

    View full text