QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: dmd.105.006437v1 34/11/1935    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Zhang, Y. Articles by Madhu, C. Search for Related Content PubMed PubMed Citation Articles by Zhang, Y. Articles by Madhu, C.

Porcine Brain Microvessel Endothelial Cells as an in Vitro Model to Predict in Vivo Blood-Brain Barrier Permeability

Departments of Pharmacokinetics, Dynamics and Metabolism (Y.Z., C.S.W.L., Y.Y., J.P., S.J., C.M.), Nonclinical Statistics (K.J.), and Drug Safety Evaluation (W.B., R.G.), Pfizer Global Research and Development, Ann Arbor, Michigan

(Received July 13, 2005; accepted August 3, 2006)

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
The objective of the study was to establish primary cultured porcine brain microvessel endothelial cells (PBMECs) as an in vitro model to predict the blood-brain barrier (BBB) permeability in vivo. The intercellular tight junction formation of PBMECs was examined by electron microscopy and measured by transendothelial electrical resistance (TEER). The mRNA expression of several BBB transporters in PBMECs was determined by reverse transcriptionpolymerase chain reaction analysis. The in vitro permeability of 16 structurally diverse compounds, representing a range of passive diffusion and transporter-mediated mechanisms of brain penetration, was determined in PBMECs. Except for the perfusion flow rate marker diazepam, the BBB permeability of these compounds was determined either in our laboratory or as reported in literature using in situ brain perfusion technique in rats. Results in the present study showed that PBMECs had a high endothelium homogeneity, an mRNA expression of several BBB transporters, and high TEER values. Culturing with rat astrocyte-conditioned medium increased the TEER of PBMECs, but had no effect on the permeability of sucrose, a paracellular diffusion marker. The PBMEC permeability of lipophilic compounds measured under stirred conditions was greatly increased compared with that measured under unstirred conditions. The PBMEC permeability of the 15 test compounds, determined under the optimized study conditions, correlated with the in situ BBB permeability with an r² of 0.60. Removal of the three system L substrates increased the r² to 0.89. In conclusion, the present PBMEC model may be used to predict or rank the in vivo BBB permeability of new chemical entities in a drug discovery setting.

Pharmaceutical companies have been actively pursuing various methods to accurately predict the brain penetration of new chemical entities (NCEs) at an early stage of drug discovery. Currently, several approaches are available for assessing brain penetration of NCEs, but they often pose a compromise between high throughput with low predictive potential and low throughput with high predictive potential. Passive diffusion permeability of NCEs may be predicted by high-throughput in silico modeling approaches based on the physicalchemical parameters (Liu et al., 2004). However, current in silico models do not account for metabolism, transporter-mediated processes, or any other drug-membrane or drug-protein interactions that may affect the ability of a drug to cross the BBB. With in vivo approaches [i.e., equilibrium brain-to-plasma distribution ratios (logBB) and BBB permeability-surface area product (logPS)] (Oldendorf, 1971; Smith, 1996), high-throughput models that can accommodate screening of a large number of compounds in early discovery are difficult to develop. Thus, a well characterized in vitro BBB cell model, which has the potential to account for the complex molecular interactions underlying BBB permeability and function as a moderate-throughput screening tool for NCEs would be highly useful in the discovery process.

A range of in vitro BBB cell models, including primary cultured/low passage brain microvessel endothelial cells and immortalized brain microvessel endothelial cell lines, have been generated by various academic laboratories. However, none of these models have been established as a widely used BBB permeability screen within the pharmaceutical industry. In recent years, the time- and resource-sparing advantages have made immortalized BBB cell models such as conditionally immortalized rat brain microvessel endothelial cells (TR-BBB cells) (Terasaki and Hosoya, 2001) attractive options. However, characterizations of immortalized BBB cell lines (Terasaki and Hosoya, 2001; Reichel et al., 2002) and initial investigations in our laboratory with TR-BBB cells (data not shown) suggest that immortalized cell cultures generally fail to form a sufficiently tight barrier for use in permeability studies. Primary cultured bovine brain microvessel endothelial cells (BBMECs), first developed by Bowman et al. (1983) and subsequently modified in Borchardt's laboratory (Audus and Borchardt, 1986a,b; Miller et al., 1992), have been the most extensively investigated in vitro BBB model. Primary cultures of BBMECs phenotypically maintain many in vivo BBB characteristics such as tight junction formation and expression of active transporter proteins and metabolic enzymes. However, despite observation of a strong qualitative correlation between in vitro BBMEC transcellular permeability and logBB (Guillot et al., 1993; Saheki et al., 1994; Cecchelli et al., 1999), a solid quantitative correlation derived from a structurally diverse set of compounds, representing passive and active brain penetration, is still lacking. Beginning in 1998, several groups have assessed the use of primary porcine brain microvessel endothelial cells (PBMECs) for in vitro permeability studies (Franke et al., 2000; Jeliazkova-Mecheva and Bobilya, 2003; Torok et al., 2003). It has been suggested that the PBMEC model may offer a more restrictive paracellular pathway compared with the BBMEC model (Franke et al., 2000). However, literature data regarding the tightness of cellular junctions of BBMEC or PBMEC models have been highly variable because of differences in cell isolation procedures, cell culture, and experimental conditions among laboratories (Fischer et al., 2000; Jeliazkova-Mecheva and Bobilya 2003; Torok et al., 2003).

The purposes of the current study were to develop a primary cultured PBMEC model under optimized study conditions and to evaluate its use as a moderate- to high-throughput screening tool to predict in vivo BBB transport of structurally diverse compounds with passive and/or active brain penetration. To achieve this goal, we first developed a PBMEC culture and evaluated how well the PBMECs retain the characteristics of in vivo BBB in terms of endothelial origin, expression of BBB transporters, and tight junction formation. After the initial characterization, we evaluated a number of culture and experimental variables to optimize study conditions. Lastly, the ability of the PBMEC model to predict in vivo BBB penetration was evaluated by comparing the transendothelial permeability measured in the PBMEC model with the BBB penetration determined by in situ brain perfusion for a set of 16 structurally diverse compounds, representing a range of passive diffusion and transporter-mediated mechanisms of brain penetration.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Materials and Reagents. [¹⁴C]Sucrose (600 mCi/mmol), [¹⁴C]diazepam (55 mCi/mmol), and [³H]dopamine (9 Ci/mmol) were obtained from Amersham Biosciences (Buckinghamshire, UK). [³H]Methotrexate (16.9 Ci/mmol), [¹⁴C]phenytoin (53.1 mCi/mmol), [¹⁴C]caffeine (53 mCi/mmol), [³H]vinblastine (5.9 Ci/mmol), [³H]metoprolol (60.4 Ci/mmol), and [³H]theophylline (13.3 Ci/mmol) were purchased from Moravek Biochemicals (Brea, CA). [³H]Mannitol (17 mCi/mmol) and [³H]taurocholic acid (2 Ci/mmol) were purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). [³H]Quinidine (20 Ci/mmol), [³H]testosterone (90 Ci/mmol), [³H]leucine (60 Ci/mmol), and [³H]phenylalanine (80 Ci/mmol) were obtained from American Radiolabeled Chemicals Inc. (St. Louis, MO). [³H]Gabapentin (44.15 Ci/mmol) was synthesized in-house.

Minimal essential medium (MEM), Ham's F-12 medium (F12), Dulbecco's modified Eagle's medium with low glucose, and rat-tail collagen were purchased from Fisher Scientific (St. Louis, MO). Fibronectin and equine serum were purchased from Sigma Chemical Co. (St. Louis, MO). Dispase and collagenase/dispase were obtained from Roche Diagnostics (Indianapolis, IN). Tissue culture Transwell inserts (24-mm diameter, 0.4-µm pore size) were obtained from Costar Corp. (Cambridge, MA). Side-by-side horizontal diffusion chambers, chamber clamps, and H1 stirrer were purchased from Perme-Gear, Inc. (Bethlehem, PA).

Selection of Compounds for Method Development. Sixteen structurally diverse compounds were selected for the evaluation of the present PBMEC model. As shown in Table 1, the compounds chosen are of diverse chemical classes, and represent both passive diffusion and carrier-mediated BBB uptake or efflux processes (McCall et al., 1982; Hargreaves and Pardridge, 1988; Su et al., 1995; Drion et al., 1996; Martel et al., 1996; Kusuhara et al., 1997; Kitazawa et al., 1998; Miller et al., 2000; Potschka and Loscher, 2001). In addition, the BBB permeability of these compounds determined by in situ brain perfusion ranged from low to high (Murakami et al., 2000; Liu et al., 2004). The logPS values of eight test compounds (taurocholic acid, methotrexate, theophylline, phenytoin, caffeine, dopamine, phenylalanine, and testosterone) were reported by Liu et al. (2004). The logPS values of sucrose, mannitol, vinblastine, and quinidine were reported by Murakami et al. (2000), and the logPS values of gabapentin, leucine, and metoprolol were determined in our laboratory. The highly permeable compound diazepam was used as the regional flow rate marker in brain perfusion studies (Liu et al., 2004); therefore, the logPS value of diazepam was not calculable.

Isolation of PBMECs. PBMECs were isolated from fresh porcine brains using a combination of enzyme digestion and ultracentrifugation approaches as described previously (Miller et al., 1992), with minor modifications. In brief, fresh porcine brains were obtained from a local slaughterhouse. The surface vessels and meninges of the brains were removed under aseptic conditions. The gray matter was collected by aspiration and then homogenized by sequential passing of the brain materials through a 1000-µm and then a 710-µm screen. The filtrate was digested with 12.5% (w/v) dispase for 3 h followed by centrifugation at 1570g for 10 min. The pellet was resuspended in 13% (w/v) dextran followed by a centrifugation at 9170g for 10 min. After centrifugation, the supernatant was discarded, and the dark red pellet was collected and subjected to further enzyme digestion in 0.52% (w/v) collagenase/dispase for 3.5 h. After the incubation, the endothelial cells were separated from cellular debris and red blood cells by Percoll gradient centrifugation at 1700g for 10 min. The average yield of PBMECs was approximately 10 x 10⁶ cells/brain. For cell preservation, cells were frozen in MEM/F12 medium supplemented with 10% dimethyl sulfoxide, 20% horse serum, 50 µg/ml gentamicin, 2.5 µg/ml amphotericin B, and 100 µg/ml heparin and stored in liquid nitrogen.

Isolation of Rat Astrocytes. Rat astrocytes were isolated from the cerebral cortex of newborn pups (1–2 days old) according to the method of McCarthy and de Vellis (1980). In brief, the cortices, free of meninges, were dissected in Hanks' balanced salt solution (Ca²⁺- and Mg²⁺-free) and incubated in 0.125% trypsin at 37°C for 10 min, followed by three washes with DMEM supplemented with 10% FBS, 1 U/ml penicillin, and 1 mg/ml streptomycin. After the washing, cells were resuspended in the same medium and filtered through a nylon mesh with a pore size of 100 µm. Filtered cells were seeded on poly-D-lysine-coated culture flasks at a density of 1.5 x 10⁵ cells/cm². On culture day 9, the oligodendrocytes and microglia were removed from the astrocyte cultures by vigorous shaking at a speed of 250 rpm at 37°C overnight. Astrocytes were seeded at a density of 2.4 x 10⁵ cells/cm² in poly-D-lysine-coated culture flasks for collection of astrocyte-conditioned medium (ACM). ACM was collected after 48 h of incubation and was sterile-filtered. The ACM was either used immediately or stored at –20°C for future use.

Culture of Cells. The freshly isolated or cryopreserved PBMECs were seeded (7.5 x 10⁴ cells/cm²) on collagen-coated, fibronectin-treated polycarbonate Transwell membranes and cultured in MEM/F12 medium supplemented with 10% horse serum, 50 µg/ml gentamicin, 2.5 µg/ml amphotericin B, and 100 µg/ml heparin. For PBMEC culture in the absence of ACM, the MEM/F12 growth medium was added to the top and bottom wells, and was renewed every other day until the cells reached confluence (typically 5–6 days). For PBMEC culture with ACM, the MEM/F12 growth medium from the bottom well of the Transwell plates was replaced by the ACM once the PBMECs reach near-confluence around day 4 in culture. Fresh MEM/F12 growth medium for PBMECs was added to the top well of the Transwell plates. The PBMECs were then incubated for another 2 to 4 days in the humidified 37°C incubator with 5% CO₂ supply before beginning the permeability studies.

Electron Microscopy for PBMECs. PBMEC monolayers grown in collagen-coated, fibronectin-treated Transwell inserts were fixed with 2% (w/v) paraformaldehyde and 2.5% (v/v) glutaraldehyde in 0.1 M sodium cacodylate (pH 7.4) containing 5% (w/v) sucrose. The cell monolayers were then post-fixed in 0.1 M sodium cacodylate buffer containing 1% (w/v) osmium tetroxide, 0.5% (w/v) potassium ferricyanide, and 5% (w/v) sucrose for 1 h and embedded. The cells were visualized through an FEI-Philips CM100 BioTwin electron microscope operated at 60 kV. Imaging was performed with a Kodak 4.2i digital camera from Advanced Microscopy Techniques, Inc. (Danvers, MA).

Histocytochemistry for PBMECs. PBMECs grown on LabTek glass chambered slides (Nalge Nunc International Corp., Naperville, IL) were fixed in ice-cold 80% methanol for 20 min and air-dried for 1 h. Nonspecific staining was blocked by incubation with serum-free protein blocker (DakoCytomation, Carpinteria, CA) for 30 min. The cells were first incubated with primary antibodies (1:100 dilution) of either rabbit polyclonal anti-von Willebrand factor (factor VIII) (DakoCytomation) or normal rabbit IgG as negative controls for 1 h at room temperature. Cells were then incubated with a secondary anti-rabbit biotinylated antibody (Vector Laboratories, Burlingame, CA) for another 45 min, followed by application of streptavidin-Alexa Fluor 488 (Molecular Probes, Eugene, OR) for 35 min. Cells were counterstained with the nuclear dye DAPI (Molecular Probes) for 10 min and then mounted with ProLong mounting medium (Molecular Probes). Cell staining was visualized using a Zeiss Axiovert 200 fluorescent microscope (Zeiss, Thornwood, NY) equipped with fluorescent filters for fluorescein isothiocyanate and DAPI visualization. Images for the two fluorescent channels were collected separately and combined using the Axiovision software package (Zeiss).

Reverse Transcription-Polymerase Chain Reaction (RT-PCR) in PBMECs. The mRNA expression of various transporters was determined by RT-PCR using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control. The primers used for each gene are listed in Table 2. Total RNA from confluent PBMEC monolayers was isolated using an RNeasy Mini Kit (QIAGEN, Valencia, CA) according to the manufacturer's instructions. RT-PCR was accomplished by using SuperScript One-Step RT-PCR with Platinum^R Taq (Invitrogen, Carlsbad, CA). The reaction of RT was carried out at 50°C for 30 min. For all transporters and the reference gene GAPDH, identical thermal cycling conditions for PCR were used: 2 min at 94°C to denature, followed by 35 cycles of 94°C for 15 s, 55°C for 30 s, and 68°C for 45 s. The products (20 µl) from the PCR amplification were separated on 2% agarose gel and stained with ethidium bromide to check product integrity. The identity of the selected positive RT-PCR products was verified with sequence analysis.

Transendothelial Electrical Resistance (TEER) Measurement with PBMECs. The TEER was used as a measure of tight junction formation by the PBMEC monolayers. The TEER value () was determined using an EndOhm chamber connected to an EVOM resistance meter (World Precision Instruments, Inc., Sarasota, FL) and reported as · cm² after correcting for the surface area of the membrane (4.5 cm²). The EndOhm system provided a more reproducible TEER measurement compared with the traditional "chopstick" electrode system. The variation of resistance measured from the same membrane insert was approximately 3 and 30 · cm² using the EndOhm and "chopstick" systems, respectively (data not shown). The TEER value of a blank membrane was subtracted from the apparent TEER values of membranes with cells to obtain the effective TEER value of the PBMEC monolayer. To use cells with ideal tight junction formation, TEER values were monitored from day 4 to day 13 for various PBMEC cultures. No permeability studies were performed when the TEER values were below 300 · cm².

In Vitro Permeability Studies in PBMECs. The in vitro permeability studies in PBMECs were performed as described by Mark and Miller (1999) with minor modifications. Confluent PBMECs were preincubated with assay buffer (122 mM NaCl, 25 mM Na₂CO₃, 10 mM D-glucose, 3 mM KCl, 1.2 mM MgSO₄·7H₂O, 0.4 mM K₂HPO₄, 1.4 mM CaCl₂, 10 mM HEPES, pH 7.4) for 30 min at 37°C. After preincubation, the filter membranes, with or without a PBMEC monolayer, were cut off from the Transwell inserts and placed between donor and receiver chambers in a side-by-side diffusion chamber. At time 0, 3 ml of assay buffer containing ¹⁴C-radiolabeled sucrose (1 µCi/ml) as a paracellular diffusion marker and ³H-radiolabled test compounds (0.2 µCi/ml) were added into the donor chamber facing the apical side of the cell monolayer. When using ¹⁴C-caffeine, ¹⁴C-phenytoin, or ¹⁴C-diazepam as test compounds, ³H-mannitol was used as the paracellular diffusion marker instead. The temperature of the medium in the donor and receiver chambers was maintained at 37°C by a continuous flow of water generated from a temperature-regulating circulating bath (Lauda E100; Brinkmann Instruments Inc., Westbury, NY). At 0, 5, 10, 20, and 40 min after adding test compound, 10-µl and 3-ml samples were taken from donor and receiver chambers, respectively. For the substrates of efflux transporters, including vinblastine, quinidine, and methotrexate, the last sample was taken at 60 min. The receiver chamber was immediately replaced with 3 ml of blank assay buffer after removal of each sample. The experiment was performed at 37°C under either unstirred or well stirred (600 rpm) conditions. Radioactivity levels of the samples collected from the donor (10 µl) and receiver (50 µl) chambers were determined using a liquid scintillation counter (Tri-Carb; Packard, Meriden, CT).

The apparent in vitro permeability coefficient of a compound (P_app) was obtained from the slope of the linear plot of compound flux into the receiver chamber over time according to the following equation:

where C_D,t is the concentration of the compound in the donor chamber at time t, M_R,t is the amount of compound in the receiver chamber at time t, and A is the surface area of the membrane exposed between the side-by-side chamber (0.636 cm²).

The effective permeability of a compound (P_e) was determined according to the following equation:

where P_m is the permeability of the compound across a collagen-coated, fibronectin-treated membrane without cells.

Because both P_app and P_m followed a normal distribution for the tested compounds, the standard error (SE) of the P_e was approximated using the Delta method as follows (Casella and Berger, 1990):

where d equals the number of days on which experiments were conducted to determine P_app and P_m.

View larger version (64K): [in this window] [in a new window]   FIG. 1. Phase contrast microscopy of primary cultured PBMECs on culture day 2 (a), day 4 (b), and day 6 (c) (original magnification 100x).

View larger version (48K): [in this window] [in a new window]   FIG. 2. Immunostaining of PBMEC monolayers with anti-von Willebrand factor VIII polyclonal antibody (green fluorescence) (a). The negative controls were incubated with normal rabbit IgG (b). Cell nuclei are counterstained with DAPI (blue fluorescence).

where A_br is the brain tissue concentration (dpm/g), C_pf is the perfusion fluid concentration (dpm/ml), T is the perfusion time (s), and V_v is the brain vascular volume (ml/g). K_in was estimated from the slope of the initial linear portion of A_br/C_pf versus T plot.

The permeability surface product (PS) (µl/s/g) was calculated from the following equation:

where F is the regional flow rate estimated from diazepam K_in data (70 µl/s/g) (Liu et al., 2004).

View larger version (99K): [in this window] [in a new window]   FIG. 3. Transmission electron microscopy of PBMECs grown on collagen-coated, fibronectin-treated chamber slides. Two adjacent cells with slightly overlapping membrane (a, original magnification 2100x; b, original magnification 11,000x). Two adjacent cells with direct contact (c, original magnification 3800x; d, original magnification 15,000x). Arrows indicate the tight junction formation.

	Results

Top Abstract Materials and Methods Results Discussion References

Immunocytochemical Characterization of PBMECs. The PBMECs demonstrated extensive positive immunostaining for factor VIII, an endothelial cell marker, predominantly in the cytoplasm, with a characteristic granular and diffused staining pattern (Fig. 2a). There was no fluorescent staining in the negative control, where primary anti-rabbit IgG was applied instead of anti-factor VIII antibody (Fig. 2b). Electron microscopic examination of confluent PBMECs showed that the cells grew into a confluent monolayer on top of the collagen/fibronectin-coated Transwell inserts with apparent intercellular tight junction formation (Fig. 3, a–d, indicated by arrows). The cells were linked to each other either by an overlapping membrane (Fig. 3, a and b) or by direct contact (Fig. 3, c and d). Regardless of the type of connection, the tight junctions were consistently localized toward the apical side of the cell membranes (Fig. 3, a–d, indicated by arrows). Furthermore, the in vitro cultured PBMECs resembled in vivo BMECs in that there were no fenestrations, few pinocytic vesicles, and an abundance of mitochondria.

View larger version (29K): [in this window] [in a new window]   FIG. 4. RT-PCR analysis of transporter mRNA and tight junction protein mRNA expression in total RNA samples prepared from confluent PBMECs. The 100-base pair (bp) DNA ladders are shown in lanes 1 and 13. Lane 2, ZO-1; lane 3, GLUT1; lane 4, LAT1; lane 5, LAT2; lane 6, MRP1; lane 7, MRP2; lane 8, MRP4; lane 9, MRP5; lane 10, MDR1; lane 11, BCRP; lane 12, GAPDH.

View larger version (18K): [in this window] [in a new window]   FIG. 5. TEER values of PBMEC monolayers on various days in cultures. Each data point represents the mean ± S.D. of six cell monolayers each day, and results are from 13 isolations. The bar indicates TEER at 300 · cm2, under which no permeability studies were performed.

TEER of PBMEC Monolayer. Before each transendothelial permeability study, the "tightness" of the PBMEC monolayer was assessed by TEER measurement. Figure 5 demonstrates the change of TEER values with duration of culture. The TEER values were slightly varied from one primary culture to another. The TEER values typically ranged from 300 to 550 · cm² on day 5 to day 9 postseeding. After 9 days in culture, the TEER values started to decline.

TEER values obtained from PBMECs cultured with ACM of one representative study are shown in Fig. 6. The results showed that culture with ACM significantly enhanced the TEER values of PBMECs by 10 to 25% on culture days 6 through 9, postseeding.

View larger version (14K): [in this window] [in a new window]   FIG. 6. Comparison of the TEER values between PBMECs treated with and without ACM. Each data point represents the mean ± S.D. of six cell monolayers each day in culture. *, p < 0.05, Student's t test.

View larger version (17K): [in this window] [in a new window]   FIG. 7. Comparison of average logPe in freshly isolated PBMECs with cyropreserved PBMECs by the TDI approach. Studies were performed under stirred conditions. Each data point represents the mean of at least three cell monolayers per experimental group. TDI 80% refers to the boundary such that at least 80% of the measurements are within 0.1 logPe unit of each other. TDI 95% refers to the boundary such that at least 95% of the measurements are within 0.16 log Pe unit of each other.

In Vitro-to-in Situ Correlation. Table 4 summarizes the in vitro permeability of the 16 selected compounds as determined by the PBMEC model. The PBMEC cultures demonstrated a large dynamic range, with P_e values ranging from 92 to 3702 x 10^–6 cm/s for the 16 compounds. In addition, the rank order of the PBMEC permeability was comparable with that of the in situ BBB penetration for the data set except for the three LAT substrates (leucine, gabapentin, and phenylalanine), which were underestimated by the in vitro PBMECs. The compounds that were highly permeable in situ (diazepam and testosterone) demonstrated a high permeability in the PBMEC model. The compounds with moderate permeability in situ (theophylline, quinidine, metoprolol, phenytoin, dopamine, and caffeine) were found to have a moderate permeability in PBMECs. Lastly, a similar trend emerged for the low-permeability compounds (sucrose, mannitol, taurocholic acid, methotrexate, and vinblastine). Furthermore, Fig. 8 depicts the correlation between the logP_e values measured in the PBMEC model and the logPS values observed in situ for the 15 test compounds (except for diazepam). The PBMEC logP_e correlated with the in situ logPS with an r² of 0.60 for the 15 compounds. When the three system L substrates were excluded from the correlation, the r² increased to 0.89.

View larger version (21K): [in this window] [in a new window]   FIG. 8. Correlation between in vitro logPe from PBMECs and logPS determined by in situ brain perfusion in rats. The logPe value represents the mean ± S.E. of at least six cell monolayers for each test compound. The full test set correlated with an r2 of 0.60. The r2 increased to 0.89 when the three system L substrates were excluded from the correlation (shown with the dotted line in the figure).

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
The main objective of the present study was to develop a simple yet robust in vitro BBB cell model to predict the net permeability (passive and active transported processes) of NCEs across the BBB in vivo.

Our initial characterization of the PBMEC model suggested that it maintained many important BBB properties. First, the extensive staining with factor VIII signified the homogeneity of the endothelial cells. Second, visualization of the intercellular tight junctions and the moderate to high TEER values (300–550 · cm²) indicated an acceptable "tightness" of the PBMEC monolayer. Third, the PBMECs expressed the mRNA of several BBB uptake and efflux transporters, including GLUT1, LAT1, MRP1, MRP4, MRP5, Pgp, and BCRP. Moreover, preliminary study of the functional activity of Pgp in the PBMECs showed that the basolateral-to-apical/apical-to-basolateral (B-to-A/A- to-B) permeability ratio of rhodamine 123, a Pgp substrate, was 1.73. Addition of a Pgp inhibitor, cyclosporin A, reduced the B-to-A/A- to-B ratio of rhodamine 123 to 1. The B-to-A/A-to-B permeability ratio of sucrose, the paracellular marker, was 1 under both conditions (data not shown). This apparent polarized transport of the Pgp substrates suggested that Pgp was functionally active and was predominantly located on the apical membrane in the PBMECs. Lastly, the PBMEC model took only 5 to 6 days to reach confluence compared with 10 to 12 days with the BBMEC model, which added to the advantage of the PBMECs as an in vitro BBB screening tool.

One of the continuing challenges to develop an in vitro BBB permeability model is the lack of standardized study conditions, which makes it difficult to interpret and compare models from laboratory to laboratory. Protocols in the literature describing various in vitro BBB models varied in many respects, including cell origins, cell isolation procedures, cell culture conditions, and permeability study conditions (Audus and Borchardt, 1986b; Rubin et al., 1991; Cecchelli et al., 1999; Franke et al., 2000; Jeliazkova-Mecheva and Bobilya, 2003). Therefore, the TEER and sucrose permeability values reported for the primary cultures of BBMECs or PBMECs were highly variable among different laboratories (Pardridge et al., 1990; Rubin et al., 1991; Jeliazkova-Mecheva and Bobilya, 2003; Torok et al., 2003) and difficult to compare. Thus, initial experiments were conducted in our laboratory to develop the optimal culture and permeability assay conditions for the present PBMEC model.

First, we tested the influence of a stirred versus unstirred condition on the permeability measurements in the PBMEC model. Physiologically, there is virtually no unstirred water layer (UWL) at the surface of the endothelium of the in vivo brain (Pardridge, 1991). However, many in vitro permeability studies were conducted under unstirred conditions (Pardridge et al., 1990; Rubin et al., 1991; Franke et al., 2000). It has been reported that in Caco-2 cells, the thickness of the UWL can be greater than 1000 µm under unstirred conditions, which is greater than the thickness of the UWL in the in vivo intestinal lumen (30–100 µm) (Lennernas, 1998). The increased thickness of the UWL in vitro conceivably presents an artificial, rate-limiting barrier to passive diffusion across the cell monolayer. In the present PBMEC model, stirring in both donor and receiver chambers (600 rpm) greatly increased the transendothelial flux of the moderate- and high-permeability compounds, but had little effect on the low-permeability compounds. In turn, the dynamic range (i.e., the ratio of P_e of high-permeability compounds to the P_e of low-permeability compounds) of PBMEC permeability was greatly increased. These data suggest that permeability studies performed under stirred conditions reduce the influence of the UWL and, therefore, better mimic the in vivo BBB environment compared with unstirred conditions.

Second, considering the close apposition of astrocytes to brain capillary endothelial cells in vivo, some researchers speculated that factor(s) produced by astrocytes are important for the establishment of a fully functional BBB in vitro (Raub et al., 1992). In fact, coculture with astrocytes or the culture of in vitro BBB cells in ACM has been shown to enhance the tight junction formation and decrease sucrose permeability in several in vitro BBB models (Cecchelli et al., 1999; Megard et al., 2002). However, these results varied from one laboratory to another. Torok et al. (2003) reported that ACM led to a 60% increase in TEER of their PBMEC model, but had no effect on the paracellular permeability of sucrose. In addition, it was found that there was only a slight (14%) reduction of the sucrose permeability in the presence of ACM in a BBMEC model (Pardridge et al., 1990). In the present study, PBMECs cultured with ACM resulted in only a 10 to 25% enhancement of TEER values and showed no effect on the permeability of sucrose. Although still lower than the TEER value in vivo (>1000 · cm²) (Butt et al., 1990), the TEER values of the PBMEC model (300–550 · cm²) were higher than that reported for most other solo-cultures of PBMECs (400 · cm²) (Fischer et al., 2000; Jeliazkova-Mecheva and Bobilya, 2003; Torok et al., 2003) and BBMECs (80–140 · cm²) (Rubin et al., 1991), as well as for TR-BBB cells (5–20 · cm²) (unpublished observation in-house). Although the paracellular restrictiveness of the PBMEC model may still warrant further improvement to be close to that reported in vivo [e.g., in vitro permeability of sucrose of 90 x 10^–6 cm · s^–1 much higher than that estimated in vivo (0.03–0.1 x 10^–6 cm · s^–1) (Levin, 1980; Rechthand et al., 1987)], a much tighter monolayer may not be critical for the PBMECs to serve as a BBB permeability screen for assessing low, medium, and high brain penetration. Furthermore, the advantages gained under influence of astrocytes may be offset by a loss of culture precision since concentrations of astrocyte-related factors may differ from study to study.

Third, the PBMEC model from the current study remains viable after more than eight months of storage in liquid nitrogen without significant loss of functionality. The average logP_e of the 16 compounds tested in fresh PBMECs correlated well with those measured from cryopreserved cells, indicating both freshly isolated and cryopreserved PBMECs can be used to perform permeability studies of NCEs.

Upon defining the study conditions in the current PBMEC model (non-ACM culture, stirred condition, and using either fresh or cryopreserved cells), the primary goal of the present study was to evaluate whether this PBMEC model could predict the in vivo BBB permeability of structurally diverse compounds despite the mechanism of BBB penetration. Correlation between in situ or in vivo permeability and in vitro permeability from primary cultured BMEC models has been reported for a variety of compounds, including small molecules (Guillot et al., 1993; Saheki et al., 1994; Cecchelli et al., 1999) and proteins (Pardridge et al., 1990). However, most previous studies tested either a small number of passive diffusion compounds (Saheki et al., 1994; Cecchelli et al., 1999) or a set of structural analogs (Guillot et al., 1993). In the present study, eight of the compounds in the test set are known to be influenced by active uptake or efflux transport processes. A strong correlation (r² = 0.89) between the in vitro logP_e and the in situ logPS was observed with the three LAT substrates excluded. Consistent with this strong correlation, the rank order of PBMEC permeability of these compounds, except for the system L substrates, was comparable to that of the in situ BBB permeability in rats, with sucrose, mannitol, taurocholic acid, methotrexate, and vinblastine at the low end of the permeability range, theophylline, quinidine, phenytoin, metoprolol, dopamine, and caffeine in the moderate permeability range, and testosterone and diazepam at the high end of the permeability range. Moreover, this is the first study (to our knowledge) to demonstrate a significant quantitative correlation between the in vitro and in situ permeabilities of compounds that are known substrates of active efflux transporters at the BBB, including Pgp (vinblastine and quinidine) and MRP (taurocholic acid, phenytoin, and methotrexate). These data provide clear evidence of the net activity of major BBB transporters in the PBMEC model.

However, the PBMECs underestimated the in situ BBB permeability for the three LAT substrates (phenylalanine, leucine, and gabapentin). One of the possible explanations is species difference in LAT expression level between rat and porcine at the BBB. Another explanation for this discrepancy may be down-regulation of LAT expression and/or function in vitro, which has been suggested by Pardridge et al. (1990). Preliminary data in our laboratory have demonstrated that the uptake of radiolabeled leucine into the PBMEC monolayers is a saturable process with an apparent K_m of 170 µM, which is close to that reported both in vivo and in vitro (Cardelli-Cangiano et al., 1981; Audus and Borchardt, 1986a). However, RT-PCR analysis shows mRNA expression of both LAT1 and LAT2 in the PBMEC model. In vivo, LAT1 protein is highly expressed in the BBB with little, if any, expression of LAT2 (Kageyama et al., 2000). The apparent expression of LAT2 in the current PBMEC model may suggest modulation of LAT expression in the PBMECs in vitro. Nevertheless, although amino acid transporter activity would be a beneficial feature of the PBMEC model, it may not be essential for the potential application of the PBMEC model as a screening tool for the majority of non-amino acid central nervous system compounds.

In summary, our data show that the PBMEC model maintains the complexities of the in vivo BBB such as complex tight junctions, moderate to high electrical resistance, and mRNA expression of active transporters. Most importantly, the current PBMEC model demonstrates the ability to predict the net passive and transporter-mediated BBB penetration with high success, medium throughput, and high reproducibility, thus allowing the PBMEC model to be used as an effective BBB permeability screening tool in pharmaceutical research.

	Acknowledgments

 
We thank Dr. Xinrong Liu from Groton PDM for assisting in the transfer of the in situ brain perfusion assay to our site. We also thank Dr. Susan Buist and Dr. Joe Ware for reviewing and providing helpful comments during the preparation of the manuscript.

	Footnotes

 
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.

doi:10.1124/dmd.105.006437.

¹ Current affiliation: Drug Metabolism and Biopharmaceutics, Incyte Corporation, Wilmington, DE.

² Current affiliation: Drug Metabolism and Pharmacokinetics, Abbott Bioresearch Center, Worcester, MA.

Address correspondence to: Dr. Cheryl S. W. Li, Department of Pharmacokinetics, Dynamics & Metabolism, Global Research & Development, Pfizer Inc., 2800 Plymouth Road, Ann Arbor, MI 48105. E-mail: cheryl.li{at}pfizer.com

	References

Top Abstract Materials and Methods Results Discussion References

 


Audus KL and Borchardt RT (1986a) Characteristics of the large neutral amino acid transport system of bovine brain microvessel endothelial cell monolayers. J Neurochem 47: 484–488.[Medline]

Audus KL and Borchardt RT (1986b) Characterization of an in vitro blood-brain barrier model system for studying drug transport and metabolism. Pharm Res (NY) 3: 81–87.

Bowman PD, Ennis SR, Rarey KE, Betz AL, and Goldstein GW (1983) Brain microvessel endothelial cells in tissue culture: a model for study of blood-brain barrier permeability. Ann Neurol 14: 396–402.[CrossRef][Medline]

Butt AM, Jones HC, and Abbott NJ (1990) Electrical resistance across the blood-brain barrier in anaesthetized rats: a developmental study. J Physiol 429: 47–62.[Abstract/Free Full Text]

Cardelli-Cangiano P, Cangiano C, James JH, Jeppsson B, Brenner W, and Fischer JE (1981) Uptake of amino acids by brain microvessels isolated from rats after portacaval anastomosis. J Neurochem 36: 627–632.[CrossRef][Medline]

Casella G and Berger R (1990) Statistical Inference. Duxbury Press, Belmont, CA.

Cecchelli R, Dehouck B, Descamps L, Fenart L, Buee-Scherrer VV, Duhem C, Lundquist S, Rentfel M, Torpier G, and Dehouck MP (1999) In vitro model for evaluating drug transport across the blood-brain barrier. Adv Drug Deliv Rev 36: 165–178.[CrossRef][Medline]

Drion N, Lemaire M, Lefauconnier JM, and Scherrmann JM (1996) Role of P-glycoprotein in the blood-brain transport of colchicine and vinblastine. J Neurochem 67: 1688–1693.[Medline]

Fischer S, Wobben M, Kleinstuck J, Renz D, and Schaper W (2000) Effect of astroglial cells on hypoxia-induced permeability in PBMEC cells. Am J Physiol 279: C935–C944.

Franke H, Galla H, and Beuckmann CT (2000) Primary cultures of brain microvessel endothelial cells: a valid and flexible model to study drug transport through the blood-brain barrier in vitro. Brain Res Brain Res Protoc 5: 248–256.[CrossRef][Medline]

Guillot F, Misslin P, and Lemaire M (1993) Comparison of fluvastatin and lovastatin blood-brain barrier transfer using in vitro and in vivo methods. J Cardiovasc Pharmacol 21: 339–346.[Medline]

Hargreaves KM and Pardridge WM (1988) Neutral amino acid transport at the human blood-brain barrier. J Biol Chem 263: 19392–19397.[Abstract/Free Full Text]

Jeliazkova-Mecheva VV and Bobilya DJ (2003) A porcine astrocyte/endothelial cell co-culture model of the blood-brain barrier. Brain Res Brain Res Protoc 12: 91–98.[CrossRef][Medline]

Kageyama T, Nakamura M, Matsuo A, Yamasaki Y, Takakura Y, Hashida M, Kanai Y, Naito M, Tsuruo T, Minato N, et al. (2000) The 4F2hc/LAT1 complex transports L-DOPA across the blood-brain barrier. Brain Res 879: 115–121.[CrossRef][Medline]

Kitazawa T, Terasaki T, Suzuki H, Kakee A, and Sugiyama Y (1998) Efflux of taurocholic acid across the blood-brain barrier: interaction with cyclic peptides. J Pharmacol Exp Ther 286: 890–895.[Abstract/Free Full Text]

Kusuhara H, Suzuki H, Terasaki T, Kakee A, Lemaire M, and Sugiyama Y (1997) P-glycoprotein mediates the efflux of quinidine across the blood-brain barrier. J Pharmacol Exp Ther 283: 574–580.[Abstract/Free Full Text]

Lennernas H (1998) Human intestinal permeability. J Pharm Sci 87: 403–410.[CrossRef][Medline]

Levin VA (1980) Relationship of octanol/water partition coefficient and molecular weight to rat brain capillary permeability. J Med Chem 23: 682–684.[CrossRef][Medline]

Lin LI (2000) Total deviation index for measuring individual agreement with applications in laboratory performance and bioequivalence. Stat Med 19: 255–270.[CrossRef][Medline]

Liu X, Tu M, Kelly RS, Chen C, and Smith BJ (2004) Development of a computational approach to predict blood-brain barrier permeability. Drug Metab Dispos 32: 132–139.[Abstract/Free Full Text]

Mark KS and Miller DW (1999) Increased permeability of primary cultured brain microvessel endothelial cell monolayers following TNF-alpha exposure. Life Sci 64: 1941–1953.[CrossRef][Medline]

Martel CL, Mackic JB, Adams JD Jr, McComb JG, Weiss MH, and Zlokovic BV (1996) Transport of dopamine at the blood-brain barrier of the guinea pig: inhibition by psychotropic drugs and nicotine. Pharm Res (NY) 13: 290–295.

McCall AL, Millington WR, and Wurtman RJ (1982) Blood-brain barrier transport of caffeine: dose-related restriction of adenine transport. Life Sci 31: 2709–2715.[CrossRef][Medline]

McCarthy KD and de Vellis J (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. J Cell Biol 85: 890–902.[Abstract/Free Full Text]

Megard I, Garrigues A, Orlowski S, Jorajuria S, Clayette P, Ezan E, and Mabondzo A (2002) A co-culture-based model of human blood-brain barrier: application to active transport of indinavir and in vivo-in vitro correlation. Brain Res 927: 153–167.[CrossRef][Medline]

Miller DS, Nobmann SN, Gutmann H, Toeroek M, Drewe J, and Fricker G (2000) Xenobiotic transport across isolated brain microvessels studied by confocal microscopy. Mol Pharmacol 58: 1357–1367.

Miller DW, Audus KL, and Borchardt RT (1992) Application of cultured bovine brain endothelial cells in the study of the blood-brain barrier. J Tissue Culture Methods 14: 217–224.[CrossRef]

Murakami H, Takanaga H, Matsuo H, Ohtani H, and Sawada Y (2000) Comparison of blood-brain barrier permeability in mice and rats using in situ brain perfusion technique. Am J Physiol 279: H1022–H1028.

Oldendorf WH (1971) Brain uptake of radiolabeled amino acids, amines, and hexoses after arterial injection. Am J Physiol 221: 1629–1639.[Free Full Text]

Pardridge WM (1983) Brain metabolism: a perspective from the blood-brain barrier. Physiol Rev 63: 1481–1535.[Free Full Text]

Pardridge WM (1991) Peptide Drug Delivery to the Brain. Raven Press, New York.

Pardridge WM, Triguero D, Yang J, and Cancilla PA (1990) Comparison of in vitro and in vivo models of drug transcytosis through the blood-brain barrier. J Pharmacol Exp Ther 253: 884–891.[Abstract/Free Full Text]

Potschka H and Loscher W (2001) Multidrug resistance-associated protein is involved in the regulation of extracellular levels of phenytoin in the brain. Neuroreport 12: 2387–2389.[CrossRef][Medline]

Raub TJ, Kuentzel SL, and Sawada GA (1992) Permeability of bovine brain microvessel endothelial cells in vitro: barrier tightening by a factor released from astroglioma cells. Exp Cell Res 199: 330–340.[CrossRef][Medline]

Rechthand E, Smith QR, and Rapoport SI (1987) Transfer of nonelectrolytes from blood into peripheral nerve endoneurium. Am J Physiol 252: H1175–H1182.

Reichel A, Abbott NJ, and Begley DJ (2002) Evaluation of the RBE4 cell line to explore carrier-mediated drug delivery to the CNS via the L-system amino acid transporter at the blood-brain barrier. J Drug Target 10: 277–283.[CrossRef][Medline]

Rubin LL, Hall DE, Porter S, Barbu K, Cannon C, Horner HC, Janatpour M, Liaw CW, Manning K, Morales J, et al. (1991) A cell culture model of the blood-brain barrier. J Cell Biol 115: 1725–1735.[Abstract/Free Full Text]

Saheki A, Terasaki T, Tamai I, and Tsuji A (1994) In vivo and in vitro blood-brain barrier transport of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors. Pharm Res (NY) 11: 305–311.

Schinkel AH, Wagenaar E, Mol CA, and van Deemter L (1996) P-glycoprotein in the blood-brain barrier of mice influences the brain penetration and pharmacological activity of many drugs. J Clin Investig 97: 2517–2524.[Medline]

Smith QR (1991) The blood-brain barrier and the regulation of amino acid uptake and availability to brain. Adv Exp Med Biol 291: 55–71.[Medline]

Smith QR (1996) Brain perfusion systems for studies of drug uptake and metabolism in the central nervous system. Pharm Biotechnol 8: 285–307.[Medline]

Su TZ, Lunney E, Campbell G, and Oxender DL (1995) Transport of gabapentin, a gamma-amino acid drug, by system l alpha-amino acid transporters: a comparative study in astrocytes, synaptosomes, and CHO cells. J Neurochem 64: 2125–2131.[Medline]

Terasaki T and Hosoya K (2001) Conditionally immortalized cell lines as a new in vitro model for the study of barrier functions. Biol Pharm Bull 24: 111–118.[CrossRef][Medline]

Torok M, Huwyler J, Gutmann H, Fricker G, and Drewe J (2003) Modulation of transendothelial permeability and expression of ATP-binding cassette transporters in cultured brain capillary endothelial cells by astrocytic factors and cell-culture conditions. Exp Brain Res 153: 356–365.[CrossRef][Medline]

Zhang Y, Han H, Elmquist WF, and Miller DW (2000) Expression of various multidrug resistance-associated protein (MRP) homologues in brain microvessel endothelial cells. Brain Res 876: 148–153.[CrossRef][Medline]

This article has been cited by other articles:

				X. Liu, C. Chen, and B. J. Smith Progress in Brain Penetration Evaluation in Drug Discovery and Development J. Pharmacol. Exp. Ther., May 1, 2008; 325(2): 349 - 356. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: dmd.105.006437v1 34/11/1935    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar