• blood–brain barrier;
  • glucocorticoids;
  • in vitro model;
  • puromycin


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In vitro blood–brain barrier (BBB) models using primary rat brain microvessel endothelial cells (BMEC) are often hampered by a lack of culture purity and poor barrier properties. To address these problems, the translation inhibitor puromycin was used to purify rat BMEC cultures. BMEC purities of 99.8% were routinely attained using puromycin treatment, and this technique proved to be far superior to other purification methods of similar difficulty. In contrast to cultures without puromycin treatment, purity of puromycin-treated cultures was unaffected by initial seeding density. Next, rat BMEC monolayer transendothelial electrical resistance (TEER) was increased by glucocorticoid treatment with either corticosterone (CORT) or hydrocortisone (HC), and a corresponding decrease in monolayer permeability to small molecules was observed. Importantly, cultures treated with both puromycin and glucocorticoid attained significantly higher TEER values (CORT 168 ± 13 Ω × cm2; HC 218 ± 66 Ω × cm2) than those treated by the glucocorticoid alone (CORT 57 ± 5 Ω × cm2; HC 70 ± 2 Ω × cm2). Glucocorticoid induction resulted in BMEC morphological changes that accompanied the increases in TEER, and BMEC tight junctions exhibited improved integrity as visualized by the localization of tight junction proteins zonula occluden-1, occludin and claudin-5. The combined use of puromycin and glucocorticoid therefore provides an in vitro system that is well suited for molecular level BBB investigations.

Abbreviations used

blood–brain barrier


brain microvascular endothelial cells








days in vitro


Dulbecco's modified Eagle's medium


endothelial cells


enhanced chemiluminescent


fetal bovine serum






phosphate-buffered saline


phosphate-buffered saline containing 40% goat serum and 1% Trition X-100


plasma-derived serum


transendothelial electrical resistance


zonula occluden-1

The cerebral microvasculature separates the brain interior from the bloodstream and has been termed the blood–brain barrier (BBB) as a result of its impermeable properties. The BBB assists in maintaining brain homeostasis and protects the brain against harmful blood-borne substances. A single layer of brain microvascular endothelial cells (BMEC) is responsible for the limited solute transfer between blood and brain, and these specialized endothelial cells (EC) display distinctive attributes when compared with peripheral endothelium. Low BMEC permeability results from continuous tight junctions between adjoining ECs (Reese and Karnovsky 1967), low levels of pinocytosis and a general lack of fenestrae (Brightman and Reese 1969; Joo 1971).

Because of the impermeable phenotype, the BBB plays major roles in disease pathology and hinders drug delivery efforts. Because of the inherent difficulties in performing molecular level studies of disease pathology in vivo, and the fact that prediction of BBB drug permeability prior to animal studies would be highly advantageous, a representative in vitro model would be of high utility. Unfortunately, when BMECs are cultured in vitro, cellular de-differentiation results in a loss of many of the functions observed in vivo. Much of this change can be ascribed to the removal of BMECs from their local brain microenvironment. Astrocytes, whose foot processes are highly invested in the abluminal surface of capillaries (Kacem et al. 1998), are thought to confer many of the unique BBB properties (Stewart and Wiley 1981; Janzer and Raff 1987). Pericytes and neurons are also intimately associated with BMECs and, although less well studied, appear to play an important role in the differentiation and regulation of the BBB (Hatashita and Hoff 1990; Lee et al. 1999; Petty and Wettstein 2001; Hori et al. 2004). Considerable effort has been focused on regaining in vivo properties in vitro by reconstitution of the microenvironment (Deli and Joo 1996). Examples include the co-culture of BMECs with astrocytes, which has been shown to increase the transendothelial electrical resistance (TEER) and tight junction complexity, while reducing the paracellular passage of molecules such as sucrose (Tao-Cheng et al. 1987; Giese et al. 1995; Mertsch et al. 1997). The addition of neurons to cultures containing BMECs caused an increase in the enzymatic activities of γ-glutamyl transpeptidase and Na+ K+ ATPase, and induced the correct localization of the tight-junction protein occludin by these cells (Tontsch and Bauer 1991; Savettieri et al. 2000). The role of pericytes has been less well characterized, but in vitro studies indicate that they may have an ability to induce BBB properties (Hori et al. 2004; Dohgu et al. 2005). Additional studies highlight the importance of fluid flow (Stanness et al. 1996), and soluble factors such as cAMP (Rubin et al. 1991; Deli et al. 1995), hydrocortisone (HC) (Hoheisel et al. 1998; Weidenfeller et al. 2005) and dexamethasone (Grabb and Gilbert 1995), that have been shown to assist in partial restoration of in vivo characteristics such as barrier tightness. Although substantial progress has been made to date, in vitro models of the BBB continue to be labor-intensive and difficult to reproduce. In addition, in vitro BBB models tend to exhibit reduced permeability characteristics, a significant shortcoming for studying molecular or cellular trafficking across the BBB.

Many of the problems encountered in developing in vitro models with in vivo permeability characteristics stem from the difficulty of obtaining pure endothelial cultures. While the basement membrane surrounding the microvascular endothelium facilitates the isolation of these structures from total brain material (Panula et al. 1978), it also encloses pericytes that are difficult to remove from the endothelial fraction. As a result, these cells often contaminate BMEC cultures in an uncontrollable manner and have been shown to prevent complete barrier formation (Parkinson and Hacking 2005). Previous research indicates that pericytes in contact with microvascular endothelial cells produce an active form of transforming growth factor type-β which inhibits endothelial growth (Orlidge and D'Amore 1987; Antonelli-Orlidge et al. 1989). Therefore, pericytes often become the predominant cell type in such cultures, increasing the permeability of the BMEC monolayer (Parkinson and Hacking 2005; Perriere et al. 2005). The need for methods that can achieve pure BMEC cultures in a simple and reproducible manner still remains.

Techniques for achieving pure BMEC cultures continue to appear in the literature and these studies demonstrate a correlation between EC purity and improved barrier properties (Parkinson and Hacking 2005; Perriere et al. 2005). Methods for the isolation of microvessel fragments usually include mechanical and/or enzymatic dissociation of the brain material followed by filtration and/or density centrifugation steps (Bowman et al. 1981; Gordon et al. 1991; Deli et al. 2003). Other methods have incorporated magnetic bead separation (Song and Pachter 2003; Parkinson and Hacking 2005) and fluorescence-activated cell sorting after labeling with an endothelial specific marker (Sahagun et al. 1989) to increase endothelial cell purity. As described above, pericytes often constitute a major contaminating cell type in BMEC cultures, but smooth muscle cells, astrocytes, and fibroblasts can also be found (Greenwood 1992; Perriere et al. 2005). Techniques for minimizing contaminating cell types in cultured BMECs include physical separation (Abbott et al. 1992), selective toxicity (Risau et al. 1990), and the appropriate control of serum and growth supplement concentrations (Bowman et al. 1982; Carson and Haudenschild 1986; Gordon et al. 1991; Abbott et al. 1992). Another method of comparative ease employs chemical mediators for achieving culture purity. BMECs express p-glycoprotein (p-gp), an efflux transporter that recognizes a variety of small molecule substrates, some of which are cytotoxic (Litman et al. 2001). BMECs can therefore survive relatively high concentrations of molecules that would be toxic to many cell types. This strategy has been exploited by including vincristine, a p-gp substrate, in the culture medium to purify murine and porcine BMEC cultures (Chen et al. 1998; Igarashi et al. 1999), and puromycin was especially well suited for purifying rat BMEC cultures (Perriere et al. 2005).

We have also demonstrated that puromycin treatment of rat BMEC cultures routinely yields nearly 100% pure EC cultures. The method clearly outperformed other facile purification methods and was reproducible regardless of initial BMEC plating density. To further the impermeability characteristics of these cultures, glucocorticoids were added to BMEC monolayers. While corticosterone is the major physiological glucocorticoid found in rodents, hydrocortisone is a more potent glucocorticoid with respect to its anti-inflammatory and glycogen-deposition effects (Haynes and Murad 1985). Both corticosterone and hydrocortisone yielded improvements in BMEC impermeability, and pure cultures responded much more favorably to glucocorticoid treatment than untreated cultures. Taken together, this study demonstrates the utility of straightforward chemical treatments for significantly improving the permeability characteristics of a highly reproducible in vitro BBB model.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References


Unless otherwise specified below, reagents were obtained from Sigma (St Louis, MO, USA). Whatman 3MM chromatography blotting paper was obtained from Whatman Inc. (Florham Park, NJ, USA). Type 2 collagenase and DNase I were purchased from Worthington Biochemical Corp. (Lakewood, NJ, USA). Collagenase/dispase and basic fibroblast growth factor were obtained from Roche Molecular Biochemicals (Indianapolis, IN, USA). Transwell-Clear® (polyester) permeable supports, 0.4-μm pore size, were acquired from Corning (Acton, MA, USA). Bovine platelet-poor plasma-derived serum (PDS) was purchased from Biomedical Technologies Inc. (Stoughton, MA, USA). Fetal bovine serum (FBS), concentrated (100 ×) antibiotic–antimycotic solution and Ham's F-12 nutrient mixture were obtained from Invitrogen (Carlsbad, CA, USA). Percoll, enhanced chemiluminescent (ECL) western blotting detection reagents and ECL hyperfilm were obtained from Amersham Biosciences (Piscataway, NJ, USA). Polymount was acquired from Polysciences Inc. (Warrington, PA, USA). Anti-human von Willebrand factor and rabbit IgG fraction negative control were acquired from Dako (Carpinteria, CA, USA). Mouse anti-α-actin was obtained from American Research Products Inc. (Belmont, MA, USA). Mouse anti-occludin, rabbit anti-zonula occluden-1 (ZO-1) and mouse anti-claudin-5 were acquired from Invitrogen. Texas Red® goat anti-rabbit IgG antibody, Alexa Fluor goat anti-mouse IgG antibody, Texas Red®-X phalloidin and 4′,6-diamidino-2-phenylindole (DAPI) dihydrochloride nuclear stain were also purchased from Invitrogen. The 5-bromo-2′-deoxy-uridine (BrdU) Labeling and Detection Kit was acquired from Roche Molecular Biochemicals. Bicinchoninic acid protein assay reagent kit was purchased from Pierce (Rockford, IL, USA). Nitrocellulose membrane and colloidal gold total protein stain were acquired from Bio-Rad Laboratories (Hercules, CA, USA). Mouse IgG1κ was obtained from BD Biosciences (San Jose, CA, USA). An EVOM voltohmmeter was obtained from World Precision Instruments (Sarasota, FL, USA).

Continuous 33% Percoll gradient formation

Percoll was mixed in a 9 : 1 ratio with 10 × concentrated (0.1 m) phosphate-buffered saline (PBS). This solution was diluted 1 : 3 in 0.01 m PBS (8.77 g/L sodium chloride, 0.26 g/L sodium phosphate monobasic, and 2.17 g/L sodium phosphate dibasic, pH 7.40) containing 5% v/v heat-inactivated FBS. The mixture was sterilized using a 0.2-μm syringe filter, 30 mL were transferred to a 50-mL centrifuge tube, and the tube was centrifuged in a fixed-angle rotor for 60 min at 30 000 g and 4°C for Percoll gradient formation.

Isolation of rat brain microvessel endothelial cells

This procedure was performed essentially as described previously (Deli et al. 2003). Brains were removed from male rats (220–250 g) and stored in Dulbecco's modified Eagle's medium (DMEM) on ice. The remainder of the isolation took place under aseptic conditions. Each brain was cut sagittally into two halves and rolled on Whatman 3MM chromatography blotting paper to remove the meninges. The cortices were dissected away from the surrounding tissue and much of the brain white matter was subsequently removed. The cortices were mashed with forceps and thoroughly triturated, followed by a 1.25-h digestion at 37°C with 0.7 mg/mL type 2 collagenase and 39 U/mL Dnase I in DMEM. Afterwards, the enzyme solution was diluted with DMEM and centrifuged at 1000 g for 8 min at 4°C in a swinging bucket rotor. The pellet was re-suspended in a 20% w/v bovine serum albumin and DMEM solution and centrifuged for 20 min at 1000 g and 4°C to obtain a microvessel enriched cell pellet. Then, the pellet was further digested in 1 mg/mL collagenase/dispase and 39 U/mL Dnase I in DMEM for 1 h at 37°C. The digested microvessel solution was diluted with DMEM and centrifuged at 700 g and 4°C for 6 min. The pellet was re-suspended and layered over a 33% continuous Percoll gradient and centrifuged at 1000 g for 10 min at 4°C in a swinging bucket rotor. Subsequently, the microvessel layer was removed using a 5-inch, 16-gauge hypodermic needle and diluted into DMEM. After centrifuging at 700 g for 10 min and re-suspending in complete culture medium, cells were counted as described in the next section and seeded on either collagen IV/fibronectin coated plastic 24-well tissue culture plates or 1.12 cm2 Transwell-Clear® permeable supports (0.4 μm pore size). Culture medium consisted of DMEM supplemented with 20% PDS, 1 ng/mL basic fibroblast growth factor, 100 μg/mL heparin, 2 mm l-glutamine, and an antibiotic–antimycotic solution (100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin). Culture medium was changed within 24 h of initial plating to remove unattached cells or vessel fragments. Cultures were maintained in a 37°C incubator under humidified 5% CO2/95% air and culture medium was changed every 2–3 days.

Primary culture of rat BMECs

Transwell-Clear® filters were coated, using 130 μL of a solution of 400 μg/mL type IV collagen from human placenta and 100 μg/mL fibronectin from bovine plasma in each 1.12-cm2 filter, for several hours in a 37°C incubator before removing and allowing them to dry. Prior to plating, cells excluding Trypan Blue were counted using a Bright-Line hemacytometer. Both single cells and cells within microvessel fragments were included in the count. Subsequently, culture plates were seeded at varying densities and Transwell-Clear® filters were seeded at a density of 0.17 brains/cm2 (depending on BMEC isolation yield this corresponded to cell densities between 6.2 × 105 and 7.2 × 105 cells/cm2). Primary (zero passage) cultures were used for all experiments described in this study.

Puromycin treatment

An initial titration experiment examined various puromycin–dihydrochloride concentrations (0.4, 4, 40 μg/mL) for their ability to purify BMEC cultures. The lowest concentration failed to remove contaminating cells, while the highest concentration inhibited BMEC growth. A concentration of 4 μg/mL removed contaminating cell types without any noticeable BMEC growth inhibition. Thus, for the experiments described here, puromycin was included in complete culture medium at a concentration of 4 μg/mL. The microvessel fragments were seeded onto the culture wells with puromycin-containing medium and treated with puromycin for 2 days, after which time normal culture medium was used.

BMEC pre-plating

Isolated microvessel fragments were initially plated onto uncoated polystyrene culture wells and placed in a 37°C incubator under humidified 5% CO2/95% air for 4 h. The unattached cells and culture medium were then removed, plated onto collagen IV/fibronectin-coated culture wells and returned to the incubator for growth of BMECs.

Ca2+- and Mg2+-free saline washes

Washes consisted of replacing the culture medium with Ca2+- and Mg2+-free Hank's balanced salt solution (8.00 g/L sodium chloride, 0.09 g/L sodium phosphate dibasic, 0.40 g/L potassium chloride, 0.06 g/L potassium phosphate monobasic, 0.035 g/L sodium bicarbonate, and 1.00 g/L d-glucose, pH 7.40) for a 15-min interval at 37°C, followed by fresh culture medium. Cultures were treated twice in this manner after 1.5 and 2.5 days in vitro (DIV).

Corticosterone and hydrocortisone treatments

An initial titration experiment examined various physiologically relevant concentrations of corticosterone (150, 550, 900, 1400 nm) for TEER induction in BMEC monolayers. The induction of TEER was found to increase steadily in this range of concentrations, with 1400 nm resulting in the highest TEER values. The maximal CORT concentration of 1400 nm was therefore used in the experiments described here in order to elicit a maximal response. Confluent BMEC monolayers (after 3.5 DIV) were treated with serum-free medium consisting of a 1 : 1 ratio by volume of DMEM and Ham's F-12 nutrient mixture, 2 mm l-glutamine, and 1400 nm CORT or 550 nm hydrocortisone. When applicable, the number of proliferating cells in the confluent monolayer was determined by the administration of BrdU labeling reagent (10 μm) at 3.5 DIV. BrdU was removed from cultures at 4.5 DIV, and the cultures were fixed and stained according to the manufacturer's instruction.

Western blotting

Twenty-four hours after the addition of serum-free medium to BMEC monolayers, Laemmli sample buffer was used to collect cell lysates. Protein concentration was determined using the bicinchoninic acid protein assay reagent kit. Ten micrograms of protein lysate was resolved by sodium dodecyl sulfate – polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. Membranes were blocked overnight at 4°C in 5% non-fat dry milk and probed for 1 h at room temperature (20°C) with mouse anti-occludin (1 μg/mL) or mouse anti-claudin-5 (3 μg/mL) antibodies. Following washing steps, the membranes were incubated with goat anti-mouse horseradish peroxidase-conjugated antibody (1 : 2000 dilution). Proteins were detected by enhanced chemiluminescence with ECL western blotting detection reagents and exposure to ECL hyperfilm. Total protein loading was verified by colloidal gold total protein stain. Three independent Transwell filters were analyzed for each culture condition.


After gently washing the BMEC cultures three times with 0.01 m PBS, fixative was carefully added and the cultures were incubated on ice for 30 min. Ice-cold methanol (100%) was used as the fixative for rabbit anti-human von Willebrand factor (14 μg/mL) and mouse anti-α-actin (670 μg/mL). Paraformaldehyde (4% w/v at 4°C) was used as the fixative for mouse anti-occludin (2.5 μg/mL), rabbit anti-ZO-1 (2.5 μg/mL) and mouse anti-claudin-5 (3 μg/mL). BMEC cultures were then blocked and permeabilized at room temperature for 30 min using 40% goat serum and containing 1% Trition X-100 (PBSG). Primary antibodies were diluted to the concentrations shown above in PBSG and were incubated with samples overnight at 4°C. Secondary antibodies (Texas Red goat anti-rabbit IgG antibody, 2 μg/mL; Alexa Fluor goat anti-mouse IgG antibody, 2 μg/mL) were diluted in PBSG and incubated with samples for 1 h at room temperature (20°C). DAPI nuclear stain at a concentration of 300 nm in PBS was added to the wells for 5 min. Samples were then post-fixed with paraformaldehyde (4% w/v at 4°C) for 15 min at room temperature. Texas Red®-X phalloidin was incubated with fixed cells at a concentration of 165 nm for 1 h at room temperature (20°C). After staining, Transwell-Clear® filters were removed from their plastic supports using a razor blade and sealed between a glass slide and coverslip using Polymount.

Control samples were treated exactly as described above, except that the primary rabbit anti-human von Willebrand factor and rabbit anti-ZO-1 antibodies were replaced by rabbit IgG fraction negative control, mouse anti-occludin and mouse anti-claudin-5 antibodies were replaced by mouse IgG1κ, and mouse anti-α-actin antibody was replaced with mouse IgG2a.

Microscopy and quantitative analysis of culture purity

Photographs were taken using an Olympus (Melville, NY, USA) fluorescence microscope connected to a Diagnostic Instruments camera (Sterling Heights, MI, USA) run by MetaVue Version 5.0r1 software (Molecular Devices Corp., Sunnyvale, CA, USA). Statistics regarding culture purity were obtained by counting α-actin-positive and von Willebrand factor-negative cells (at least 2400 cells total) from multiple wells (three wells for Fig. 2 and two wells for Fig. 3). Counting was performed after overlaying von Willebrand factor, anti-α-actin, and DAPI nuclear stain images from the same field in Adobe Photoshop Elements 2.0. Percentage values of culture purity were arrived at by using the number of α-actin-positive cells, unstained cells and the total number of cells (DAPI stain) in each image. Unstained cells were classified as non-endothelial because of their lack of von Willebrand factor staining. Therefore, % EC = (total cells −α-actin-positive cells − unstained cells)/total cells * 100%. Because of the high purity of puromycin-treated cultures including PDS, counting for these wells was performed by scanning the entire well by eye, counting all α-actin-positive cells, and then percentage EC = (total cells − α-actin-positive cells)/total cells * 100%.


Figure 2.  A comparison of methods used for purifying primary rat BMEC cultures. Cells were grown for a total of 5 DIV under various culture conditions. FBS, fetal bovine serum; puro, puromycin treatment; PDS, bovine platelet-poor plasma derived serum; wash, Ca2+- and Mg2+-free saline wash treatment (cultured using FBS); pre-plate, preplating treatment (cultured using FBS). Data is representative of two independent culture experiments. The results are the mean ± SD of three culture wells. Statistical analysis of the data was performed by unpaired Student's t-tests and revealed significant differences (p < 0.03) between the purities of all samples, except between the FBS (without puromycin), wash, and pre-plate samples. The PDS plus puromycin-treated condition yielded a BMEC purity of 99.8 ± 0.04%.

Download figure to PowerPoint


Figure 3.  Purity of primary rat brain microvessel endothelial cell cultures as a function of initial cell seeding density after 5 DIV. Open circles represent data obtained from cells treated for 2 days with 4 μg/mL puromycin, while filled circles represent data from untreated cultures. Data is representative of two independent culture experiments, and the results are the means ± SD of two culture wells. Error bars for most of the data are smaller than the plot symbols. Culture medium contained PDS and all cultures that were seeded at or above 1.6 × 105 cells/cm2 were confluent after 5 DIV.

Download figure to PowerPoint

Quantitative analysis of frayed junctions, average BMEC area, and percent BMEC coverage

The percentage of cells containing frayed junctions over a significant fraction (greater than 10%) of their total cell border was determined after analyzing immunocytochemical images of cultures probed for ZO-1, occludin or claudin-5. Junctions between adjacent cells were defined as frayed if immunolabeling illuminated tight junction protrusions that are not parallel to the cell–cell border (Fig. 5). Total cell numbers were obtained from the corresponding DAPI stained images. A minimum of 850 cells from two separate wells was analyzed to determine this percentage. Average BMEC area and percent BMEC coverage was assessed by measuring the specific area and substrate coverage of at least 1000 cells.


Figure 5.  Immunocytochemistry of tight junction proteins in the primary rat microvessel endothelial cell cultures. Cells were grown on Transwell-Clear® filters in culture medium until 3.5 DIV and then switched to serum-free medium with or without the specified glucocorticoid for an additional 1.4 DIV. (a–d) Untreated primary cultures. (e–h) Primary cultures treated with 4 μg/mL puromycin for 2 days. (i–l) Primary cultures treated with 4 μg/mL puromycin and 550 nm hydrocortisone exhibiting smaller, spindle-shaped cells with little fraying at the cell borders. (a, e, i) ZO-1 immunolabeling. (b, f, j) Occludin immunolabeling. (c, g, k) Claudin-5 immunolabeling. (d, h, l) F-actin labeling using phalloidin. Cultures used for this figure were the same cultures as those analyzed in Fig. 4 and Table 1. Inset in (e) is a close-up of a frayed cell border, and arrows indicate a sampling of other regions having frayed borders. Scale bars represent 10 μm.

Download figure to PowerPoint

Resistance measurements

Transendothelial electrical resistance was measured using an EVOM voltohmmeter. Final resistances (Ω × cm2) were calculated after subtracting the resistance of an empty filter. Three TEER measurements were taken for each Transwell-Clear® filter at each time point and the values averaged. TEER measurements of triplicate filters for each culture condition were used to compute the mean and standard deviations reported.

Fluorescein permeability measurements

Prior to permeability studies, fresh serum-free medium was administered to BMEC monolayers and allowed to equilibrate for 1 h. A solution of sodium fluorescein (376 Da) in serum-free medium was then added to the upper compartment of the Transwell-Clear® filters to yield a final concentration of 1 μm. Two hundred-microliter aliquots were removed from the lower compartment and the volume replaced with pre-warmed serum-free medium after 0, 15, 30, 45, and 60 min. Based on the rate of influx of fluorescein into the lower compartment, permeability coefficients were calculated as previously described (Perriere et al. 2005). Permeability measurements of triplicate filters for each culture condition were performed.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Comparison of methods for controlling rat BMEC culture purity

The use of puromycin was compared with several other facile methods for its ability to achieve high-purity rat BMEC cultures. Freshly isolated rat brain microvessel fragments initially attached to the collagen IV/fibronectin-coated plastic culture wells, and cells began spreading within several hours of seeding. While untreated cultures initially contained predominately ECs, contaminating cells became increasingly apparent (between 2 and 3 DIV) and by 5 DIV their numbers represented greater than 15% of the total cells. Endothelial cells expressed the vascular von Willebrand factor marker, while the major contaminating cell type was the pericyte, which expressed α-actin in vitro(Figs 1a and b). Initial titration experiments with the translation inhibitor puromycin indicated that a culture medium concentration of 4 μg/mL was optimal for reducing the level of contaminating cells while not affecting EC proliferation rate (Figs 1c and d). Thus, this concentration of puromycin was used in all subsequent experiments.


Figure 1.  Primary cultures of rat brain microvessel endothelial cells after being cultured for 5 DIV. (a) and (c) are phase images and (b) and (d) are the corresponding immunofluorescent images probed for von Willebrand factor (red), α-actin (green), and DAPI nuclear stain (blue). (a, b) Cells were cultured in medium containing FBS; (c, d) cells were treated for 2 days with 4 μg/mL of puromycin and cultured in medium containing PDS, resulting in a dramatic reduction in the number of α-actin positive pericytes. Samples were initially seeded at an approximate cell density of 3.2 × 105 cells/cm2. Scale bars represent 50 μm.

Download figure to PowerPoint

Various methods were tested for their ability to minimize the extent of contamination in rat BMEC cultures (Fig. 2). The use of bovine PDS rather than untreated FBS led to significant increases in culture purities. Von Willebrand factor-positive BMEC percentages were raised from 69 ± 5% for FBS to 84 ± 1% for PDS. Cells grown in FBS-containing culture medium that were treated with puromycin had purities of 97.4 ± 0.2%, where the contaminating fraction consisted mostly of cells having spindle-shape EC morphology that expressed both von Willebrand factor and α-actin. In contrast, this population of putative de-differentiated ECs was absent from cultures containing PDS. The use of PDS and puromycin treatment resulted in cultures with extremely high purities of 99.8 ± 0.04% (Figs 1c and d, and Figure 2). In our hands, the use of FBS and Ca2+- and Mg2+-free saline washes, or a pre-plating step previously reported to limit BMEC culture contamination (Gordon et al. 1991; Abbott et al. 1992) did not provide any increase in rat BMEC culture purity. Therefore, the properties of the high-purity PDS and puromycin-treated cultures were chosen for further evaluation throughout the rest of this study.

Effects of initial cell seeding density on rat BMEC purity

After identifying puromycin as the most effective method for obtaining pure BMEC cultures, the effects of initial BMEC seeding densities were compared for cultures with or without puromycin treatment. Freshly isolated rat brain microvessel fragments were seeded at densities ranging between 0.5 × 105 and 6.4 × 105 cells/cm2 and allowed to grow for 5 DIV, with or without an initial 2-day puromycin treatment (Fig. 3). The purity of the cultures lacking puromycin treatment was clearly a function of the initial seeding density. Untreated cultures seeded at a higher initial density typically achieved a more complete monolayer of ECs (data not shown). However, these higher density cultures also contained more contaminating cells types, predominately present as a second layer of α-actin-positive pericytes residing on top of the EC monolayer. Thus, an overall decrease in percent EC was observed in higher seeding density cultures. Untreated cultures with low seeding densities (< 1.2 × 105 cells/cm2) failed to form complete monolayers of ECs and typically had a higher percentage of contaminating cell types. There existed a small range of seeding densities (1.2 × 105 to 1.6 × 105 cells/cm2) where a complete monolayer of ECs formed and minimal contaminating cells (∼15%) were present. In addition, without puromycin treatment, cultures had a significant population (up to 8%) of cells that were unlabeled by anti-von Willebrand factor or anti-α-actin. This population of cells increased proportionally with initial cell seeding density (data not shown). In contrast, cultures treated with puromycin attained purities of nearly 100% regardless of initial seeding density (Fig. 3). These cultures also formed complete monolayers even at lower initial seeding densities, allowing for the routine culture of 27 cm2 of ECs per rat brain after 5 DIV.

Effects of glucocorticoid treatment on TEER, permeability and cellular morphology

Because puromycin treatment yielded extremely pure BMEC monolayers, glucocorticoids were used to test whether or not these cells responded more favorably to BBB inductive factors. Physiologically relevant concentrations of the glucocorticoids CORT and HC were added to confluent rat BMEC monolayers, and the TEER was measured. Initial experiments revealed that CORT increased the TEER of puromycin-treated monolayers in a dose dependent manner between the concentrations of 150 and 1400 nm (data not shown). A CORT concentration of 1400 nm was used in the experiments described here to elicit a maximal response. A HC concentration of 550 nm was used as previously described by Hoheisel et al. (1998). Prior to glucocorticoid treatment (time 0), puromycin-treated cultures had TEER values reproducibly higher than those for untreated cultures (68 ± 8 Ω × cm2 vs. 37 ± 4 Ω × cm2), indicating a tighter barrier with restricted paracellular ion flux (Fig. 4, time 0). Unlike the highly pure puromycin-treated cultures, monolayers lacking puromycin treatment contained contaminating cell types, as evidenced by their morphology and α-actin staining. The contaminating cells could be seen both on the filter surface and overlaying ECs. When CORT or HC was added to rat BMEC cultures without puromycin treatment, the TEER values increased and remained improved over a 24-h period (Fig. 4). The maximum resistance in CORT-treated (57 ± 5 Ω × cm2) or HC-treated (70 ± 2 Ω × cm2) cultures was significantly higher than that seen in cultures lacking glucocorticoids (37 ± 4 Ω × cm2). In addition, the maximum TEERs achieved by cultures treated with both puromycin and glucocorticoid (168 ± 13 Ω × cm2 for CORT; 218 ± 66 Ω × cm2 for HC) were approximately 3-fold higher than those treated with the glucocorticoid alone (Fig. 4). Correspondingly, the sodium fluorescein permeability of HC-treated cultures (1.1 × 10−6 cm/s) was found to be more than 7-fold lower than that of puromycin-treated cultures (8.3 × 10−6 cm/s) and dramatically lower than that of untreated cultures (91.5 × 10−6 cm/s) (Table 1). While puromycin-treated cultures tended to maintain EC confluence, cultures lacking puromycin treatment exhibited gaps in the monolayer devoid of cells or filled with pericytes. Finally, the average area occupied by each EC when treated with both puromycin and glucocorticoids decreased (4.4 × 10−5 cm2/cell for CORT; 3.2 × 10−5 cm2/cell for HC) and the BMECs became more spindle shaped as compared with BMECs treated solely with puromycin (6.2 × 10−5 cm2/cell) (Fig. 5; Table 1). This trend of decreased cellular area was roughly inversely proportional to the respective maximum TEER measurements for each condition (Table 1). However, the phenomenon of reduced EC area was not nearly as evident in the absence of puromycin.


Figure 4.  Transendothelial electrical resistance (TEER) of primary rat microvessel endothelial cell cultures. Cells were grown on Transwell-Clear® filters in culture medium until 3.5 DIV and then switched to serum-free medium with or without the specified glucocorticoid (time 0). Filled-plot symbols represent cultures that were not treated with puromycin, while open-plot symbols represent cultures that were treated for 2 days with 4 μg/mL puromycin. (a) Effects of corticosterone treatment (1400 nm). (b) Effects of hydrocortisone treatment (550 nm). Data is from a single experiment but is split between two plots for clarity. Puromycin and untreated culture data is identical in both (a) and (b). Data is representative of two independent culture experiments. Results are the means ± SD of four Transwell-Clear® filters.

Download figure to PowerPoint

Table 1.   Effects of puromycin and glucocorticoid treatment on cellular morphology and monolayer permeability
PuromycinGlucocorticoidFrayeda,d (%)Cell areab,d (10–5 cm2)Coveragec (%)Max TEER (Ω × cm2)FL Pee (10−6 cm/s)
  1. a The percentage of cells containing frayed edges over a significant fraction (greater than 10%) of their total cell border. b Average cell areas were determined using a combination of EC-specific immunocytochemical and DAPI nuclear staining images to measure the area occupied by a given number of cells. c The percentage of available culture surface covered by ECs. d Except for ( *) values, all pairwise comparisons within a given category are statistically significant with p < 0.05 as determined by the unpaired Student's t-test. Results are the mean  ± SD of at least six images from at least two different culture wells. e Fluorescein permeability values (FL Pe) were obtained from an isolation of primary BMECs independent from that used for the rest of the data in the table. At 24 h after HC induction, TEER values were measured and permeability data acquired: 179 ± 14 Ω × cm2 for the puromycin- and HC-treated samples, 43 ± 2 Ω × cm2 for the puromycin-treated samples, and 13 ± 5 Ω × cm2 for the untreated samples (compare with the 24-h TEER values from the independent experiment represented in Fig. 4b). Results are the means ± SD of three Transwell-Clear® filters. ND, not determined. Permeability values are statistically significant with p < 0.05.

+Hydrocortisone12 ± 23.1 ± 0.199.7 ± 0.1218 ± 661.1 ± 0.3
+Corticosterone21 ± 24.4 ± 0.199.3 ± 0.2168 ± 13ND
+37 ± 66.2 ± 0.298.3 ± 0.394 ± 58.3 ± 1.2
Hydrocortisone19 ± 24.8 ± 0.5*81 ± 570 ± 4ND
Corticosterone29 ± 65.2 ± 0.2*71 ± 1257 ± 5ND
39 ± 85.6 ± 0.429 ± 1537 ± 491.5 ± 26.6

To investigate the origins of decreased cellular area upon glucocorticoid treatment of puromycin-purified monolayers, a cell proliferation assay was performed. HC treatment of puromycin-treated monolayers did not result in a statistically significant change in the total number of proliferating, BrdU-positive cells (5000 ± 1000 cells/cm2) compared with a puromycin-treated monolayer lacking glucocorticoid (6000 ± 1000 cells/cm2).

Effects of glucocorticoid treatment on tight junctions and the actin cytoskeleton

To more fully understand the structural basis for the increased TEER values upon glucocorticoid treatment, the tight junctions between adjoining BMECs were examined by immunocytochemistry. The localization of various tight junction proteins in untreated rat BMEC cultures was compared with puromycin and/or glucocorticoid-treated monolayers. Puromycin treatment did not affect the localization of tight junction proteins as compared with untreated cultures (Fig. 5). The localization of ZO-1, occludin, and claudin-5 can be seen predominately at the cell–cell junctions. A significant percentage of the cells exhibited tight junctions having a frayed appearance at high magnification (Fig. 5, Table 1). The addition of CORT or HC to the cultures resulted in fewer frayed junctions and a more uniform distribution of tight junction protein at the cell borders as compared with cultures without glucocorticoid treatment (Fig. 5). Similar to the correlation between cellular area and TEER, the number of frayed edges decreased under conditions exhibiting increased TEER (Table 1). In addition, western blotting indicated that the per cell expression levels of the paracellular cleft-resident proteins occludin and claudin-5 were the same in puromycin-treated cultures, whether or not HC was present (data not shown). Finally, the cellular distribution of F-actin filaments was examined following puromycin treatment or glucocorticoid induction (Fig. 5). No differences in actin cytoskeleton organization were apparent as peri-junctional actin and cytoplasmic actin fibers were distributed similarly under each condition.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This study has focused on the development of a facile, reproducible method for cultivating rat BMECs as in vitro BBB models having improved permeability properties. To this end, we identified puromycin treatment as being the most robust method for purifying rat BMEC cultures. This method is insensitive to seeding density effects and routinely generates 99.8% EC cultures. Importantly, puromycin-purified cultures respond much more favorably to glucocorticoid induction than do untreated cultures. The impermeability as measured by TEER and fluorescein permeability is greatly improved and is a consequence of well-developed tight junctions.

Puromycin is an antibiotic produced by Streptomyces alboniger that inhibits the growth of Gram-positive bacteria and various animal and insect cells. This aminonucleoside antibiotic functions by specifically inhibiting peptidyl transfer in prokaryotic and eukaryotic ribosomes. BMECs can withstand a relatively high concentration of this molecule because of their expression of the efflux transporter p-gp (MDR1), which recognizes puromycin and effluxes it from these cells. Recently Perriere et al. (2005) demonstrated that puromycin was not toxic to endothelial cells at the concentrations used here, and that chronic treatment could induce mdr1b expression. Inclusion of puromycin in the medium of rat BMEC cultures for the first 2 DIV dramatically increased culture purity, as also reported by this previous study (Fig. 2). Perriere et al. (2005) also demonstrated that puromycin treatment yielded the purest rat BMEC (98.8% pure) monolayers with the least toxicity compared with other p-gp substrates, vincristine, vinblastine, colchicine, and doxorubicin. Similarly, we have shown that puromycin yields 99.8% pure BMEC cultures. In our hands, this treatment was the vastly superior method for attaining culture purity.

Another technique compared in the purity analysis was the use of Ca2+- and Mg2+-free saline washes of cultures after 1.5 and 2.5 DIV, the period during which contaminating cell outgrowth became significant. The brief treatment of cultures with Ca2+- and Mg2+-free saline has been reported to preferentially remove contaminating cell types leaving the ECs behind (Abbott et al. 1992). This technique promoted little increase in culture purity (Fig. 2). Pre-plating of freshly isolated capillary fragments onto uncoated culture surfaces was also attempted in an effort to remove contaminating cell types. The basis of this method is rooted in the observation that ECs have been shown to adhere and spread well to collagen and fibronectin but poorly to uncoated plastic or glass (Gordon et al. 1991; Abbott et al. 1992). As pericytes and smooth muscle cells adhere relatively well to uncoated plastic, the strategy involves pre-plating the cell solution onto an uncoated plastic surface for several hours (Gordon et al. 1991; Dore-Duffy 2003). The pre-plating step removes the faster adhering cells from the ECs, which are then seeded onto a coated culture surface. Similar to the conclusions found in Parkinson and Hacking (2005), this method was unsuccessful despite these previous reports (Gordon et al. 1991; Dore-Duffy 2003). The failure of this method to increase purity was likely because of the loss of cellular material as a result of the pre-plating step, resulting in a very low relative seeding density and therefore a higher contaminating cell fraction after 5 DIV (see below). Thus, these cultures did not attain confluency and the method lacked reproducibility.

Besides the use of puromycin, a major key to achieving high-purity BMEC cultures was the replacement of FBS with PDS as a source of serum. By switching the FBS medium supplement to PDS in cultures lacking puromycin treatment, purities of near 84% were obtained, representing a 15% increase in purity. This increase can be attributed to the lack of platelet-derived growth factor in PDS, which reduces the growth rate of pericytes and smooth muscle cells while having little effect on the EC growth rate (Bowman et al. 1982; Carson and Haudenschild 1986; Gordon et al. 1991; Abbott et al. 1992). The use of PDS and puromycin treatment provided for culture purities routinely as high as 99.8%, while cultures grown in FBS with puromycin treatment were not quite as pure (97.4%). The lower purity of FBS-containing cultures was mainly because of the presence of a small population of cells expressing both von Willebrand factor and α-actin that were therefore not categorized as ECs. Morphologically, these cells appeared to be of endothelial origin, as evidenced by their spindle shape and their tight packing into the monolayer. Indeed, there is previous evidence of porcine brain ECs reversibly expressing α-actin in culture (Amberger et al. 1991). Therefore, the use of FBS may either promote the growth of a subpopulation of ECs that express α-actin or induce the expression of α-actin in a subpopulation of ECs. Either way, an EC expressing α-actin represents a de-differentiated form, so the use of PDS was preferred to assist in maintaining BMEC phenotype.

In order to test the experimental robustness of the puromycin method, cultures were seeded at varying initial cell densities. Cultures treated with puromycin were essentially pure regardless of the initial seeding density (Fig. 3). Thus, the use of puromycin provides a high degree of reproducibility between experiments by consistently attaining high purity despite the inevitable differences between microvessel preparations. In contrast, in untreated cultures, purity was found to be a function of initial seeding density with a very small window of seeding densities achieving maximal (> 85%) purity and confluence (Fig. 3). In low seeding density cultures where a confluent monolayer of BMECs was not formed by 5 DIV (< 1.2 × 105 cells/cm2), contaminating cell types grew on the substrate surface in addition to growing on top of existing BMEC colonies. The production of active transforming growth factor-β by such co-cultures can inhibit EC growth (Orlidge and D'Amore 1987; Antonelli-Orlidge et al. 1989) and prevent the formation of confluent monolayers. Accordingly, for cultures where high purity and yield are needed, the use of puromycin and PDS is preferred, as a broad range of initial cell seeding densities can result in confluent monolayers.

TEER is a general measure of paracellular ion flux, and serves as a proxy measurement for assessing EC barrier function (Hoheisel et al. 1998). When primary rat BMECs were cultured on Transwell-Clear® filters with or without puromycin, TEER measurements revealed a statistically significant increase in TEER values for cultures having puromycin treatment (Fig. 4). The study by Perriere et al. (2005) also reported an increase in TEER for puromycin-treated cultures as compared with untreated cultures. This can likely be attributed to the decrease in the number of pericytes and other contaminating cells, a notion supported by the recent findings of Parkinson and Hacking (2005), who showed that pericyte reduction caused a decrease in monolayer permeability to sucrose.

In order to further improve barrier properties, serum-free medium containing CORT or HC in the range of physiological concentrations was added to BMEC monolayers and resulted in significantly increased TEER (Fig. 4) [1400 nm for corticosterone (Karlson et al. 1994), 550 nm for hydrocortisone (Vahl et al. 2005)]. To our knowledge, this is the first demonstration of using the predominant rodent glucocorticoid, corticosterone, to improve in vitro BBB barrier properties. Interestingly, glucocorticoid addition to puromycin-treated samples resulted in greater TEER values than those seen without puromycin treatment. These results indicate that puromycin-treated monolayers respond more favorably to glucocorticoid treatment than untreated monolayers, and demonstrate that the use of both puromycin and glucocorticoids can allow for the reproducible formation of ‘in vivo-like’ BBB models. The most impermeable rat BBB models typically yield TEER values in the same range as those reported here (218 ± 66 Ω × cm2 for HC; de Vries et al. 1996; Demeuse et al. 2002), while in vitro models using BMEC from other species can at times yield higher TEER [bovine 600 Ω × cm2 (Zysk et al. 2001), and porcine 1800 Ω × cm2 (Nitz et al. 2003)]. After the 24-h time period, the effects of serum-free culture begin to lessen the benefits of glucocorticoid induction, and the TEER drops (Fig. 4). However, serum-free conditions were used in this study as the addition of glucocorticoids to serum-containing medium (10% serum) results only in minimal BBB induction (Hoheisel et al. 1998).

The BBB models described here responded to glucocorticoid treatment on a reliable timescale and maintained their maximal barrier properties for hours, enabling their use in permeability studies. As an example, the functional permeability of monolayers was determined using the normally BBB-impermeant small molecule, fluorescein. Mapping with elevations in TEER, puromycin treatment significantly lowered monolayer permeability, while HC induction further decreased the permeability (Table 1). The fluorescein permeability values determined for the puromycin (8.3 × 10−6 cm/s) and HC-treated cultures (1.1 × 10−6 cm/s) compare favorably with the lowest fluorescein permeability values reported for rat BMEC–astrocyte co-culture BBB models (0.75–4.2 × 10−6 cm/s; Kis et al. 2001; Perriere et al. 2005). The fluorescein permeability for the puromycin and HC system also compares well with that observed for bovine (2.2 × 10−6 cm/s; Gaillard and de Boer 2000) or human BBB models (53 × 10−6 cm/s; Muruganandam et al. 2002). In addition to soluble mediators like CORT and HC, puromycin-treated BMECs have also been shown to respond to astrocyte induction (Perriere et al. 2005), raising the possibility that puromycin-treated BMECs can be used in studies focused on intercellular interactions that are prevalent at the BBB.

The postulated physical basis for the observed improvements in barrier properties after puromycin and glucocorticoid treatment includes morphological changes and alterations in junctional structure. First, an immunocytochemical analysis revealed that more complete monolayers formed in cultures treated with puromycin, and that these cultures were populated with far fewer pericytes than in untreated cultures (Fig. 4, time 0; Table 1). Parkinson and Hacking (2005) described a similar finding noting the superior barrier properties of BMEC cultures with fewer pericytes. Next, upon glucocorticoid induction, a switch in EC morphology from a larger, slightly cobblestone appearance to a smaller, spindle shape took place leading to a monolayer having increased EC density (Table 1). Proliferation assays suggested that the increased cell density was not attributable to an increase in the proliferation rate as a consequence of HC treatment, but instead was because of a decrease in the rate of cell detachment in HC-treated cultures. This result is consistent with a previous study illustrating the anti-apoptotic effects of hydrocortisone in a porcine in vitro BBB model (Arndt et al. 2004). It is of note that, despite the higher putative cell detachment rate of puromycin-treated cultures as compared with cultures treated with both puromycin and HC, cultures lacking HC were still capable of maintaining confluent monolayers (Table 1).

The changes in morphology were accompanied by higher TEER values, lower permeability coefficients, and a less frayed appearance along the cell borders, as evidenced by staining for ZO-1, occludin and claudin-5 tight junction proteins (Table 1; Fig. 5). It is also interesting to note that gross changes in the cellular actin distribution did not accompany the observed refinement in junctional structure. The association between frayed tight-junction structure and TEER was previously noted by Weidenfeller et al. (2005). Their study employed transmission electron microscopy to demonstrate an HC-mediated morphological switch from frayed to continuous tight junctions, and this switch corresponded to increases in TEER. Taken together with our observations, these studies indicate that glucocorticoid treatment results in tighter packing (reduced cellular area) and fewer frayed junctions that serve to promote higher TEER values.

Recent evidence suggested that HC increases occludin mRNA and protein levels in an immortalized mouse brain endothelial cell line by activating the glucocorticoid receptor and its binding to putative glucocorticoid responsive elements in the occludin promoter (Forster et al. 2005). In addition, ZO-1 expression was induced in an immortalized rat brain endothelial cell line (GPNT) upon dexamethasone addition (Romero et al. 2003). In contrast to these systems, the study described here employed non-transformed primary endothelial cell cultures, and the net per cell expression of occludin and claudin-5 did not change following hydrocortisone treatment. Although net expression per cell does not change, it is still possible that tight junction protein turnover may be increased during HC-induced junction rearrangements and that the synthesis of new tight junction proteins may be up-regulated to assist in remodeling cell–cell junctions.

Methods for isolating brain microvessels invariably result in the co-isolation of some level of contaminating cells. Our results validate the use of puromycin for the reproducible attainment of essentially pure rat BMEC cultures. These purified cells respond optimally to glucocorticoid induction by the formation of a tighter barrier, as evidenced by increased TEER and reduced fluorescein permeability, without the use of additional cell types or conditioned media. While not all applications require a pure EC system, our model would be well suited for several experimental goals, including drug permeability screening and leukocyte trafficking. Furthermore, as a consequence of the BMEC purity, this model may be extremely helpful in ascribing observed attributes specifically to endothelial cells and not to cellular culture contaminants. This benefit may be particularly important for genomic and proteomic analyses focused on deconvoluting the complex roles that different cell types or soluble factors have in conferring BBB properties.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was funded with National Institutes of Health grant AA013834. ARC was the recipient of a University of Wisconsin Biotechnology Training Program Graduate Fellowship. ARJ was the recipient of a National Science Foundation Graduate Research Fellowship.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Abbott N. J., Hughes C. C., Revest P. A. and Greenwood J. (1992) Development and characterisation of a rat brain capillary endothelial culture: towards an in vitro blood–brain barrier. J. Cell Sci. 103, 2337.
  • Amberger A., Bauer H., Tontsch U., Gabbiani G., Kocher O. and Bauer H. C. (1991) Reversible expression of sm α-actin protein and sm α-actin mRNA in cloned cerebral endothelial cells. FEBS Lett. 287, 223225.
  • Antonelli-Orlidge A., Smith S. R. and D'Amore P. A. (1989) Influence of pericytes on capillary endothelial cell growth. Am. Rev. Respir. Dis. 140, 11291131.
  • Arndt S., Seebach J., Psathaki K., Galla H. J. and Wegener J. (2004) Bioelectrical impedance assay to monitor changes in cell shape during apoptosis. Biosens. Bioelectron 19, 583594.
  • Bowman P. D., Betz A. L., Ar D., Wolinsky J. S., Penney J. B., Shivers R. R. and Goldstein G. W. (1981) Primary culture of capillary endothelium from rat brain. In Vitro 17, 353362.
  • Bowman P. D., Betz A. L. and Goldstein G. W. (1982) Primary culture of microvascular endothelial cells from bovine retina: selective growth using fibronectin-coated substrate and plasma-derived serum. In Vitro 18, 626632.
  • Brightman M. W. and Reese T. S. (1969) Junctions between intimately apposed cell membranes in the vertebrate brain. J. Cell Biol. 40, 648677.
  • Carson M. P. and Haudenschild C. C. (1986) Microvascular endothelium and pericytes: high yield, low passage cultures. In Vitro Cell Dev. Biol. 22, 344354.
  • Chen Z., Zandonatti M., Jakubowski D. and Fox H. S. (1998) Brain capillary endothelial cells express MBEC1, a protein that is related to the Clostridium perfringens enterotoxin receptors. Lab. Invest. 78, 353363.
  • Deli M. A. and Joo F. (1996) Cultured vascular endothelial cells of the brain. Keio J. Med. 45, 183199.
  • Deli M. A., Dehouck M. P., Abraham C. S., Cecchelli R. and Joo F. (1995) Penetration of small molecular weight substances through cultured bovine brain capillary endothelial cell monolayers: the early effects of cyclic adenosine 3′,5′-monophosphate. Exp. Physiol. 80, 675678.
  • Deli M. A., Abraham C. S., Niwa M. and Falus A. (2003) N,N-diethyl-2-[4-(phenylmethyl) phenoxy]ethanamine increases the permeability of primary mouse cerebral endothelial cell monolayers. Inflamm. Res. 52, S39S40.
  • Demeuse P., Kerkhofs A., Struys-Ponsar C., Knoops B., Remacle C. and Van Den Bosch de Aguilar P. (2002) Compartmentalized co-culture of rat brain endothelial cells and astrocytes: a syngenic model to study the blood–brain barrier. J. Neurosci. Meth. 121, 2131.
  • Dohgu S., Takata F., Yamauchi A., Nakagawa S., Egawa T., Naito M., Tsuruo T., Sawada Y., Niwa M. and Kataoka Y. (2005) Brain pericytes contribute to the induction and up-regulation of blood–brain barrier functions through transforming growth factor-β production. Brain Res. 1038, 208215.
  • Dore-Duffy P. (2003) Isolation and characterization of cerebral microvascular pericytes. Meth. Mol. Med. 89, 375382.
  • Forster C., Silwedel C., Golenhofen N., Burek M., Kietz S., Mankertz J. and Drenckhahn D. (2005) Occludin as direct target for glucocorticoid-induced improvement of blood–brain barrier properties in a murine in vitro system. J. Physiol. 565, 475486.
  • Gaillard P. J. and De Boer A. G. (2000) Relationship between permeability status of the blood–brain barrier and in vitro permeability coefficient of a drug. Eur. J. Pharm. Sci. 12, 95102.
  • Giese H., Mertsch K. and Blasig I. E. (1995) Effect of MK-801 and U83836E on a porcine brain capillary endothelial cell barrier during hypoxia. Neurosci. Lett. 191, 169172.
  • Gordon E. L., Danielsson P. E., Nguyen T. S. and Winn H. R. (1991) A comparison of primary cultures of rat cerebral microvascular endothelial cells to rat aortic endothelial cells. In Vitro Cell Dev. Biol. 27A, 312326.
  • Grabb P. A. and Gilbert M. R. (1995) Neoplastic and pharmacological influence on the permeability of an in vitro blood–brain barrier. J. Neurosurg. 82, 10531058.
  • Greenwood J. (1992) Characterization of a rat retinal endothelial cell culture and the expression of P-glycoprotein in brain and retinal endothelium in vitro. J. Neuroimmunol. 39, 123132.
  • Hatashita S. and Hoff J. T. (1990) Brain edema and cerebrovascular permeability during cerebral ischemia in rats. Stroke 21, 582588.
  • Haynes R. C. and Murad F. (1985) Adrenocorticotropic Hormone; Adrenocortical steroids and their synthetic Analogs; Inhibitors of Adrenocortical Steroid Biosynthesis, in The Pharmacological Basis of Therapeutics., 7th edn. (Gilman, A. G. G. L., Rall, T. W., Murad, F., eds), pp. 14591489. MacMillan, New York.
  • Hoheisel D., Nitz T., Franke H., Wegener J., Hakvoort A., Tilling T. and Galla H. J. (1998) Hydrocortisone reinforces the blood-brain properties in a serum free cell culture system. Biochem. Biophys. Res. Commun. 247, 312315.
  • Hori S., Ohtsuki S., Hosoya K., Nakashima E. and Terasaki T. (2004) A pericyte-derived angiopoietin-1 multimeric complex induces occludin gene expression in brain capillary endothelial cells through Tie-2 activation in vitro. J. Neurochem. 89, 503513.
  • Igarashi Y., Utsumi H., Chiba H. et al. (1999) Glial cell line-derived neurotrophic factor induces barrier function of endothelial cells forming the blood–brain barrier. Biochem. Biophys. Res. Commun. 261, 108112.
  • Janzer R. C. and Raff M. C. (1987) Astrocytes induce blood–brain barrier properties in endothelial cells. Nature 325, 253257.
  • Joo F. (1971) Increased production of coated vesicles in the brain capillaries during enhanced permeability of the blood–brain barrier. Br. J. Exp. Pathol. 52, 646649.
  • Kacem K., Lacombe P., Seylaz J. and Bonvento G. (1998) Structural organization of the perivascular astrocyte endfeet and their relationship with the endothelial glucose transporter: a confocal microscopy study. Glia 23, 110.
  • Karlson P., Doenecke D. and Kooman J. (1994) Biochemie Fur Mediziner und Naturwissenschaftler, pp. 417467. Georg Thieme Verlag, Stuttgart.
  • Kis B., Deli M. A., Kobayashi H. et al. (2001) Adrenomedullin regulates blood–brain barrier functions in vitro. Neuroreport 12, 41394142.
  • Lee E. J., Hung Y. C. and Lee M. Y. (1999) Early alterations in cerebral hemodynamics, brain metabolism, and blood–brain barrier permeability in experimental intracerebral hemorrhage. J. Neurosurg. 91, 10131019.
  • Litman T., Druley T. E., Stein W. D. and Bates S. E. (2001) From MDR to MXR: new understanding of multidrug resistance systems, their properties and clinical significance. Cell Mol. Life Sci. 58, 931959.
  • Mertsch K., Haseloff R. F. and Blasig I. E. (1997) Investigations of Radical Scavengers by Using an In Vitro Model of the Blood–Brain Barrier, pp. 881886. Elsevier, Amsterdam, the Netherlands.
  • Muruganandam A., Tanha J., Narang S. and Stanimirovic D. B. (2002) Selection of phage-displayed llama single-domain antibodies that transmigrate across human blood–brain barrier endothelium. FASEB J. 16, 240242.
  • Nitz T., Eisenblatter T., Psathaki K. and Galla H. J. (2003) Serum-derived factors weaken the barrier properties of cultured porcine brain capillary endothelial cells in vitro. Brain Res. 981, 3040.
  • Orlidge A. and D'Amore P. A. (1987) Inhibition of capillary endothelial cell growth by pericytes and smooth muscle cells. J. Cell Biol. 105, 14551462.
  • Panula P., Joo F. and Rechardt L. (1978) Evidence for the presence of viable endothelial cells in cultures derived from dissociated rat brain. Experientia 34, 9597.
  • Parkinson F. E. and Hacking C. (2005) Pericyte abundance affects sucrose permeability in cultures of rat brain microvascular endothelial cells. Brain Res. 1049, 814.
  • Perriere N., Demeuse P., Garcia E. et al. (2005) Puromycin-based purification of rat brain capillary endothelial cell cultures. Effect on the expression of blood–brain barrier-specific properties. J. Neurochem. 93, 279289.
  • Petty M. A. and Wettstein J. G. (2001) Elements of cerebral microvascular ischaemia. Brain Res. Brain Res. Rev. 36, 2334.
  • Reese T. S. and Karnovsky M. J. (1967) Fine structural localization of a blood–brain barrier to exogenous peroxidase. J. Cell Biol. 34, 207217.
  • Risau W., Engelhardt B. and Wekerle H. (1990) Immune function of the blood–brain barrier: incomplete presentation of protein (auto-) antigens by rat brain microvascular endothelium in vitro. J. Cell Biol. 110, 17571766.
  • Romero I. A., Radewicz K., Jubin E., Michel C. C., Greenwood J., Couraud P. O. and Adamson P. (2003) Changes in cytoskeletal and tight junctional proteins correlate with decreased permeability induced by dexamethasone in cultured rat brain endothelial cells. Neurosci. Lett. 344, 112116.
  • Rubin L. L., Hall D. E., Porter S. et al. (1991) A cell culture model of the blood–brain barrier. J. Cell Biol. 115, 17251735.
  • Sahagun G., Moore S. A., Fabry Z., Schelper R. L. and Hart M. N. (1989) Purification of murine endothelial cell cultures by flow cytometry using fluorescein-labeled griffonia simplicifolia agglutinin. Am. J. Pathol 134, 12271232.
  • Savettieri G., Di Liegro I., Catania C. et al. (2000) Neurons and ECM regulate occludin localization in brain endothelial cells. Neuroreport 11, 10811084.
  • Song L. and Pachter J. S. (2003) Culture of murine brain microvascular endothelial cells that maintain expression and cytoskeletal association of tight junction-associated proteins. In Vitro Cell Dev. Biol. Anim. 39, 313320.
  • Stanness K. A., Guatteo E. and Janigro D. (1996) A dynamic model of the blood–brain barrier ‘in vitro. Neurotoxicology 17, 481496.
  • Stewart P. A. and Wiley M. J. (1981) Developing nervous tissue induces formation of blood–brain barrier characteristics in invading endothelial cells: a study using quail–chick transplantation chimeras. Dev. Biol. 84, 183192.
  • Tao-Cheng J. H., Nagy Z. and Brightman M. W. (1987) Tight junctions of brain endothelium in vitro are enhanced by astroglia. J. Neurosci. 7, 32933299.
  • Tontsch U. and Bauer H. C. (1991) Glial cells and neurons induce blood–brain barrier-related enzymes in cultured cerebral endothelial cells. Brain Res. 539, 247253.
  • Vahl T. P., Ulrich-Lai Y. M., Ostrander M. M., Dolgas C. M., Elfers E. E., Seeley R. J., D'Alessio D. A. and Herman J. P. (2005) Comparative analysis of ACTH and corticosterone sampling methods in rats. Am. J. Physiol. Endocrinol. Metab. doi: 10.1152/ajpendo.00122.2005.
  • De Vries H. E., Blom-Roosemalen M. C., Van Oosten M., De Boer A. G., Van Berkel T. J., Breimer D. D. and Kuiper J. (1996) The influence of cytokines on the integrity of the blood–brain barrier in vitro. J. Neuroimmunol. 64, 3743.
  • Weidenfeller C., Schrot S., Zozulya A. and Galla H. J. (2005) Murine brain capillary endothelial cells exhibit improved barrier properties under the influence of hydrocortisone. Brain Res. 1053, 162174.
  • Zysk G., Schneider-Wald B. K., Hwang J. H., Bejo L., Kim K. S., Mitchell T. J., Hakenbeck R. and Heinz H. P. (2001) Pneumolysin is the main inducer of cytotoxicity to brain microvascular endothelial cells caused by Streptococcus pneumoniae. Infect. Immun. 69, 845852.