Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Brain capillary endothelial cells, i.e., the BBB, are well known to function as a dynamic interface with regard to the transfer of nutrients and drugs from circulating blood to brain interstitial fluid and vice versa (Cornford, 1985; Pardridge, 1993). For the development of new drugs and their therapeutic use, one of the most important issues is to evaluate and/or predict the BBB permeability. Lipophilicity and the molecular weight are, in general, known to be important physicochemical properties affecting BBB transport, whereas drugs, having analogous structures to those of nutrients, have been demonstrated to be transported via a carrier-mediated systems at the BBB (van Bree et al., 1988; Terasaki et al., 1991). There have been many reports showing a reasonably good correlation between the BBB PS and the octanol/water partition coefficient divided by the square root of the molecular weight of substrates (Levin, 1980; Cornford et al., 1982); however, several drugs have been also shown to exhibit markedly lower BBB PS value, e.g. vincristine, adriamycin and cyclosporin A (Levin, 1980; Cefalu and Pardridge, 1985). One hypothesis for these drugs showing restricted distribution has been proposed: namely, that the BBB has a molecular weight threshold, e.g., 400 Da (Levin, 1980) or 600 Da (Pardridge, 1994), to restrict lipophilic drug uptake into the brain. However, it has been also reported that the octanol-water partition coefficient is a good predictor for several peptides with molecular weights greater than 1000 except N-tyrosinated peptides which have been shown to be transported into the brain by a carrier-mediated system (Banks and Kastin, 1985). Moreover, recent studies have revealed that MDR1 (P-glycoprotein), acts as an efflux pump to transport drugs from brain to the circulating blood (Tatsuta et al., 1992; Tsuji et al., 1992, 1993; Schinkel et al., 1994).

It is, therefore, important to examine whether the BBB can transport a substrate having a molecular weight greater than 600 Da or not. BCECF-AM, which has been used as a fluorescent reagent for the measurement of intracellular pH (Fig. 1), has a molecular weight of 809 Da and is very rapidly hydrolyzed to BCECF, an acidic form of BCECF-AM (fig. 1), in cells by an enzymatic reaction. As BCECF has a hydrophilic nature and is well known to be trapped by cells (Paradiso et al., 1984; Weiner and Hamm, 1989), it is possible to assume that BCECF-AM is hydrolyzed to BCECF immediately after crossing the BBB and BCECF is retained in the brain. Accordingly, we have chosen BCECF-AM as a model substrate to test the molecular weight threshold hypothesis.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Chemical structures and values of the molecular weight and octanol/water partition coefficients of BCECF-AM and BCECF.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. [³H]-Water (925 MBq/ml), [³H]-mannitol (832.5 GBq/mmol), [³H]-3-O-methyl-D-glucose (2782.4 GBq/mmol), [¹⁴C]-carboxyl-inulin (0.093 GBq/g) and [¹⁴C]-butanol (0.059 GBq/mmol) were purchased from New England Nuclear, Boston, MA. [¹⁴C]-Diazepam (1.96 GBq/mmol) was obtained from Amersham International, Ltd., Buckinghamshire, UK. BCECF-AM, (2,7-bis(2-carboxyethyl)-5(6)-carboxyfluorescein tetraacetoxymethyl ester) and BCECF and HEPES were provided by Dojindo Chemicals, Kumamoto, Japan. All isotopes, were stored at -20°C until required, except [³H]-water which was maintained at 4°C. Horse serum and Trypsin-EDTA were purchased from Gibco, Grand Island, NY, collagenase/dispase and dispase were obtained from Boehringer Mannheim GmbH, Mannheim, Germany and Godoshuzo, Tokyo, respectively, and human fibronectin was from Iwaki Glass, Tokyo. Amphotericin B, dextran (industrial grade, MW 87,000), gentamycin sulfate and percoll were obtained from Sigma Chemical Co., St. Louis, MO, heparin (1000U/ml) was from Nihon Upjohn, Tokyo, polymyxin B sulfate was from Wako Pure Industries Ltd., Osaka, Japan and Hionic Fluor (liquid scintillation cocktail) was purchased from Packard Instrument Co., Netherlands. Xylazine and Ketaral 50 (ketamine hydrochloride) were purchased from Sigma Chemical Co., St. Louis, MO. and Sankyo Co., Tokyo. Cellbanker was purchased from Nihon Zenyaku Kogyo Co., Fukushima, Japan. All other chemicals were of reagent grade and used without further purification.

Animals. Male Wistar rats (Nisseizai, Co., Tokyo, Japan), weighing 250 to 300 g were used throughout this study. They had free access to food and water. For the carotid artery injection and the in situ brain perfusion studies, rats were anesthetized with intramuscular ketamine hydrochloride, 235 mg/kg, and xylazine, 2.3 mg/kg.

Isolation and culture of bovine BCEC. Capillary endothelial cells were isolated from bovine brains by the method reported previously (Hughes and Lantos, 1989; Terasaki et al., 1991) with minor modifications. In brief, two fresh bovine brains were rinsed with MEM containing 26 mM NaHCO₃. The cerebral gray matter was scraped off and then minced with a razor blade. After treatment with 0.5% dispase solution, the cerebral microvessels were obtained by centrifugation in 13% dextran. Then, the cell mixture was incubated with 1% collagenase/dispase solution and the crude cell suspension was centrifuged on a continuous 15 to 75% Percoll gradient to separate endothelial cells from contaminating red blood cells, fat and cell debris. The endothelial cells thus obtained were seeded in 10 collagen-coated dishes (100 mm diameter, Iwaki Glass, Tokyo, Japan) which were also coated with fibronectin. The cells were cultured for 3 days in culture medium containing MEM/Ham F12(50/50), 10 mM HEPES, 13 mM NaHCO₃, 50 mg/liter gentamycin, 2.5 mg/liter amphotericin B, 10% horse serum and 50 mg/liter polymixin B at 5% CO₂-95% air using an N₂-O₂-CO₂ Incubator BNP-110 M (Tabai Espec Co., Osaka, Japan) at 37°C. Then, the cells were cultured for 5 or 6 days in the same culture medium without polymixin B. The cells were treated with 2.5% trypsin/1 mM EDTA for 2 min at 37°C and then collected. The cells were suspended in the Cellbanker solution at a concentration of 2 × 10⁶ cells/ml and stored at -100°C until required. The cells were seeded at 2 to 5 × 10⁵ cells/cm² in collagen-coated dishes (Iwaki Glass, Tokyo) which were also coated with fibronectin and cultured for 7 days in the culture medium without polymixin B.

In vitro uptake studies. The uptake of BCECF-AM into cultured monolayers of BCEC was performed using the method reported previously (Terasaki et al., 1991). Cultured cells were washed twice with 1 ml PBS(-), pH 7.40, 137 mM NaCl, 2.7 mM KCl, 1.5 mM KH₂PO₄, 8.1 mM Na₂HPO₄) at 37°C. Uptake was initiated by adding 250 µl 5 µM BCECF-AM containing 122 mM NaCl, 3 mM KCl, 25 mM NaHCO₃, 1.2 mM MgSO₄, 0.4 mM K₂HPO₄, 1.4 mM CaCl₂, 10 mM HEPES, 10 mM D-glucose and 0.5% of DMSO to cells. At designated times after incubation, uptake was terminated by washing the cells three times with 1 ml ice-cold PBS. Intracellular BCECF-AM concentrations were determined as follows. To solublize cells, 200 µl 0.1% SDS was added to each dish and, after standing at room temperature for 10 min, 200 µl 0.1 N NaOH was added to hydrolyze BCECF-AM. Following a further 5 min incubation at room temperature, 0.80 ml 100 mM HEPES was added to adjust the pH to 7.6. The fluorescence intensity was measured by the spectrofluorometric method described below. Protein content was determined by the Lowry method (Lowry et al., 1951), using bovine serum albumin as a standard. Net uptake was expressed as µl medium per mg protein. To estimate the uptake rate, linear regression was performed. As the surface area of brain capillary endothelial cells has been reported to be 100 cm²/g brain (Bradbury, 1979), the in vivo BBB PS value was also calculated from the in vitro uptake study using BCEC. The transport rates are presented as means ± S.E., statistical differences were assessed by Student's t test.

Carotid artery injection technique. The apparent PS value and efflux rates of BCECF-AM at the BBB were determined using the carotid artery injection technique (Oldendorf, 1970) with a slight modification. Briefly, after exposure of the right carotid artery, 250 µl Ringer's-HEPES buffer (141 mM NaCl, 4.0 mM KCl, 2.8 mM CaCl₂, 10 mM HEPES, pH 7.40) containing 100 µM BCECF-AM, 4 µCi/ml [¹⁴C]-butanol and 10% DMSO were rapidly injected into the carotid artery. At designated times, the rats were decapitated and the hemisphere ipsilateral to the injected side was immediately excised and rinsed with saline. Accurately weighed tissue (0.6 g) was homogenized with 2.4 ml 10 mM HEPES (pH 7.4), using a homogenizer (Ultra-Turrax, Ika-Werk, Janke & Kunkel, Breisgau, Germany). 0.3 mL homogenate was solubilized in 1 ml 2 N NaOH for the assay of [¹⁴C]-butanol. In addition, 2 ml of homogenate was used for the assay of BCECF-AM.

In situ brain perfusion. The in vivo transport experiments were also carried out by the in situ brain perfusion technique reported previously (Takasato et al., 1984; Saheki et al., 1994). After exposure of the right carotid artery, the occipital and superior thyroid arteries were ligated and cut, and the right pterygopalatine artery was ligated. Then, the right external carotid artery was catheterized for perfusion using 22 cm of polyethylene tubing (PE-50) filled with sodium heparin (100 IU/ml). Both BCECF-AM(5 µM) and [¹⁴C]-carboxyl-inulin(0.25 µCi/ml) were dissolved in perfusion medium containing 128 mM NaCl, 24 mM NaHCO₃, 4.2 mM KCl, 2.4 mM NaH₂PO₄, 1.5 mM CaCl₂, 0.9 mM MgCl₂, 9 mM D-glucose and 0.5% DMSO (pH 7.40). The perfusion medium was freshly prepared, oxygenated with 95% O₂/5% CO₂ and heated to 37°C by a temperature-regulating circulator equipped with a peristatic pump (Masterflex, Cole-Parmer Instrument Co., Chicago, IL) and incubator (SM-05, Taitec Co., Saitama, Japan).

(1)

Determination of apparent partition coefficient. The apparent partition coefficients of BCECF-AM and BCECF were determined at 37°C by a published method (Leo et al., 1971; Tsuji et al., 1977). n-Octanol and 10 mM HEPES (pH 7.4) were used for the organic and the aqueous phases, respectively. The initial BCECF-AM and BCECF concentrations in the aqueous phase were 5 µM. Equal (1 ml) volumes of octanol and aqueous phase were used for the determination of BCECF-AM whereas 10 and 0.5 ml volumes, respectively, were used for the determination of BCECF. After shaking the mixture vigorously for 2 hr at 37°C and standing for 1 hr at 37°C, the aqueous phase was separated by centrifugation at 3000 rpm for 10 min at 37°C (Refrigerated Centrifuge RL-100, Tomy Seiko, Co., Tokyo, Japan). For the determination of BCECF-AM, 0.6 ml 0.1 N of NaOH was added to 0.6 ml aqueous phase and kept for 2.5 min at room temperature. Then, 1.8 ml 0.1 M HEPES was added to the solution to adjust the pH to 7.4. For the determination of BCECF, 2.4 ml 10 mM HEPES was added to 0.6 ml aqueous phase. The concentration was determined by spectrofluorometric assay as described below.

Analytical procedures: spectrofluorometric assay of BCECF-AM. The concentration of BCECF-AM in the brain was determined by spectrofluorometric assay as follows. Four volumes (ml) of 10 mM HEPES (pH 7.4), relative to the brain weight (g), were added to the brain tissue. Using a homogenizer (Ultra-Turrax), a brain homogenate was prepared and incubated for 30 min at 37°C to hydrolyze BCECF-AM. The efficiency of hydrolysis was investigated and found to be almost 100%. Then 1/10 the volume of 11% TCA was added to the homogenate and kept for 30 min on ice. Then, the sample was centrifuged at 13,000 rpm for 15 min and 2 M HEPES was added to the supernatant to adjust the pH to 7.4. The fluorescence intensity of BCECF was determined by spectrofluorometry (F-2000, Hitachi, Tokyo, Japan), the wavelengths were set at 500 nm for excitation and 530 nm for emission, respectively.

Radiochemical assay. The radioactivity of [¹⁴C]-butanol, [¹⁴C]-inulin, [³H]-water, [³H]-mannitol, [¹⁴C]-3-O-methyl-D-glucose and [¹⁴C]-diazepam in the brain was measured as follows. The brain or brain homogenate were dissolved in 1.0 ml 2 N NaOH at 50°C for 3 h, Hyonic Fluor (Packard, Groningen, Netherlands) was then added and the radioactivity was determined using a model LC-6000 liquid scintillation counter (Beckmann Instruments Co., Fullerton, CA).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

In vitro uptake of BCECF-AM into cultured bovine BCEC. The time-course of BCECF-AM uptake into cultured monolayers of BCEC at 37°C was linear with time up to 60 sec (fig. 2). The uptake value extrapolated to zero time, which was considered to represent the adsorption to the cell surface, was small (3.84 µl/mg protein). The uptake rate of BCECF-AM obtained from the slope by linear regression analysis was found to be 151 ± 2 (µl/min/mg protein, mean ± S.E.). Since the surface area of the cultured BCEC was 29.4 cm²/mg protein, i.e., 68.0 µg/well (2 cm²), in our study and the surface area of rat brain capillary has been reported to be 100 cm²/g brain (Bradbury, 1979), the in vivo BBB PS value based on this in vitro rate would be predicted to be 446 ± 7 µl/min/g brain.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Time-course of BCECF-AM uptake by cultured monolayers of BCEC at 37°C. Each point represents the mean ± S.E. of three or four experiments.

In vivo brain perfusion study of BCECF-AM. The BBB PS value of BCECF-AM was found to be 251 ± 41 µl/min/g brain for a 45 sec perfusion and 354 ± 99 µl/min/g brain for a 60 sec perfusion (table 1). No significant difference was observed between these two perfusion periods and the averaged BBB PS value of BCECF-AM was calculated as 295 ± 48 µl/min/g brain. The apparent vascular volume of the perfused brain was determined using [¹⁴C]-carboxyl-inulin and found to be 10.1 ± 2.7 µl/g brain after 45 sec perfusion and 11.1 ± 2.6 µl/g brain after 60 sec perfusin (mean ± S.E.).

In vivo carotid artery injection of BCECF-AM. The carotid artery injection technique provided an estimate of 132 ± 25 µl/min/g brain for the in vivo BBB PS value of BCECF-AM. Because 10% DMSO was added to dissolve BCECF-AM in the injectate, the effects of DMSO on nonspecific permeability at the BBB, cerebrovascular volume, carrier-mediated transport and cerebral blood flow rate were also examined. The carotid artery injection of a mixture of [³H]-3OMG and [¹⁴C]-butanol or [¹⁴C]-mannitol and [³H]-water were carried out, with or without 10% DMSO in the injectate. As shown in table 2, no significant effect of DMSO was observed on the apparent uptake of any of the compounds.

In vivo efflux of BCECF-AM from the brain. The in vivo efflux of BCECF-AM was studied by the carotid artery injection technique. The percentage of the injected dose remaining in the brain was determined for [¹⁴C]-butanol and BCECF-AM at designated times after injection. As shown in figure 3, significant efflux of [¹⁴C]-butanol from the brain was observed and the apparent elimination rate constant, kel, of butanol was determined to be 0.34 ± 0.02 min¹. By contrast, significant efflux was not observed for BCECF-AM over a 120-sec washout period (fig. 3).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Semilogarithmic plots of the residual amount of total BCECF (BCECF + BCECF-AM) () and [14C]-butanol () in the brain vs. time after carotid artery injection. The solid line was obtained by nonlinear regression analysis of the data using a monoexponential equation. Each point represents the value of a single experiment.

Prediction of the BBB PS value of BCECF-AM based on lipophilicity and molecular weight. Using the previous reported values (Levin, 1980; Cornford et al., 1982), we have obtained the following relationship between in vivo BBB PS value (PS_BBB) and the octanol/water partition coefficient divided by the square root of the molecular weight (PC/MW^1/2) from the linear regression analysis of the data.

(2)

As adriamycin, eppipodophylotoxin, vincristine, bleomcin and cyclosporin A exhibited a significantly lower BBB permeability than expected from their lipophilicity (Levin, 1980; Cefalu and Pardridge, 1985), these drugs were not used for the regression analysis.

3

Comparison of predicted and observed values for the BBB PS value of BCECF-AM. As shown in figure 4, the BBB PS value determined by the in vitro uptake study using BCEC, the in vivo carotid artery injection and the in situ brain perfusion methods were plotted against the value of the octanol/water partition coefficient divided by the square root of the molecular weight together with previously reported values. The BBB PS value obtained were either similar or greater than that predicted from equation 2.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Comparison of BBB permeability PS between the predicted values from octanol/water partition coefficients and the experimental values, determined using in vitro cultured monolayers of BCEC (open square), in vivo carotid artery injection (open circle) or brain perfusion methods (open triangle). Closed circles and closed triangles were obtained from the values reported previously (Levin, 1980, Cornford, 1985; Cefalu and Pardridge, 1985). Open diamond represents the value of erythrosine B reported previously (Levitan et al., 1984). The solid line was obtained by linear regression using the data in the closed circles. Drugs shown as closed triangles are adriamycin, bleomycin, epipodophylotoxin, vincristine and cyclosporin A that exhibit a significantly lower BBB permeability than expected from their lipophilicity. The broken line represents the value of the octanol-water partition coefficient divided by the square root of the molecular weight of BCECF-AM.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Because BCECF-AM is susceptible to extensive and rapid enzymatic hydrolysis during the uptake study and DMSO was added to the medium, we performed three different experiments to determine the BBB PS value. The apparent in vitro BBB PS value of BCECF-AM determined was 446 ± 7 µl/min/g brain (fig. 2), indicating significant passage of BCECF-AM at the BBB. As no BCECF, an acid form of BCECF-AM, was observed in the uptake medium during the study (data not shown), the apparent PS value of BCECF-AM does not involve that of BCECF. The in vitro uptake study shown in figure 2 suggests no adsorption of BCECF-AM to the surface of bovine brain capillary endothelial cell membrane, despite the fact that BCECF-AM has a hydrophobic nature, i.e., the octanol-water partition coefficient determined was 5.66 ± 0.27. Therefore, the apparent PS value of BCECF-AM determined by the in situ brain perfusion and the in vivo carotid artery injection may reflect membrane permeation at the BBB. The in vivo BBB PS value of BCECF-AM determined by the in situ brain perfusion was 295 ± 48 µl/min/g brain (table 1), similar to that determined in vitro, 446 ± 7 µl/min/g brain. This result agrees with previous report demonstrating that there is reasonably good agreement between the in vitro and in vivo BBB PS value for HMG-CoA reductase inhibitors (Saheki et al., 1994).

Regarding the transport mechanism of BCECF-AM at the BBB, the concentration- and temperature-dependence were examined for the initial uptake rate of BCECF-AM into the cultured monolayer of BCEC. Because the solubility of BCECF-AM is limited, we could not perform the uptake study at concentrations greater than 5 µM. However, the fact that no significant saturation of the uptake was observed over the concentration range 0.1 to 5.0 µM suggests that BCECF-AM is transported into brain capillary endothelial cells by passive diffusion. Moreover, the apparent activation energy obtained for the initial uptake of BCECF-AM was 5.09 kcal/mol, showing only a limited energy requirement for the transport process. As this small activation energy corresponds to that of passive diffusion (Martin et al., 1983), it is likely that BCECF-AM is transported at the BBB by passive diffusion. Although we could not rule out the possibility that BCECF-AM is transported by carrier-mediated transport and the expression of carrier protein is significantly reduced in the cultured brain capillary endothelial cells, the PS value of BCECF-AM obtained in the in vivo experiments would also reflect the passive diffusion across the BBB.

The in vivo BBB PS value obtained by the carotid artery injection, 132 ± 24 µl/min/g brain, was less than half that in the in situ brain perfusion and the in vitro uptake studies. It has been reported that the injectate mixes partially with the circulating blood after a rapid administration via the carotid artery (Pardridge et al., 1985). Moreover, in preliminary studies, BCECF-AM was hydrolyzed immediately in whole blood at 37°C (data not shown). Therefore, BCECF-AM would be hydrolyzed partially in the internal carotid artery before uptake at the BBB, resulting in an apparently lower estimate for the permeability rate. The other possible explanation is that DMSO, which was added at a concentration of 10% to the injection solution, causes the reduced permeability of BCECF-AM. However, the observation of no difference in the apparent uptake of [¹⁴C]-mannitol, [³H]-water, [¹⁴C]-butanol and [³H]-3OMG in the presence and absence of 10% DMSO in the injectate (table 2) suggests that the carotid artery injection of DMSO does not cause any significant change in cerebrovascular volume, apparent cerebral blood flow rate and hexose transport rate at the BBB.

The apparent elimination rate constant of [¹⁴C]-butanol, determined in the carotid artery injection washout study, was 0.34 ± 0.02 min¹ (fig. 3), which was similar to the previously reported value (Pardridge and Fierer, 1985), suggesting again that the carotid artery injection of 10% DMSO does not cause any difference in cerebral blood flow rate. No apparent efflux of BCECF-AM was demonstrated (fig. 3). Based on the BBB PS value of 132 ± 24 µl/min/g brain in the carotid artery injection study, the efflux rate constant is predicted as 0.132 min¹, assuming that BCECF-AM is effluxed from the brain across the BBB at the same rate as the influx process (passive diffusion) and it is not bound to any constituents of brain tissue, i.e., the cerebral distribution volume equals 1.0 ml/g brain. A large difference in the apparent efflux rate constant between observed and predicted values would be explained by the hypothesis that BCECF-AM is hydrolyzed to BCECF immediately after crossing the BBB, resulting in no apparent efflux of BCECF from the brain. Because BCECF has been reported to be transported by an ATP-dependent transporter in tumor cells (Allen et al., 1990; Collington et al., 1991, 1992) and isolated rat hepatocytes (Takeguchi et al., 1993), it is possible that BCECF is pumped out of the brain to a significant extent by the efflux transport system at the BBB. However, no apparent efflux was demonstrated. It is also possible that BCECF-AM is hydrolyzed in the extracellular fluid space of the brain and/or brain parenchymal cells and BCECF does not permeate through the abluminal membrane of endothelial cells.

As far as the development of new drugs is concerned, cerebral distribution is one of the most important pharmacokinetic characteristics governing pharmacological and toxicological responses in the central nervous system. Although lipophilicity has been believed to play a dominant physicochemical role in determining BBB PS value, several drugs have been found to exhibit a remarkably low permeability compared with their significant lipophilicity (Levin, 1980; Cefalu and Pardridge, 1985). To explain this contradiction, one possible explanation has been that the brain capillary endothelial cell membrane has a molecular weight threshold, e.g., 400 Da (Levin, 1980) or 600 Da (Pardridge, 1994) restricting permeation of substrate by passive diffusion. Although these apparent molecular threshold phenomena for the BBB transport could not be explained by the simple diffusion model, a "pore" or "kink" model of membrane lipid-mediated transport has been proposed previously (Trauble, 1971; Pardridge, 1991). In this model, it is considered that kinks are created by rotations about the C-C bonds of the hydrocarbon chains of membrane phospholipids and a drug molecule having small molecular radius can enter into the kinks formed at the surface of the membrane and then be transported through a kink-mediated "molecular hitch-hiking" mechanism. Therefore, a drug molecule having a greater molecular radius than that of the pore would be significantly restricted to transport through the membrane lipid bilayer (Trauble, 1971, Pardridge, 1991). However, as shown in figure 4, BCECF-AM, with a molecular weight of 808.7 Da, permeated the BBB at a comparable or greater rate than that estimated from its lipophilicity and molecular weight, using the previously mentioned relationship (Levin, 1980; Cornford et al., 1982), suggesting that the BBB does not restrict the permeation of substrate with a molecular weight of 800 Da or less. Previously, Levitan et al. (1984) have examined the BBB transport of erythrosine B, a fluorescent dye, having a molecular weight of 879.9. As shown in figure 4, the in vivo BBB PS value of erythrosine B, i.e., 39 µl/min/g brain, agreed well with the predicted BBB PS value using equation 2 and the molecular weight and the octanol-water partition coefficient, 0.713, reported previously (Levitan et al., 1984). These results also support the notion that a drug molecule having a molecular weight of more than 600 can cross the BBB at a rate predicted from its lipophilicity and molecular weight.

In conclusion, our study demonstrates that the BBB does not restrict the permeation of BCECF-AM, a model substrate with a molecular weight of more than the putative threshold of 600 Da.