Abstract
The purpose of the present study was to characterize rat organic anion transporter (Oat) 3 (Oat3, Slc22a8) in the efflux transport at the inner blood-retinal barrier (BRB). Reverse transcription-polymerase chain reaction analysis showed that rat (r) Oat3 mRNA is expressed in retinal vascular endothelial cells (RVECs), but not rOat1 and rOat2 mRNA. The expression of Oat3 in the retina and human cultured retinal endothelial cells was further confirmed by Western blot analysis. Immunohistochemical staining in RVECs showed that rOat3 is colocalized with glucose transporter 1, but not P-glycoprotein, suggesting that rOat3 is possibly located at the abluminal membrane of the RVEC. The contribution of rOat3 to the efflux of [3H]p-aminohippuric acid ([3H]PAH), [3H]benzylpenicillin ([3H]PCG), and [14C]6-mercaptopurine ([14C]6-MP), substrates of rOat3, from the vitreous humor/retina to the circulating blood across the inner BRB was evaluated using the microdialysis method. [3H]PAH, [3H]PCG, [14C]6-MP, and [14C] or [3H]d-mannitol, a bulk flow marker, were biexponentially eliminated from the vitreous humor after vitreous bolus injection. The elimination rate constant of [3H]PAH, [3H]PCG, and [14C]6-MP during the terminal phase was approximately 2-fold greater than that of d-mannitol. This efflux transport was reduced in the retinal presence of probenecid, PAH, and PCG, whereas it was not inhibited by digoxin. In conclusion, rOat3 is expressed at the inner BRB and involved in the vitreous humor/retina-to-blood transport of PAH, PCG, and 6-MP. This transport system is one mechanism to limit the retinal distribution of PAH, PCG, and 6-MP.
The inner blood-retinal barrier (BRB) is a selectively permeable interface between the circulating blood and neural retina and is essential for the maintenance of neural functions (Cunha-Vaz, 2004; Hosoya and Tomi, 2005). Although the inner BRB forms complex tight junctions of retinal capillary endothelial cells to restrict nonspecific transport between the circulating blood and neural retina (Cunha-Vaz, 2004), retinal capillary endothelial cells express a variety of unique transporters that play essential roles in supplying nutrients to the retina and are responsible for the efflux of endobiotics and xenobiotics (Hosoya and Tomi, 2005).
Understanding the transport mechanisms of organic an-ions at the inner BRB will provide pharmacologically important information about the effective delivery of many anionic drugs to the retina. Moreover, the transport processes for organic anions at the inner BRB are physiologically important because hormones and neurotransmitters are mostly metabolized in the form of organic anions in the retina and need to undergo efflux transport from the retina to the circulating blood. We recently reported the use of a microdialysis method that showed that estradiol 17-β glucuronide undergoes efflux transport via organic anion transporter polypeptide (oatp) 1a4 (Slco1a4; oatp2) at the BRB (Katayama et al., 2006).
Organic anion transporter(s) (Oat, SLC22a) play essential roles in the disposition of clinically important anionic drugs, including antibiotics, antitumor drugs, and anti-HIV and anti-inflammatory agents in the body. Oat1-3 are mainly localized at the basolateral membrane of the kidney and use a tertiary transport mechanism to move organic anions across the basolateral membrane into the proximal tubule cells for subsequent exit/elimination across the apical membrane into the urine (Rizwan and Burckhardt, 2007). As far as these Oats are concerned, Kusuhara et al. (1999) have provided evidence using Northern blot analysis that Oat3 mRNA is present in the rat retina.
Although antibiotics are used for the treatment of bacterial endophthalmitis, most antibiotics exhibit poor penetration of the vitreous humor and retina after systemic administration. However, the elimination rate of β-lactam antibiotics, such as carbenicillin and cefazolin, after intravitreal injection in rhesus monkeys is three times faster than that of aminoglycoside antibiotics, such as gentamicin, and this elimination is inhibited in the presence of probenecid (Barza et al., 1983). Betz and Goldstein (1980) demonstrated that the uptake of p-aminohippuric acid (PAH) by isolated retinal capillaries is slightly greater than that of the extracellular marker, sucrose, and is inhibited by fluorescein and penicillin. Although these pieces of evidence suggest that Oat(s) is involved in efflux transport of PAH and anionic β-lactam antibiotics across the inner BRB because Oats accept PAH, and cefazolin as a substrate (Rizwan and Burckhardt, 2007), currently, very little is known about their expression and transport functions at the inner BRB.
The purpose of the present study was to investigate the vitreous humor/retina-to-blood transport of Oat3-mediated drugs, such as PAH, benzylpenicillin (PCG), and 6-mercaptopurine (6-MP), and clarify the Oats involved in anionic drug efflux transport at the BRB. The in vivo efflux transport of anionic drugs was characterized using microdialysis (Katayama et al., 2006). The expression and localization of Oat3 at the inner BRB was determined by reverse transcription (RT)-polymerase chain reaction (PCR), Western blot, and immunohistochemical analyses.
Materials and Methods
Animals. Male Wistar rats, weighing 250 to 300 g, were purchased from Nippon SLC (Hamamatsu, Japan). The investigations using animals described in this report conformed to the provisions of the Animal Care Committee, University of Toyama (no. 2006-3, 4), and the Association for Research in Vision and Ophthalmology Statement on the Use of Animals in Ophthalmic and Vision Research.
Reagents. [3H]PAH (4.18 Ci/mmol), d-[1-14C]mannitol ([14C]d-mannitol, 56 mCi/mmol), and d-[1-3H(N)]mannitol ([3H]d-mannitol, 14.2 Ci/mmol) were purchased from PerkinElmer Life and Analytical Sciences (Waltham, MA). [3H]PCG (17.0 Ci/mmol) and [14C]6-MP (51 mCi/mmol) were obtained from GE Healthcare (Chalfont St. Giles, UK) and Moravek Biochemicals (Brea, CA), respectively. All other chemicals were of reagent grade and available commercially.
Cell Culture. Primary cultured human retinal endothelial cells were obtained from Dainippon Pharmaceutical (Osaka, Japan) and cultured in endothelial cell basal medium containing growth supplement (Cell Applications, Inc., San Diego, CA) at 37°C. All cells were seeded onto rat tail collagen type I-coated tissue culture plates (BD Biosciences, Bedford, MA) and cultured in a humidified atmosphere of 5% CO2/air.
Isolation of Rat Retinal Vascular Endothelial Cells. Magnetic beads coated with anti-rat CD31 antibodies were used to collect purified rat retinal vascular endothelial cells (RVECs), as described previously (Tomi and Hosoya, 2004). In brief, mouse anti-rat CD-31 antibodies (Millipore Bioscience Research Reagents, Temecula, CA) were incubated with Dynabeads pan mouse IgG (Invitrogen, Carlsbad, CA) overnight at 4°C to obtain magnetic beads coated with anti-rat CD31 antibodies. Rat retinas were minced and digested in 0.1% collagenase type I (Invitrogen) and 0.01% DNase I (Roche Diagnostics, Mannheim, Germany) in Ca2+- and Mg2+-free Hanks' balanced salt solution for 30 min at 37°C with agitation. The digests were filtered through a 30-μm nylon mesh and then centrifuged at 200g for 10 min. The pellets were resuspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Moregate Biotech, Bulimba, Australia) and incubated with magnetic beads coated with anti-rat CD31 antibodies for 1 h at room temperature. RVECs labeled with the magnetic beads were positively selected by affinity binding to a magnet.
RT-PCR Analysis. Total cellular RNA was prepared from phosphate-buffered saline-washed cells using an RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany). Single-strand cDNA was made from 1 μg of total RNA by RT using oligo(dT) primer. The PCR was performed using a gene amplification system (GeneAmp PCR system 9700; Applied Biosystems, Foster City, CA) with rOat1, rOat2, rOat3, and β-actin-specific primers through 40 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min. The sequences of the specific primers were as follows: the sense sequence was 5′-GGC ACC TTG ATT GGC TAT GTC TAC-3′, and the antisense sequence was 5′-CAG CAT GGA GAG ACA GAG GAA GA-3′ for rat Oat1 (SLC22a6, GenBank accession no. NM_017224); the sense sequence was 5′-ATC AGC ACC GTC TTC TGG TCG G-3′, and the antisense sequence was 5′-CAT CAT GCA GCA CAG TGA GAT ATG-3′ for rat Oat2 (SLC22a7, GenBank accession no. AY816233); the sense sequence was 5′-ATC TCA TCA ACA TCT ATT GGG TAC TG-3′, and the antisense sequence was 5′-CAG AGA GAG ACA GAA GGT CAC AC-3′ for rat Oat3 (SLC22a8, GenBank accession no. AB017446); and the sense sequence was 5′-ATC TCA TCA ACA TCT ATT GGG TAC TG-3′, and the antisense sequence was 5′-GTG TGA CCT TCT GTC TCT CTC TG-3′ for the β-actin (GenBank accession no. NM_031144). The PCR products were separated by electrophoresis on an agarose gel in the presence of ethidium bromide and visualized under ultraviolet light. The molecular identity of the resultant product was confirmed by sequence analysis using a DNA sequencer (ABI PRISM 310; Applied Biosystems).
Quantitative Real-Time PCR. Quantitative real-time PCR was performed using an ABI PRISM 7700 sequence detector system (Applied Biosystems) with 2× SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's protocol. To quantify the amount of specific mRNA in the samples, a standard curve was produced for each run using the plasmid (pGEM-T Easy Vector; Promega, Madison, WI) containing the gene of interest. This enabled standardization of the initial mRNA content of cells relative to the amount of β-actin. PCR was performed using rat Oat3 or β-actin-specific primers.
Western Blot Analysis. Proteins were obtained by dissolving cells in sample buffer consisting of 5% SDS, 250 mM Tris-HCl, pH 6.8, 10% glycerol, 6% 2-mercaptoethanol, and 0.01% bromphenol blue, followed by heating for 10 min at 95°C and centrifugation for 10 min at 4°C and 9000g. Supernatants were separated and used as whole-cell extracts. The protein (20 μg for retina and 5 μg for other) was electrophoresed on an SDS-polyacrylamide gel and subsequently electrotransferred to a polyvinylidene difluoride membrane. After incubation with blocking agent solution (Block Ace; Dainihon Pharmaceutical, Osaka, Japan), the membranes were incubated with rabbit anti-rOat3 antibody (1.0 μg/ml) (Mori et al., 2003) for 16 h at 4°C. The membranes were subsequently incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG. The bands were visualized using an enhanced chemiluminescence kit (GE Healthcare).
Immunohistochemical Analysis. Under deep pentobarbital anesthesia (50 mg/kg body weight i.p.), rats were perfused transcardially with 4% formaldehyde in 0.1 M phosphate buffer. The eyeball was isolated and immersed in 0.5 M sucrose-0.1 M phosphate buffer. Sections (14 μm in thickness) were cut from the frozen eye using a Leica cryostat (CM1900; Leica, Wetzlar, Germany) and mounted onto silanized glass slides (Dako North America, Inc., Carpinteria, CA). After incubation with 10% goat serum (Nichirei, Tokyo, Japan) for 1 h at room temperature, sections were incubated with rabbit anti-rOat3 antibody (1.0 μg/ml) (Mori et al., 2003) or guinea pig anti-GLUT1 antibody (0.5 μg/ml) (Sakai et al., 2003) or mouse anti-P-glycoprotein antibody (C219; Signet Laboratories, Dedham, MA; 1.0 μg/ml) overnight at 4°C. Sections were subsequently incubated with Alexa 488-conjugated goat anti-rabbit IgG antibodies (1:200; Invitrogen) and Cy3-conjugated anti-mouse or anti-guinea pig IgG antibodies (1:100; Millipore Bioscience Research Reagents) for 1 h at room temperature. Sections were then mounted on coverslips using Vectashield mounting medium (Vector Laboratories, Burlingame, CA) and viewed using a confocal laser microscope (LSM 510; Carl Zeiss GmbH, Jena, Germany).
Microdialysis Study. The rats were anesthetized with an intraperitoneal injection of sodium pentobarbital (60 mg/kg), and their heads were mounted on a stereotaxic frame (SR-6; Narishige, Tokyo, Japan). Their eyelids were locally anesthetized by instillation of 2% Xylocaine (lidocaine) and fixed using surgical sutures to prevent blinking. A 25-gauge needle was inserted approximately 1 mm below the corneal scleral limbus through the pars plana at a depth of 3.0 mm. The needle was removed, and the [3H]test compound (5 μCi) and [14C]d-mannitol (1 μCi) or the [14C]test compound (1 μCi) and [3H]d-mannitol (5 μCi), dissolved in 1.0 μl of Ringer-HEPES solution (141 mM NaCl, 4.0 mM KCl, 2.8 mM CaCl2, 10.0 mM HEPES, pH 7.4), were administered using a 10-μl microsyringe (Hamilton Co., Reno, NV) at a depth of 3.0 mm from the surface of the eye. The microdialysis probe (TEP-50; Eicom, Kyoto, Japan) was implanted immediately into the vitreous chamber and fixed with surgical glue (Aron Alpha A, Sankyo, Tokyo) on the surface of the eye. Ringer-HEPES solution at 37°C was delivered to the probe continuously at 2 μl/min via polyethylene tubing (SP19; inner diameter, 0.35 mm; outer diameter, 1.05 mm; Natsume, Tokyo, Japan) connected to an infusion pump (model 11; Harvard Apparatus Inc., Holliston, MA). The dialysate sampling and determination of radioactivity were performed according to a previous report (Katayama at al., 2006).
For the experiment conducted in the presence of inhibitors, each inhibitor was dissolved in Ringer-HEPES solution and delivered to the probe as described above. The retinas were isolated at different times, and the concentrations of PAH, PCG, sulfobromophthalein (BSP), and probenecid in the retina were determined by high-performance liquid chromatography using a 4.6 × 15-mm Inertsil ODS-2 column (GL Sciences, Tokyo, Japan) and a UV monitor. The concentration of digoxin in the retina was determined using a kit (TDXTM-digoxin; Abbott Laboratories, Abbott Park, IL). The plateau concentration of inhibitors in the retina was reached within 1 h.
The probe recovery was assessed in the in vitro medium containing a constant level of compounds and calculated as: 
where CT (disintegrations per minute per milliliter) is the concentration in the dialysate solution, and CV (disintegrations per minute per milliliter) is the concentration in the test solution. The lag time for the passage through the polyethylene tubing connected to the distal end of the microdialysis probe was approximately 6 min. The recovery of [3H]PAH, [3H]PCG, and [14C]6-MP from the solution was 4.89 ± 0.06, 6.13 ± 0.07, and 7.08 ± 0.10% (n = 3), respectively. The recovery of [14C]d-mannitol and [3H]d-mannitol from the solution was 5.51 ± 0.07 (n = 6) and 5.03 ± 0.06% (n = 3), respectively. Recovery values were constant over 180 min after the lag time.
Data Analysis. The vitreous concentrations normalized by the injected dose [CP (percentage dose per milliliter)] were estimated from the radioactivities in the dialysate using eq. 2, and the CP at time t, Cp(t) was fitted to a biexponential equation (eq. 3) by nonlinear least-squares regression analysis using the computer program MULTI (Yamaoka et al., 1981): 
where CT is the concentration in the dialysate (disintegrations per minute per milliliter), and Dosetracer (disintegrations per minute) is the total radioactivity of substance after intravitreal injection. The constants A and B are intercepts on the y-axis for each exponential segment of the curve in eq. 3. The constants α and β are the apparent first order rate constants for the initial and terminal phase, respectively. The pharmacokinetic parameters (A, α, B, and β) were successfully estimated by the damping Gauss-Newton method.
Unless otherwise indicated, all data represent means ± S.E.M. An unpaired, two-tailed Student's t test was used to determine the significance of differences between two groups. Statistical significance of differences among means of several groups was determined by one-way analysis of variance followed by the modified Fisher's least-squares difference method.
Results
The Expression of rOat3 in Retinal Capillaries. The expression of rOat1, rOat2, and rOat3 in rat RVECs was examined using RT-PCR analysis (Fig. 1A). rOat3 mRNA was amplified at 372 bp in rat RVECs and total retina and in the kidney, which is a positive control, whereas rOat1 and rOat2 mRNA were not. The mRNA expression level of rOat3 in the RVECs and non-RVECs was determined by quantitative real-time PCR. The quantity of rOat3 mRNA in the RVECs compensated with β-actin (rOat3/β-actin mRNA) was 1.51 × 10-4 ± 0.38 × 10-4, whereas it was not detected in the non-RVECs (<1.52 × 10-5). In accordance, the expression of rOat3 mRNA in the RVECs was at least more than 10-fold greater than that in the non-RVECs.
The expression of Oat3 protein was also determined in the rat retina and primary cultured human retinal endothelial cells by Western blot analysis as shown in Fig. 1B. Rat kidney and brain were used as a positive control. The band of rOat3 in the retina and Oat3 in primary cultured human retinal endothelial cells were detected at 50 kDa and were identical to the reported value (Hasegawa et al., 2002; Mori et al., 2003). The localization of rOat3 in the retinal capillaries was determined by immunohistochemical analysis (Fig. 1C). Positively stained signals were observed around the retinal capillaries for rOat3 (green, Fig. 1, C1, C3, and C7), P-glycoprotein (red, Fig. 1C4), and GLUT1 (red, Fig. 1C8). The signal for rOat3 (green, arrowhead) did not overlap with that for P-glycoprotein, which is a marker for the luminal side of the retinal capillary (red, arrow, Fig. 1C6) (Hosoya and Tomi, 2005) and overlapped with that for GLUT1, which is localized on both the luminal and abluminal sides of the retinal capillary (red, arrowhead, Fig. 1C10) (Takata et al., 1992). rOat3 immunoreactivity was also detected in the ganglion cell layer (Fig. 1C1). No significant fluorescence was observed in retinal capillaries in the absence of the anti-rOat3 antibody (Fig. 1C2). These results support the hypothesis that rOat3 is mainly localized on the abluminal side of the inner BRB.
Time Profile of the Efflux of [3H]PAH, [3H]PCG, and [14C]6-MP from the Vitreous Humor. The time profile of the remaining percentage of [3H]PAH, [3H]PCG, and [14C]6-MP in the vitreous humor after bolus injection is shown in Fig. 2. [14C]d-Mannitol or [3H]d-mannitol were coinjected with [3H] and [14C]radiolabeled test compounds, respectively, as bulk flow markers. [3H]PAH, [3H]PCG, [14C]6-MP, and [14C]/[3H]d-mannitol were biexponentially eliminated from the vitreous humor. Although the initial rapid decline of [3H]PAH, [3H]PCG, and [14C]6-MP was almost the same as that of [14C]/[3H]d-mannitol, the second decline in [3H]PAH, [3H]PCG, and [14C]6-MP was significantly greater than that of [14C]/[3H]d-mannitol (Fig. 2). The apparent elimination rate constant (β) during the terminal phase was 0.0124 ± 0.0006 min-1 for PAH, 0.0156 ± 0.0008 min-1 for PCG, and 0.0197 ± 0.0014 min-1 for 6-MP and 1.68-, 1.71-, and 1.97-fold greater than that of d-mannitol, respectively (Fig. 3).
Expression of rOat3 at the inner BRB. A, RT-PCR analysis of rOat1, rOat2, and rOat3 in the RVEC fraction, non-RVEC fraction, and retina. Lane 1, RVEC fraction; lane 2, non-RVEC fraction; lane 3, rat retina; lane 4, positive control of rat kidney for rOat1 and rOat3 and rat liver for rOat2. + and -, presence or absence of RT, respectively. B, Western blot analysis of Oat3 in the retina and human cultured retinal endothelial cells. Lane 1, rat kidney; lane 2, rat brain; lane 3, rat retina; lane 4, primary cultured human retinal endothelia cells. Rat kidney and brain were used as a positive control. C, rOat3 immunoreactivities were detected in rat retinal capillary endothelial cells by immunofluorescence. Rat retina was stained with anti-rOat3 antibody (green, arrowhead; C1, C3, C5, C6, and C7), anti-P-glycoprotein antibody (red; C4-C6), and anti-GLUT1 antibody (red; C8,-C10). Low (C1 and C2), middle (C3-C5 and C7-C9), and high (C6 and C10) magnification views under confocal scanning microscopy. The signal for rOat3 (green, arrowhead) did not overlap with that for P-glycoprotein (red, arrow; C6) and overlapped with that for GLUT1 (red, arrowhead; C10). C2, no significant fluorescence was observed in retinal capillaries in the absence of the anti-rOat3 antibody. Scale bar, 10 μm.
Time profile of [3H]PAH (A), [3H]PCG (B), and [14C]6-MP (C) in the vitreous humor after vitreous bolus injection. A, outflow pattern of [3H]PAH and [14C]d-mannitol from the microdialysis probe. Open circles and triangles, concentration in dialysate of [14C]d-mannitol and [3H]PAH, respectively. Each point represents the mean ± S.E.M. (n = 16). B, outflow pattern of [3H]PCG and [14C]d-mannitol from the microdialysis probe. Open circles and triangles, concentration in dialysate of [14C]d-mannitol and [3H]PCG, respectively. Each point represents the mean ± S.E.M. (n = 6). C, outflow pattern of [14C]6-MP and [3H]d-mannitol from the microdialysis probe. Open circles and triangles, concentration in dialysate of [3H]d-mannitol and [14C]6-MP, respectively. Each point represents the mean ± S.E.M. (n = 5).
Elimination rate constants (β) of [3H]PAH, [3H]PCG, [14C]6-MP, and [14C]/[3H]d-mannitol during the terminal phase. Each column represents the mean ± S.E.M. (n = 5-16).
Inhibitory Effect of Several Organic Anions on Efflux of [3H]PAH, [3H]PCG, and [14C]6-MP from the Vitreous Humor. To examine whether the difference between the β value of [3H]PAH, [3H]PCG, and [14C]6-MP and [14C]/[3H]d-mannitol contributes to carrier-mediated efflux transport, an inhibition study was performed. Table 1 shows the elimination rate differences between test compounds and d-mannitol in the presence or absence (control) of organic anions during the terminal phase. The efflux of [3H]PAH was significantly inhibited in the presence of 1.5 mM PAH, 1.6 to 8.3 mM PCG, 1.0 mM probenecid, and 0.6 mM BSP, whereas 0.35 μM digoxin had no effect. The efflux of [3H]PCG was significantly inhibited in the presence of 1.6 to 8.3 mM PCG, 3.3 mM PAH, and 1.0 mM probenecid, whereas 0.35 μM digoxin had no effect. Moreover, the efflux of [14C]6-MP was significantly inhibited in the presence of 3.3 mM PAH, 1.6 mM PCG, and 1.0 mM probenecid, whereas 0.35 μM digoxin had no effect.
Effect of several inhibitors on the elimination rate difference between test compounds ([3H]PAH, [3H]PCG, or [14C]6-MP) and bulk flow markers ([14C]D-mannitol or [3H]D-mannitol) during the terminal phase Each inhibitor was perfused in the microdialysis probe. Each item of data represents the mean ± S.E.M. (n = 3-16). Percentage of control was calculated as follows: (βvalue of test compound - β value of [14C]/[3H]D-mannitol in the presence of inhibitor)/(β value of test compound - β value of [14C]/[3H]D-mannitol in the absence of inhibitor) × 100.
Discussion
The present study provides, for the first time, evidence that rOat3 is expressed in retinal capillaries and demonstrates that PAH, PCG, and 6-MP, which are substrates of rOat3 (Mori et al., 2004; Rizwan and Burckhardt, 2007), undergo efflux from the vitreous humor and these efflux transports are inhibited in the presence of rOat3 substrates and inhibitors. In the RT-PCR analysis (Fig. 1A), the amplified product of rOat3 was detected in RVECs and total retina and in the kidney, suggesting rOat3 expression in the rat retina. In contrast, no expression of rOat1 and rOat2 mRNA in RVECs was detected in this study. Moreover, quantitative real-time PCR revealed that the rOat3 mRNA level in RVECs was more than 10-fold greater than that in non-RVECs, suggesting that rOat3 is predominantly expressed in RVECs. The expression of Oat3 protein in retinal endothelial cells was further confirmed by Western blot analysis (Fig. 1B). Because the RVEC protein obtained by magnetic isolation was not enough to perform Western blotting (Tomi and Hosoya, 2004), primary cultured human retinal endothelial cells were used to examine whether inner BRB expresses Oat3 protein. The bands at 50 kDa were detected in rat retina and primary cultured human retinal endothelial cells and in the rat kidney and brain, where rOat3 is reported to be expressed (Kikuchi et al., 2003; Mori et al., 2003). Immunohistochemical staining demonstrated that rOat3 is most probably expressed on the abluminal side of the retinal capillaries because the signal for rOat3 was located slightly outside that of P-glycoprotein and merged with that of GLUT1 (Fig. 1C). In the blood-brain barrier, rOat3 is expressed on the abluminal side of the brain endothelial cells and plays a role in taking up endogenous and exogenous organic anions (Kikuchi et al., 2003; Mori et al., 2003; Ohtsuki et al., 2004).
To discover whether rOat3 at the inner BRB is involved in efflux transport of anionic drugs, a microdialysis study was performed using [3H]PAH, [3H]PCG, and [14C]6-MP as substrates of Oat3 (Mori et al., 2004; Rizwan and Burckhardt, 2007). The dialysate concentration of [3H]PAH, [3H]PCG, and [14C]6-MP and the bulk flow marker, [14C]/[3H]d-mannitol, in the vitreous humor via the microdialysis probe decayed in a biexponential manner (Fig. 2). The initial slope of the drug concentration-time profile was steeper than that of the later slope, supporting the hypothesis that the first and second declines reflect diffusion into the vitreous humor, including the microdialysis tube, after vitreous bolus injection and elimination from the vitreous humor of both test compound and d-mannitol, respectively (Katayama et al., 2006). The β-values of [3H]PAH, [3H]PCG, and [14C]6-MP were 1.68- to 1.97-fold greater than that of [14C]/[3H]d-mannitol, which is a bulk flow marker for passage from the vitreous humor to Schlemm's canal and/or the uveoscleral outflow route (Fig. 3), showing that [3H]PAH, [3H]PCG, and [14C]6-MP undergo efflux transport across the BRB in addition to elimination from the vitreous humor via bulk flow and passive diffusion. Moreover, the β-values of [3H]PAH, [3H]PCG, and [14C]6-MP were reduced in the presence of 1.5 to 3.3 mM PAH, 1.6 to 8.3 mM PCG, and 1.0 mM probenecid, suggesting that PAH, PCG, and 6-MP are transported through a probenecid-sensitive and carrier-mediated efflux transport process(es) across the BRB. The efflux transport of [3H]PAH, [3H]PCG, and [14C]6-MP from rat vitreous/retina was significantly inhibited by 1.0 mM probenecid or 0.6 mM BSP and by PAH and PCG in a concentration-dependent manner, whereas 0.35 μM digoxin had no effect (Table 1). Probenecid and BSP are inhibitors of organic anion transporters, including rOat3 (Kusuhara et al., 1999). PAH is a common substrate of rOat1 (Michaelis constant, Km = 14 μM) and rOat3 (Km = 65 μM) (Sekine et al., 1997; Kusuhara et al., 1999). PCG is a relatively specific substrate of rOat3 with a Km of 83 μM (Kikuchi et al., 2003). On the other hand, digoxin is a specific substrate of oatp 1a4, with a Km of 0.24 μM (Noé et al., 1997). The 50% inhibition concentration of PCG and PAH for [3H]PAH and [3H]PCG efflux transport was around 3.3 mM. The affinity for rOat3 of PAH and PCG is somewhat lower than that reported based on Km and inhibition constant values. Because the inhibitor concentration in the retina was determined using retinal homogenate, it is conceivable that the inhibitor concentration in interstitial fluid in the retina is smaller than that in the retinal homogenate. Nevertheless, this evidence suggests that rOat3 is involved in the efflux transport of PAH and PCG across the inner BRB.
Relapse of childhood acute lymphoblastic leukemia involving the eye is a rare but challenging problem. One possible reason may involve the restricted distribution of 6-MP, which is frequently used in patients with acute lymphoblastic leukemia, in the eye (Somervaille et al., 2003). The distribution of β-lactam antibiotics in the vitreous humor/retina is restricted, resulting in reduced neurotoxicity and difficulties in treating bacterial endophthalmitis. In contrast, β-lactam antibiotics are rapidly eliminated from vitreous humor compared with gentamicin and capable of being inhibited by probenecid (Barza et al., 1983). The present study shows [3H]PCG and [14C]6-MP efflux transport from the vitreous humor/retina, which can be inhibited by probenecid. These results support the hypothesis that rOat3, which is mostly localized on the abluminal side of the inner BRB, contributes to the restricted distribution and efflux transport of 6-MP and β-lactam antibiotics in the retina. It is likely that some β-lactam antibiotics are substrates of oatp 1a4 (Nakakariya et al., 2008), which is present at the rat inner BRB and apical side of outer BRB (Gao et al., 2002), and it is possible that other transporters are also involved in the uptake of anionic β-lactam antibiotics.
The elimination of compounds from the retina into the systemic circulation consists of two steps, i.e., uptake across the abluminal membrane and subsequent excretion across the luminal membrane. Regarding the transport at the luminal membrane, it is conceivable that ATP-binding cassette (ABC) transporter C [multidrug resistance protein (MRP)] removes anionic drugs, including β-lactam antibiotics, in the retinal endothelial cells. We previously used mouse RVECs to show that ABCC3 (MRP3), C4 (MRP4), C6 (MRP6), and ABCG2 mRNAs are predominantly expressed at the inner BRB (Tachikawa et al., 2008). In particular, MRP4 is a very likely candidate for the transport of these anionic drugs because MRP4 transports several anionic drugs, such as PAH, and β-lactam antibiotics, such as PCG (Smeets et al., 2004; Uchida et al., 2007).
From a physiological viewpoint, rOat3 at the inner BRB could be responsible for efflux transport of metabolites of neurotransmitters from the neural retina. Homovanillic acid, an end metabolite of dopamine, undergoes efflux transport from the brain to the circulating blood via rOat3 of the blood-brain barrier (Mori et al., 2003). Moreover, vanillyl-mandelic acid, 3,4-dihydroxymandelic acid, and 4-hydroxy-3-methoxyphenylglycol derived from norepinephrine and epinephrine, 5-hydroxyindoleacetic acid and 5-methoxytryptophol derived from serotonin, and imidazole-4-acetic acid and 1-methyl-4-imidazolic acid derived from histamine are potential substrates of rOat3 (Mori et al., 2003). Because these end products of neurotransmitters are mostly present in the retina (Pourcho, 1996; Witkovsky, 2004), rOat3 at the inner BRB could play a key role in regulation of their retinal concentration.
In conclusion, this study provides the first evidence that rOat3 is most probably expressed on the abluminal side of the retinal capillaries and plays an important role in the efflux transport of PAH, PCG, and 6-MP from the vitreous humor/retina to blood across the inner BRB. These findings provide important information about the physiological role of the inner BRB and how the inner BRB restricts the distribution of anionic drugs and 6-MP to the retina.
Acknowledgments
We thank Masahiko Watanabe for the anti-GLUT1 antibody.
Footnotes
-
This work was supported by the Japan Society for the Promotion of Science grant-in-aid for Scientific Research.
-
doi:10.1124/jpet.108.146381.
-
ABBREVIATIONS: BRB, blood-retinal barrier; Oatp, organic anion transporter polypeptide; Oat, organic anion transporter; PAH, p-aminohippuric acid; PCG, benzylpenicillin; 6-MP, 6-mercaptopurine; RT, reverse transcription; PCR, polymerase chain reaction; RVEC, retinal vascular endothelial cell; r, rat; GLUT, glucose transporter; BSP, sulfobromophthalein; ABC, ATP-binding cassette; MRP, multidrug resistance protein.
- Received September 18, 2008.
- Accepted December 29, 2008.
- The American Society for Pharmacology and Experimental Therapeutics
