Mechanisms of Nucleobase Transport in Rabbit Choroid Plexus EVIDENCE FOR A Na (cid:49) -DEPENDENT NUCLEOBASE TRANSPORTER WITH BROAD SUBSTRATE SELECTIVITY*

The overall goal of this study was to determine the mechanisms by which nucleobases are transported in the choroid plexus. Choroid plexus tissue slices were obtained from the lateral ventricles of rabbit brains and depleted of ATP with 2,4-dinitrophenol. In the presence of an initial inwardly directed Na (cid:49) gradient, hypoxanthine accumulated in the tissue slices against a concen- tration gradient. Na (cid:49) -stimulated hypoxanthine uptake was saturable with a K m of 31.1 (cid:54) 9.71 (cid:109) M and a V max of 2.69 (cid:54) 0.941 nmol/g/s (mean (cid:54) S.E.). Na (cid:49) -stimulated hypoxanthine uptake was inhibited by (100) (cid:109) M naturally occurring purine and pyrimidine nucleobases (adenine, cytosine, guanine, hypoxanthine, thymine, uracil, and xanthine) as well as by the nucleoside analog, dideoxya- denosine. The stoichiometric coupling ratio between Na (cid:49) and hypoxanthine was 1.7:1. The data demonstrate the presence of a novel Na (cid:49) -dependent nucleobase transporter in the choroid plexus, which is distinct from the previously described Na (cid:49) -nucleoside transporter in choroid plexus and from Na (cid:49) -dependent nucleobase transporters in other tissues in terms of its kinetics, substrate selectivity, and Na (cid:49) -nucleobase stoichiometry. This transporter

Nucleobases and their structural analogs have become important drugs in the treatment of a number of viral infections of the central nervous system such as cytomegalovirus retinitis, herpes simplex encephalitis, and AIDS-related dementia complex. Therefore, nucleobase transport mechanisms at the major barriers of the central nervous system, namely the bloodbrain barrier and the blood-cerebrospinal fluid barrier (choroid plexus epithelium), may play a critical role in targeting nucleobase analogs to the affected tissues. Moreover, such transport systems may be important in the salvage of purines for nucleic acid synthesis in the brain.
Saturable, low affinity nucleobase transporters have been identified in the blood-brain barrier (1,2). However, it appears that these transporters may not be sufficient to mediate the flux of important quantities of hypoxanthine, the principal salvageable nucleobase, or other nucleobases into the brain. As a result, attention has focused on the choroid plexus as a possible route of entry for hypoxanthine and other nucleobases into the brain (3).
In mammalian cells, transcellular flux of nucleobases is mediated by specific transporters in the plasma membranes. Two major classes of nucleobase transporters have been characterized: equilibrative and concentrative. Equilibrative transporters are present in a number of cell types and are broadly selective for purine and pyrimidine nucleobases. Recent studies with acyclovir and ganciclovir, nucleobase/nucleoside analogs, demonstrate that these agents permeate the human erythrocyte by way of equilibrative mechanisms (14,15).
Recently, concentrative nucleobase transporters have been identified in LLC-PK 1 cells (4), guinea pig kidney (5,13), rat jejunal tissue (6), and guinea pig placenta (7). These transporters are secondary active and Na ϩ -dependent and mediate the influx of specific nucleobases. The substrate selectivity of the transporter in LLC-PK 1 cells has been studied in considerable detail and includes selected purine and pyrimidine nucleobases.
The goal of this study was to determine the mechanisms of nucleobase transport in the choroid plexus. Our data demonstrate the presence of a novel Na ϩ -dependent nucleobase transporter in the choroid plexus, which is distinct from previously described Na ϩ -driven nucleobase transporters in other tissues in terms of its substrate selectivity and kinetics. This transporter may play an important role in the delivery of both salvageable and therapeutic nucleobases to the central nervous system.

EXPERIMENTAL PROCEDURES
Preparation of ATP-depleted Choroid Plexus Tissue Slices-Choroid plexus tissue slices from rabbit were ATP-depleted by the method of Whittico (8). Choroid plexuses were obtained from the lateral ventricles of New Zealand White rabbits. The choroid plexus tissue was placed in KCl buffer (37°C) of the following composition: KCl (150 mM), mannitol (40 mM), and HEPES (25 mM), pH 7.4, with 1 M Tris. Choroid plexus tissue was cut into 2-3-mm pieces and ATP-depleted by incubating at 37°C for 20 min in 2,4-dinitrophenol (250 M) in KCl buffer. Under these conditions, the ATP concentration is reduced to less than 10% of control (9). Following this incubation, the choroid plexus slices in buffer were stored on ice until uptake experiments were performed.
Accumulation Studies-Uptake of [ 3 H]hypoxanthine was studied by methods published previously (10). Briefly, individual choroid plexus tissue slices were incubated with 140 l of reaction mixture containing whether metabolism or degradation of hypoxanthine had occurred. Procedures used were as described previously (10).
Data Analysis-The radioactive content from each choroid plexus tissue slice was expressed as a volume of distribution (V d ) as described previously (8,10). 3 H͔hypoxanthine/g choroid plexus dpm͓ 3 H͔hypoxanthine/ml media Ϫ dpm͓ 14 C͔mannitol/g choroid plexus dpm͓ 14 C͔mannitol/ml media Statistical analysis was carried out by a Student's unpaired t test. A probability, p, of less than 0.05 was considered significant. Data points were determined in triplicate for each experiment. Data, unless mentioned otherwise, are expressed as the mean Ϯ S.E. of data obtained from three experiments in choroid plexus tissue from separate animals. Standard methods were used to determine IC 50 values and Michaelis-Menten kinetics (10).
To determine the stoichiometric coupling between Na ϩ and hypoxanthine, a modified version of the Hill equation was used, in the reaction mixture. The data were transformed with a logarithm and linearly regressed to obtain a and n. This equation has been used previously in stoichiometry studies (12). Materials-[ 3 H]Hypoxanthine (11.6 Ci/mmol), [ 3 H]thymidine (65 Ci/ mmol), and [ 14 C]mannitol (56 mCi/mmol) were purchased from either Amersham Life Science, Arlington Heights, IL or Moravek Biochemicals, Inc., Brea, CA. Hypoxanthine, adenine, adenosine, dideoxyadenosine (ddA), guanine, xanthine, caffeine, cytosine, thymine, uracil, thymidine, and proline were purchased from either Sigma or Aldrich. All other chemicals were purchased from either Sigma, Fisher Scientific, or Aldrich. New Zealand White rabbits were purchased from Nitabell Rabbitry, Hayward, CA. Cytoscint ES scintillation fluid was purchased from ICN Biomedical Inc.

RESULTS
In the absence of a Na ϩ gradient, hypoxanthine accumulated in the tissue slices and reached an equilibrium (V d ϭ 2.57 Ϯ 0.29) in approximately 5 min (Fig. 1). This uptake was not reduced by any compound (including unlabeled hypoxanthine) in the concentration range used in these studies. These results suggest that in the absence of a Na ϩ gradient, the uptake of hypoxanthine represents a nonselective or low affinity binding process or transport process.
In contrast, in the presence of an initial inwardly directed Na ϩ gradient (150 mM), hypoxanthine accumulated temporarily ("overshoot phenomenon") in the tissue slices above the equilibrium value (V d ϭ 6.11 Ϯ 0.91 at 1 min) (Fig. 1). Thin layer chromatography (TLC) studies indicated that hypoxanthine was not significantly metabolized at 30 s or 5 min (data not shown).
Kinetic experiments (Fig. 2) were performed in which the rate of hypoxanthine uptake (at 30 s) as a function of concentration was determined in the presence of an inwardly directed Na ϩ gradient (150 mM). The data are consistent with a single saturable process. The data from each of the three experiments were fit to an appropriate Michaelis-Menten equation, which included a linear component (8). The K m and V max (mean Ϯ S.E.) for Na ϩ -stimulated hypoxanthine uptake were 31.1 Ϯ 9.71 M and 2.69 Ϯ 0.94 nmol/g/s, respectively.
To determine the stoichiometry (Fig. 3) of the Na ϩ -dependent nucleobase transport system, the Na ϩ -dependent uptake of hypoxanthine (0.24 M) was examined in the presence of increasing Na ϩ concentrations (0 -140 mM). The uptake of hypoxanthine (V d ) was sensitive to Na ϩ concentration. The data were fit to a Hill equation as described under "Experimental Procedures." The Hill coefficient was 1.7 Ϯ 0.4 (mean Ϯ S.E.) for Na ϩ :hypoxanthine and was significantly different from 1 but not from 2.
The effect of the nucleobases, hypoxanthine and thymine, on the Na ϩ -dependent uptake of the nucleoside, thymidine, was examined (Fig. 7). At concentrations of 100 M, hypoxanthine and thymine did not significantly inhibit Na ϩ -dependent thy- midine uptake at 30 s. These data suggest that nucleobases and nucleosides do not share the same Na ϩ -stimulated transporter in the choroid plexus. Further studies using NEM, an irreversible sulfhydryl modifying agent, significantly inhibited Na ϩthymidine uptake but not Na ϩ -hypoxanthine uptake (data not shown). This further demonstrates that these two compounds do not share the same transporter. DISCUSSION Previous studies demonstrated that hypoxanthine accumulates and is metabolized in rabbit choroid plexus tissue slices (3). However, the driving force of the transport mechanism(s) was not determined. By using ATP-depleted choroid plexus slices, we experimentally imposed an initial inwardly directed Na ϩ gradient and directly demonstrated that hypoxanthine transport is coupled to a Na ϩ gradient (Fig. 1). These data provide the first demonstration of a Na ϩ -nucleobase transporter in the choroid plexus.
The data in this study suggest that the Na ϩ -dependent hypoxanthine transporter in the choroid plexus is distinct from Na ϩ -dependent hypoxanthine transporters in other tissues. First, Na ϩ -dependent hypoxanthine uptake in the choroid plexus has a lower affinity (31.1 M) than Na ϩ -stimulated hypoxanthine uptake in guinea pig kidney (4.4 M) (5, 13) and LLC-PK 1 cells (0.79 M) (4) but a higher affinity than the saturable uptake system for hypoxanthine (400 M) identified in the blood brain barrier (2).
Second, the substrate selectivity of the Na ϩ -stimulated hypoxanthine transporter in the choroid plexus differs from that of Na ϩ -stimulated hypoxanthine transporters in other tissues. In the choroid plexus, the purine nucleobases, adenine, guanine, and xanthine, and the pyrimidine nucleobases, cytosine, thymine, and uracil, were potent inhibitors of Na ϩ -dependent hypoxanthine uptake (Figs. 4 and 5). In contrast, in LLC-PK 1 cells the nucleobases, adenine, cytosine, and xanthine, did not inhibit Na ϩ -stimulated hypoxanthine uptake (4). The transporter also appears to differ in its substrate selectivity from the Na ϩ -dependent nucleobase transporter in renal brush border membrane vesicles from guinea pig, which excludes adenine (5,13). A Na ϩ -dependent nucleobase transporter for pyrimidines in rat jejunal tissue has been characterized (6); however, the interactions of purines with the transporter were not examined. The interaction of ddA with the Na ϩ -dependent nucleobase transporter suggests that the choroid plexus may play a role in the transport of certain clinically relevant nucleoside analogs.
Our data are consistent with a 2:1 coupling ratio for Na ϩ - hypoxanthine transport. Na ϩ -hypoxanthine transport exhibits a 1:1 stoichiometry in LLC-PK 1 cells and a 2:1 stoichiometry in the guinea pig kidney (13). Similar data have been obtained for Na ϩ -nucleoside transport in choroid plexus. That is, a 2:1 Na ϩnucleoside stoichiometry has been obtained for the N-3, Na ϩnucleoside transporter in choroid plexus (10).
The Na ϩ -nucleobase transporter in the choroid plexus also appears to be distinct from the Na ϩ -nucleoside transporter in the same tissue. Neither cytidine (Fig. 5) nor thymidine (data not shown) inhibited Na ϩ -stimulated hypoxanthine uptake in the choroid plexus. Conversely, neither hypoxanthine (100 M) nor thymine (100 M) inhibited Na ϩ -stimulated thymidine uptake (Fig. 7). Moreover, previous studies in this laboratory have demonstrated that the sulfhydryl modifier, NEM, irreversibly inhibits Na ϩ -nucleoside transport in choroid plexus (11). However, NEM did not inhibit Na ϩ -hypoxanthine uptake in rabbit choroid plexus.
In conclusion, a Na ϩ -dependent nucleobase transporter in choroid plexus from rabbit has been characterized. This transporter is broadly selective for both purine and pyrimidine nucleobases and appears to differ from the recently characterized Na ϩ -nucleobase transporters in guinea pig kidney and LLC-PK 1 based on its kinetics and substrate selectivity. However, this cannot be confirmed until the transport has been cloned and sequenced. Further studies are also needed to determine the relative contribution of the choroid plexus (in comparison with the blood brain barrier) in transporting physiologically relevant quantities of nucleobases into the brain and the role of the transporter in targeting nucleobase and nucleoside analogs to the central nervous system.