Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Human immunodeficiency virus 1 (HIV-1) infection of the brain causes HIV-1 encephalopathy and AIDS dementia complex (ADC), a syndrome characterized by cognitive, motor, and behavioral disturbances (Navia and Price, 1998). Patients with severe ADC usually present microglial activation and multinucleated giant cells, the hallmarks of HIV encephalitis. In the central nervous system (CNS), it appears that only macrophages and microglia harbor the virus particles and the number of proviruses correlates with HIV encephalitis (Navia and Price, 1998).

Microglia are a distinct population of non-neuronal cells that are involved in maintenance of neuronal homeostasis, synaptic plasticity, and repair. These cells play a key role in CNS immune and inflammatory reactions, and are susceptible to diverse insults, ranging from physical damage to infectious agents (Gonzalez-Scarano and Baltuch, 1999). At the two extremes, microglia morphology is 1) spheroid/activated, a state that appears early in development and after brain lesions, and 2) ramified/resting microglia, present in the normal adult brain. Microglial activation, a process that often includes changes in morphology and surface expression of immune-related molecules, has been described in virtually all human neuropathologies, including neurodegenerative diseases, multiple sclerosis, and HIV-1 infection (Gonzalez-Scarano and Baltuch, 1999). The microglia used in our in vitro cell system, are spheroid, proliferating, and capable of producing nitric oxide (Hong et al., 2000; C. A. Colton and L. C. Schlichter, unpublished data).

For the effective treatment of AIDS-ADC or AIDS encephalopathy, anti-HIV drugs need to reach the CNS in significant amounts. ZDV (3'-azido-3'-deoxythymidine, AZT, Retrovir), a thymidine analog, is a potent inhibitor of the in vitro replication and cytopathic effect of HIV. Although limited CNS access has been reported for most antiretroviral drugs, after ZDV administration, substantial improvement was noted in ADC patients (Fischl et al., 1987). The clinical benefits include increased survival, reduced number of opportunistic infections, and partial improvement in neurological defects. However, ZDV treatment is limited by its hematological toxicity (anemia, neutropenia), the development of HIV resistance, and its limited accumulation in certain tissues, including the CNS, which is an important site for HIV replication.

ZDV can enter the CNS by passive diffusion across the blood-brain barrier (BBB) and, more significantly, the blood-cerebrospinal fluid (CSF) barrier (Thomas and Segal, 1997). However, efflux via a probenecid-sensitive transport system at the BBB contributed to the restricted distribution of ZDV in brain tissue (Takasawa et al., 1997b). Low CSF-to-plasma concentration ratios of ZDV after i.v. infusion have been reported in rats (0.15) (Galinsky et al., 1990), rabbits (0.26) (Wong et al., 1993), and humans (0.5) (Blum et al., 1988). Thus, the low steady-state ZDV levels in the CNS depend on both diffusive and active processes.

Because ZDV can act either as an organic anion or cation, consistent with the resonance structures of the azido moiety and pK_a value of 9.68 (Henry et al., 1988), in principle, its transport could involve nucleoside transporters, organic anion or organic cation transporters. Interestingly, ZDV interacts with an organic anion transporter at the basolateral membrane (Griffiths et al., 1991) and an organic cation transporter at the brush-border membrane (Griffiths et al., 1992) of renal cortical vesicles. Consistent with such transport, an energy-dependent organic cation carrier is inhibited by ZDV in opossum kidney cells (Chen et al., 1999). Yao et al. (1996) provided direct evidence for Na⁺-dependent ZDV transport using heterologous expression of the intestinal/kidney nucleoside transporter rCNT1 in Xenopus oocytes. The lipophilicity imparted by the azido group allows nonfacilitated diffusion of ZDV into human erythrocytes, lymphocytes, macrophages, bone marrow progenitor cells, and intestinal epithelial cells.

Our previous study identified a Na⁺-dependent nucleoside transporter highly sensitive to ZDV in microglia cells (Hong et al., 2000). However, it is not known which, if any, transporters for ZDV exist in microglia, the primary target of HIV infection in the CNS. The goal of this study was to determine the in vitro disposition of ZDV by a brain parenchymal cell line (MLS-9 microglia cells) to characterize its mechanism of transport, and identify potential interactions with other compounds at the carrier site(s).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

MLS-9 Culture. Cultures of MLS-9 microglia cells were prepared as previously described (Zhou et al., 1998). Microglia were originally enriched from neopallia of 2- or 3-day-old Wistar rats (Schlichter et al., 1996) to >98% purity, as judged by staining with isolectin B4, OX-42, and ED-1 antibody (Sigma Chemical Co., St. Louis, MO). Colony-stimulating factor 1 was used to induce proliferation and the continuous microglia cell line MLS-9 was established from one of the colonies that arose after several weeks of culturing. These cells also stained with isolectin B4 (100%), OX-42 antibody (98%), and ED-1 antibody (99%), but were negative for the astrocyte marker glial fibrillary acidic protein, and the fibroblast protein fibronectin (Zhou et al., 1998).

Uptake of ZDV. For uptake experiments, MLS-9 cells were seeded on 2-cm², 24-well plates at 5 × 10⁵ cells/well, and only those grown to a uniform, confluent monolayer (within 4-5 days) were used. The integrity of cell monolayers was evaluated microscopically, before and after the uptake measurements. Most importantly, the morphology and uptake values were the same among the cell passages used.

Metabolism of ZDV. Intracellular metabolism of [³H]ZDV during the uptake assay was monitored by thin-layer chromatography. MLS-9 cells were incubated for 1, 10, and 30 min with 0.5 mM [³H]ZDV (100 µCi/ml) at 37°C. After the uptake reaction these cell suspensions, including 1 mM solutions of the standards thymine and ZDV, were spotted and chromatographed for 4 h on 250-µm-thick silica gel-coated plates (silica gel 60; Sigma Chemical Co.) that were impregnated with a fluorescent indicator. The solvent was butan-1-ol saturated with water. When the plate was dried, the zones bearing the standards were located under UV light. The R_F values corresponding to ZDV, thymine, and ZDV nucleotides (mono-, di-, and triphosphate metabolites) were 0.83, 0.72, and 0, respectively. Each zone was scraped from the plate and analyzed for radioactivity by standard liquid scintillation counting methods.

Intracellular pH Measurements. Intracellular pH was measured fluorimetrically using the pH-sensitive carboxyfluorescein derivative 2',7'-bis(carboxyethyl)-5(6'-carboxyfluorescein (BCECF) (Molecular Probes, Eugene, OR). Single-cell measurements were performed with a Zeiss Axiovert 100TV microscope equipped with a Cohu (model 1412) charge-coupled device camera, which is controlled by the operational software Axon Imaging Work Bench, 2.1. The nonfluorescent, membrane-permeant acetoxymethyl ester BCECF-AM enters cells readily and is cleaved by cytosolic esterases to yield the nonpermeant form whose fluorescence is proportional to pH. Confluent monolayers of MLS-9 cells grown in 24-well tissue culture plates were preincubated (40 min, 37°C) in EBBS containing 20 µM BCECF-AM and 0.1% BSA. Cells were washed and preincubated in 30 mM NH₄Cl solution for 30 min, and then followed by its removal and replacement with either EBSS or 1 mM amiloride. Baseline measurements were recorded in EBSS. Fluorescence was detected with dual excitation at 440 and 495 nm, and emission at 535 nm. Emission ratios for 440- versus 495-nm excitation were computed by the software. Intracellular pH was calibrated by comparing the fluorescence intensity of cells with 100 µM BCECF acid in EBSS solution, in which the pH was buffered with 0.1 mM MES (pH 6.1), BES (pH 6.7 and 7.1), HEPES (pH 7.4 and 7.7), and Tris (pH 8.1 and 8.6). All buffers were from Sigma Chemical Co.

Data Analysis. Experiments were repeated at least two times using cells from different passages. Data points in each experiment represent quadruplicate measurements. Results are presented as mean ± S.D. from a minimum of two separate experiments. The Michaelis-Menten kinetic parameters (K_m and V_max) for ZDV transport were determined by a nonlinear least-squares analysis in Sigma Plot 4.0. To estimate the inhibitory constant (K_i) of the inhibitor verapamil, a least-squares regression analysis was used to determine the linear correlation between V and V/[ZDV], the Eadie-Hofstee linear transformation. The IC₅₀ or concentration of various organic cationic drugs causing a 50% reduction in ZDV uptake was estimated by a sigmoidal, four-parameter inhibition model. Statistical significance was assessed by the Student's t test for unpaired experimental values, the test of repeated measures of ANOVA, and/or the post hoc multiple-comparison Bonferroni t test, as appropriate. A p value <0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Metabolism of ZDV. The intracellular metabolism of ZDV has been examined extensively in cell lines, in retrovirus-infected mouse tissues, and in human peripheral blood mononuclear cells (Furman et al., 1986). To exert its antiviral activity, ZDV is converted by several intracellular kinases to its phosphorylated anabolites. That is, thymidine kinase, thymidylate kinase, and pyrimidine nucleoside diphosphate kinase consecutively convert ZDV to ZDV-monophosphate, ZDV-diphosphate, and ZDV-triphosphate, respectively. Because metabolism can complicate the interpretation of transport studies, we determined the extent of ZDV metabolism by microglia cells. After the uptake reaction, chromatographic analysis of the cell indicated no significant ZDV metabolism before 30 min. For instance, in cells incubated with 0.5 mM [³H]ZDV for 30 min at 37°C, 87% of the intracellular radioactivity was associated with the unmodified form ZDV, 5% cochromatographed with thymine, and 8% with the ZDV nucleotides (mono-, di-, and triphosphate metabolites). After a 10-min incubation, 94, 4, and 2% of the radioactivity was recovered in the ZDV, thymine, and ZDV nucleotide fractions, respectively. Similarly, after a 1-min incubation, 96, 3, and 1% of the radioactivity comigrated with ZDV, thymine, and ZDV nucleotides, respectively (data not shown).

Time Course of Specific ZDV Uptake. Figure 1 shows that specific ZDV (0.5 mM) uptake by MLS-9 cells at 37°C was essentially linear for the first 20 min and began to approach an equilibrium value by ~30 min. Therefore, a 10-min incubation time was used to represent the initial rate of ZDV influx into the cells. The ZDV concentration (0.5 mM) was chosen after considering the kinetics of the system involved (see below), to approximate the apparent K_m value and better describe ZDV uptake properties.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time course of specific ZDV uptake by MLS-9 cells. Specific ZDV (0.5 mM) uptake was measured in standard EBSS medium over 90 min at 37°C. ZDV uptake in the presence of 3 mM verapamil was used to correct for nonspecific ZDV uptake. Results are expressed as mean ± S.D. of three separate experiments. Within each experiment, data points represent the average of quadruplicate measurements.

Kinetics of ZDV Uptake. The initial rate of ZDV influx into MLS-9 cells was measured at 10 min (37°C) with ZDV concentrations from 20 µM to 2.5 mM. The total ZDV uptake rate (V) can be described by the following equation:

(1)

where V_max is the maximum uptake rate; K_m is the affinity constant, and D is the coefficient for nonspecific, diffusive process. The specific ZDV uptake rate was calculated by subtracting from the total rate, the nonspecific uptake measured with a high concentration of the most potent inhibitor we could identify (3 mM verapamil). Figure 2 shows evidence for a specific, saturable carrier-mediated component, which conformed to simple Michaelis-Menten kinetics. The linear correlation (Eadie-Hofstee plot, r = 0.97; data not shown) between V and V/[ZDV] of the specific ZDV uptake values further confirms the single nature of the transport system. The kinetic constants, determined by nonlinear regression analysis, were 1024 ± 102 µM, 326 ± 14 pmol/mg/min, and 0.2 ± 0.1 pmol/mg/min/µM for the K_m, V_max, and D, respectively.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Kinetics of ZDV uptake. ZDV uptake was measured at various ZDV concentrations (20 µM to 2.5 mM) in standard EBSS medium at 10 min (37°C), with nonspecific uptake measured in the presence of 3 mM verapamil. The specific uptake rate was calculated by subtracting the nonspecific component from the total. The kinetic constants (Km, Vmax) were determined by a nonlinear least-squares fit of the simple Michaelis-Menten equation to the specific data. Results are expressed as mean ± S.D. of six separate experiments, with quadruplicate measurements in each. All subsequent uptake measurements were corrected by subtracting nonspecific uptake.

Specificity Studies. ZDV has previously been shown to inhibit an organic cation transporter in rat renal brush-border membrane vesicles (Griffiths et al., 1992). Thus, to determine the selectivity of the transporter, we tested several organic cations (1 mM) for their ability to inhibit the influx of 0.5 mM [³H]ZDV, measured at 10 min and 37°C (Fig. 3). Among these compounds, verapamil, mepiperphenidol, quinidine, cimetidine, and N¹-methylnicotinamide, known inhibitors of organic cation transport systems, were the most effective inhibitors (70-79% inhibition). Thiamine, 1-methyl-4-phenylpyridinium (MPP⁺), guanidine, tetraethylammonium (TEA), and trimethoprim (49-67% inhibition) were weak inhibitors. Because these results suggested the possible involvement of an organic cation transport (OCT) system for ZDV, we also explored the effect of ZDV on the uptake of radiolabeled TEA, a standard organic cation probe. As opposed to findings with all the other cloned OCT systems (i.e., OCT1, OCT2, OCT3, OCTN1, and OCTN2), which are readily known to transport TEA (Gorboulev et al., 1997; Kekuda et al., 1998; Wu et al., 1998b,c), the uptake of TEA by microglia cells was low and not sensitive to ZDV (data not shown).

View larger version (61K): [in this window] [in a new window]   Fig. 3.   Organic cations inhibit specific ZDV uptake. Uptake of ZDV (0.5 mM in standard EBSS medium at 37°C) was measured at 10 min in the presence of 1 mM of an organic cation: MPP+, trimethoprim, TEA, guanidine, thiamine, choline, colchicine, quinine, N1-methylnicotinamide (NMN), cimetidine, quinidine, mepiperphenidol (MEP), verapamil (VER). Results are expressed as mean ± S.D. of three separate experiments, with each experimental point done in quadruplicate. **, significantly different from the control value (p < 0.01).

(2)

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Competitive inhibition by verapamil on ZDV uptake. A, dose dependence of verapamil (VER) in inhibiting ZDV uptake. Specific ZDV (0.5 mM) uptake was measured at 10 min, 37°C, and in the presence of increasing verapamil concentrations (20 µM-3 mM). The data were fit to a sigmoidal inhibition model (under Materials and Methods). Results are expressed as percentage of the control uptake ± S.D. for three separate experiments. B, Eadie-Hofstee plots of verapamil inhibition of ZDV uptake. ZDV (100, 250, and 500 µM, and 1 and 2 mM) uptake was measured in standard EBSS medium at 10 min, 37°C, in the absence (control) or presence of various concentrations of verapamil. Least-squares regression analysis was used to determine the Eadie-Hofstee linear correlations. Results are expressed as mean ± S.D. of two separate experiments, with quadruplicate measurements. (Some error bars are smaller than the symbols.)

(3)

pH and Membrane Potential Effects. Several studies have identified organic cation/proton exchange mechanisms in brush-border membrane vesicles from kidney and intestine, and in the basolateral membranes of hepatocytes (Zhang et al., 1998). The driving force, a proton gradient, is known to be functionally linked to both the Na⁺/H⁺ antiporter and the H⁺-ATPase (Bendayan et al., 1994). To assess the presence of such a mechanism, we measured ZDV uptake in the presence of an outwardly directed proton gradient, generated by a standard NH₄Cl acidification procedure (Jans et al., 1987; Bendayan et al., 1994). Acidification was confirmed by monitoring the changes in intracellular pH with the fluorescent probe BCECF (Fig. 5). Addition of NH₄Cl caused an initial intracellular alkalinization (to pH 7.5), followed by a dramatic acidification upon NH₄Cl removal (from pH 7.1 to 6.3). The pH spontaneously recovered to its initial value by ~20 min. After a second NH₄Cl treatment and washout, the pH recovery was abolished by 1 mM amiloride, an inhibitor of the Na⁺/H⁺ antiporter. ZDV uptake was enhanced for the first 5 min, and then decreased to about control values by 10 min, presumably due to the dissipation of the proton gradient (Fig. 6). The converse effect of a proton gradient in the opposite direction (Fig. 7) further supports a mechanism whereby protons exchange with ZDV in microglia cells. That is, the initial uptake value was increased by extracellular alkalinization (pH_i ~ 7.1 < pH_o 8.0), and decreased by extracellular acidification (pH_i ~ 7.1 > pH₀ 6.0). Thus, ZDV uptake appears to be increased by an outwardly directed proton gradient and decreased by an inwardly directed one. The requirement for an outwardly directed proton gradient was further supported by the decrease in uptake (by 35%) after exposure to 5 µM monensin (Fig. 8), a Na⁺/H⁺ ionophore that collapses the pH gradient. Taken together, these results suggest that proton efflux is required to drive the entry of ZDV into MLS-9 cells.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Intracellular pH changes. MLS-9 cells were preincubated for 40 min at 37°C in EBSS containing 20 µM BCECF-AM ester with 0.1% BSA, and baseline measurements were recorded in standard EBSS. The cells were initially exposed to 30 mM NH4Cl in EBSS for 30 min, washed, and placed in EBSS for 30 min, followed by a second NH4Cl treatment, removal, and replacement with EBSS + 1 mM amiloride. Results represent the combined data from four separate experiments.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   An outwardly directed proton gradient increases specific ZDV uptake. Specific ZDV (0.5 mM) uptake was measured over 10 min at 37°C, in the absence (control) or presence of an outwardly directed proton gradient (pHi < pHo). The gradient was generated by exposing the cells to 30 mM NH4Cl, washing, and placing in uptake buffer. Results are expressed as mean ± S.D. of three different experiments, with quadruplicate measurements. **, values that are significantly different from control value (p < 0.01).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Effect of various proton gradients on specific ZDV uptake. MLS-9 cells preincubated with standard EBSS (pH 7.4) were incubated at 37°C over 10 min in either standard EBSS medium (pH 7.4, control ) or EBSS medium containing Tris (pH 8.0 ) or MES (pH 6.0 ). Each experimental point represents the mean ± S.D. of quadruplicate measurements from two separate experiments. Significant differences from the control value are indicated (*p < 0.05; **p < 0.01).

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Monensin decreases specific ZDV uptake. MLS-9 cells were incubated over 30 min at 37°C in standard EBSS medium (control) or in the presence of 5 µM monensin, an Na+/H+ ionophore. Results are expressed as mean ± S.D. of three different experiments, with quadruplicate measurements. Values that are significantly different from controls are indicated (*p < 0.05; **p < 0.01).

View larger version (26K): [in this window] [in a new window]   Fig. 9.   Effect of membrane potential on specific ZDV uptake. ZDV (0.5 mM) uptake was measured at 37°C over 10 min in the presence of 138 mM Na+ (), 138 mM K+ (), 138 mM K+ + 1 µM valinomycin (), and 2 mM K+ + 1 µM valinomycin (), a potassium ionophore. Results are expressed as mean ± S.D. of two separate experiments; each experimental point represents quadruplicate measurements. Values significantly different from control values are indicated (**p < 0.01; *p < 0.05).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

ZDV, a thymidine analog, is considered one of the first-line therapeutic agents for AIDS and AIDS dementia complex (Fischl et al., 1987). To be active, intracellular ZDV must be phosphorylated to its triphosphate derivative and incorporated into the viral DNA by HIV reverse transcriptase, resulting in the termination of DNA elongation (Furman et al., 1986). At 1 to 5 µM concentrations, ZDV is an effective and selective inhibitor of viral DNA replication in vitro.

Penetration of anti-HIV drugs into the CNS is of clinical concern due to the neurological effects of HIV (Navia and Price, 1998). Results from in vivo studies examining ZDV uptake into brain, after a single pass through the cerebral circulation, have demonstrated nonsignificant net transport of ZDV across the BBB (Terasaki and Pardridge, 1988; Wu et al., 1998a). Moreover, Thomas and Segal (1997) have determined, using a ventriculocisternal perfusion technique, that ZDV influx across the blood-CSF and BBB occurs primarily by a diffusional process. Because at present no information is available on the mechanism of transport of nucleoside analog drugs within brain parenchyma, the purpose of this study was to characterize the properties of ZDV uptake by microglia, the major target of HIV infection in the brain.

Kinetic analysis of ZDV uptake by microglia shows a saturable, low-affinity system (K_m = 1 mM). Similarly, weak interactions have been reported for ZDV in rat renal membrane vesicles, where ZDV inhibited transport of the organic cation N¹-methylnicotinamide at the brush-border membrane and transport of the organic anion p-aminohippuric acid at the basolateral site, with IC₅₀ values of 2.5 mM and 225 µM, respectively (Griffiths et al., 1991, 1992). These in vitro experiments further support in vivo results that showed that in the kidney, ZDV is transported by a basolateral probenecid-sensitive organic anion transporter, and an apical cimetidine-sensitive organic cation transporter (Aiba et al., 1995). Interestingly, a high K_m (0.5 mM) has been reported in the transport of ZDV by a recombinant Na⁺-dependent nucleoside transporter, functionally expressed in Xenopus oocytes (Yao et al., 1996).

Specific transporters responsible for handling cationic compounds have been reported mainly in the kidney, liver, and intestine. Three polyspecific, potential-sensitive organic cation transporters (OCT1, rOCT2/hOCT2, OCT3) cloned originally from rat kidney, and placenta, are differentially expressed in brain (Zhang et al., 1998). Sensitive reverse transcription-polymerase chain reaction and Northern blot studies reported low brain expression of OCT1 and OCT2, although several cationic neurotoxins and neurotransmitters are accepted as substrates by both transporters (Gorboulev et al., 1997; Grundemann et al., 1997). In contrast, mRNA transcripts specific for OCT3 were detected in significant amounts by Northern analysis in brain (i.e., cerebral cortex, cerebellum, hippocampus) and several other tissues (Kekuda et al., 1998; Wu et al., 1998b). Not only is OCT3 capable of transporting various cationic neurotoxins (i.e., MPP⁺) and neurotransmitters but also it has been shown to exhibit transport properties of an extraneuronal monoamine transporter (Wu et al., 1998b). Recently, two other cloned, electroneutral, H⁺-driven organic cation transporters, OCTN1 and OCTN2, were shown to be expressed widely in human tissues, including the brain (Tamai et al., 1997; Wu et al., 1998c).

Results from our substrate specificity studies show that several endogenous and exogenous organic cations (i.e., N¹-methylnicotinamide, quinidine, mepiperphenidol, verapamil) inhibit specific ZDV uptake in a concentration-dependent manner with verapamil being a competitive inhibitor. However, the inhibition is modest (IC₅₀ range = 156-200 µM) and the system was weakly inhibited by the OCT probes TEA and MPP⁺. Furthermore, we observed that the cell line was unable to efficiently transport TEA. This specificity pattern is not shown for other OCT members such as OCT1, OCT2, OCT3, OCTN1, and OCTN2, which are all known to readily transport TEA (Gorboulev et al., 1997; Grundemann et al., 1997; Kekuda et al., 1998; Wu et al., 1998b,c). Thus, despite some OCT specificity similarities, the characterized transporter for ZDV in microglia appears to be distinct from any of the other cloned members of the OCT family.

Several studies have characterized the involvement of a H⁺/organic cation transporter in the renal luminal transport of organic cations. The driving force, a proton gradient, is known to be functionally linked to both the Na⁺/H⁺ antiporter and the H⁺-ATPase (Bendayan et al., 1994). To elucidate the energetics in ZDV transport by microglia, we investigated the effect of an outwardly directed proton gradient generated by a standard NH₄Cl loading procedure (Jans et al., 1987; Bendayan et al., 1994; Faff et al., 1996). The 20-min recovery of internal pH to baseline 7.1 from an intracellular acidification (pH 6.3, a decrease of 0.8 units) was confirmed by microfluorimetry. Moreover, the diuretic drug amiloride, a known inhibitor of the Na⁺/H⁺ antiporter, abolished the pH recovery and maintained the cell acidification, suggesting functional activity of a Na⁺/H⁺ antiporter. It is well established that amiloride rapidly and competitively inhibits the Na⁺ site of the Na⁺/H⁺ antiporter without affecting other Na⁺-coupled transport processes, shown, for instance, by its selective inhibition of the Na⁺/H⁺ exchanger in rabbit renal microvillus membrane vesicles (Kinsella and Aronson, 1981). Our studies show that with intracellular acidification, ZDV uptake is significantly enhanced within the first 5 min. The accumulation of ZDV cannot be explained by nonionic diffusion because the zwitterionic nature of the molecule is not modified within the pH range studied (6.3-8.0). Furthermore, ZDV uptake was significantly reduced (i.e., by 35% at 30 min) in the presence of the Na⁺/H⁺ ionophore monensin, thus suggesting the involvement of a H⁺ exchanger. In mouse microglia, several pH regulatory systems (i.e., Na⁺/HCO₃ cotransporter, Cl/HCO₃ exchanger, H⁺/K⁺- ATPase, Na⁺/H⁺ antiporter) are functionally expressed (Faff et al., 1996; Shirihai et al., 1998).

To test for membrane potential effects, ZDV uptake was investigated in high- and low-K⁺ medium in the presence of the potassium ionophore valinomycin. Membrane depolarization enhanced ZDV uptake, whereas hyperpolarization caused a decrease in ZDV uptake. Unlike other potential-driven OCT systems (Kekuda et al., 1998; Wu et al., 1998b), when depolarized with high extracellular K⁺, an inside positive potential difference drives an increased uptake of ZDV in microglia cells. Thus, although ZDV uptake system in microglia has some specificity features of an organic cation transporter, our results suggest a system that is novel in its sensitivity to membrane potential and pH. Various normal mechanisms that depolarize microglia could enhance ZDV uptake and, ultimately, increase its pharmacological activity. Activation of a number of channels has been reported to depolarize microglia, i.e., Ca²⁺ (Colton et al., 1994), Na⁺ (Korotzer and Cotman, 1992), anion channels (Schlichter et al., 1996), or purinergic receptor-gated channels (Illes et al., 1996).

In our cell system, the prototypical organic anion p-aminohippuric acid and the inhibitor probenecid had no effect on ZDV uptake, implying the lack of involvement of an organic anion transporter in MLS-9 cells. In addition, ZDV accumulation by MLS-9 cells was insensitive to a number of nucleosides, nucleoside analog drugs, and standard nucleoside transport inhibitors, suggesting that nucleoside transporters are not involved. These findings are consistent with reports documenting that the absence of the 3'-hydroxy group of the ribose moiety greatly reduces the ability of compounds to be transported by equilibrative nucleoside transporters. Furthermore, ZDV, 2',3'-dideoxyadenosine, and 2',3'-dideoxycytidine (zalcitabine) did not have measurable affinity for the BBB nucleoside transporter (Terasaki and Pardridge, 1988) or the Na⁺-dependent nucleoside transporter at the choroid plexus (Wu et al., 1994). In microglia, we have previously observed that ZDV was a potent noncompetitive inhibitor of a Na⁺-nucleoside transporter (Hong et al., 2000).

In summary, a novel, electrogenic, proton-driven ZDV transporter has been characterized in microglia, the target and reservoir of HIV infection in the brain. This system may play a significant role in the transport of other weak organic cation substrates and metabolites in brain parenchyma. The low affinity of ZDV for the rat microglia transporter suggests that drug-drug interactions may readily occur at the transporter site(s) with other weak organic cations, most likely causing either a decrease in therapeutic efficacy, an increase of toxicity, or both in AIDS patients receiving routinely multiple drug regimens. Moreover, because interspecies differences in the kinetic and selectivity properties of organic cation transporter homologs from rodent, rabbit, and human have been reported (Dresser et al., 2000), it is possible that a ZDV transporter with different kinetics may be expressed in humans. This could result in substantial differences in the in vivo microglial handling of drugs (i.e., higher affinity transport system for ZDV), basic metabolites, and toxins. It remains to be established whether other physiological roles can be attributed to this transport system.