Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The acidic amino acids, glutamate and aspartate, are the predominant excitatory neurotransmitters in the mammalian central nervous system (for review, see Mayer and Westbrook, 1987). In addition to their essential role as mediators of rapid signaling, these excitatory amino acids (EAAs) also contribute to brain damage observed in acute insults to the nervous system, including hypoxemia, head trauma, and seizure disorders (for review, see Choi, 1992). The levels of EAAs approach 10 mmol/kg in the brain whereas extracellular levels are maintained at micromolar or submicromolar concentrations (for review, see Schousboe, 1981). Because there is no evidence for extracellular metabolism, it is generally assumed that low extracellular concentrations of the EAAs are maintained by Na⁺-dependent transport activity (Schousboe, 1981). Malfunction and/or reverse operation of transporter proteins results in an accumulation of extracellular EAAs and excessive activation of EAA receptors contributing to excitotoxicity (for review, see Attwell et al., 1993).

In the early 1990s, three cDNA clones that express Na⁺-dependent glutamate transport in heterologous expression systems were identified and named GLAST (Storck et al., 1992), GLT-1 (Pines et al., 1992), and EAAC1 (Kanai and Hediger, 1992). Using strategies based on the homology of these transporters, three human homologs were identified and named EAAT1-3 (Arriza et al., 1994) and then subsequently two additional glutamate transporters (neuronal EAAT4 and retinal EAAT5) were identified (Fairman et al., 1995; Arriza et al., 1997).

Before the cloning of individual transporters, pharmacological studies provided evidence for multiple subtypes of transporters that are differentially expressed in various brain regions (for review, see Robinson and Dowd, 1997). For example, transport in crude synaptosomal membranes prepared from forebrain (cortex, hippocampus, or striatum) regions and midbrain is inhibited by dihydrokainate (DHK) with IC₅₀ values of approximately 100 µM, whereas transport in cerebellar preparations is essentially insensitive to inhibition by DHK. L--aminoadipate (L-AAD) displays the opposite pattern of inhibition. The cloning of transporter subtypes led to studies of their pharmacological properties and localization. Based on its sensitivity to inhibition by L-AAD and cerebellar localization, it is assumed that EAAT4 may be sufficient to explain the predominant transport activity observed in cerebellar crude synaptosomes (Fairman et al., 1995). The other neuronal transporter, EAAC1, has pharmacological properties that differ from those observed for cortical crude synaptosomes (Dowd et al., 1996). The glial transporter, GLAST, has pharmacological properties that are dramatically different from those observed in crude cortical synaptosomes and are similar to those of transport observed in undifferentiated astrocyte cultures that express GLAST (Arriza et al., 1994; Stoffel and Blau, 1995; for review, see Robinson and Dowd, 1997). The only other cloned transporter that could explain the forebrain pharmacology is the glial transporter GLT-1 (corresponding to the human homolog EAAT2). The pharmacological properties of the human homolog appear to agree with those observed in forebrain/cortical crude synaptosomes (Arriza et al., 1994), but the pharmacological characterization of the rat homolog is limited and is not consistent between studies (Pines et al., 1992; Wang et al., 1998).

The availability of cDNA clones and specific antibodies has also facilitated studies of the regulation of these glutamate transporters. There is evidence that GLAST-mediated transport can be rapidly down-regulated by activation of protein kinase C (PKC) and subsequent phosphorylation of the transporter (Conradt and Stoffel, 1997). Recent studies suggest that EAAC1-mediated transport is increased by activation of PKC or decreased by an inhibitor of phosphatidylinositol 3-kinase (Davis et al., 1998). These effects on activity are correlated with altered cell surface expression and are consistent with regulation of activity through trafficking to and from the cell surface. There is also evidence that GLT-1 may be rapidly regulated by activation of PKC (Casado et al., 1993; Ganel and Crosson, 1998). In the earlier study, PKC activation increased glutamate transport; in the later study, PKC activation decreased transport activity by increasing the K_m value.

The goal of the present study was to examine the pharmacological properties of the rat homolog, GLT-1, and the possible mechanisms of rapid regulation by PKC. To accomplish these goals, GLT-1 cDNA was stably transfected into cell lines that express low levels of endogenous glutamate transport activity. In two separate cell lines, the pharmacological properties of GLT-1 parallel those observed in crude cortical synaptosomes. These two cell lines were used to study the regulation of GLT-1 by PKC. In both cell lines, glutamate transport activity was unaffected by activation of PKC. Western blotting and analyses of transport activity suggest that the previously reported regulation of GLT-1 may be attributed to regulation of endogenously expressed EAAC1 in the cell lines used in earlier studies.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. L-[³H]Glutamate (40-60 Ci/mmol) was purchased from DuPont (Boston, MA). Nonradioactive L-glutamate (Sigma Chemical Co., St. Louis, MO) was used to dilute the specific activity. The sources for all the EAA analogs are listed in Table 1. MCB 3901 (a Syrian hamster adenovirus type 12-induced tumor, catalog no. CRL-9595), L-M (TK-) (a subline of 5-bromo-2-deoxyuridine-resistant strain of the L-M mouse fibroblast cell line, catalog no. CCL-1.3), and HeLa (epithelioid carcinoma, catalog no. CCL-2) cell lines were obtained from the American Type Culture Collection (Rockville, MD). The GLT-1 cDNA in pBluescript SK- was a generous gift from Dr. Baruch Kanner (Hebrew University, Jerusalem, Israel). Phorbol 12-myristate 13-acetate (PMA) was purchased from Sigma (St. Louis, MO). Forskolin, bisindolylmaleimide II (Bis), okadaic acid (Prorocentrum concavum), and A23187-free acid (streptomyces chartreusensis) were purchased from Calbiochem (La Jolla, CA).

Subcloning, Stable Transfection, and Maintenance of Cell Lines. L-M (TK-) and MCB cell lines were maintained in Dulbecco's modified Eagle's medium (catalog no. 11960-051, no added glutamine; Gibco BRL, Gaithersburg, MD), 10% heat-inactivated defined fetal bovine serum (catalog no. SH30070.03; Hyclone, Logan, UT), and 1% penicillin (100 U/ml)/streptomycin (100 µg/ml; catalog no. 15140-122; Gibco BRL) in a 7% CO₂, 37°C incubator. HeLa cells were maintained in the same media supplemented with 2 mM L-glutamine, and at 5% CO₂.

Measurement of Na⁺-Dependent Glutamate Transport Activity. Transport assays were performed as described previously (Davis et al., 1998). In a 37°C water bath, cultures (plated in 12-well dishes) were prerinsed two times with 1 ml of prewarmed sodium- or choline-containing buffer (Tris base, 5 mM; HEPES, 10 mM; NaCl or choline chloride, 140 mM; KCl, 2.5 mM; CaCl₂, 1.2 mM; MgCl₂, 1.2 mM; K₂HPO₄, 1.2 mM; and dextrose, 10 mM). Uptake was initiated by the addition of 1 ml of prewarmed Na⁺- or choline-containing buffer, which contained L-[³H]-glutamate with or without EAA analogs. After 5 min, cultures were rinsed three times with 1 ml of ice-cold choline buffer to stop transport activity. Cells were dissolved in 1 ml of 0.1 M NaOH and a 500 µl aliquot was used to quantitate radioactivity by scintillation spectrometry. Protein was also measured in an aliquot of solubilized cells using a commercial Bradford kit (Bio-Rad protein assay, Bio-Rad Laboratories, Hercules, CA). Na⁺-dependent transport activity was calculated as the difference in accumulated radioactivity in the Na⁺-containing and choline-containing buffer. In these assays, less than 10% of the substrate was accumulated and transport was measured under conditions of initial velocity (transport was linear with time beyond 5 min).

Western Analysis. After washing twice with ice-cold PBS, cells were harvested using 1 ml per 10 cm dish of Na⁺-HEPES buffer (20 mM, pH 7.5), containing MgCl₂ (0.4 mM), EDTA (0.2 mM), phenylmethylsulfonyl fluoride (20 µM), leupeptin (2 µg/ml), and aprotinin (0.22 µg/ml). After sonication and centrifugation at 14,000 RPM in a microcentrifuge for 10 min, an aliquot of the supernatant was used for analysis of protein content and an aliquot was diluted 1:2 in solubilizing buffer (2% SDS, 10% mercaptoethanol, 5% glycerol, 0.005% bromphenol blue, and 50 mM Tris-Cl, pH 7.0). The samples were frozen at 20°C. Cell culture samples or control brain specimens, prepared as described previously (Schlag et al., 1998), were resolved using 10% polyacrylamide gel electrophoresis. After transfer to polyvinylidine fluoride membranes (Immobilon P, Millipore, Bedford, MA), immunoblots were probed with anti-GLT-1, anti-GLAST, anti-EAAC1, or anti-EAAT4 antibodies (Furuta et al., 1997). Immunoreactivity was visualized by enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Curve Fitting and Statistical Analysis. All values reported are the mean ± S.E.M. of at least three independent observations. Except where noted, IC₅₀ values were obtained by fitting experimental data to a theoretical curve with a Hill coefficient of one using One Site Competition method in GraphPad Prism software (GraphPad Prism, version 2.0, GraphPad Software, Inc., San Diego, CA). When two groups were compared, a Student's unpaired t test was used to compare the values. Multiple groups were compared by ANOVA with post hoc analysis. A p < .05 was considered significant.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Selection and Characterization of Cell Lines Stably Transfected with GLT-1. One problem with choosing cell lines for stable expression of glutamate transporters is a high level of endogenous activity observed in several of the cell types regularly used for stable transfection, including CHO, COS, or HEK293 (M.B.R., unpublished observations). MCB (Desai et al., 1995), and L-M (TK-) (J.D.R., unpublished observations) express undetectable to low levels of glutamate transport and HeLa cells were previously used for transient transfection of GLT-1 (Pines et al., 1992; Casado et al., 1993). To examine endogenous expression of transporter subtypes in these cell lines, transporter specific antibodies were used for Western analysis with cortical or cerebellar protein as a positive control (Fig. 1). A strong immunoreactive band for GLT-1 was observed with a small amount of cortical membrane homogenate, but no GLT-1 immunoreactivity was observed when 80 times more cell line protein was analyzed. Similarly, except for HeLa cells, no EAAT4 or GLAST immunoreactivity was observed in these cell lines; weak immunoreactive bands for these transporters were observed in HeLa cells. No EAAC1 immunoreactivity was observed in L-M (TK-) cells, an extremely faint immunoreactive band was consistently observed in MCB cells (at approximately 64 kDa), and a robust EAAC1 immunoreactive band was observed in HeLa cells (Fig. 1).

View larger version (73K): [in this window] [in a new window]   Fig. 1.   Western blot analysis of Na+-dependent glutamate transporters expressed in untransfected cell lines (L-M (TK-), MCB, and HeLa) and in transfected cell lines (LM-GLT1-8, MCB-GLT1-6, and MCB-GLT1-12). For the GLT-1 immunoblot, 80 µg of protein was loaded in each lane, except for cortical membrane homogenate (1 µg protein), HeLa cells (50 µg protein), and MCB-GLT1-12 cells (20 µg). For the other antibodies, 80 µg of protein was loaded in each lane except for HeLa cells and cortical or cerebellar tissue (50 µg protein). All cell homogenates were harvested from cells grown to 80 to 90% confluence (similar to that used for analyses of transport activity). For each antibody, all of the lanes are from the same gel with the same exposure. Each of these Western blots were repeated in at least two additional independent experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Transport activity in untransfected (stippled bars) and transfected (solid bars) L-M (TK-) and MCB cell lines. The accumulation of L-[3H]-glutamate (0.5 µM) was measured in the presence (+Na+) and absence of Na+ (Na+) as described in Experimental Procedures. Na+-dependent (Na+-dep.) accumulation was calculated as the difference between these two values. Data are the mean ± S.E.M. of at least three independent observations each made in triplicate.

Kinetic and Pharmacological Properties of L-[³H]-Glutamate Transport Activity in Cell Lines Stably Transfected with GLT-1. The concentration dependence of Na⁺-dependent L-[³H]-glutamate transport activity was examined in both LM-GLT1-8 and MCB-GLT1-6. Between the concentrations of 1 and 1000 µM L-glutamate, transport activity was consistent with a single saturable process with K_m values of 37 ± 4 µM (n = 4) in LM-GLT1-8 cells and 18.6 ± 0.3 µM (n = 3) in MCB-GLT1-6 cells (Fig. 3). The V_max values were 3.1 ± 0.4 nmol/mg protein per min in LM-GLT1-8 cells and 2.7 ± 0.3 nmol/mg protein per min in MCB-GLT1-6 cells, indicating that both cell lines express a comparable capacity for transport. These capacities for transport are comparable to those observed in cortical, hippocampal, or striatal crude synaptosomes (Robinson et al., 1991).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Eadie-Hofstee plot of the concentration dependence of Na+-dependent L-[3H]-glutamate transport in LM-GLT1-8 (A) and MCB-GLT1-6 (B). Inset: Saturation isotherm of the same data (Michaelis-Menten). Data are the mean ± S.E.M. values of at least three independent observations, each performed in triplicate. For LM-GLT1-8 cells, the Vmax value was 3.1 ± 0.4 nmol/mg protein per min and the Km value was 37 ± 4 µM. For MCB-GLT1-6 cells, the Vmax value was 2.7 ± 0.3 nmol/mg protein per min and the Km value was 18.6 ± 0.3 µM.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Sensitivity of GLT-1-mediated transport to inhibition by L-trans-PDC and DHK. Na+-dependent L-[3H]-glutamate transport was measured in the absence and presence of increasing concentrations of inhibitors in MCB-GLT1-6 () and LM-GLT1-8 (). Data are mean ± S.E.M. values of three independent observations, each performed in triplicate. The solid lines represent the theoretical curves with Hill coefficient of 1. IC50 values are presented in Table 1. The dotted line represents the previously reported sensitivity of cerebellar glutamate transport to the inhibitor and the dashed line represents the previously reported sensitivity of cortical glutamate transport to the inhibitor (Robinson et al., 1991; 1993).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Correlation between the potency of EAA analogs for inhibition of Na+-dependent GLT-1-mediated transport and their potency for inhibition of Na+-dependent glutamate transport in cortical (A) or cerebellar (B) synaphosomes. The log of the IC50 value for inhibition of cortical or cerebellar glutamate transport (see Table 1) was plotted as a function of the log IC50 value for inhibition of GLT-1-mediated transport. The IC50 values used for GLT-1 were the averages of the IC50 values for inhibition of LM-GLT1-8 and MCB-GLT1-6. For compounds with data for inhibition that are best fit to two sites in either cerebellar or cortical tissue, the Log IC50 value for the predominant component was used. For compounds without full inhibition curves, the IC50 values were calculated from the percent inhibition assuming a theoretical curve with a Hill coefficient of 1. The numbers refer to each compound used for this correlation (see Table 1 for the names of each compound).

Effects of Over-Expression of Glutamate Transporters in Cell Lines. One approach to limiting the impact of endogenous glutamate transporter expression in cell lines is to express higher levels of transfected transporter relative to that observed in untransfected cells (Matthews et al., 1997). In our initial studies, we decided to attempt a similar approach with an MCB clone (MCB-GLT-12) that expressed much higher levels of GLT-1 protein than MCB-GLT1-6 (Fig. 1). In this cell line, the V_max value for L-[³H]-glutamate uptake was 24 ± 5 nmol/mg protein per min (data not shown, n = 4), which was almost 10-fold higher than that observed in MCB-GLT1-6. In this cell line, the K_m value for glutamate transport was four times higher than that observed in MCB-GLT1-6 even though less than 10% of the glutamate was transported from the extracellular space (K_m = 78 ± 30 µM). Similar shifts in the potencies of compounds cleared by transporters have been observed in brain slices and cells in culture (for a recent discussion, see Speliotes et al., 1994), and have been attributed to the rapid clearance of glutamate from the local environment near the transporters. This could explain the lower apparent affinity of the transporter for glutamate, but the pharmacological properties of transport were also affected by this high level of expression. The sensitivity of transport to DHK was much lower in MCB-GLT-12 (IC₅₀ = 588 ± 89 µM) than in MCB-GLT1-6 (IC₅₀ = 32 ± 7 µM) as was the sensitivity to inhibition by L-trans-PDC (112 ± 42 µM, n = 3), kainate (1140 ± 220 µM, n = 3), and L-anti-endo-3,4-methanopyrrolidine dicarboxylate (108 ± 9 µM, n = 3). To rule out the possibility that confluent MCB-GLT1-12 cells express higher levels of EAAC1, GLAST, or EAAT4, protein sample were harvested from confluent cultures. We found no evidence for higher levels of any of these proteins in confluent MCB-GLT1-12 cells than in confluent MCB-GLT1-6 (data not shown, n  2 independent experiments). Although we cannot rule out the possibility that confluence induces expression of EAAT5 or an uncloned transporter, the observation that all of these inhibition data conform to theoretical curves with a Hill slope of 1 suggests that a single population of sites mediates activity in these confluent cultures. To test the possibility that high expression might be influencing the pharmacological properties of GLT-1 by changing the local environment close to the cells, the sensitivity of transport to DHK was examined using MCB-GLT1-12 cells grown to approximately 30% confluence. Under these conditions, the IC₅₀ value for inhibition by DHK was 148 µM (n = 2), a value which approaches that observed in 80 to 90% confluent LM-GLT1-8 and MCB-GLT1-6 cells. Thus, it appears that high levels of activity can influence the pharmacological properties.

Effects of Modulation of PKC on GLT-1-Mediated Transport. Using HeLa cells, Casado and her colleagues published a report that suggested that GLT-1-mediated transport could be rapidly increased by activation of PKC (Casado et al., 1993). Because this type of rapid regulation could have significant implications for the control of extracellular EAAs, we pursued this regulation of GLT-1 in both stably transfected cell lines. In both cell lines, preincubation with the phorbol ester (PMA) for 30 min had no effect on transport activity (Fig. 6A, mean of at least nine independent observations). As a positive control, each of the stock solutions of PMA was tested for activation of endogenous EAAC1-mediated transport in C6 glioma cells. All stocks used caused an increase in transport activity (data not shown, see Davis et al., 1998 for description of assay). Higher concentrations of PMA (up to 1 µM) and longer incubations (up to 1 h) were also tested and neither caused an increase in GLT-1-mediated transport activity in transfected cells (data not shown, n = 2).

View larger version (41K): [in this window] [in a new window]   Fig. 6.   Effect of modulation of PKC on GLT-1-mediated transport in LM-GLT1-8 and MCB-GLT1-6 cells. A, cells were pretreated with vehicle dimethyl sulfoxide (DMSO), 100 nM PMA, or 10 µM Bis for 30 min before measurement of Na+-dependent glutamate (0.5 µM) transport. Data are the mean ± S.E.M. of at least nine independent observations with PMA and at least three independent observations with Bis. There was no significant effect of either treatment on Na+-dependent glutamate transport activity. B, cells were pretreated with either the phosphatase inhibitor okadaic acid (1 µM) or the Ca2+ ionophore A23187 (100 nM) alone and in combination with PMA (100 nM) for 30 min. Data are the mean ± S.E.M. of at least three independent observations. There was no significant effect of any of the treatments.

View larger version (39K): [in this window] [in a new window]   Fig. 7.   Western blot analysis of EAAC1 and GLT-1 expressed in untransfected and transfected HeLa cells. For the GLT-1 immunoblots (A and B), 50 µg of HeLa protein and 0.25 µg of cortical membrane homogenate protein was loaded in each lane. For the EAAC1 immunoblots (C and D), 50 µg of protein was loaded in each lane. Note that these Western blots were also probed with an antiactin antibody and that the immunoblot for Hela-GLT1-14 is a darker exposure than HeLa-GLT1-7. All cell homogenates were harvested from cells grown to 80 to 90% confluence (similar to that used for analyses of transport activity). For each panel, all of the lanes are from the same gel with the same exposure. Each of these Western blots was repeated with at least two additional independent experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Analyses of Na+-dependent L-[3H]-glutamate transport activity in HeLa cells stably transfected with GLT-1 cDNA. A, Eadie-Hofstee plot of the concentration dependence of Na+-dependent L-[3H]-glutamate transport in untransfected HeLa (O) and HeLa-GLT1-7 cells (). Data are the mean ± S.E.M. values of at least three independent observations, each performed in triplicate. For untransfected HeLa cells, the Vmax value was 212 ± 74 pmol/mg protein per min and the Km value was 37 ± 20 µM. For HeLa-GLT1-7 cells, the Vmax value was 1300 ± 230 pmol/mg protein per min and the Km value was 16 ± 1 µM. B, sensitivity of glutamate transport activity (0.5 µM) to inhibition by DHK in HeLa-GLT1-7. Transport activity was measured in the absence and presence of increasing concentrations of DHK. Data are mean ± S.E.M. values of three independent observations, each performed in triplicate. Data were fit to two sites using GraphPad Prism. The equation used to fit two sites was constrained such that the sum of the fraction of high-affinity sites and the fraction of low-affinity sites equaled 1. The mean IC50 values were 16 ± 5 and 1,200 ± 100 µM and the mean fraction of the high-affinity component was 0.31 ± 0.04. The solid line represents the theoretical curve for the two-site fit of these data and the dashed line represents a theoretical curve for a one-site fit of these data (IC50 value = 378 µM). C, effect of modulation of PMA on glutamate transport activity in HeLa-GLT1-7. Cells were pretreated with vehicle (DMSO) or 100 nM PMA for 30 min before measurement of Na+-dependent glutamate (0.5 µM) transport activity. Transport activity was measured in the absence or presence of 300 µM DHK. Delta represents the difference in activity measured in cells treated with DMSO or PKC. Data are the mean ± S.E.M. of three independent observations each performed in triplicate. Although PMA caused a significant increase in transport activity (*p < .001), the PMA-induced increase in transport activity was not affected by DHK.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In the present study, three different cell lines were stably transfected with the cDNA for the GLT-1 subtype of glutamate transporter. Stably transfected cell lines have been used as ideal model systems to study the function, pharmacology, and post-translational regulation of receptors, channels, and other cell surface proteins. In most cases, the corresponding cDNA is transfected into a cell line that does not endogenously express the protein itself or functional homologs. Choosing a cell line to stably express glutamate transporters presents a challenge because many cells endogenously express these transporters, presumably to provide glutamate and/or aspartate for metabolic purposes. One approach to overcoming the impact of endogenous expression might be to select a clone that expresses high levels of transport activity. In the present paper, we provide evidence that high expression can affect the kinetic and pharmacological properties of transport. To overcome these problems, we chose cell lines that express undetectable to low levels of endogenous transport activity and characterized transfected clones that express moderate levels of GLT-1-mediated activity.

The properties (K_m and sensitivity to inhibitors) of glutamate transport activity are similar in both the MCB-GLT1-6 and LM-GLT1-8 cell lines. This suggests that the low level of endogenous transport in untransfected MCB cells did not significantly affect the properties of GLT-1. Some of the compounds tested in the present study have been tested as inhibitors of GLT-1 in earlier studies. In the original paper that described the cloning and expression of the rat homolog of GLT-1, HeLa cells were used as an expression system and both DHK and L-AAD were reported to potently inhibit transport activity (IC₅₀ values < 8 µM; Pines et al., 1992). Arriza and his colleagues expressed the human homolog of GLT-1 (EAAT2) in COS-7 cells and found that it is blocked by DHK with an IC₅₀ value of 23 µM and is not blocked by 1 mM L-AAD (Arriza et al., 1994). While the present study was underway, Wang and colleagues reported that the rat homolog of GLT-1 expressed in oocytes is inhibited by DHK with an IC₅₀ value of 8 µM and that the IC₅₀ value for L-AAD is greater than 1 mM (Wang et al., 1998). Dunlop and his colleagues reported that the human homolog of GLT-1 (EAAT2) expressed in Madin Darby Canine kidney cells is inhibited by DHK with an IC₅₀ value of 15 µM and that the IC₅₀ value for L-AAD is greater than 1 mM (Dunlop et al., 1998). In the present study, DHK inhibited GLT-1-mediated transport with an IC₅₀ value of approximately 30 to 50 µM and L-AAD inhibited transport with an IC₅₀ value of approximately 2 to 3 mM. Thus, the sensitivity to L-AAD observed in the original study has not been observed by four other groups. It is possible that this difference is because the transporter was solubilized and reconstituted in the original study.

The sensitivity of GLT-1-mediated transport to inhibition by a number of compounds is remarkably similar to that observed in cortical tissue. Because the pharmacologies of the brain glutamate transporters, GLAST, EAAC1, and EAAT4, are different from those of cortex (see introduction), the simplest interpretation of these results is that GLT-1 mediates the bulk of cortical glutamate transport activity. Although the pharmacology of transport in hippocampus, striatum, and midbrain has been examined with only a limited number of compounds, the pharmacological properties of GLT-1 are also consistent with transport in these brain regions, suggesting that GLT-1 mediates the bulk of transport activity in forebrain and midbrain. Although one cannot rule out the possibility that an uncloned transporter mediates activity in these brain regions, GLT-1 gene deletion and knock-down studies also support this conclusion. Infusion (ICV) of antisense oligonucleotides specific for GLT-1 decreases protein expression by 60% and causes a 50% decrease in brain transport activity (Rothstein et al., 1996). Genetic deletion of GLT-1 (Tanaka et al., 1997) and pathologic loss of GLT-1 in amyotrophic lateral sclerosis (Rothstein et al., 1995) are also associated with a significant loss (up to 90%) of transport activity in brain tissues. One limitation of these studies is that the neurodegeneration that accompanies these decreases in GLT-1 may lead to an overestimation of the contribution of GLT-1 to transporter activity.

If GLT-1 is indeed the predominant cortical (forebrain) transporter, it raises the possibility that nerve terminals do not accumulate significant levels of glutamate after release, but there are early lesioning studies that originally suggested that there was significant transport into neurons. In these studies, several groups demonstrated that lesioning of afferents decreases transport activity in the target area (for review, see Fagg and Foster, 1983). Although the original interpretation of these data was that transporters were localized to nerve terminals, more recent studies have demonstrated that these lesions result in decreased expression of the glial transporters (Ginsberg et al., 1995; Levy et al., 1995). This suggests that neurons participate in the regulation of expression of the glial transporters in vivo. In fact, in vitro studies have demonstrated that GLT-1 expression in astrocytes is dependent on the presence of neurons (Gegelashvili et al., 1997; Swanson et al., 1997; Schlag et al., 1998). Thus, some of the early observations that suggested that neurons may be important for the clearance of glutamate can be explained by glial transport.

Still, there are observations that cannot be easily explained. First, GLT-1 protein levels in cerebellar synaptosomes are comparable to that observed in forebrain synaptosomes (Robinson, 1998). Yet, no DHK sensitivity is observed in cerebellar synaptosomal membrane preparations (for review, see Robinson and Dowd, 1997). Second, transport in neuron-enriched cultures, which express only low levels of GLT-1 immunoreactivity, has a pharmacology that is comparable to that observed in cortical synaptosomes (Wang et al., 1998). Although this suggests that there may be another transporter with properties similar to GLT-1, the authors did not exclude the possibility that this low level of GLT-1 expression may explain the observed pharmacology.

In an earlier report, Casado et al. (1993) concluded that GLT-1 can be regulated by activation of PKC. In the present study, we found no evidence for regulation of GLT-1-mediated transport activity by PKC. Activation of PKC, inhibition of PKC, and activation of PKC in the presence of a phosphatase inhibitor had no detectable effect in two independent cell lines (LM-GLT1-8 and MCB-GLT1-6). In this earlier report, C6 glioma cells were used to demonstrate phosphorylation of an immunoprecipitable band. We have found that C6 glioma cells express EAAC1 but not GLT-1 or the other transporters (Dowd et al., 1996, Davis et al., 1998). Using reverse transcription-polymerase chain reaction another group (Palos et al., 1996) also has concluded that C6 glioma express only EAAC1. Because EAAC1 is rapidly increased in response to activation of PKC (Davis et al., 1998), it is possible that either the antibody used for immunoprecipitations cross-reacts with EAAC1 or that different sublines of C6 glioma express different subtypes of transporters. Casado and her colleagues also transiently expressed GLT-1 in HeLa cells using vaccinia virus. They found that phorbol esters increase activity of the wild-type transporter, but not of a mutated GLT-1 (S133N) (Casado et al., 1993). This provides fairly convincing evidence that GLT-1 can be regulated by PMA. In the present study, the effects of PMA on transport activity were examined in untransfected HeLa and in GLT-1-transfected HeLa cells (clone HeLa-GLT1-7). Although the increase in transport activity caused by PMA was greater in GLT-1-transfected cells, the level of EAAC1 immunoreactivity was higher in this transfected cell line, making it difficult to directly determine if GLT-1 is regulated. We attempted to indirectly examine the possible regulation of GLT-1 by PMA using a strategy to selectively inhibit GLT-1-mediated transport activity using a concentration of DHK that would be predicted to inhibit GLT-1-mediated transport up to 90% and inhibit EAAC1-mediated transport approximately 20%. Although this experiment demonstrates that the increase in activity caused by PMA was not inhibited by 300 µM DHK, we cannot rule out the possibility that PMA changes the potency of DHK for inhibition of GLT-1. It is difficult to reconcile the results of the present study and those previously published by Casado and her colleagues. It is possible that regulation by PKC is influenced by the level of transporter expression and that the higher expression presumed to occur with vaccinia virus may explain the difference. Alternatively, it is possible that vaccinia virus induces expression of a protein (possibly a subtype of PKC) that is required for regulation of GLT-1.

In conclusion, examination of the properties of the GLT-1 transporter reveals that its pharmacological characteristics are sufficient to explain the majority of activity in crude synaptosomes prepared from cortex and possibly other forebrain regions. Because GLT-1 is generally thought to be expressed by mature astrocytes in vivo (for review, see Robinson and Dowd, 1997), this suggests that extracellular EAAs released from the nerve terminal recycle through the astrocyte rather than through reuptake directly into the nerve terminal. Although there is substantial evidence for post-translational regulation of other transporters and cell surface proteins by PKC (for references, see Davis et al., 1998), we were unable to detect any effect of PKC stimulation on GLT-1 activity.