Excessive accumulation of glutamate in the CNS leads to excitotoxic neuronal damage. However, glutamate clearance is essentially mediated by astrocytes through Na+-dependent high-affinity glutamate transporters (excitatory amino acid transporters (EAATs)). Nevertheless, EAAT function was recently shown to be developmentally restricted in astrocytes and undetectable in mature astrocytes. This suggests a need for other cell types for clearing glutamate in the brain. As blood monocytes infiltrate the CNS in traumatic or inflammatory conditions, we addressed the question of whether macrophages expressed EAATs and were involved in glutamate clearance. We found that macrophages derived from human blood monocytes express both the cystine/glutamate antiporter and EAATs. Kinetic parameters were similar to those determined for neonatal astrocytes and embryonic neurons. Freshly sorted tissue macrophages did not possess EAATs, whereas cultured human spleen macrophages and cultured neonatal murine microglia did. Moreover, blood monocytes did not transport glutamate, but their stimulation with TNF-α led to functional transport. This suggests that the acquisition of these transporters by macrophages could be under the control of inflammatory molecules. Also, monocyte-derived macrophages overcame glutamate toxicity in neuron cultures by clearing this molecule. This suggests that brain-infiltrated macrophages and resident microglia may acquire EAATs and, along with astrocytes, regulate extracellular glutamate concentration. Moreover, we showed that EAATs are involved in the regulation of glutathione synthesis by providing intracellular glutamate. These observations thus offer new insight into the role of macrophages in excitotoxicity and in their response to oxidative stress.
In the CNS, glutamate plays a major role as a neurotransmitter. At high extracellular concentrations, glutamate is also a powerful neurotoxin capable of inducing severe excitotoxic damage to neurons (1).
Extracellular glutamate concentration is regulated by transporter proteins primarily observed in neurons and astrocytes. These transporters are essential for ensuring a high signal-to-noise ratio and preventing neuronal damage (for review, see Ref. 2). Many neurological diseases may be associated with glutamate transport failure (3, 4, 5). Five subtypes of high-affinity glutamate transporters (excitatory amino acid transporters 1–5 (EAAT1–5)3) have been cloned from mammalian tissues (6, 7, 8, 9, 10, 11). They form a new family of molecules, with 50–55% amino acid sequence identities. This transport system, XAG−, transports l-Asp, d-Asp, and l-Glu with similar affinities and couples the electrochemical gradient of three cotransported sodium ions and one countertransported potassium ion with that of the amino acids (12). EAAT1 and EAAT2 were primarily observed in astrocytes, and EAAT3 is a neuronal transporter with a somatodendritic location (13). EAAT gene knockout showed that the astroglial transporters EAAT1 and EAAT2 are involved in protection against excitotoxicity by clearing extracellular glutamate, whereas EAAT3 is not (14, 15). Nevertheless, Stanimirovic et al. (16) recently showed that EAAT function is developmentally restricted in cultured astrocytes. Indeed, embryonic and early postnatal astrocytes (P0) express high EAAT levels in vitro, but glutamate uptake drops in P10–P21 astrocytes and becomes undetectable in P50 astrocytes (16). This finding suggests that glutamate clearance in mature brain would need the contribution of other cell types.
During brain injury, the CNS parenchyma is open to infiltration by blood cells, resulting in a mixed population of inflammatory cells in the damaged tissue. Many studies have shown that most of the macrophages present in brain lesions originate from blood monocytes (17). Brain macrophages are important effectors of the local immune response, although they are thought to contribute to neurotoxicity by producing inflammatory cytokines, quinolic acid, and also glutamate (18, 19, 20, 21). Glutamate release by stimulated macrophages and microglia is mediated by a cystine/glutamate transport system other than EAATs. This system, Xc−, is a Na+-independent anionic amino acid transport present in numerous cell types both in the CNS and the periphery. Generally, cystine is taken up by this transporter in exchange for intracellular glutamate and is then reduced to cysteine. Thus, this transport system is important for maintaining intracellular glutathione (GSH) levels (22, 23). An increase in extracellular glutamate concentration could thus deplete intracellular GSH by competing with cystine uptake (24, 25). Recent studies have demonstrated the presence of EAATs in nonneural cells (26, 27, 28), suggesting that extracellular glutamate clearance is required both centrally and outside the brain.
As tissue macrophages and microglia release glutamate upon in vitro stimulation (20, 21, 29, 30), we addressed the question of whether these cells also possess Na+-dependent high-affinity glutamate transporters for clearing extracellular glutamate. Using human monocyte-derived macrophages, we showed that these cells do have an XAG− glutamate transport system, which could change current thinking about the role of macrophages in the brain and raises the question as to the role of these transporters in the regulation of glutamate-driven immunoregulation in the periphery.
Materials and Methods
Human monocyte isolation and differentiation
Human PBMC were isolated from the blood of healthy HIV-seronegative donors by Ficoll-Hypaque density gradient centrifugation. Monocytes were separated from PBMC by countercurrent centrifugal elutriation. Monocytes (2 × 106 cells/well) were seeded in 48-well plates in RPMI 1640 medium (Boehringer Mannheim, Mannheim, Germany) supplemented with 10% heat-inactivated (56°C for 30 min) FCS (Boehringer Mannheim), 2 mM l-glutamine (Boehringer Mannheim), and 1% antibiotic mixture (penicillin, streptomycin, and neomycin; Life Technologies, Grand Island, NY). Cells were maintained at 37°C in a humidified 5% CO2 atmosphere. In our hands, blood monocytes (≥95% enriched after elutriation) became adherent after 1 h of culture and then spontaneously detached from the plastic after 24 h and retained a monocyte-like appearance for 5 days (Fig. 1⇓A). Monocytes were then washed with PBS and distributed in 48-well plates (0.5 × 106 cells/well) in 10% FCS culture medium supplemented with 15% human PBMC-conditioned medium (7 days of culture). At day 7–8 of culture, cells tightly adhered to the plastic, and morphological differentiation occurred such that the monocyte-macrophages became fibroblast-like (Fig. 1⇓B). At day 9–12, the cells became large, well-dispersed rounded macrophages, and they retained this appearance for about 25 days (Fig. 1⇓C).
Flow cytometric analysis of cell surface molecule expression
Adherent cells were detached from the plastic by a 20-min incubation at 37°C in nonenzymatic cell dissociation solution (Sigma, Saint Quentin Fallavier, France). The cells were incubated for 30 min at 4°C with FITC- or PE-conjugated mAbs against CD14 (Becton Dickinson, Mountain View, CA), HLA-DR (Immunotech, Marseille, France), CD11b (Immunotech), CD16 (Immunotech), or irrelevant isotype-matched controls. The cells were washed twice with PBS, fixed in 200 μl PBS/1% paraformaldehyde (weight to volume ratio), and analyzed for fluorescence using a FACScan flow cytometer (Becton Dickinson). Viable cells were gated using a forward- and side-scatter pattern. HLA-DR and CD16 expression increased with cell differentiation (Table I⇓). CD14 expression was up-regulated until day 5, was undetectable in fibroblast-like macrophages (day 7), and was highly expressed again in a small subpopulation of differentiated amoeboid macrophages (day 12). Similarly, the CD11b adhesion receptor was absent in day 5 monocytes but was strongly expressed in adherent macrophages.
Obtention of tissular macrophages
Macrophages were purified from bronchoalveolar lavages performed in cynomolgus macaques after local anesthesia with lignocaide (Xylovet, Sanofi, France), as previously described (31). Spleens were obtained from C57BL/6 mice. We also obtained spleen tissue from two splenectomized patients, one after idiopatic thrombopenic purpura and one after Minkowski-Chauffard hemolytic anemia. Spleens were gently dissociated in isotonic NaCl with forceps and were sieved (100 μm). Mononucleated cells were isolated by Ficoll-Hypaque density gradient centrifugation. Simian alveolar and murine splenic macrophages were purified by adhesion for 1 h to 48-well plates in RPMI 1640 with 10% FCS, 2 mM l-glutamine, and 1% antibiotic mixture (106 cells/wells). Human splenic cells were cultured for 5 days in RPMI 1640 with 10% FCS, 2 mM l-glutamine, and 1% antibiotic mixture, and then nonadherent cells were removed. Glutamate uptake experiments were performed at days 9–12.
Microglial cell cultures
Microglia were purified from mixed glial cultures from neonatal C57BL/6 mice, as previously described (32). Briefly, pieces of cortex from postnatal 1-day-old mice were incubated in trypsin and mechanically dissociated. Cells were plated and fed weekly with DMEM (Life Technologies), 4.5 g/L glucose, Glutamax-I, and 10% FCS. After 13–15 days of culture, microglial cells were dislodged from mixed glial cultures by shaking for 2 h at 220 rpm. Microglial cells were allowed to settle in 48-well plastic dishes for 30 min, after which the supernatant was replaced with DMEM, 10% FCS, and 2% B27 supplement (Life Technologies). Experiments were performed 6 days later, that is, 19–21 days after brain removal. Microglia-enriched cultures were more than 98% pure, as assessed by immunocytochemistry (Mac-1) and isolectin (B4) staining (data not shown).
Primary mouse cortical neuron cultures
Primary mouse cortical cells were cultured from 15-day C57BL/6 mouse embryos. Cortices were dissected under a binocular microscope, carefully freed of meninges, and incubated in trypsin/EDTA for 10 min at 37°C. Trypsin was inactivated by incubation in DMEM, 4.5 g/L glucose, Glutamax-I, and 1% FCS. Cells were then dissociated mechanically in DMEM, 1% FCS, with a flame-narrowed Pasteur pipette. Cells were pelleted by centrifugation and resuspended in DMEM, 2% B27 and 3% FCS. Ninety-six-well plates coated with poly d-lysine were seeded at 7 × 104 cells per well in 150 μl of medium (DMEM/B27/FCS, with antibiotics). Cultures were kept at 37°C, 5% CO2 for 2 days. The medium was then replaced with serum-free DMEM/B27/antibiotics. After 1 or 2 wk in culture, cells were immunocytochemically assessed to be >95% neurons (according to microtubule associated protein-2 immunolabeling) and less than 6% glial cells (5% glial fibrillary acidic protein-positive cells and less than 1% MAC-1- or IB4-positive cells).
Monocyte-derived macrophage (MDM; 8–12 days) were cultured in DMEM containing 4.5 g/L glucose Glutamax-I without FCS at 37°C and 5% CO2 in the presence or absence of 100 μM, 300 μM, or 1 mM glutamate. Culture medium containing the same concentrations of glutamate was incubated in the same plate but in the absence of MDM as a control. Supernatants were harvested at 6 h, 20 h, or 4 days, centrifuged to eliminate cell debris, and stored at −20°C. Aliquots (100 μl) were tested in triplicate by incubation with 7-day-old primary mouse cortical cells for 24 h. Neuron viability was then measured by using the MTT assay (Sigma). Results were expressed in OD540–630. The percentage of neuroprotection afforded by MDM against glutamate was calculated as follows: % protection = 100 × [(MDM with glutamate) − (medium with glutamate)/(MDM without glutamate) − (medium with glutamate)].
Glutamate uptake was determined for MDM, simian alveolar macrophages, murine microglia, and murine and human spleen macrophages, seeded in 48-well plates. The uptake medium was 137 mM NaCl, 0.7 mM K2HPO4, 1 mM CaCl2, 1 mM MgCl2, 5 mM glucose, and 10 mM HEPES (pH 7.4). We assessed Na+ dependence by replacing the NaCl (137 mM) with 137 mM choline chloride (Sigma). Cells were washed with 1 ml PBS and incubated for 20 min at 37°C in 200 μl uptake medium with ionic modifications or inhibitors, if necessary, such as dl-threo-β-hydroxyaspartic acid (THA), l-trans-pyrrolidine-2,4-dicarboxylic acid (trans-PDC), dihydrokainate (DHK), l-cystine, quisqualic acid, l-homocysteate, or l-α-aminoadipate (Sigma). The medium was aspirated and replaced with 100 μl uptake medium (with ionic modifications or inhibitors, if necessary) containing l-[2,3-3H]glutamic acid (30–60 Ci/mmol; ICN, Irvine, CA). For concentrations above 50 μM, [3 6 cells/tube), and washed once with PBS. Starvation and uptake were done as they were for macrophages. Uptake was stopped by washing twice with 5 ml of cold PBS.
mRNA levels were assessed by a noncompetitive RT-PCR method routinely used in our laboratory (33). Briefly, RNA was extracted using RNAble (Eurobio, les Ulis, France) according to the manufacturer’s instruction. Total RNA was treated with 5 U RNase-free DNase (Boehringer Mannheim) for 45 min at room temperature, and DNase was then inactivated by heating for 5 min at 95°C. RNA was reverse-transcribed in optimal conditions, as previously defined (34, 35). Primers were as follows: EAAT-1 sense, 5′-GCTAGATAGTAAGGCATCAGGGAA-3′; EAAT-1 antisense, 5′-AAGCACATGGAGAAGACAACTAGA-3′ (amplicon size, 429 bp); EAAT-2 sense, 5′-TGGATGCTAAGGCTAGTGGC-3′; EAAT-2 antisense, 5′-GCACCTCAGTCACAGTCTCG-3′ (amplicon size, 345 bp); EAAT-3 sense, 5′-TTCTAGGTATTGTGCTGGTGGTGA-3′; EAAT-3 antisense, 5′-TCCAAAGACAAGGCAAAAGACAAT-3′ (amplicon size, 350 bp); GAPDH sense, 5′-ACCACCATGGAGAAGGCTGG-3′; GAPDH antisense, 5′-CTCAGTGTAGCCCAGGATGC-3′ (amplicon size, 509 bp).
Primer specificity was confirmed by both amplicon size assessment and restriction analysis with StuI, HindIII, and AVAII. RT-PCR amplicons were resolved in a 1.5% agarose gel by electrophoresis, and signal was quantified with densitometric analysis software (NIH Image 1.2; W. Rasband, National Institutes of Health, Bethesda, MD). The relative abundance of mRNA species was determined using a standard curve for each PCR run. PCR was performed with three or four, one in four dilutions of each sample, giving a semilog range of amplification. Each amplification was repeated at least twice. Data are expressed as the ratio of the signal obtained for each glutamate transporter divided by that obtained for GAPDH in the same sample, to permit the comparison of RNA species between samples.
Intracellular glutathione content
MDM were cultured overnight in DMEM without cystine, glutamine, and glutamate (DMEM Cyst−/Gln−/Glu−; Life Technologies), supplemented with 0.1% FCS. Cells were then washed with PBS and incubated with 300 μl DMEM Cyst−/Gln−/Glu− supplemented with 0.1% FCS in the presence or absence of cystine, glutamate, or THA for 4.5 h. MDM were washed with PBS and lysed with 150 μl PBS, 0.1% Tween for 1 h. GSH content was measured using an enzymatic assay (Cayman Chemicals, Ann Arbor, MI) as specified by the manufacturer. Protein content of cell lysates was determined by the Bradford method.
Glutamate uptake by MDM
MDM were incubated with various concentrations of [3H]glutamate for 5 min at 4°C or at 37°C, and cell-associated radioactivity was measured (Fig. 2⇓A). Cell-associated radioactivity was about 90% lower for incubations at 4°C than for those at 37°C, consistent with glutamate transport by MDM rather than receptor binding. At 37°C, intracellular radioactivity linearly increased with time for at least 5 min and was within 10 and 20% of linearity after 10 min (data not shown). The values obtained at 5 min were thus considered to be satisfactory approximations of initial uptake rates. We determined the kinetic parameters of glutamate uptake by MDM by measuring initial uptake velocities at 37°C for glutamate concentrations of 1–400 μM. Glutamate uptake was found to be saturable and approached saturation at 150 μM. Lineweaver-Burk plots (Fig. 2⇓B) revealed a Michaelis constant (Km) of 77 ± 6 μM and a maximum velocity (Vmax) of 2044 ± 181 pmol/mg protein/min.
Ionic requirements of glutamate uptake by MDM
We determined the glutamate transport in uptake medium with ionic modifications. We first showed that the absence of sodium inhibited glutamate uptake by 80 and 70% for glutamate concentrations of 1 and 200 μM, respectively (Fig. 3⇓A). The mean ± SEM sodium dependence for nine donors was 68 ± 4.7%, with a range of 50–88% (see Fig. 7⇓). The effects on glutamate uptake of changing K+ and Ca2+ concentration were also evaluated. Increasing K+ concentration from 1.4 mM to 68.5 mM reduced uptake by 60 ± 5% compared with the value obtained when K+ was replaced by an equimolar concentration (68.5 mM) of choline chloride (Fig. 3⇓B). These observations suggest that 60–80% glutamate uptake by MDM is mediated by the Na+/K+-dependent high-affinity glutamate transport (XAG−). The absence of Ca2+ had no effect on glutamate uptake, suggesting that this transport mechanism is different from the Ca2+-dependent glutamate uptake also described for brain homogenates (36).
Inhibition of MDM glutamate uptake by EAA analogues
We assessed the potential of EAA analogues for inhibiting glutamate uptake by MDM (Table II⇓). THA and trans-PDC, two competitive inhibitors specific for EAATs, efficiently inhibited 1 μM glutamate uptake with inhibition constant (Ki) values of 15 ± 5 and 56 ± 33 μM, respectively. DHK and l-α-aminoadipate, which inhibit EAAT2 but not EAAT1, inhibited uptake by only 25.6 ± 2.5% and 29.5 ± 10%, respectively, when present in a 1000-fold excess over glutamate. l-cystine, l-homocysteate, and quisqualate reduced glutamate transport by 16.6 ± 11%, 13.7 ± 5.7%, and 19.9 ± 5%, respectively. Thus, glutamate uptake by MDM is mostly mediated by a THA- and trans-PDC-sensitive transport system, such as EAAT1 or EAAT3, but probably not by EAAT2. At low glutamate concentrations, only 15% of glutamate uptake was abolished by blocking the cystine/glutamate transporter, demonstrating the higher affinity of glutamate for EAATs over the cystine/glutamate antiporter.
Time-course of Na+-dependent high-affinity glutamate transporter expression and function during culture of monocyte-macrophages
EAAT1 and EAAT2 genes were weakly expressed or undetectable on freshly sorted monocytes (Figs. 4⇓ and 5A). EAAT1 and EAAT2 mRNA expression levels markedly increased after 1 h in culture, reached a maximum by day 2 (120,000- and 55,000-fold increases, respectively), and then slowly decreased until day 12 (Fig. 5⇓A). We found no change in EAAT3 mRNA levels over time (weak signal, data not shown). On the day of elutriation, the monocytes did not transport significant amounts of glutamate (71 ± 3 pmol/mg protein/min) (Fig. 5⇓B). An increase in Na+-independent glutamate uptake was observed after 20 h of culture (580 ± 122 and 536 ± 196 pmol/mg protein/min for total and Na+-independent glutamate uptake, respectively). Na+-independent glutamate transport then decreased slowly with time, reaching 237 ± 10 pmol/mg protein/min after 5 days. Na+-dependent glutamate transport (1,347 ± 107 pmol/mg protein/min) and an increase in Na+-independent transport (1,271 ± 295,107 pmol/mg protein/min) began on day 8, concomitant with the morphological differentiation of monocytes into fibroblast-like macrophages. Total glutamate uptake then slowly decreased between days 8 and 14, and there were no further changes in the two systems of glutamate transport from day 14 to day 30 (data not shown).
Effect of TNF-α on glutamate transport by monocytes
We investigated whether TNF-α, an inflammatory cytokine present at high concentration in both the periphery and CNS in many diseases, could induce glutamate transport by monocytes. Three-day stimulation of freshly elutriated monocytes induced a dose-dependent increase in Na+-independent glutamate transport (152 ± 9 vs 1123 ± 4 pmol/mg protein/min for 0 and 100 ng/ml TNF-α, respectively) and the appearance of the Na+-dependent glutamate transport (859 pmol/mg protein/min for monocytes stimulated with 100 ng/ml TNF-α) (Fig. 6⇓). This suggests that circulating monocytes may rapidly acquire functional EAATs in inflammatory conditions.
Glutamate uptake by in vivo-differentiated macrophages and microglia
We investigated whether tissue macrophages and microglia transport glutamate (Fig. 7⇓). No significant glutamate transport was detected for murine splenic and simian alveolar macrophages the day of isolation (about 250 and 160 pmol/mg protein/min for total uptake and glutamate binding, respectively). Nevertheless, we detected efficient glutamate transport in 9- to 12-day cultured human splenic macrophages and in 20-day cultured murine microglia that was 79 ± 3.2% and 76 ± 7% Na+-dependent, respectively, and had a velocity similar to that of human MDM (1469 ± 92 and 2230 ± 433 pmol/mg protein/min for 100 μM glutamate, respectively). This shows that tissue macrophages originating from both the CNS and peripheral organ do express functional EAATs after some days in culture.
Effect of MDM on glutamate-induced neurotoxicity
We tested whether MDM could regulate extracellular glutamate concentration by culturing MDM for 6 h, 20 h, or 4 days in the presence of various concentrations of glutamate. Supernatants were harvested and tested for neurotoxicity using primary mouse neuronal cultures. We also assessed the neurotoxicity of these glutamate concentrations in culture medium without MDM as a control (Fig. 8⇓). As previously described, we observed that supernatants from macrophages without glutamate induced a neurotoxicity, which was maximal after 6 h of culture (36%) and decreased to 26% and 18% after 20 h and 4 days of culture, respectively. In our culture conditions, 100 μM glutamate was sufficient for maximal toxicity to neurons (50%). No significant degradation of glutamate was observed in culture medium at 37°C (in the absence of MDM), as neurotoxicity was similar for all incubation times. MDM induced a time- and dose-dependent protection against glutamate neurotoxicity. After 6 h of culture (Fig. 8⇓A), MDM reduced the neurotoxicity induced by 100 μM glutamate by 43%. After 20 h of culture (Fig. 8⇓B), MDM reduced glutamate-induced neurotoxicity by 62.5, 77.5, and 36% for 100, 300, and 1000 μM glutamate, respectively. After 4 days of culture (Fig. 8⇓C), maximum protection against glutamate neurotoxicity was obtained: 100, 81, and 41% for 100, 300, and 1000 μM glutamate, respectively. Vmax for Na+-dependent glutamate uptake by MDM in this experiment was 3000 pmol/mg protein/min (uptake measured for 5 min, data not shown). This velocity would result in the clearance of 100% of 100 and 300 μM glutamate and 32% of 1 mM glutamate after 4 days of culture. These values are consistent with the observed kinetics of the neuroprotective effect of MDM. Fig. 8⇓D shows the mean ± SEM percentage of neuroprotection induced by MDM against glutamate toxicity in two independent experiments. The results are similar to those shown in Fig. 8⇓, A–C.
Effect of extracellular glutamate and glutamate transport on intracellular GSH concentration
Because cystine and glutamate are precursors for GSH synthesis, we tested the ability of these two amino acids to modulate GSH synthesis. A weak level of intracellular GSH (about 40 nmol/mg protein) was measured in MDM cultured in the absence of cystine. Cystine (100 μM) induced GSH synthesis (73 ± 2 nmol/mg protein), and this level was increased by 43% with 100 μM exogenous glutamate (104 ± 3 nmol/mg protein). The level of intracellular GSH in cystine- and glutamate-incubated cells returned to a value of 35 nmol/mg protein when MDM were incubated with 5 mM of l-buthionine-[S,R]-sulfoximine, a blocker of γ-glutamyl-cysteine-synthetase (data not shown). When MDM were cultured in the presence of cystine, glutamate, and THA (1 mM), intracellular GSH level returned to a value of 58 ± 6 nmol/mg protein. This demonstrates that EAATs are indeed involved in the regulation of intracellular GSH synthesis by MDMs by providing intracellular glutamate.
We demonstrated in this study that macrophages derived from human blood monocytes have both a Na+-independent and a Na+-dependent (67.7 ± 4.7%) glutamate transport system. This latter was inhibited by an increase in extracellular K+ concentration. This is consistent with the mechanism proposed by Kanner (37) for the Na+/K+ high-affinity uptake of glutamate in the CNS. Glutamate uptake would indeed be driven by a Na+ concentration gradient and by a K+ concentration gradient in the opposite direction. However, we cannot exclude that increased extracellular K+ may also act through disturbance of the Na+ gradient as it depolarizes the membrane potential and reduces the electrical component of the electrochemical gradient for Na+ entry. Because uptake was insensitive to Ca2+, it clearly differs from the Ca2+/Cl−-dependent glutamate uptake previously described in brain homogenates (36). At low glutamate concentration (1 μM), glutamate uptake was dose-dependently inhibited by the EAAT inhibitors THA and Trans-PDC, whereas a 1000-fold molar excess of l-cystine (or l-homocysteate) over glutamate inhibited glutamate uptake by only 15–20%. This demonstrates that MDM express EAATs and that extracellular glutamate is mostly taken up by the XAG− system rather than the Xc− system.
In this study, numerous lines of evidence suggest that EAAT1 would be largely responsible for Na+-dependent glutamate transport by MDM: 1) DHK, an EAAT2-specific inhibitor, and l-α-aminoadipate, an EAAT2 and EAAT4-specific inhibitor weakly inhibited glutamate uptake (≤30%) (9, 38, 39); and 2) EAAT3 mRNA was barely detectable in MDM. However, we cannot rule out the involvement of the recently cloned EAAT5 because specific inhibitors of this transporter have not been yet identified. The Km and Vmax for glutamate transport were 77 ± 6 μM and 2044 ± 181 pmol/mg protein/min, respectively, for total uptake and 58.5 ± 23 μM and 1333 ± 471 pmol/mg protein/min for Na+-dependent uptake. These values are very similar to those obtained by others for cortical neurons, cortical synaptosomes, and glial cultures (40, 41, 42). Therefore, MDM may be as efficient as neural cells (astrocytes and neurons) in clearing extracellular glutamate, suggesting an important role in the CNS.
Time-course studies demonstrated a large increase in EAAT1 and EAAT2 mRNA levels at an early stage of monocyte culture. Cultured monocytes did not display Na+-dependent glutamate transport until day 5, but we detected a Na+-independent glutamate transport after 20 h of culture that was probably mediated by the cystine/glutamate transporter. Na+-dependent transport coincided with the morphological differentiation of monocytes into fibroblast-like cells, and a high level of transport was maintained until day 30 of culture. Such a dissociation of EAAT gene expression and glutamate transporter activity has already been described (for review, see Ref. 43). It would be of interest to investigate the possible posttranscriptional regulation or regulation of signal transduction mechanisms, accounting for differences between gene expression and transporter activity in MDM. Indeed, Casado et al. (44) showed that protein kinase C-dependent phosphorylation induced an increase of EAAT activity, and Dowd and Robinson (45) described the PMA induction of EAATs in cycloheximide-treated C6 cells, which was therefore independent of protein synthesis.
We did not detect any glutamate transport by in vivo-differentiated macrophages from spleen or lung on the day of cell isolation. This is consistent with previous reports (29, 30) suggesting that resting tissue macrophages do not constitutively express functional EAATs. Nevertheless, our in vitro data show that MDM activated by adhesion to plastic before differentiation, as well as cultured human splenic macrophages and cultured microglia, do have a highly efficient XAG− transport system. This provides evidence that, in the CNS, both resident (microglia) and infiltrating macrophages acquire high-affinity Na+-dependent glutamate transporters and lower excitotoxicity. Alternatively, XAG− transport may also be induced in the periphery by specific stimulation by cytokines or inflammatory mediators. This second possibility is supported by our results demonstrating that TNF-α stimulation induces Na+-dependent glutamate transport by monocytes and that 9- to 12-day cultured splenic macrophages do have efficient XAG− transport systems. Additional studies of the stimulation requirements for XAG− system expression or regulation in monocytes and macrophages and in situ expression of EAATs are required to support this observation.
It is known from antisense-based knockout studies that EAAT1 and EAAT2 are critical for the regulation of extracellular glutamate concentrations in the brain (14, 15). The extracellular concentration of glutamate significantly increases in many neurological disorders (for review, see Ref. 2). This may be due to an increase in intrathecal glutamate production (including by activated macrophages and microglia) and/or to the down regulation of astroglial glutamate transporters, especially through macrophage-produced mediators such as arachidonic acid, oxygen-free radicals, or TNF (19, 20, 46, 47, 48, 49). We indeed observed that MDM constitutively produce neurotoxins, but we also showed that MDM time- and dose-dependently clear glutamate from culture medium, thereby reducing excitotoxicity. Our results thus suggest that during brain injury associated with an increase in extracellular glutamate concentration, infiltrating and resident macrophages, although producing neurotoxins, may also, together with astrocytes, regulate extracellular glutamate levels. These observations are in line with a recent study demonstrating that EAAT activity, although detectable in embryonic and early postnatal astrocytes, decline to undetectable levels in mature astrocytes (16). This suggests that the relative roles of astrocytes and macrophages/microglia in the regulation of extracellular glutamate concentrations might be reconsidered and should be more precisely studied with a developmental perspective.
Cystine, cysteine, and glutamate are GSH precursors, and Reichelt et al. (50) reported in retinal Muller glial cells that GSH synthesis could be limited by the capacity of EAATs to provide intracellular glutamate for both cystine uptake and direct insertion into GSH. The presence of both the cystine/glutamate transporter and EAATs on MDM, the higher affinity of glutamate for EAATs, and our data (Fig. 9⇓) demonstrating that uptake of glutamate via EAATs indeed increases GSH synthesis induced by cystine suggest that there is a continuous glutamate exchange between intra and extracellular media. Glutamate is probably excreted through the cystine/glutamate transporter in exchange for cystine, leading to a decrease in its intracellular concentration. In turn, EAATs may take up extracellular glutamate, thus limiting the extracellular competition between glutamate and cystine for the cystine/glutamate antiporter and maintaining intracellular glutamate availability. Our data are in accordance with the observations reported on retinal Muller glial cells (50) and suggest that cooperation between the cystine/glutamate transporter and EAATs may thus play a role in the antioxidant functions of macrophages by regulating intracellular concentration of the radical scavenger GSH.
The involvement of EAATs in controling cystine availability may also help macrophages in providing cysteine to lymphocytes that do not possess the cystine/glutamate antiporter (Ref. 51 ; and A.-C. Rimaniol and G. Gras, unpublished data). In addition, lymphocytes have glutamate receptors (52), and high glutamate concentrations depress lymphocyte function, as described in HIV infection and in patients with bronchial carcinoma (53, 54). Thus, EAAT expression in peripheral macrophages should be evaluated in pathological conditions because this would make it possible to locally restrain the detrimental effects of glutamate. Amino acid metabolism has already been implicated in the interactions between macrophages and lymphocytes in that suppressive macrophages deplete extracellular tryptophan, leading to cycle arrest in activated lymphocytes (55). Thus, amino acid transporters may be effectors of fine mechanisms for local immune response modulation.
We thank Dr. Raphaël Szymocha, Franck Mouthon, and Thierry Lévêque for helpful scientific advice. We are indebted to Dr. Thierry de Revel (Service d’Hématologie, Hôpital Percy, Clamart, France) for providing human spleen.
↵1 This work was supported in part by grants from the Agence Nationale de Recherches sur le SIDA (ANRS) and Sidaction. A.-C.R. is a recipient of a fellowship from the ANRS.
↵2 Address correspondence and reprint requests to Dr. Gabriel Gras, Service de Neurovirologie, DSV/DRM, Commissariat à l’Energie Atomique, BP 6, 60–68 avenue de la division Leclerc, 92265 Fontenay-aux-Roses, France. E-mail address:
↵3 Abbreviations used in this paper: EAAT, excitatory amino acid transporter; GSH, glutathione; MDM, monocyte-derived macrophage; THA, dl-threo-β-hydroxyaspartic acid; DHK, dihydrokainate; trans-PDC, l-trans-pyrrolidine-2,4-dicarboxylic acid; Vmax, maximum velocity, Km, Michaelis constant; Ki, inhibition constant.
- Received October 25, 1999.
- Accepted March 7, 2000.
- Copyright © 2000 by The American Association of Immunologists