Address correspondence and reprint requests to Roberto Paes-de-Carvalho, Departamento de Neurobiologia, Instituto de Biologia, Caixa Postal 100180, Centro, Niterói, RJ 24001-970, Brasil. E-mail: email@example.com
Vitamin C is transported in the brain by sodium vitamin C co-transporter 2 (SVCT-2) for ascorbate and glucose transporters for dehydroascorbate. Here we have studied the expression of SVCT-2 and the uptake and release of [14C] ascorbate in chick retinal cells. SVCT-2 immunoreactivity was detected in rat and chick retina, specially in amacrine cells and in cells in the ganglion cell layer. Accordingly, SVCT-2 was expressed in cultured retinal neurons, but not in glial cells. [14C] ascorbate uptake was saturable and inhibited by sulfinpyrazone or sodium-free medium, but not by treatments that inhibit dehydroascorbate transport. Glutamate-stimulated vitamin C release was not inhibited by the glutamate transport inhibitor l-β-threo-benzylaspartate, indicating that vitamin C release was not mediated by glutamate uptake. Also, ascorbate had no effect on [3H] d-aspartate release, ruling out a glutamate/ascorbate exchange mechanism. 2-Carboxy-3-carboxymethyl-4-isopropenylpyrrolidine (Kainate) or NMDA stimulated the release, effects blocked by their respective antagonists 6,7-initroquinoxaline-2,3-dione (DNQX) or (5R,2S)-(1)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801). However, DNQX, but not MK-801 or 2-amino-5-phosphonopentanoic acid (APV), blocked the stimulation by glutamate. Interestingly, DNQX prevented the stimulation by NMDA, suggesting that the effect of NMDA was mediated by glutamate release and stimulation of non-NMDA receptors. The effect of glutamate was neither dependent on external calcium nor inhibited by 1,2-bis (2-aminophenoxy) ethane-N′,N′,N′,N′,-tetraacetic acid tetrakis (acetoxy-methyl ester) (BAPTA-AM), an internal calcium chelator, but was inhibited by sulfinpyrazone or by the absence of sodium. In conclusion, retinal cells take up and release vitamin C, probably through SVCT-2, and the release can be stimulated by NMDA or non-NMDA glutamate receptors.
Many organisms synthesize ascorbate from glucose in the liver but some vertebrate species, including humans, are unable to produce this vitamin because of the lack of a functional gene for the enzyme l-gulono-γ-lactone oxidase, which converts l-gulono-γ-lactone to l-ascorbate (Rice 2000; Wilson 2005); therefore, a dietary source of ascorbate is required.
Differences between SVCT-1 and 2 include organ and tissue distribution. For instance, isoform 1 is present in epithelial organs such as liver and kidney, whereas SVCT-2 is found in brain and skeletal muscle (Rice 2000; Takanaga et al. 2004; Wilson 2005). Moreover, SVCT-2 appears to have higher affinity for l-ascorbate when compared to SVCT-1 (Daruwala et al. 1999; Wilson 2005).
Glutamate is the major excitatory neurotransmitter in the CNS. This neurotransmitter is involved in many aspects of normal brain functions such as memory and learning as well as in the processing of light stimuli (Kolb 2003; Wassle 2004) in the retina. Glutamate can also trigger processes leading to neuronal death (Danbolt 2001; Ferreira and Paes-de-Carvalho 2001). Glutamate activates metabotropic and ionotropic receptors. The metabotropic receptors (mGluR) consist of eight subtypes, subdivided into group I (mGluR1 and mGluR5), group II (mGluR2 and mGluR3) and group III (mGluR4, 6, 7 and 8) on the basis of sequence homology and G-protein coupling (Danbolt 2001). Ionotropic receptors are classified in three subtypes, based on their selective agonists N-methyl-d-aspartate (NMDA), α-amino-3-hydroxy-5-methyl-4-isoazolepropionic acid (AMPA) and 2-carboxy-3-carboxymethyl-4-isopropenylpyrrolidine (kainate) receptors (Danbolt 2001).
The chick retina is an useful model for neurochemical and developmental studies in the CNS since it is accessible during development and several of its neurotransmitter systems have already been well characterized (de Mello 1978; Large et al. 1985; do Nascimento et al. 1998; Cossenza and Paes de Carvalho 2000; Paes-de-Carvalho 2002; Paes-de-Carvalho et al. 2005; Calaza et al. 2006; Cossenza et al. 2006; Magalhaes et al. 2006; Mejia-Garcia and Paes-de-Carvalho 2006). In the present work we have investigated the expression of SVCT-2 in the retina and functions of a transport system for ascorbate in cultured chick retinal cells. We found that retinal cells take up ascorbate, in a time- and dose-dependent manner, and that activation of ionotropic glutamate receptors induces the release of vitamin C, an effect abolished by their respective antagonists. Interestingly, the stimulation by glutamate or NMDA was blocked by the AMPA/kainate receptor antagonist 6,7-initroquinoxaline-2,3-dione (DNQX). Moreover, both the uptake and release of vitamin C were sodium-dependent and inhibited by sulfinpyrazone, indicating that these processes are mediated by the SVCT-2 protein, which we could detect in the retina and in mixed cultures. The results indicate the expression of an ascorbate transport system and its regulation by glutamate in the retina.
Material and methods
Fertilized White Leghorn chicken eggs were obtained from a local hatchery and incubated at 38°C in a humidified atmosphere. [14C] ascorbic acid (50 mCi/mmol), polyvinylidene difluoride membranes and enhanced chemiluminescence (ECL) kits were obtained from GE Healthcare (Buckinghanshire, UK). d-[2,3 3H]Aspartic Acid (12.2 Ci/mmol) was from PerkinElmer (Waltham, MA, USA). HEPES, EGTA, sulfinpyrazone, ascorbic acid, penicillin, streptomycin, glutamic acid, kainic acid, dimethyl sulphoxide, DNQX, (5R,2S)-(1)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate (MK-801), 1-aminocyclopentane-trans-1,3-dicarboxylic acid, l-β-threo-benzylaspartate (T-BOA), 4′,6-diamidino-2-phenylindole (DAPI), 1,2-bis (2-aminophenoxy) ethane-N′,N′,N′,N′,-tetraacetic acid tetrakis (acetoxy-methyl ester) (BAPTA-AM), bovine serum albumin (BSA) and poly-l-lysine hydrobromide were purchased from Sigma/RBI (St. Louis, MO, USA). LY294002, 2-amino-5-phosphonopentanoic acid (APV) and NMDA were from Biomol (Plymouth Meeting, PA, USA). Minimum essential medium, fetal bovine serum (FBS), glutamine and trypsin were supplied by GIBCO (Grand Island, NY, USA). (R,S)-1-aminoindan-1-5-dicarboxylic acid was purchased from Tocris Cookson Inc. (Ellisville, MO, USA). Goat anti-SVCT-2 antibody and the specific blocking peptide were from Santa Cruz Biotecnology (Santa Cruz, CA, USA) and rabbit anti-mouse secondary antibody coupled to Alexa568 and donkey anti-goat secondary antibody coupled to Alexa488 were obtained from Molecular Probes (Eugene, OR, USA). Horseradish peroxydase-conjugated secondary antibody was from Invitrogen (Carlsbad, CA, USA). Mouse monoclonal 2M6 antibody was kindly supplied by B. Schlosshauer (NMI Naturwissenschaftliches und Medizinisches Institut an der Universität Tübingen, Markwiesenstr. 55, D-72770 Reutlingen, Germany). Biotinylated anti-goat antibody and avidin-biotin complex (Vectastain Elite) were from Vector and OCT medium was from Sakura Finetek (Torrance, CA, USA). All other reagents were of analytical grade.
A total of three post-hatched White Leghorn chicks were used in this study. The procedures for the use of animals were in accordance with the ‘NIH guide for the Care and Use of Laboratory Animals’. Animals between post-hatching days 0 and 5 were killed by decapitation, the eyes enucleated and the posterior part of the eye containing the retinal pigment epithelium and the neural retina was fixed by immersion with 4% paraformaldehyde in 0.16 M phosphate buffer (PB), pH 7.2, for 2 h. The tissues were washed three times for 10 min with PB and submitted to a sucrose gradient (15% and 30%) overnight for cryoprotection. In the next day, the tissues were immersed in OCT medium, frozen and cryosectioned. Sections of 10–12 μm thickness were made perpendicular to the vitreal surface and collected on poly-l-lysine hydrobromide-coated slides (200 μg/mL). Twenty four hours later, sections were submitted to an antigenic recovery by immersing the slides in boiled citrate buffer (10 mM, pH 6.0) for 1 min. After this procedure, sections were washed with phosphate-buffered saline (PBS), pH 7.4, incubated for 1 h with a blocking solution (5% BSA plus 5% FBS plus 0.25% Triton X-100 in PBS) and then overnight with the goat anti-SVCT-2 primary antibody (1 : 50). After further washes with PBS, sections were incubated for 2 h with biotinylated anti-goat antibody (1 : 200). Another series of washes was performed and the sections incubated with the avidin-biotin complex (1 : 50) for 90 min. All reagents mentioned above were diluted in the blocking solution. The sections were rinsed twice for 10 min with PBS and once for 10 min with Tris-buffered saline, pH 7.6, before incubation with 0.05% 3,3-diaminobenzidine and 0.01% hydrogen peroxide for 10 min. After several washes in Tris-buffered saline, slides were mounted with 40% glycerol in PB. Control sections were incubated with the goat anti-SVCT-2 primary antibody plus the blocking peptide diluted in blocking solution, or incubated only with the blocking solution in the absence of primary antibody. All incubations were performed at 25°C. All sections were examined and photographed using a Zeiss Axioskop light microscope.
Cell cultures were washed with Hank’s balanced salt solution (HBSS) and fixed for 10 min at 25°C with 1% paraformaldeyde in 0.16 M phosphate buffer. After blocking non-specific sites with PBS containing 5% BSA, 5% FBS and 0.1% Triton X-100 (blocking solution) for 2 h, cultures were washed and incubated with blocking solution containing primary antibody (SVCT-2 1 : 100) overnight. To label Müller glial cells, we incubated cultures overnight with an antibody against 2M6 (1 : 200), which was previously characterized and recognizes a Müller glia cell-specific antigen (Schlosshauer et al. 1991). When both antibodies anti-SVCT-2 and anti-2M6 were used together they were incubated consecutively. Cultures were next washed and incubated with secondary antibodies (Alexa 488 and Alexa 588 1 : 200) for 2 h. Next, cultures were incubated with 4′,6-diamidino-2-phenylindole (1 : 1000) for 30 s and washed twice for 5 min. All incubations were performed at 25°C.
Preparation of cultures
Monolayer cultures of chick retinal cells were prepared as previously described (de Mello 1978). Briefly, retinas from 8-day-old chick embryos were dissected from other ocular tissues, including the retinal pigment epithelium, and digested with 0.2% trypsin in calcium and magnesium-free Hank’s balanced salt solution, for 15 min at 37°C. Cells were then suspended in minimum essential medium supplemented with 3% heat-inactivated FBS, penicillin (100 U/mL), streptomycin (100 mg/mL) and glutamine (2 mM) and seeded in tissue culture plastic dishes at a density of 2 × 104 cells/mm2. Cultures were maintained at 37°C in a humidified incubator with 95% air and 5% CO2. The medium was changed after 1 day in culture (C1) and experiments were performed at C3–C4. Cultures of rat cells were obtained after dissection of retinas from 2-day-old Lister Hooded rats and incubation in medium 199 supplemented as in chick cultures except that FBS concentration was 5% and the cells were seeded over poly-l-ornithine-coated dishes (50 μg/mL).
[14C] vitamin C uptake
Before initiating experiments, the medium was removed and dishes were rinsed twice with HEPES-buffered salt solution, containing NaCl 140 mM; KCl 5 mM; HEPES 20 mM; glucose 4 mM; MgCl2 1 mM and CaCl2 2 mM, pH7.4. For uptake experiments, cultured cells were pre-incubated 10 min at 37°C with HBSS in the absence or presence of different drugs, incubated with [14C] Ascorbate (0.1 μCi/mL) during the indicated periods, rinsed twice and lysed with 5% trichloroacetic acid for determination of intracellular radioactivity. In some experiments, sodium-free HBSS was prepared in which sodium chloride was substituted by lithium chloride.
[14C] vitamin C release
For release experiments, after incubation with [14C] ascorbate (0.3 μCi/mL) for 40 min at 37°C, cultures were rinsed twice with HBSS and further incubated with HBSS during four periods of 3 min to completely remove extracellular radioactivity. Cells were then incubated for 10 min with HBSS in order to estimate the basal release. Cultures were incubated for another period of 10 min with HBSS containing different antagonists. A third period of 10 min of incubation was performed with HBSS containing glutamate, NMDA or Kainate in the absence or presence of different antagonists, and cells were then lysed with 5% trichloroacetic acid. For all 10 min incubation periods, samples were reserved to estimate the radioactivity. In all experiments using NMDA, HBSS did not contain magnesium chloride. In calcium-free experiments, HBSS did not contain calcium chloride and 2 mM EGTA was added to chelate residual calcium. In sodium-free experiments, HBSS did not contain sodium chloride, which was replaced by lithium chloride. The radioactivity of all these samples was measured by liquid scintillation spectroscopy. Results were normalized to percent of control after calculation of the percent fractional release, that is, the percent of radioactivity released compared to intracellular radioactivity at each period of time.
Western blot analysis
For detection of SVCT-2, samples were washed with HBSS, sample buffer (4.4 mL Milli-Q water; 1 mL Tris-HCl 0.5 M pH6,8; 0.8 mL glycerol; 1.6 mL 10% sodium dodecyl sulfate; 0.4 mL β-mercaptoethanol) was added and the material boiled for 5 min. Samples containing 60 μg protein were submitted to 9% SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) and proteins transferred to polyvinylidene difluoride membranes membranes which were incubated overnight with antibodies that specifically recognize SVCT-2 (1 : 800 dilution), washed, incubated with anti-goat peroxidase-conjugated secondary antibody (1 : 1000 dilution) and developed by ECL kit. The total amount of protein in each sample was determined using the Bradford reagent, with bovine serum albumin as a standard.
Statistical analyses were performed using one-way anova followed by the Bonferroni post-test using the software GraphPad Software Inc. (San Diego, CA, USA).
Presence of SVCT-2 in chick retina
To observe the retinal vitamin C transporters, chick retina sections were immunostained with an antibody against SVCT-2 and processed for immunohistochemistry. As observed in the central panel of Fig. 1(a), specific labeling was observed mainly over cells in the inner portion of the inner nuclear layer and in the ganglion cell layer. The staining was blocked by pre-incubating sections with the specific SVCT-2 anti-peptide (second panel from the left) or in the absence of primary antibody (extreme left panel). Retinas from 14-day-old chick embryos (E14) and from 2 day-old neonatal rats (P2) were also dissected and processed for western blot analysis. For both retina types, intense bands for SVCT-2 were observed (Fig. 1b), as previously observed for the rat retina (Tsukaguchi et al. 1999). Intense bands for SVCT-2 were also observed in samples from cultures of retinal cells from 8-day-old chick embryos incubated for 3 days (E8C3) and cultures from 2-day-old rats incubated for 2 days (P2C2) (Fig. 2a). A blocking peptide caused the disappearance of bands, confirming their specificity (Fig. 2a). Immunofluorescence studies were performed with E8C3 cultures using the same SVCT-2 antibody (Fig. 2b and c). A stained population of cultured cells was observed and this staining was completely blocked by pre-incubation with the blocking peptide (Fig. 2b, right panel). Since these cultures contain neurons growing over a layer of flat glial cells, as observed in the phase contrast image (Fig. 2b, left panel), we performed immunocytochemistry experiments to identify whether neurons or glial cells were expressing the SVCT-2 transporter. For this purpose we used an antibody against a specific glial cell marker (2M6) (red) simultaneously with an antibody against SVCT-2 (green). As shown in Fig. 2(c), no double labeling was found, implying that SVCT-2 is expressed in cultured retinal neurons but not in glial cells. Indeed, the fluorescence associated with SVCT-2 was restricted to the cell body of neurons as observed in other neuronal populations (Mun et al. 2006). The lack of specific staining in cells incubated in presence of the inhibitory peptide (Fig. 2b, right panels and Fig. 1a, second panel from the left), as well as in cells incubated in absence of primary antibodies for SVCT-2 and 2M6 (Fig. 2c, right panels and Fig. 1a, extreme left panel), confirmed the validity of these observations.
The uptake of [14C] vitamin C in the cultures was time-dependent, reaching a plateau in approximately 55 min (Fig. 3a). Using an incubation time in the linear phase (5 min), the concentration curve showed saturation in concentrations close to 50 μM (Fig. 3b). The Eadie-Hofstee plot was curvilinear (Fig. 3b, inset), indicating the existence of at least two uptake components (R2 = 0.992), with Kms of 6.5 ± 1.4 and 37.9 ± 7.4 μM and Vmaxs of 103.9 ± 10.0 and 270.2 ± 31.7 pmols/mg protein.minute.
Evaluation of oxidation state of taken up vitamin C
Several studies have shown that while the transport of ascorbate is mediated by SVCTs, glucose transporters 1, 3 and 4 mediate DHA transport (Rumsey et al. 1997; Siushansian et al. 1997; Qutob et al. 1998; Korcok et al. 2003; Wilson 2005). Since ascorbate can be spontaneously oxidized to DHA, we performed experiments to verify the chemical species that was being taken up by cultured cells. The incubation of cultures with 20 mM glucose, a treatment known to block DHA uptake, had only a marginal effect on the uptake (Fig. 4a). It is also well established that phosphoinositide 3-kinase (PI3K) is involved in the recruitment of glucose transporters to the cell membrane (Song et al. 2001; Tengholm and Meyer 2002). Thus, when this kinase is blocked, the uptake of DHA should be reduced. We therefore used LY294002 (50 μM), a PI3K blocker, to evaluate vitamin C uptake by the cells in culture. As observed in Fig. 4(a), no significant reduction in the uptake was detected. These results suggest that the observed transport is of ascorbate rather than its oxidized form. In contrast, 1 mM sulfinpyrazone, a non-selective inhibitor of vitamin C transporter (Tsukaguchi et al. 1999; Wang et al. 2000; McNulty et al. 2005; Davis et al. 2006; May et al. 2006), strongly decreased the uptake of [14C] ascorbate (Fig. 3b). The substitution of sodium chloride by lithium chloride also inhibited the uptake (Fig. 4b), suggesting that [14C] ascorbate is taken up by the SVCT-2.
Stimulation of vitamin C release by glutamate
As observed in other CNS tissues (Grünewald and Fillenz 1984; Cammack et al. 1991; Rebec and Pierce 1994; Wilson et al. 2000), glutamate stimulated [14C] vitamin C release in retinal cultures. This effect occurred in a dose-dependent manner, reaching a plateau with approximately 100 μM glutamate (Fig. 5a). Significantly, cultures responded to a second stimulus of glutamate (Fig. 5b), indicating that the release of vitamin C is not because of cell lysis and is regulated by glutamate in a reversible manner.
Two different mechanisms were evaluated to explain the effect of glutamate on ascorbate release. One possible mechanism is the heteroexchange of glutamate/ascorbate (Rebec and Pierce 1994), in which the influx of glutamate induces an efflux of ascorbate or vice-versa through a carrier protein that carries out the exchange. To test this possibility, we loaded cultured retinal cells with [3H] d-aspartate, which is known to substitute for glutamate in transport mechanisms (Kanai and Hediger 1992; Danbolt 2001), and stimulated the cells with different ascorbate concentrations. As shown in Fig. 6(a), concentrations of ascorbate from 10 to 500 μM did not promote [3H] d-aspartate release, in contrast to the stimulation of the cells with glutamate, which did release [3H] d-aspartate. This indicates that the heteroexchange mechanism is not playing any role in the glutamate effects on vitamin C release in our system. Another possibility is that glutamate uptake promotes cell swelling and triggers the activation of volume-sensitive anion channels capable of releasing ascorbate (Wilson et al. 2000). This possibility was also ruled out because T-BOA, a non-permeant glutamate uptake blocker, had no effect on glutamate-stimulated vitamin C release (Fig. 6b). Similar results were obtained using threo-b-hydroxyaspartate (THA), another potent competitive glutamate transport blocker (data not shown).
Stimulation of vitamin C release by glutamate ionotropic agonists
To investigate whether glutamate receptors were involved in the effect of glutamate on [14C] vitamin C release, cultured retinal cells were treated with agonists and antagonists of glutamate ionotropic receptors. NMDA (50 μM) or Kainate (50 μM) increased [14C] vitamin C release and these effects were respectively blocked by 10 μM MK-801 or 50 μM DNQX (Fig. 7a and b). MK-801 was used to evaluate if the stimulation by kainate was mediated by release of glutamate and activation of NMDA receptors. No effect was observed (Fig. 7b), indicating that NMDA or kainate independently promote vitamin C release. To exclude the involvement of metabotropic receptors, we treated the cultures with 100 μM 1-aminocyclopentane-trans-1,3-dicarboxylic acid Trans-ACPD, a metabotropic agonist and found that it did not stimulate release. Moreover, addition of 100 μM (R,S)-1-aminoindan-1-5-dicarboxylic acid (AIDA), a metabotropic receptor antagonist, did not block the effect of 100 μM glutamate (data not shown).
Effect of ionotropic antagonists on vitamin C release stimulated by glutamate
The increased release induced by 100 μM glutamate was completely blocked by 200 μM DNQX (Fig. 8a) but neither by the non-competitive NMDA antagonist MK-801 (100 μM) nor the competitive NMDA antagonist APV (Fig. 8b). This raised the possibility that the stimulation by glutamate was mediated by activation of AMPA/kainate receptors rather than NMDA receptors. One possible explanation for the fact that stimulation of NMDA receptors induces vitamin C release is that NMDA itself promotes glutamate release from one cell, which in turn activates AMPA/kainate receptors present in a second cell type. To test this hypothesis, we stimulated cultures with 50 μM NMDA after addition of 50 μM DNQX and observed that DNQX was able to completely inhibit NMDA effect (Fig. 8c).
Involvement of calcium on vitamin C release
As observed above, activation of glutamate receptors could stimulate vitamin C release, raising the possibility that this process is mediated by calcium influx induced by channel opening by glutamate or kainate. To test if the release was calcium-dependent, cultures were stimulated with either 1 mM glutamate or 50 μM kainate in calcium-free saline. Under these conditions, the results of stimulation with either glutamate or kainate were very similar to that observed in normal saline (Fig. 9a and b). To evaluate if internal calcium was involved in this process, we added BAPTA-AM (50 μM) to HBSS and observed that it had no effect on glutamate-stimulated vitamin C release (Fig. 9c).
Effect of sodium removal and sulfinpyrazone on vitamin C release
The results presented above suggest that the release of vitamin C was mediated by a calcium-independent (non-vesicular) mechanism. In order to investigate the possible participation of the SVCT-2 transporter in the release process, we tested the effect of 1 mM sulfinpyrazone, a known blocker of ascorbate transport (Tsukaguchi et al. 1999; Wang et al. 2000; McNulty et al. 2005; Davis et al. 2006; May et al. 2006). As observed in Fig. 10(a), this drug was able to block both the basal and glutamate-stimulated release. Sulfinpyrazone also blocked vitamin C release mediated by 50 μM NMDA or 50 μM kainate, as shown in Fig. 10(b) and (c). Moreover, removal of sodium ions from the saline inhibited the release stimulated by 1 mM glutamate (Fig. 10a) or 50 μM kainate (Fig. 10d), further corroborating the hypothesis that vitamin C release was mediated by SVCT-2.
The presence of vitamin C transporters is well documented in many tissues, including brain and retina (Tsukaguchi et al. 1999; Takanaga et al. 2004; Mun et al. 2006). The main subtype present in the CNS is SVCT-2 (Mun et al. 2006) but the regulation of these transporters in neurons and glial cells is poorly understood. We have demonstrated that SVCT-2 is present in the mature chick retina and is also expressed in cultured chick retinal neurons, but not in glial cells. We have also studied the properties of ascorbate uptake and release by chick retinal cells in culture. Cells that have previously taken up [14C] ascorbate were able to release vitamin C in response to glutamate and glutamate ionotropic agonists such as NMDA or kainate, in a dose and sodium-dependent but calcium-independent manner. Interestingly, NMDA-stimulated release is blocked by the AMPA/kainate antagonist DNQX suggesting the involvement of endogenous glutamate. The effect of glutamate is not mediated by glutamate uptake or by an ascorbate/glutamate heteroexchange mechanism. Moreover, both the uptake and release are inhibited by removal of sodium ions from the incubation medium or by sulfinpyrazone, indicating the involvement of SVCT-2.
Properties of [14C] ascorbate uptake by chick retinal cells in culture
When [14C] ascorbate is added to chick retinal cells in culture, the uptake of radioactivity showed a time and concentration-dependent profile. The Eadie-Hofstee plot was curvilinear indicating the presence of high and low affinity components, as observed in other embrionary systems (Castro et al. 2001). The uptake was not inhibited by high glucose or the PI3K inhibitor LY294002, indicating that the uptake was because of ascorbate and not DHA transport (Qutob et al. 1998; Song et al. 2001; Tengholm and Meyer 2002). Moreover, the uptake showed an absolute requirement for extracelluar sodium but not for extracellular calcium ions, since it was completely blocked in absence of extracellular sodium ions but was not modified in absence of extracellular calcium and presence of the extracellular calcium chelating agent EGTA. Together with the data showing a strong inhibition of uptake by sulfinpyrazone, the present results strongly indicate that the uptake of [14C] ascorbate by cultured retinal cells is mediated by SVCT transporters, probably of SVCT-2 subtype. Recent work has suggested that the transport mediated by SVCT-2 is absolutely dependent on bivalent cations, specifically calcium and magnesium (Godoy et al. 2007). Our present data show that calcium is not necessary for the uptake but we were unable to test if magnesium could substitute for calcium in the uptake since the cultures showed an extensive cell death when incubated in the absence of both ions (data not shown).
Release of [14C] ascorbate by glutamate and ionotropic agonists
The radioactivity incorporated by retinal cells in culture after incubation with [14C] ascorbate could be released in basal or glutamate-stimulated conditions. Glutamate did promote a stimulation of approximately 100% in the release, an effect that could be blocked when sodium ions were substituted by lithium ions, but not by chelating extracellular calcium with EGTA or intracellular calcium with BAPTA-AM. Interestingly, sulfinpyrazone was able to block both basal and glutamate-stimulated release, strongly indicating that the SVCT is bidirectional and that glutamate is inducing the transporter to work preferentially in an outward direction, promoting [14C] ascorbate efflux. Glutamate-stimulated release of vitamin C was demonstrated in many cell types but the mechanism is not totally clear. One possibility is that a putative glutamate-ascorbate exchanger (Rebec and Pierce 1994; Wilson 2005) promotes an influx of glutamate and an efflux of ascorbate or vice-versa. Our present experiments seem to rule out this possibility, since ascorbate is unable to promote [3H] d-aspartate efflux even in concentrations as high as 500 μM. One possible alternative would be that the incubation with glutamate and the uptake of this amino acid induces cell swelling and activation of volume-sensitive anion channels able to release ascorbate or metabolites (Wilson et al. 2000). However, our data show that blocking glutamate transporters with T-BOA had no effect on glutamate-stimulated vitamin C release, indicating that this mechanism is not involved. On the other hand, both hypotheses (hetero-exchange and cell swelling induced by glutamate uptake) are made less likely, at least in our culture system, by the fact that glutamate ionotropic agonists such as NMDA or kainate were also able to promote vitamin C release in these cultures. Each of these agonists was able to induce a release that could be blocked by their respective antagonists MK-801 or DNQX, indicating that the effects were respectively mediated by activation of NMDA or AMPA/kainate receptors. One possibility is that the activation of each class of glutamate receptors promotes neuronal depolarization and reversal of SVCT in an outward direction. This idea is corroborated by the fact that depolarization of cultures with veratridine also promoted a release of vitamin C that could be blocked by the voltage-sensitive sodium channel inhibitor tetrodotoxin (data not shown). γ-amino-butyric acid (GABA) transport appears to have a similar behavior, and reversal of GABA transport induced by depolarization was previously reported (Yazulla and Kleinschmidt 1983). A transporter-mediated release of GABA in retinal cultures was also shown to be strictly sodium-dependent, induced by veratridine and blocked by tetrodotoxin (do Nascimento and de Mello 1985; do Nascimento et al. 1998). In that study, as well as in the present work, the release was blocked when sodium ions were substituted by lithium ions, indicating that the release is dependent on sodium influx and not on depolarization, since lithium ions are able to substitute for sodium in action potential events but not transporter-mediated events. The mechanisms involved in reversal of transport activity are not known, but we favor the idea that changes in local sodium concentrations are able to control the transport direction. A plausible mechanism to explain this phenomenon was suggested by Godoy et al. (2007). Under normal conditions, sodium ions are present at low concentrations in the intracellular medium resulting in a high Km (low affinity) of ascorbate for interaction with SVCT. On the other hand, when the cell is depolarized, an increase of intracellular sodium concentration is observed in the microdomains where SVCT is present, lowering the Km for ascorbate and increasing the affinity for interaction. Further studies are necessary to test this hypothesis.
The effect of NMDA is dependent on the release of endogenous glutamate: a hypothesis
As discussed above, activation of glutamate ionotropic receptors induces an increase of vitamin C release in the cultures. Glutamate produces a similar increase of release, but this effect is only blocked by the non-NMDA antagonist DNQX and not by the NMDA receptor channel blocker MK-801 or the competitive NMDA antagonist APV. The release stimulated by kainate was blocked by DNQX but not by MK-801. However, the release stimulated by NMDA was also completely blocked by DNQX. One possible explanation is that the activation of NMDA receptors promotes the release of glutamate from a glutamatergic neuron that, in turn, activates non-NMDA receptors in another cell and induces the release of vitamin C. The release of endogenous glutamate by kainate was previously demonstrated (Nduaka et al. 1999; Hashimoto et al. 2000). An alternative hypothesis is that NMDA is inducing the release of endogenous glutamate and this amino acid is promoting the release of vitamin C by a receptor-independent mechanism. However, the stimulation by glutamate or NMDA is completely blocked by DNQX, confirming that the stimulation of vitamin C release by both agents is mediated by activation of AMPA/kainate receptors. Moreover, our results also show that the effect of glutamate is not dependent on glutamate uptake. Based on these data, a model for a retinal circuit involved in vitamin C release in retinal cultures would take into account: i) one first cell type in culture is glutamatergic and expresses NMDA receptors which, when activated, are able to promote the release of glutamate; ii) a second cell type expresses AMPA/kainate receptors and SVCT-2, and consequently it is able to take up and release ascorbate; iii) glutamate released from the first cell after NMDA stimulation activates AMPA/kainate receptors in the second cell type and promotes the release of ascorbate. The existence of neurons expressing only AMPA or kainate, but not NMDA receptors, is controversial. One possible candidate for such cell in the CNS is the glial cell. However, we have shown in the present work that SVCT-2 is not expressed in retinal glial cells in culture.
In conclusion, we have demonstrated a high-affinity and bidirectional transport system for ascorbate in cultures of chick retinal cells. The release of ascorbate was stimulated by glutamate in a receptor-dependent manner, raising the possibility that glutamate triggers important homeostatic and protective mechanisms by regulating the release of this antioxidant.
We greatly acknowledge Dr. Edward Ziff for his critical review of the manuscript and Ms. Luzeli R. de Assis for the technical assistance. This work was supported by grants from CNPq, CAPES, FAPERJ and PRONEX/MCT. CCP, VSM and RAM were recipients of undergraduate and graduate student fellowships from CNPq and CAPES, and RPC is a research fellow from CNPq.