SEARCH

SEARCH BY CITATION

Keywords:

  • ascorbic acid;
  • glucose;
  • glucose transporters;
  • glutamate;
  • lactate;
  • sodium–vitamin C cotransporter 2

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

It has been demonstrated that glutamatergic activity induces ascorbic acid (AA) depletion in astrocytes. Additionally, different data indicate that AA may inhibit glucose accumulation in primary cultures of rat hippocampal neurons. Thus, our hypothesis postulates that AA released from the astrocytes during glutamatergic synaptic activity may inhibit glucose uptake by neurons. We observed that cultured neurons express the sodium-vitamin C cotransporter 2 and the facilitative glucose transporters (GLUT) 1 and 3, however, in hippocampal brain slices GLUT3 was the main transporter detected. Functional activity of GLUTs was confirmed by means of kinetic analysis using 2-deoxy-d-glucose. Therefore, we showed that AA, once accumulated inside the cell, inhibits glucose transport in both cortical and hippocampal neurons in culture. Additionally, we showed that astrocytes are not affected by AA. Using hippocampal slices, we observed that upon blockade of monocarboxylate utilization by α-cyano-4-hydroxycinnamate and after glucose deprivation, glucose could rescue neuronal response to electrical stimulation only if AA uptake is prevented. Finally, using a transwell system of separated neuronal and astrocytic cultures, we observed that glutamate can reduce glucose transport in neurons only in presence of AA-loaded astrocytes, suggesting the essential role of astrocyte-released AA in this effect.

Abbreviations used
2-DOG

deoxyglucose

4-CIN

α-cyano-4-hydroxycinnamate

AA

ascorbic acid

ACSF

artificial CSF

DTT

dithiothreitol

EPSC

excitatory post-synaptic current

GFAP

glial fibrillary acidic protein

MAP2

microtubule-associated protein-2

GLUT

glucose transporter

SVCT

sodium–vitamin C transporter

In situ hybridization and immunohistochemical analyses have demonstrated a high sodium–vitamin C cotransporter 2 (SVCT2) expression in neurons from different regions of the brain (Tsukaguchi et al. 1999; Astuya et al. 2005; Garcia et al. 2005; Mun et al. 2006). SVCT2 is not expressed in astrocytes; however, it has been postulated that SVCT2 is induced after ischemia or in culture conditions (Berger et al. 2003; Astuya et al. 2005; Garcia et al. 2005). Kinetic analyses have shown that ascorbic acid (AA), the reduced form of vitamin C, is taken up into neurons using SVCT2 (Castro et al. 2001). Additionally, functional data indicate that astrocytes transport mainly dehydroascorbic acid, the oxidized form of vitamin C, using the glucose transporter (GLUT1) (Korkok et al. 2003; Astuya et al. 2005). Thus, astrocytes concentrate vitamin C (after reduction to AA) at extremely high concentrations. Regarding these physiological characteristics, it has been postulated that astrocytes may be involved in vitamin C recycling in the nervous system, avoiding the hydrolysis of dehydroascorbic acid produced by antioxidative protection (Hediger 2002).

It has been shown that extracellular concentration of AA increases during cerebral activity (O’Neil et al. 1982; Boutelle et al. 1989; Pierce and Rebec 1990) and that AA released from astrocytes is closely related to the activity of glutamatergic neurons (O’Neill et al. 1984;Basse-Tomusk and Rebec 1991;Ghasemzadeh et al. 1991;Rebec and Pierce 1994). In fact, AA is able to modulate glutamatergic transmission (Cammack et al. 1991; Rebec et al. 2005) and glutamate triggers the release of AA in primary cultures of rat cerebral astrocytes (Wilson et al. 2000). On the other hand, it has been postulated that AA inhibits glucose accumulation in primary cultures of rat hippocampal neurons (Patel et al. 2001). Furthermore, under circumstances of high extracellular AA concentrations, inhibition of glucose uptake by neurons may adaptively facilitate the uptake of lactate instead. Different studies have indicated that glutamate released from neurons stimulates aerobic glycolysis and lactate exportation in astrocytes (Pellerin and Magistretti 1994; Pellerin et al. 1998, 2005; Schurr et al. 1999). However, different studies have indicated that neurons use ambient glucose, and not glial-derived lactate, as the major substrate during activity, thus the lactate utilization is a controversial issue at the present (Chih and Roberts 2003;Dienel and Cruz 2004).

In the present study, our hypothesis is that AA released from astrocytes inhibits glucose uptake in neurons during extended glutamatergic synaptic activity. We demonstrated the expression of SVCT2 and GLUT1 and 3 in cultured neurons, however, in brain sections the main GLUT detected in hippocampal CA1 neurons was GLUT3. Additionally, we found that intracellular AA is able to inhibit deoxyglucose (2-DOG) uptake in cortical and hippocampal neurons. Using hippocampal slices treated with α-cyano-4-hydroxycinnamate (4-CIN), we defined that glucose is able to support synaptic activity only when AA uptake is inhibited. Results from transwell system experiments, using astrocyte and neuron primary cultures, suggest that in presence of glutamate, AA is preferably released from astrocytes. Thus, during glutamatergic synaptic activity, neuronal 2-DOG uptake may be inhibited by AA released from astrocytes.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Neuronal and astrocyte primary cultures

Neurons and astrocytes were obtained from 17-day-old embryos and 3-day-old neonate Wistar rats, respectively. The forebrains were removed and neocortex was dissected. Tissue was digested with 0.12% trypsin (wt/vol; Gibco Co., Rockville, MD, USA) in phosphate buffer 0.1 mol/L (phosphate-buffered saline, pH 7.4, 320 mOsm) and triturated with a fire-polished Pasteur pipette. For neurons, cells were plated at 0.3 × 106 cells/cm2 in plates coated with poly-l-lysine (mol. wt > 350 kDa; Sigma-Aldrich Corporation, St Louis, MO, USA). After 20 min, floating cells were removed and attached cells cultured for 5 days in Neurobasal Medium (Gibco) supplemented with B27 (Gibco), 100 U/mL penicillin, streptomycin 100 μg/mL, 2.5 μg/mL amphotericin B, and 293 mg/mL l-glutamine (Nalgene, Rochester, NY, USA) (Castro et al. 2001). Astrocyte cultures were established from five brains plated in 10 culture dishes (100 × 15 mm) and grown in Minimal Essential Medium containing 10% fetal bovine serum (Hyclone, Logan, UT, USA) and 293 mg/mL l-glutamine (Garcia et al. 2003). After 3–4 weeks, the cultured cells were shaken for 12 h, the floating cells were removed and attached cells were cultured for an additional 2 weeks under the same conditions.

Immunocytochemistry and confocal microscopy

Cells were grown on coverslips coated with poly-l-lysine and fixed with Histochoice (Sigma-Aldrich). Additionally, Wistar rat brains (17-day old) were dissected and fixed directly by immersion in Bouin’s solution. Thick (40 μm) transverse sections were cut with a cryostat and the sections were processed free-floating. Tissue sections or cultured cells were incubated overnight with the following antibodies: anti-GLUT1 (1 : 300), anti-GLUT3 (1 : 100), anti-βIII-tubulin (1 : 500; Alpha Diagnostic International Inc., San Antonio, TX, USA), anti-SVCT2 (1 : 100; Santa Cruz Biotechnology, Santa Cruz, CA, USA), monoclonal anti-glial fibrillary acidic protein (GFAP) (1 : 200; Dako, Carpintera, CA, USA) or monoclonal anti-microtubule-associated proteins-2 (MAP2) (1 : 50; Chemicon International Inc., Temecula, CA, USA). The cells were washed and incubated with anti-rabbit or anti-goat IgG-Alexa 488 (1 : 300; Invitrogen, Carlsbad, CA, USA) and propidium iodide (1.7 μg/mL; Sigma-Aldrich) subsequently washed and mounted. For double-label immunofluorescence analysis, brain sections were incubated with donkey anti-rabbit-Cy2 (1 : 200; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) and donkey-anti-mouse-Cy3 (1 : 200; Jackson ImmunoResearch Laboratories Inc.). Each section was counter-stained with TOPRO-3 (Invitrogen) a DNA dye.

Reverse transcription-polymerase chain reaction

Total RNA from cells and tissues was isolated using Trizol (Invitrogen). For RT-PCR, 1 μg of RNA was incubated in 10 μL reaction volume containing 10 mmol/L Tris–HCl (pH 8.3), 50 mmol/L KCl, 5 mmol/L MgCl2, RNase inhibitor 20 U, 1 mmol/L dNTPs, 2.5 μmol/L of oligod(T) primers, and 50 U of MuLV reverse transcriptase (New England Biolabs, Ipswich, MA, USA) for 10 min at 23°C followed by 30 min at 42°C and 5 min at 94°C. Parallel reactions were performed in the absence of reverse transcriptase to control for the presence of contaminant DNA. For amplification, a cDNA aliquot in a volume of 12.5 μL containing 20 mmol/L Tris–HCl (pH 8.4), 50 mmol/L KCl, 1.6 mmol/L MgCl2, 0.4 mmol/L dNTPs, 0.04 U of Taq DNA polymerase (Gibco-BRL, Carlsbad, CA, USA), and 0.4 mmol/L primers was incubated at 94°C for 4 min, 94°C for 15–50 s, 55°C (PCR for SVCT2) or 60°C (PCR for GLUT1 and GLUT3) for 30–50 s, and 72°C for 30–135 s for 35 cycles. PCR products were separated by 1.2–1.5% agarose gel electrophoresis and visualized by staining with ethidium bromide. The following primers were used to analyze the expression of (i) GLUT1: sense 5′-CATGTATGTGGGGGAGGTGT-3′, antisense 5′ GACGAACAGCGACACCACAG-3′ (expected product 559 bp); (ii) GLUT3: sense 5′-GGGCATGATTGGCTCTTTTT-3′, antisense 5′-GGGCTGCGCTCTGTAGGATA-3′ (expected product 366 bp); (iii) SVCT2: sense 5′-ACGTTTGGATGCAGGTTACCC-3′, antisense 5′-GTATCCTGGCTGTCTGTTCA-3′ (expected product 517 bp).

Immunoblotting

Total protein extracts were obtained from 5 day in vitro rat cortical neurons by homogenizing the cells in buffer A (0.3 mmol/L sucrose, 3 mmol/L dithiothreitol (DTT), 1 mmol/L EDTA, 100 μg/mL phenylmethylsulfonyl fluoride, 2 μg/mL pepstatin A, 2 μg/mL leucopeptin, and 2 μg/mL aprotinin) and sonicated three to five times for 10 s at 4°C. Proteins were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (70 μg/lane) in a 10% (w/v) polyacrylamide gel, transferred to polyvinylidene difluoride membranes (0.45-μm pore; Amersham Pharmacia Biotech., Piscataway, NJ, USA) and probed with the affinity-purified antibodies: anti-GLUT1 (Alpha Diagnostics International), anti-GLUT3 (Alpha Diagnostic International), and anti-SVCT2 (Santa Cruz Biotechnology). Pre-absorbed antibodies incubated with the peptides used to elicit them were used as negative controls. The reaction was developed using the enhanced chemiluminescence western blotting analysis system (Amersham Biosciences, Pittsburgh, PA, USA).

Uptake analysis

Cultures were carefully selected under the microscope to ensure that only plates showing uniformly growing cells were used. In each experiment, cells from 12 wells were removed and used to estimate the average total cell number in each well. Uptake assays were performed in 400 μL of incubation buffer (15 mmol/L HEPES pH 7.4, 135 mmol/L NaCl, 5 mmol/L KCl, 1.8 mmol/L CaCl2, and 0.8 mmol/L MgCl2) containing 1–1.2 μCi of 2-deoxy-d-[1,2-(N)3H]glucose (26.2 Ci/mmol; Dupont NEN, Boston, MA, USA). Uptake was stopped by washing the cells with ice-cold incubation buffer containing 0.2 mmol/L HgCl2. Cells were dissolved in 200 μL of lysis buffer (10 mmol/L Tris–HCl pH 8.0 and 0.2% sodium dodecyl sulfate), and the incorporated radioactivity was measured by liquid scintillation spectrometry (Castro et al. 2001; Astuya et al. 2005). The Michaelis constant, Km and Vmax, were calculated using non-linear regression by Michaelis–Menten equation (hyperbola single rectangular with two parameters). To investigate how many components are involved in 2-DOG transport Lineweaver–Burk analysis were used.

In AA inhibition experiments, cells were pre-loaded with AA during 40 min at 37°C (loading media) in presence of 0.1–1 mmol/L DTT (intracellular AA assays) or co-incubated with AA and radioisotopes in presence of 0.1–1 mmol/L DTT (extracellular AA assays). All inhibition experiments were carried out under initial velocity conditions to discriminate between 2-DOG transport and phosphorylation. In inhibition experiments, statistical comparison between two or more groups of data was carried out using analysis of variance (anova, followed by Bonferroni post-test).

Hippocampal slice preparation

Hippocampal slices were obtained from Wistar rats (12 to 17-day old). Brain slices (400-μm thick) were cut with a Vibratome (Pelco 101, Series 1000; Vibratome, St Louis, MO, USA) and incubated for 1 h at 20°C in artificial CSF (ACSF; 124 mmol/L NaCl, 2.69 mmol/L KCl, 1.25 mmol/L KH2PO4, 2 mmol/L MgSO4, 26 mmol/L NaHCO3, 10 mmol/L glucose, and 2 mmol/L CaCl2, pH 7.3) gassed with 95% O2 and 5% CO2. Slices were transferred to a 2 mL chamber fixed to an inverted microscope stage Olympus (Tokyo, Japan) BX50WI and superfused continuously at 1 mL/min with gassed ACSF at 30°C (Araque et al. 2002).

Electrophysiology

Recordings from pyramidal neurons from CA1 hippocampal region were made using the whole cell configuration of the voltage clamp technique. A −70 mV holding potential was maintained during the whole experiment. Current was acquired at 10 Khz. A total of 5–7 MΩ, borosilicate glass electrodes were used. Pipette internal solution was 131 mmol/L potassium gluconate, 1 mmol/L EGTA, 1 mmol/L MgCl2, 2 mmol/L ATP-K2, 0.3 mmol/L GTP-Na, 6 mmol/L KCl, 1 mmol/L NaCl, and 5 mmol/L HEPES, pH 7.3, 290 Osm). Excitatory post-synaptic currents (EPSCs) were evoked by a tungsten electrode in the Schaffer collateral pathway. A total of 10 V stimulus pulses of 0.1–0.2 ms were applied every 200 ms. Recordings were obtained with an Axoclamp 2A amplifier (Axon Instruments, Foster City, CA, USA). PClamp 8 software (Axon Instruments) was used for stimuli generation, data display, acquisition, and storage (Araque et al. 2002). The drugs were superfused with gassed ACSF (extracellular substances) or put into the patch pipette (intracellular substances). Data represents the mean ± SD of EPSCs amplitude (in relative values).

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Neurons in culture express and use glucose and AA transporters

We observed a homogenous cellular population with neuronal characteristics in rat neuronal cultures (5 day in vitro) stained with anti-βIII-tubulin (Fig. 1a). Additionally, few GFAP-positive cells were observed (data not shown). Using immunofluorescence, RT-PCR and western blot analysis, we demonstrated that neurons in culture express GLUT1 and GLUT3 (Figs 1a and b). In our immunofluorescence analysis, GLUT3 staining was more intense than GLUT1, and in both cases the expression pattern occurred in neuronal somas and processes (Figs 1a and b). Confirming our previous reports (Castro et al. 2001), immunofluorescence and RT-PCR analysis demonstrated that our neuronal cultures express SVCT2 (Fig. 1c). Additionally, western immunoblotting studies showed a single band of 62 kDa in protein extracts isolated from cultured neurons (Fig. 1c).

image

Figure 1.  Glucose transporter (GLUT1 and GLUT3) and ascorbic acid [(sodium–vitamin C transporter (SVCT2)] transporters are expressed in cortical neurons. (a) Expression analysis of GLUT1. RT-PCR, lane 1: DNA 100-base pair (bp) standard, the numbers to the right indicate the base pairs; lanes 2 and 3: mRNA isolated from cultured neurons (2) and rat brain cortex (3); and lane 4: RT(−). Western blot, lane 1: proteins isolated from cultured neurons and lane 2: negative control. The numbers to the left indicate the kDa. Immunofluorescence, staining using anti-GLUT1 (1 : 100) and anti-βIII-tubulin (1 : 2000) antibodies (green). Nuclei were stained with propidium iodide (red). Scale bar: 50 μm. (b) Expression analysis of GLUT3. RT-PCR, lane 1: DNA 100-bp standard; lanes 2 and 3: mRNA isolated from neurons (2) and cultured Sertoli cells (3); and lane 4: RT(−). Western blot, lane 1: proteins isolated from cultured neurons and lane 2: negative control. The numbers to the left indicate the kDa. Immunofluorescence, Staining using anti-GLUT3 (1 : 100) antibody (green). Nuclei were stained with propidium iodide (red). Scale bar: 50 μm. (c) Expression analysis of ascorbic acid transporter SVCT2. RT-PCR, lane 1: DNA 100-bp standard; lanes 2, 3, and 4: mRNA isolated from rat brain (2) and cortical neurons (3 and 4); and lane 5: RT(−). Western blot, lane 1: proteins isolated from cultured neurons and lane 2: negative control. The numbers to the left indicate the kDa. Immunofluorescence, staining with anti-SVCT2 antibody (green). (a–c) Nuclei were stained with propidium iodide (red). Scale bars in (a–c): 50 μm.

Download figure to PowerPoint

We confirmed the function of GLUTs in cultured neurons by means of kinetic analysis using 2-DOG. The aim of these experiments was to establish experimental conditions that would allow transport inhibition assays and not accumulation inhibition assays. In time course studies for 2-DOG zero trans transport, the uptake was linear for the first 20 s and equilibrium was reached after approximately 30 s (Fig. 2a and inset). The initial velocity of 2-DOG uptake was 1177 ± 154 pmol/min × 106 cells. To determine the kinetic properties of the transporters involved in 2-DOG uptake (and not the glycolytic rate), we used a short uptake period of 15 s. Dose-response uptake studies revealed that 2-DOG uptake is saturated at approximately 5 mmol/L (Fig. 2b). The Michaelis–Menten analysis revealed an apparent Km of 1.4 ± 0.17 mmol/L and a Vmax of 1794 ± 79.8 pmol/min × 106 cells (R > 0.99). According to Lineweaver–Burk plot we observed only one functional component (Fig. 2c). Because glucose affinities of isoforms 1 and 3 are very similar (Maher et al. 1996; Vannucci et al. 1997), the observed Km may correspond to GLUT1 and GLUT3 function.

image

Figure 2.  Functional expression of glucose transporters in cultured cortical neurons. (a, b, and c) Deoxyglucose (2-DOG) transport analysis. (a) Time course uptake for 60 and 15 s (inset) of 0.5 mmol/L 2-DOG at 20°C. (b) Dose-response of 2-DOG transport using a 15 s uptake assay at 20°C. (c) Double-reciprocal plot of substrate dependence for 2-DOG transport. The data represent the mean ± SD of three experiments. The lines in plots correspond to non-linear regression (R > 0.99) from exponential single with two parameters (a), hyperbola single rectangular (b), and linear regression (c; R > 0.99).

Download figure to PowerPoint

Neuronal intracellular AA inhibits 2-DOG uptake in cortical and hippocampal neurons in culture. The inhibitory effect is not observed in astrocytes

Previously, we have described the functional activity of SVCT2 in cultured neurons (Castro et al. 2001). In this work, similar kinetic parameters were observed in cortical and hippocampal neurons (data not shown). Additionally, it has been observed that intracellular AA is readily lost when the external media is replaced with AA-free media (Makar et al. 1994). For that reason, AA was incubated with the cultured neurons for 40 min and then rapidly replaced with 2-DOG uptake solution. We observed that AA is able to inhibit 2-DOG uptake (15 s) at 20°C when cortical neurons were pre-loaded with different concentrations of AA (intracellular AA; Fig. 3a). On the other hand, no effect was observed when the cortical neurons were co-incubated for 15 s with AA and 2-DOG at 20°C or 4°C (extracellular AA, Fig. 3b). Similar evidence was obtained from assays performed on hippocampal neurons in culture (Figs 3c and d). On the other hand, neither intracellular nor extracellular AA affects 2-DOG uptake in cultured primary cortical astrocytes (Figs 3e and f).

image

Figure 3.  Intracellular ascorbic acid (AA) inhibits deoxyglucose (2-DOG) uptake in neurons but not in cultured astrocytes. (a and b) Semi-log plots for the inhibition of 2-DOG transport in cortical neurons at 4°C (○) and 20°C (•) by intracellular AA (a, cells were pre-loaded with AA at the indicated concentrations) or extracellular AA (b, cells co-incubated with AA and 2-DOG). In intracellular experiments, values >0.03 mmol/L at 20°C were statistically significantly from control at < 0.001. (c and d) Bars plot for the inhibition of 2-DOG transport in hippocampal neurons at 4°C (black bars) and 20°C (gray bars) by intracellular (c) or extracellular (d) AA. In intracellular experiments, 0.1 and 1 mmol/L values were statistically significantly from control at < 0.001. (e and f) Semi-log plots for the inhibition of 2-DOG uptake in cortical astrocytes at 4°C (○) and 20°C (•) by intracellular (e) or extracellular (f) AA. No statistically differences were observed. Every experiment was performed at 0.5 mmol/L 2-DOG using 15 s uptake assays. The absolute value of 2-DOG uptake in control (absence of AA) was 290.3 ± 19.1 (neurons) or 4527 ± 118 (astrocytes) at 20°C and 276.9 ± 3.3 (neurons) or 3942 ± 97 (astrocytes) pmol/min × 106 cells at 4°C. The data represent the mean ± SD of three experiments.

Download figure to PowerPoint

Neuronal intracellular AA impaired the capability of glucose to support glutamatergic synaptic activity

Recordings from pyramidal neurons from hippocampal CA1 region, in brain slices from suckling rats, were obtained using voltage clamp technique in whole cell configuration. EPSCs were evoked by stimulating the Schaffer collateral pathway. We observed that glucose deprivation produced a slow depression of synaptic responses. After deprivation, glucose was not able to restore EPSCs in the presence of external 200 μmol/L 4-CIN (Fig. 4a), a classical cellular and mitochondrial monocarboxylate transporter inhibitor. In this experimental condition, we inferred that endogenous AA accumulated inside the neuron, inhibited glucose uptake (Fig. 4a, schematic representation). Thus, in a similar deprivation experiment we observed that glucose was able to support synaptic responses only when AA uptake was inhibited in the recording neuron, by an intracellular anti-SVCT2 antibody (Fig. 4b and schematic representation). In this experimental situation, EPSC amplitude was restored to 60% of control by glucose re-administration in the presence of 4-CIN (Fig. 4b). Finally, we implemented a new experimental condition where voltage clamp technique in whole cell configuration was performed with intracellular anti-SVCT2 and using 100 μmol/L AA introduced inside the cell, in order to revert the protective effect generated by the anti-SVCT2 antibody. In this experiment, glucose was not able to support synaptic activity recovery after the deprivation period and extracellular 4-CIN (Fig. 4c and schematic representation). In conclusion, AA was able to inhibit the capability of glucose to support the recovery of synaptic activity after glucose deprivation and extracellular 4-CIN incubation. However, when AA uptake is inhibited, exogenous glucose is able to support the neuronal activity in the presence of 4-CIN.

image

Figure 4.  Effect of neuronal ascorbic acid (AA) uptake inhibition on the restoration of glucose-supported excitatory post-synaptic currents (EPSCs). (a, b, and c) Time course of EPSCs amplitude. Traces at the center are representative field EPSCs for a glucose-deprived slice. Calibration bar: 100 pA, 20 ms. Right side: Schematic representation of the different experimental conditions analyzed. (a) EPSCs after glucose deprivation in the presence of 200 μmol/L α-cyano-4-hydroxycinnamate (4-CIN). (b) EPSCs recovery after glucose deprivation in the presence of 200 μmol/L 4-CIN and intracellular anti- sodium–vitamin C transporter 2 (SVCT2; 1 : 2000 into intracellular solution). EPSCs are restored by administration of 10 mmol/L glucose. (c) EPSCs after glucose deprivation in the presence of 200 μmol/L 4-CIN, intracellular anti-SVCT2 (1 : 2000 into intracellular solution) and 100 μmol/L AA. Data represent mean values ± SD (n = 5–7). The antibody and 4-CIN were incorporated after 7 min of glucose incubation and maintained for approximately 75 min.

Download figure to PowerPoint

To define if glucose uptake inhibition by AA is mainly associated to the expression of GLUT3 and/or GLUT1 in CA1 hippocampal neurons, we performed double-label immunofluorescence analyses and brain tissue sections were counter-stained by TOPRO-3 (blue). Intense GLUT3 immunoreaction was observed associated to MAP2-positive dendritic projections (Figs 5a–d). Using high magnification we confirmed that GLUT3 immunoreaction is mainly associated to the neuropil and was absent from neuronal cell bodies (Figs 5e–h). We also analyzed the expression of GLUT1 in hippocampus. Intense GLUT1-positive reaction was observed in endothelial cells of the brain vessels, however, low immunoreaction was associated to the neuropil (Figs 5i–l). GLUT1 was not detected in CA1 pyramidal cell bodies. Finally, in the same brain zone we analyzed GLUT1 expression in GFAP-positive astrocytes. Most of GLUT1-positive neuropil reaction was not associated to astrocytes. In these cells, we only observed focal immunopositive regions that showed colocalization with GFAP (Figs 5m–p, inset). Thus, our immunohistochemical results showed that GLUT3 is the main transporter expressed in CA1 pyramidal neurons.

image

Figure 5.  Glucose transporter 3 (GLUT3) is mainly localized in CA1 pyramidal cells processes. Brain transverse sections (40 μm) were cut with a cryostat and processed for free-floating immunohistochemistry. The sections were incubated overnight with different combinations of the following antibodies: anti-GLUT1 (1 : 300), anti-GLUT3 (1 : 100), monoclonal anti-GFAP (1 : 200), or monoclonal anti-MAP2 (1 : 50). For double-label immunofluorescence, the brain sections were incubated with donkey anti-rabbit-Cy2 and donkey-anti-mouse-Cy3 (1 : 200). Each section was counter-stained with TOPRO-3 a DNA dye. (a–p) Hippocampal CA1 region. (e–h) Higher magnifications of the area showed in (a). Inset in (n–p), higher magnification of astrocyte GLUT1 and GFAP positive. V: blood vessel GLUT1 positive. Scale bar in (a–p): 20 μm.

Download figure to PowerPoint

AA from glutamate-stimulated astrocyte inhibits neuronal 2-DOG uptake

To observe that astrocytic AA is able to inhibit neuronal glucose uptake, we carried out experiments using a transwell system. Astrocytes and neurons are plated in different compartments, but share the same media (Fig. 6a). Prior to the experiment, only the astrocytes were loaded with AA. Neuronal 2-DOG uptake was inhibited when neurons and pre-loaded, but not unloaded, astrocytes were incubated with l-glutamate. After 60 min treatment with l-glutamate, we observed approximately 70% inhibition in 2-DOG transport (Fig. 6b). In unloaded astrocyte experiments neurons were co-incubated with l-glutamate and no effect was observed (Fig. 6b). Additionally, we studied the effect of l-glutamate on 2-DOG uptake, showing that different concentrations of l-glutamate does not inhibit 2-DOG uptake (Fig. 6c) in cultured cortical neurons. Therefore, l-glutamate is not a neuronal glucose transport inhibitor, but l-glutamate is a necessary factor for neuronal glucose inhibition. Every condition was compared with l-glutamate-free experiments (black bars, Fig. 6b). We observed approximately 20% inhibition in absence of l-glutamate after 60 min of incubation. Probably, spontaneous AA depletion from astrocytes may occur.

image

Figure 6.  Effect of astrocytic ascorbic acid (AA) in neuronal deoxyglucose (2-DOG) uptake. (a) Astrocytes were plated in transwell inserts and co-cultured with neurons during 48 h. Previous to the experiment, only astrocytes were pre-loaded with 1 mmol/L AA for 60 min. Astrocytes (loaded and washed) and neurons were co-incubated in presence of 0.25 mmol/L l-glutamate during 15, 30, 45, and 60 min. Measurements of 2-DOG uptake (15 s uptake assay, 20°C) were performed in neurons treated with l-glutamate. (b) Bar plot of 2-DOG uptake inhibition. 2-DOG uptake in neurons was inhibited when the cultures were treated with l-glutamate 0.25 mmol/L and co-incubated with astrocytes pre-loaded with AA. In presence of l-glutamate, values from 30, 45, and 60 min of incubation were statistically different from control at < 0.001. When the astrocytes were not pre-loaded, no effect was observed. (c) Semi-log plot for the inhibition of 2-DOG transport by l-glutamate. The data represent the mean ± SD of three experiments.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In this study, we have determined that neuronal glucose transport is inhibited by AA accumulated inside the cells. Additionally, the results suggest that AA may be released from astrocytes stimulated by extracellular glutamate. GLUT3, the predominant GLUT in neurons (Aller et al. 1997; Leino et al. 1997; Vannucci et al. 1997; Nualart et al. 1999; Godoy et al. 2006), may be directly associated to the selective AA effect in these cells (Fig. 7).

image

Figure 7.  Ascorbic acid (AA) is able to modulate neuronal glucose transport during glutamatergic neurotransmission. Neurons use glucose to support their synaptic activity only when AA is not intracellularly concentrated (1). AA could be released from nearby astrocytes in response to glutamate stimuli (2). Thus, this modulation would be driven by glutamate and astrocytes surrounding the synaptic cleft. The neuronal effect may be potentiated regarding that glucose transporter (GLUT3; the neuronal GLUT) transports glucose seven times faster than GLUT1, the main glial GLUT (Vannucci et al. 1997). Astrocytes are involved in the recycling of vitamin C. It has been postulated that intracellularly generated AA is released after glutamate stimulation.

Download figure to PowerPoint

Different data indicate that AA uptake in neurons is a sodium-dependent mechanism. Thus, using RT-PCR, western blot, and immunocytochemistry, we have confirmed the SVCT2 expression in neurons isolated from the brain cortex. Similar results have been shown in brain tissue by immunohistochemistry and in situ hybridization analyses (Tsukaguchi et al. 1999; Astuya et al. 2005; Garcia et al. 2005;Mun et al. 2006). These data indicate that SVCT2 is the main AA transporter expressed by neurons and therefore plays a central role in providing these cells with reduced vitamin C (Fig. 7). Previously, we have studied the kinetic parameters associated to the AA uptake in neurons (Castro et al. 2001), and surprisingly, we observed the presence of two affinity constants, 103 and 8 μmol/L, suggesting a regulatory mechanism associated with the uptake of AA in these cells. Recently, an apparent Km of 113 μmol/L for ascorbate has been established in neuroblastoma cells (May et al. 2006), which is different from the one observed in SVCT2-expressing Xenopus oocytes (Tsukaguchi et al. 1999) but is similar to the higher Km observed in our previous study in cultured neurons (Castro et al. 2001). These data indicate that SVCT2 transporter (apparent Km > 100 μmol/L in neurons) should be responsible for the transport and accumulation of near millimolar concentration of AA inside these cells.

Our experiments to analyze the inhibitory effect of AA on neuronal glucose uptake were performed over a few seconds (15 s). This time was determined in time-course experiments to assure initial velocity conditions. Under these conditions, we observed that intracellular AA is able to inhibit 2-DOG uptake in cultured neurons. AA oxidation was not observed until 60 min of incubation in presence of DTT at 37°C (HPLC determination, data not shown), therefore these results confirm that reduced vitamin C (AA), and not oxidized vitamin C (dehydroascorbic acid, transported by GLUTs) (Nualart et al. 2003; Astuya et al. 2005) is the inhibiting agent in neurons. Finally, we observed that the inhibition of 2-DOG transport by intracellular AA was temperature dependent; however, additional experiments are required to clarify this point.

Patel et al. (2001) observed glucose uptake inhibition in presence of extracellular AA, however, they performed accumulation experiments using incubation times of 30 min, which allows neuronal uptake of AA. In agreement with this, we observed glucose uptake inhibition in cultured neurons pre-loaded with AA during 20 min (data not shown). Moreover, the substrate concentration used by Patel was three orders of magnitude less than the Km for 2-DOG transport. Patel et al. (2001) suppose that AA may directly inhibit the GLUT by reducing sulfhydryl groups in the transporter due to AA being an electron donor. But this supposition is questionable because thermodynamically AA cannot reduce an oxidized thiol.

To directly investigate if an AA effect takes place under synaptically active conditions, we carried out experiments using rat hippocampal slices. In agreement with previous reports using extracellular recordings (Izumi et al. 1997), in presence of a MCT inhibitor (4-CIN, a cellular and mitochondrial monocarboxylate transporter inhibitor), we observed that glucose is not able to support the EPSCs recovery after the deprivation period. However, when AA uptake was inhibited in the recording neuron, by using an anti-SVCT2 antibody, we observed EPSCs recovery after deprivation period. Thus, AA entry to neurons under synaptic activity is able to modulate the use of glucose in neural cells. We were able to observe the EPSCs recovery by using an anti-SVCT2 antibody when loading AA directly into the recording cell. Therefore, the anti-SVCT2 antibody used in the present experiments was able to inhibit SVCT2 function from the intracellular site. In summary, when AA uptake and monocarboxylate uptake in recording neuron is inhibited, glucose is able to support the recovery of synaptic activity. Only when monocarboxylate transport is inhibited, glucose is not able to support the neural activity recovery. Additionally, as it was early observed by Cox and Bachelard (1988), our data indicate that the metabolism of glucose itself, not lactate or pyruvate, is necessary to sustain neuronal function under the conditions studied.

It has been proposed that glutamate directly inhibits hexose transport in neurons co-cultivated with astrocytes (Porras et al. 2004); however, we were unable to detect this glutamate effect. Perhaps the phenomenon being observed actually happens subsequently to glutamate activated release of AA from the astrocytes present in co-cultured cells. In our cultured cortical neurons, glutamate did not affect 2-DOG transport. In the same way, in transwell experiments neuronal 2-DOG transport was not inhibited in the presence of glutamate and unloaded astrocytes. Furthermore, based on our electrophysiology experiments in which AA by itself was able to inhibit the recovery of observed current it is most likely that AA and not glutamate explains the observed results. In conclusion, after glucose deprivation, glucose could rescue the neuronal response to electrical stimulation only if AA uptake is prevented.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We gratefully acknowledge the technical assistance of G. Rerhen. We also would like to acknowledge the helpful suggestions of S. Brauchi and E. Martin. MC is recipient of a CONICYT doctoral fellowship, Chile. This work was supported in part by Chilean grants FONDECYT 1050095, CONICYT-PBCT ACT02, DID-UACh 2004060, CONICYT 403124, MECESUP AUS0006, and Fondecyt 1060135.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References