SEARCH

SEARCH BY CITATION

Keywords:

  • ascorbate;
  • astrocytes;
  • astrocytosis;
  • ischemia;
  • oxidative stress;
  • SVCT2

Abstract

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

The sodium-vitamin C co-transporter SVCT2 is primarily responsible for the accumulation of the important antioxidant ascorbate into brain cells. In vitro studies have demonstrated strong expression of this transporter in cultured astrocytes, whereas in situ hybridization analysis has so far detected SVCT2 only in neurons. In the present study, we examined the response of SVCT2 mRNA expression in the brain to focal ischemia induced for 2 h by unilateral middle cerebral artery occlusion. The mRNA expression patterns of SVCT2 and the glutamate-activated immediate early gene Arc were investigated at 2 and 22 h after ischemia. SVCT2 and Arc mRNA expression was lost in the ischemic core at both time points. In areas outside the core, Arc was strongly up-regulated, primarily at 2 h, whereas SVCT2 showed an increase at 2 and 22 h. SVCT2 expression was increased in neurons as well as in astrocytes, providing the first evidence for SVCT2 expression in astrocytes in situ. These findings underscore the importance of ascorbate as a neuroprotective agent and may have implications for therapeutic strategies. In addition, the increase of SVCT2 in astrocytes after ischemia suggests that cultured astrocytes are exposed to chronic oxidative stress.

Abbreviations used
BCIP/NBT

5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium

DIG

digoxigenin

MCAO

middle cerebral artery occlusion

SVCT2

sodium-vitamin C transporter 2

Ascorbate (reduced vitamin C) is an important antioxidant, enzyme co-factor and neuromodulator in the brain (Grunewald 1993; Rebec and Pierce 1994; Rice 2000). Stored at millimolar concentrations in brain cells, ascorbate is released into the extracellular compartment upon neuronal stimulation, and protects against oxidative radicals generated by neural activity. Until recently, the identity of the ascorbate-storing cells in the brain was unknown because of two differing lines of evidence. Rice and colleagues have argued that more ascorbate is stored in neurons because the ascorbate content in developing and adult brain correlates with neuron density and not with glial cell density (Rice and Russo-Menna 1998). In contrast, other studies have suggested that the majority of ascorbate is stored in astrocytes because cultured astrocytes can accumulate ascorbate up to an intracellular concentration of 7–10 mm (Siushansian and Wilson 1995; Wilson 1997). We have recently cloned the sodium-coupled vitamin C co-transporter SVCT2, which apparently is responsible for the ascorbate uptake from the blood into brain CSF and into brain cells (Tsukaguchi et al. 1999). We have shown, using in situ hybridization, that SVCT2 mRNA expression in brain is detectable in choroid plexus, neurons and meninges, but not in astrocytes. These findings are consistent with the concept that the majority of ascorbate is stored in neurons and not in astrocytes in normal brain.

Interestingly, using northern blotting, we, and Wilson and colleagues, have found that SVCT2 mRNA expression can also be demonstrated in cultured astrocytes (Berger and Hediger 2000; Korcok et al. 2000), suggesting that SVCT2 expression may be up-regulated in astrocytes under culturing conditions. The nature of the stimulus that induces the SVCT2 expression in astrocytes in culture is unknown, but may be related to the fact that they are likely exposed to a different degree of oxidative stress than astrocytes in situ. Thus, we hypothesized that astrocytes in situ may only express ascorbate transport activity under stimulating conditions, such as during gliosis, or after ischemia. To test this hypothesis, we have previously examined SVCT2 mRNA expression in gliotic astrocytes in the penumbra of an excitotoxic brain lesion, but we have found no evidence for astrocytic SVCT2 expression under those conditions (Berger and Hediger 2000). In the present study, we have examined SVCT2 expression in rat brains subjected to 2 h of focal ischemia, a stimulus well known to cause marked release of ascorbate into the extracellular compartment (Hillered et al. 1988; Yusa 2001). SVCT2 expression was analyzed using in situ hybridization at 2 h of reperfusion, when generation of oxidative radicals should be well established, and after 22 h of reperfusion, when long-term damage is more fully developed. The SVCT2 expression was analyzed in the ischemic core, where blood flow and energy supplies are strongest affected by the occlusion, in adjacent penumbral regions, and in dentate gyrus and cingulate cortex, two peri-infarct regions shown in previous studies to exhibit changes in immediate early gene expression as a result of spreading depression (Hata et al. 2000). As a measure of glutamatergic stimulation, adjacent sections were hybridized with a probe for the immediate early gene Arc which is activated by glutamatergic receptor activation (Kunizuka et al. 1999). The results demonstrate that astrocytes in situ can indeed induce expression of SVCT2 mRNA following an ischemic injury. Moreover, we report that neurons in the penumbra and in peri-infarct areas also up-regulate their SVCT2 expression following ischemia.

Middle cerebral artery occlusion (MCAO) was performed on male Sprague-Dawley rats (260–270 g, Charles River Laboratories Inc., Wilmington, MA, USA) at Guilford Pharmaceuticals Inc., Baltimore, MD, using the intraluminal filament technique (Longa et al. 1989; Britton et al. 1997). Briefly, the rat was anesthetized with 1.5% halothane and its rectal temperature was regulated at 37.0–37.5°C using a heating blanket during the surgery. The right common carotid artery was exposed at the level of external and internal carotid artery bifurcation. The external carotid artery and its occipital branch were cauterized and cut. A piece of 3–0 monofilament nylon suture with a blunted tip and 0.1% poly l-lysine coating was advanced into the internal carotid artery via the proximal end of the external carotid artery stump to block the origin of the middle cerebral artery. Each rat was allowed to emerge from anesthesia during MCAO. Then, 2 h later, the rats were re-anesthetized for removal of the intraluminal suture to allow reperfusion. Right before reperfusion, the rat's body temperature was measured and contralateral circling behavior was recorded. These scores were used for screening non-qualified ischemia. During reperfusion, the animals were kept without anesthesia or sedation in their home cage at room temperature (25°C). No warming blanket was used during reperfusion, as the animals were mildly hyperthermic. At 2 and 22 h of reperfusion, brains were collected and immediately frozen. A total of seven ischemic and two sham-operated animals were examined for each time point, as well as six un-operated controls. For sham operation, all procedures were identical except for the application of the intraluminal suture. All procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals established by the National Institute of Health in the United States.

Brains were sectioned in the coronal plane using a Leica cryostat. Series of 14-μm thick sections were collected onto Superfrost plus microscope slides in intervals of approximately 330 μm from bregma levels 2.2 mm rostrally to − 4.1 mm caudally. The slides were stored at − 80°C for up to 2 weeks before processing.

Non-radioactive in situ hybridization was performed using digoxigenin (DIG)-labeled cRNA probes for SVCT2 and Arc as described previously (Tsukaguchi et al. 1999). Both the SVCT2 and Arc (Lyford et al. 1995) probes contained about 3 kb of sequence, and were alkali-hydrolyzed to an average length of 200–400 bases. The brain sections were fixed and acetylated, and then hybridized at 68°C overnight, or over three nights, to the probes (approximate concentrations 100 ng/mL). Hybridized probe was visualized using alkaline phosphatase-conjugated anti-DIG Fab fragments (Roche, Indianapolis, IN, USA) and 5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium (BCIP/NBT) substrate (Kierkegard and Perry Laboratories, Gaithersburg, MD, USA). Sections were rinsed several times in 100 mm Tris, 150 mm NaCl, 20 mm EDTA pH 9.5, and coverslipped with glycerol gelatin (Sigma, St Louis, MO, USA). Control sections were incubated in an identical concentration of the sense probe transcript.

Identification of cell types

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

Some sections were double-stained with an antiserum against the astrocyte marker GFAP (DAKO), or they were processed for double in situ hybridization using the DIG-labeled SVCT2 probe and an FITC-labeled probe against either the astrocytic glutamate transporter GLT-1, or the microglial marker MCSF (macrophage colony stimulating factor) (Borycki et al. 1993) as described (Berger et al. 1998). Briefly, for double in situ hybridization, the DIG-labeled probe was visualized at the bright-field level using alkaline phosphatase labeling, and the FITC-labeled probe was detected at the fluorescent level, using in sequence, mouse anti-FITC antibodies, biotinylated antimouse antibodies, horse radish peroxidase-linked streptavidin, the tyramide signal amplification using biotinylated tyramide, and CY3-conjugated streptavidin (Berger et al. 1998).

Digital pictures were taken with a Nikon E-600 microscope and a SPOT digital camera. Composites were prepared using Adobe Photoshop software. Where signal intensities were compared between ischemic and control brains, pictures were taken with the same exposure settings, and the picture contrast was subsequently identically enhanced. To allow for direct comparison of signal intensity, slides were incubated in identical concentrations of probe and were processed identically. Signal intensities were measured using NIH Image software, and analyzed statistically using Student's t-tests.

Results

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

All rats subjected to MCAO had an increased body temperature of 38.9°C or higher and an involuntary circling behavior to the contralateral side, which indicated that successful ischemia had been achieved. The extent of the ischemic area, as assessed by the changes seen in mRNA labeling, was similar to previous experiments (Slusher et al. 1999), covering most of striatum and lateral and ventral cerebral cortex. Ischemic areas extended from the frontal pole of the striatum to dorso-lateral areas of hypothalamus with some interanimal variability.

Figure 1 shows examples of a low power view of SVCT2 mRNA labeling in coronal sections through striatum at 2 and 22 h after ischemia. At 2 h (Fig. 1a), damage in the form of lost labeling is mostly evident in ipsilateral striatum and ventral forebrain while the lateral cortex still shows distinguishable, though reduced, cellular labeling. At 22 h (Fig. 2b), the labeling in cortex has also disappeared throughout most of the dorso-ventral axis. A clear demarcation line is visible between ischemic core areas where SVCT2 labeling is absent, and the adjacent areas where SVCT2 labeling appears more intense. SVCT2 labeling in sham-operated rats was similar to non-operated controls (not shown at this magnification, see below). Adjacent sections hybridized with the Arc probe showed pronounced increases at 2 h in ipsilateral cingulate cortex and penumbral areas as described previously (Kunizuka et al. 1999) (see below). This up-regulation of Arc expression had mostly subsided at 22 h after reperfusion.

image

Figure 1. Photomicrographs of SVCT2 mRNA expression in coronal rat brain sections in ischemic brains, as detected by non-isotopic in situ hybridization. MCAO was induced unilaterally for 2 h, and brains were examined at 2 h (a) and 22 h of reperfusion (b). (a) SVCT2 mRNA labeling is absent in the ischemic core in the striatum (asterisks) and the ventral pre-optic region. The labeling in overlying dorso-lateral cortex is also reduced in its intensity. (b) At 22 h the areas of reduced SVCT2 labeling extend throughout most of the ipsilateral hemisphere, with the exception of cingulate cortex. Note that the signal intensity of the labeling in penumbral areas and in cingulate cortex is increased. Bar = 200 μm.

Download figure to PowerPoint

image

Figure 2. mRNA expression for SVCT2 (a–c, g–i) and the immediate early gene Arc (d–f, j–l) in striatum (a–f) and cingulate cortex (g–l) in unoperated control animals (a, d, g, j), in ischemic animals at 2 h of reperfusion (b, e, h, k), and in sham-operated animals at 2 h (c, f, i, l). Both SVCT2 and Arc labeling are greatly reduced in the ischemic core area in striatum (asterisks in b and e). In the adjacent penumbra, Arc labeling is markedly elevated, reflecting glutamate release and glutamatergic receptor activation, whereas SVCT2 labeling appears not noticeably affected. Similarly, in cingulate cortex, Arc expression (k) is greatly increased at 2 h while SVCT2 expression (h) appears similar to unoperated controls (g). Sham-operation did not have pronounced effects on either Arc or SVCT2 mRNA expression in either region. Bar = 100 μm.

Download figure to PowerPoint

The SVCT2 mRNA-positive cells in control brain are primarily neurons, as we have shown in our earlier studies (Fig. 2). No labeling was observed using sense probe (not shown). At 2 h of reperfusion, the SVCT2 mRNA expression in the ischemic core was markedly decreased compared with non-ischemic and sham-operated controls (Figs 2a–c). The SVCT2 labeling immediately adjacent to the ischemic core in the striatum was not noticeably changed at 2 h (Fig. 2b). In contrast, the immediate early gene Arc was markedly up-regulated in the same penumbral area in an adjacent section (Fig. 2e). Ipsilateral cingulate cortex showed a similar response. The SVCT2 mRNA labeling was not noticeably different in ischemic animals compared with either unoperated or sham-operated controls (Figs 2g–i), whereas the Arc mRNA was strongly up-regulated (Fig. 2k). Arc expression in sham-operated animals was also similar to unoperated controls (Figs 2f and l).

In the dentate gyrus, Arc mRNA labeling in the granule cells was strongly up-regulated at 2 h of reperfusion, showing the characteristic expression in the dendrites of the molecular layer (Fig. 3d), which has been described previously (Kunizuka et al. 1999). The SVCT2 labeling in granule cells in an adjacent section (Fig. 3e) also appeared more intense at the early time-point (see below for quantitation). At higher magnification (Fig. 3f), no SVCT2 expression was evident in the granule cells dendrites, though faintly stained scattered cells were detectable in the molecular layer. At 22 h, the Arc expression in dentate gyrus had subsided (Fig. 3g). In contrast, a marked increase in the intensity of the SVCT2-labeled cells in the dentate gyrus molecular layer was apparent at 22 h (Fig. 3h). The distribution pattern of these SVCT2-positive cells, and their appearance at higher magnification (Fig. 3i), suggests that they are astrocytes (see below for double-labeling experiments). Up-regulation of neuronal SVCT2 mRNA expression in granule cells had mostly subsided at 22 h (Fig. 3h). Sham-operation had only mild effects on the appearance of SVCT2 or Arc labeling at 2 h (not shown) or 22 h (Figs 3j–l).

image

Figure 3. SVCT2 and Arc mRNA expression in dentate gyrus in unoperated controls (a–c), ischemic animals at 2 h of reperfusion (d–f), ischemic animals at 22 h of reperfusion (g–i), and in sham-operated animals at 22 h (j–l). At 2 h, Arc expression in granule cells (arrows) is greatly increased (d) compared with control (a), reflecting stimulation via perforant path glutamatergic projections. SVCT2 mRNA expression in these granule cells is also increased at 2 h, though to a much smaller degree (e). At 22 h, the up-regulation of Arc expression in granule cells has subsided (g) and the SVCT2 expression in these cells (h) also appears more like controls. However, SVCT2 labeling in the molecular layers of dentate gyrus has greatly increased at 22 h (arrow in h). The molecular layers contain few neurons, suggesting up-regulation of SVCT2 mRNA expression in astrocytes. (c, f, i, l) Higher magnification views of SVCT2 expression in the molecular layer. At 2 h, a few cells are detectable which are only faintly labeled (arrows). In contrast, at 22 h numerous strongly labeled cells are evident (i). (j–l) Sham controls. Bars: k = 400 μm, l = 40 μm.

Download figure to PowerPoint

As shown in Fig. 1, and at higher magnification in Fig. 4(a), SVCT2 labeling in penumbral areas in striatum and in projections areas such as cingulate cortex appeared more intense at 22 h after reperfusion. This increased intensity is due to stronger labeling of neurons and to the appearance of SVCT2 positive astrocytes. In layer I of cingulate cortex, the number of SVCT2-labeled cells was greatly increased in the ischemic rats compared with unoperated or sham-operated controls (Fig. 4b–d, see below for quantitation). In cortical layers II to VI, strongly labeled cells that present the star-like appearance of astrocytes can be clearly distinguished in ipsilateral cingulate cortex but not on the contralateral side (Figs 4h and i). The appearance of astrocyte-like labeling was less obvious in penumbral regions in the striatum (Figs 4j and k). However, neuronal SVCT2 labeling was clearly increased in intensity in ipsilateral striatum compared with the contralateral side.

image

Figure 4. SVCT2 mRNA expression in coronal sections at the level of the striatum at 22 h of reperfusion and in controls. (a) Low power view of ischemic core (asterisks), and adjacent penumbral areas (arrows) and the cingulate cortex peri-infarct area (arrowhead). SVCT2 labeling is absent in the core, but is increased in intensity compared with the contralateral side in the penumbra and in cingulate cortex. Areas highlighted in cingulate cortex and striatum indicate areas shown in h, i and j, k, respectively. (b, c, d) Labeling in top layers of cingulate cortex in unoperated controls (b), 22-h ischemic animals (c), and 22-h sham controls (d). The number of SVCT2-positive cells in layer I is markedly increased in the ischemic animal (arrow in c). Though less obvious, darkly labeled cells are also more numerous in lower cortical layers in the ischemic animal than in controls. (e) Strongly labeled cells are present in corpus callosum in the ischemic core area (arrow). Note lack of labeling in overlying cortex or in striatum. (f, g) SVCT2-positive cells are present in ipsilateral cingulum (g) but not on the contralateral side (f). (h, i) Labeling in an unstimulated area of contralateral cingulate cortex (h) and in stimulated ipsilateral cingulate cortex (i). Note darkly stained cells that have a star-like, astrocyte-type morphology (arrows) in the stimulated cortex. (j, k) Increased intensity of labeling of neurons in ipsilateral penumbra (k) compared with contralateral striatum (j) Note that in this region darkly stained astrocyte-like cells are not apparent. Bars: a = 200 μm, d = 200 μm, e = 200 μm, g = 100 μm, k = 40 μm.

Download figure to PowerPoint

In addition to these gray matter regions, SVCT2 labeling was also up-regulated in white matter areas at 22 h of reperfusion. As shown in Fig. 4(e), strongly labeled SVCT2-positive cells are present in corpus callosum in dorso-lateral regions where all labeling was lost in the overlaying cerebral cortex or in the underlying striatum (Fig. 4e). Similarly, in the cingulum, the number of SVCT2-positive cells is greatly increased on the ipsilateral side compared with the contralateral side (Figs 4f and g).

To verify the SVCT2 mRNA expression in astrocytes at 22 h after reperfusion, sections were either double-labeled with an antibody for the astrocyte marker GFAP, or they were processed for double-in situ hybridization analysis using a probe against the astrocytic glutamate transporter GLT-1 (Berger and Hediger 1998) or against the microglial marker MCSF (macrophage colony-stimulating factor) (Borycki et al. 1993). As shown for corpus callosum in Fig. 5(a, d, g), the SVCT2–mRNA-containing cells co-localize with the cell bodies of GFAP-positive cells, demonstrating that they are astrocytes. In Fig. 5(a), four SVCT2-positive cells are indicated by arrows that all show GFAP-immunoreactivity (Fig. 5g). Double-in situ hybridization of SVCT2 with GLT-1 (Figs 5b, e, g) or MCSF (Figs 5c, f, h) also confirmed that the SVCT2-positive cells in the molecular layers of hippocampus and dentate gyrus are astrocytes and not microglia. Similar results were found for the SVCT2-positive cells in layer 1 of cingulate cortex (not shown).

image

Figure 5. The SVCT2-mRNA expressing cells in corpus callosum and the molecular layer of dentate gyrus at 22 h of reperfusion are astrocytes. Sections were co-stained with an astrocyte-specific antiserum against GFAP (a, d, g) or co-hybridized with a probe for the glial glutamate transporter GLT-1 (b, e, h) or for the macroglia marker MCSF (c, f, i). a–c show brightfield views of the SVCT2 mRNA labeling, d–f the simultaneous exposure of brightfield and fluorescent signals, and g–i the fluorescent antibody staining or in situ labeling. (a, d, g) The SVCT2-expressing cells in corpus callosum (arrows in a, brightfield) are GFAP-positive (d, g). (b, e, h) In dentate gyrus molecular layer, GLT-1 mRNA is co-expressed with SVCT2 mRNA. (c, f, i) MCSF is not co-expressed with SVCT2 in the molecular layer. Bar = 10 μm.

Download figure to PowerPoint

To assess whether the changes in SVCT2 labeling intensities seen at 2 h in dentate gyrus granule cells, and at 22 h in astrocytes, were significant, pictures taken from identically processed sections were analyzed using NIH image software. Figure 6 gives an example of the areas that were analyzed in the molecular and granule cell layers of dentate gyrus and in layer I of cingulate cortex in ipsi- and contra-lateral hemisphere. As shown in Table 1, at 2 h, average signal intensities in the ipsilateral granule cell layer were significantly higher than in unoperated controls. The intensity on the contralateral side was also elevated because some animals showed bilateral up-regulation. Intensities in the two sham-operated controls were below the unoperated controls (not shown). At 22 h, the average intensity on the ipsilateral side was still elevated while that on the contralateral side was back to control levels. In the molecular layer, signal intensities were similar to controls at 2 h on both sides. However, at 22 h, ipsilateral molecular layer showed a significant increase in labeling intensity whereas the increase on the contralateral side was not significant due to a wider variation. In cingulate cortex, average intensities on ipsi- and contra-lateral sides were moderately but significantly elevated at 2 h. These increases became even more pronounced at 22 h.

image

Figure 6. Indication of the field sizes analyzed for signal intensity in the molecular and granule cell layers in dentate gyrus (a) and in layer I of cingulate cortex (b). In b, the two squares indicate contralateral (left) and ipsilateral (right) fields. Bar = 100 μm.

Download figure to PowerPoint

Table 1.  Comparison of average labeling intensities in peri-infarct areas in control, 2- and 22-h ischemic animals a
  Control2 h22 h
  • a

    Values are mean ± SD (n).

  • *

    p < 0.05 and

  • **

    p < 0.01 compared with the value on the ipsilateral side in controls.

Cingulate cortexIpsi21.67 ± 1.02 (8)25.05 ± 2.09** (8)29.83 ± 2.84** (7)
 Layer IContra20.90 ± 0.88 (8)24.39 ± 2.20** (8)26.83 ± 2.62** (7)
Dentate gyrusIpsi146.27 ± 11.94 (7)193.60 ± 12.66** (7)168.64 ± 16.94* (7)
 Granule cellsContra142.32 ± 8.44 (7)167.70 ± 17.19* (7)140.96 ± 21.63 (8)
Dentate gyrusIpsi20.64 ± 1.54 (7)22.23 ± 3.32 (7)31.07 ± 7.98** (7)
 Molecular layerContra21.08 ± 1.80 (7)21.41 ± 2.89 (7)24.14 ± 7.40 (7)

Discussion

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

The present study shows that ischemia induced by MCAO for 2 h causes profound loss of SVCT2 mRNA expression in the ischemic core, and an increase in SVCT2 in adjacent penumbral areas and in peri-infarct areas at 2 and 22 h of reperfusion. At the early time point, the observed increase of SVCT2 occurred primarily in neurons. At the later time point, SVCT2 expression was also found in astrocytes, confirming our hypothesis that astrocytic SVCT2 expression is induced by oxidative stress. These findings underscore the importance of ascorbate as an antioxidant and may have implications for therapeutic interventions.

The loss of SVCT2 and Arc mRNA expression in the ischemic core at 2 and 22 h after ischemia is due to the generalized loss of protein synthesis and energy supply in this area and correlates well with the pronounced reductions in ascorbate levels found in the core area in other ischemia models (Lyrer et al. 1991; Uemura et al. 1991). In the penumbra, SVCT2 expression is initially (at 2 h) not markedly affected, in contrast to Arc, which, as an immediate early gene, is strongly up-regulated next to the ischemic core. The strongest increase in SVCT2 at 2 h was found in the dentate gyrus granule cells, which also showed strong Arc up-regulation. This increase in dentate gyrus granule cells subsided by 22 h for both SVCT2 and Arc. At 22 h however, SVCT2 up-regulation was noticeable in neurons in penumbral areas next to the ischemic core. Also, at 22 h SVCT2 expression became obvious in astrocytes in the molecular layer of dentate gyrus, and in layer I of cingulate cortex, and in white matter areas such as subiculum and corpus callosum. SVCT2 expression was only observed in neurons and in astrocytes in the mentioned areas (besides the expression in choroid plexus and meninges). At no time point did we find evidence for expression in microglial cells or oligodendrocytes. It is not known whether the increases in SVCT2 transporter expression in penumbra and peri-infarct areas affect tissue levels of ascorbate in these regions. Studies that measured ascorbate content after ischemia do not have the required spatial resolution to make such an assessment (Lyrer et al. 1991).

The mechanism leading to increased SVCT2 expression in neurons or astrocytes likely involves the action of glutamate. The marked increases in Arc expression seen in penumbra and peri-infarct areas suggest the release of glutamate, probably via spreading depression-like activation of pathways that originate in the infarct zone, including pathways such as the entorhinal cortex projections to dentate or intercortical projections to cingulate cortex (Kunizuka et al. 1999; Hata et al. 2000). However, an excitotoxic action of glutamate alone is not the cause for the SVCT2 increase as, in our previous study, direct intrastriatal injection of quinolinic acid did not cause up-regulation of SVCT2 mRNA in astrocytes at 3 days after injection (Berger and Hediger 2000). Rather, it is likely that activation of SVCT2 requires the complex series of events that accompany ischemia/reperfusion, including glutamate release, calcium entry, and generation of reactive oxygen species through the actions of glutamate and through mitochondrial hyperoxidation (Rosenthal et al. 1995; Lipton 1999). Interestingly, ascorbate uptake in cultured rat astrocytes is stimulated by dibutyryl cAMP treatment (Wilson 1989), suggesting that a glutamate receptor-mediated mechanism involving cAMP may be involved in the up-regulation of SVCT2 expression after ischemia.

Our demonstration that astrocytes in situ show detectable levels of SVCT2 mRNA after an ischemic insult suggests that the SVCT2 expression found in cultured astrocytes (Berger and Hediger 2000; Korcok et al. 2000) is the result of chronic oxidative stress. Culturing brain cells leads to loss of ascorbate content (Makar et al. 1994), primarily because ascorbate cannot be synthesized de novo by brain cells. However, vitamin C can be supplemented in the culture medium, which causes improved histology and prevents edema formation of brain slices (Rice 1999). It remains to be established whether unstressed astrocytes in situ express a basal level of SVCT2 that is below the detection limit of in situ hybridization.

In contrast to peri-infact regions, an up-regulation of SVCT2 expression in astrocytes was not discernable at 22 h in gray matter in penumbral areas like medial striatum, or the ventral preoptic region, even though neuronal SVCT2 was increased. Using our double-labeling techniques, we could not identify astrocytes with significantly enhanced SVCT2 expression in these penumbral areas. One major distinction between penumbra and peri-infarct regions is that the blood flow and the energy supply are not directly affected by the ischemia in peri-infarct areas. Thus, it appears that the up-regulation of SVCT2 in astrocytes and neurons may be mediated via different mechanisms or may have different thresholds. The reduced energy supply in the penumbra (Hossmann 1999) may prevent the induction of SVCT2 expression in astrocytes even though it could allow an increase in neurons.

In contrast to gray matter, penumbral astrocytes in white matter did show clear up-regulation of SVCT2 expression, most notably in corpus callosum, subiculum, and ventral forebrain bundle. It is possible that these white matter astrocytes in the penumbra react differently to the ischemia because they belong to the fibrous astrocyte subtype, which is well known to have an expression profile different from that of protoplasmic astrocytes in the gray matter (Miller and Raff 1984; Kobayashi et al. 1996). Indeed, white matter does possess specialized mechanisms that make it more resistant to ischemia (Fern et al. 1996). Alternatively, it is feasible that the effects of ischemia on blood flow and energy supply in white matter are less pronounced than in gray matter, and thus allow the activation of the SVCT2 gene in astrocytes.

The increased ascorbate transporter expression by astrocytes and neurons after ischemia may have important pathological implications. Under normal conditions, neurons accumulate relatively high concentrations of vitamin C via SVCT2 for protection against oxidative stress, which occurs during neuronal activity (Hediger 2002). As a result of the scavenging of free radicals, ascorbate is oxidized to dehydroascorbate (DHA), which is then reduced back to ascorbate by glutathione and other intracellular thiols (Meister 1994). The importance of ascorbate as an intracellular antioxidant has been well documented in pond turtles, which have particularly high concentrations of vitamin C in the brain (Rice et al. 1995). These animals also have a remarkable tolerance for oxygen depletion during diving. High levels of vitamin C may represent an adaptation to prevent oxidative damage during re-oxygenation after a hypoxic dive. The intracerebral hemorrhage observed by Nussbaum and colleagues (Sotiriou et al. 2002) in SVCT2-null mice further supports the notion that vitamin C and SVCT2 protect the brain from free radical damage. During ischemia, massive amounts of ascorbate are released into the extracellular compartment in the ischemic core area (Hillered et al. 1988; Yusa 2001), but it is unclear what happens to extracellular ascorbate concentration in the penumbra or peri-infarct areas. Ascorbate can exit neurons through glutamate transporter-mediated glutamate-ascorbate heteroexchange (Cammack et al. 1991). Therefore, when ischemia causes pronounced glutamate release in penumbra and the projection areas in dentate gyrus and cingulate cortex, reuptake of that glutamate may increase extracellular ascorbate. In addition, oxygen depletion, which occurs during the ischemic event, results in neuronal depolarization and compromises ascorbate reuptake via sodium-coupled, SVCT2-mediated ascorbate transport. Finally, in areas where cell swelling occurs due to the ischemia, ascorbate may be released through volume-sensitive organic anion channels (Siushansian et al. 1996). Thus, the up-regulation of SVCT2 in neurons after the ischemic event may represent an attempt to replenish intracellular ascorbate stores for intracellular antioxidant defense. Moreover, the induction of SVCT2 in astrocytes may be a mechanism to remove excess extracellular ascorbate after the initial ischemic period had subsided. However, further studies are clearly required to define the precise role of the ascorbate transport induction in neurons and astrocytes after ischemia.

Acknowledgements

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

This study was supported by NIH grant 32001 to MH and by Guilford Pharmaceuticals Inc.

References

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