The Up-regulated Expression Of GLT-1 And Maintained Balance Of Amino Acids In Extracellular Fluid Participated In The Induction Of Brain Ischemic Tolerance In Rats

Transient sublethal cerebral ischemia could protect hippocampal neurons against delayed neuronal death (DND) induced normally by lethal ischemic insult. The transient cerebral ischemia is usually referred to as cerebral ischemic preconditioning (CIP), and the protective role is named as brain ischemic tolerance (BIT). The phenomenon was first found by Kitagawa in 1990 and proved by many other studies. It is very important to clearify the mechanisms of BIT induced by CIP for developing new therapeutic methods to enhance the tolerance of neurons to ischemia and hypoxia.Many studies have proved that excitotoxicity of glutamate is an important mechanism for DND induced by transient global ischemic insult. Glutamate uptake is transiently reduced and the extracellular glutamate is increased after hypoxia-ischemia insults in the brain. Excitatory amino acid transporters (EAATs) are essential for maintaining normal extracellular level of glutamate. Five distinct high-affinity, sodium-dependent EAATs are identified in the rat brain, which include EAAT1 (glutamate/aspartate transporter, GLAST), EAAT2 (glial glutamate transpor-1, GLT-1), EAAT3 (excitatory amino acid carrier 1, EAAC1), EAAT4 and EAAT5. GLAST and GLT-1 are localized primarily in astrocytes. EAAC1 is widely distributed in hippocampal neurons. EAAT4 is localized mainly in cerebellar Purkinje cells. EAAT5 is mainly localized in retina. EAATs normally remove glutamate into cells and are driven by sodium, potassium and possibly by hydroxide ion gradients. Conversely, When the ion gradient or membrane potential drops, for instance during ischemia or epileptic activity, EAATs may decrease the uptake of glutamate, or even reverse and release glutamate into the extracellular space in a calcium-independent manner. Therefore, extracellular glutamate becomes abnormally high and leads to excitotoxicity.Although both neurons and glia contain EAATs, it is generally accepted that the uptake capacity of astrocytes is much higher than that of neurons. Many studies have shown that GLT-1 plays a principal role in removing the released glutamate from the extracellular space and maintaining the extracellular glutamate below neurotoxic level in the brain. Considering the importance of GLT-1 in removing glutamate, it is reasonable to hypothesize that GLT-1 maybe play an important role in the acquisition of the brain ischemic tolerance induced by CIP. Unfortinately, there were limited reports until now concerning the role of GLT-1 in the induction of brain ischemic tolerance. Although they obtained a similar conclusion that GLT-1 participated in the induction of brain ischemic tolerance, the mechanisms underlying were just reversal. Additionally, they just performed the experiments in vitro. Little is known whether GLT-1 plays a role during the induction of brain ischemic tolerance in vivo. Therefore, the present study was undertaken to study whether GLT-1 participates in the induction of brain ischemic tolerance in vivo by observeing the expression of GLT-1 and glial fibrillary acidic protein (GFAP), a specific protein expressed by astrocytes, using immunohisto- chemistry and western blot analysis, and changes in concentrations of extracellular glutamate, aspartate, glycine andγ-aminobutyric acid (GABA) using brain microdialysis and high performance liquid chromatography (HPLC) during the induction of brain ischemic tolerance in rats.1 The up-regulation of GLT-1 and GFAP in the rat hippocampal CA1 subfield induced by CIPThe rat global cerebral ischemic model was established by four-vessel occlusion. To clarify the role of GLT-1 during the induction of BIT in vivo, the expression of GLT-1 and GFAP in the CA1 hippocampus during the induction of brain ischemic tolerance in rats was observed by immunohistochemistry and western blotting.1.1 Neuropathological evaluationOne hundred and forty five adult male Wistar rats were divided into 5 groups randomly:①control group (n=5);②vertebral artery occluding group (n=35): the bilateral vertebral arteries were electrocauterized permanently;③CIP group (n=35): a global brain ischemia for 3 min was given;④brain ischemic insult group (n=35): a global brain ischemic insult for 8 min was given;⑤CIP+ischemic insult group (n=35): a CIP was performed first and then a lethal global ischemic insult for 8 min was given 2 days after the CIP. The observations were performed at time 0 (immediate), 3 h, 1 d, 2 d, 3 d, 5 d and 7 d, after the last operation or treatment (n = 5 in each time point), except for the control group. At the determined endpoint of the experiment, the animals were sacrificed and the brain was removed for the neuropathological evaluation. The brain tissues were sectioned, and the delayed neuronal death (DND) was observed under staining with thionin. The histological changes of the hippocampal CA1 subfield were divided into 4 histological grade (HG) under the light microscope according to the following standard: grade 0, no neuron death; gradeⅠ, scattered single neuron death; gradeⅡ, mass neuron death; gradeⅢ, almost complete neuron death. The neuronal density (ND) of the hippocampal CA1 subfield was determined by counting the number of surviving pyramidal neurons with intact cell membrane, full nucleus and clear nucleolus within 1 mm linear length of the CA1. The average of number of pyramidal neurons in 3 areas of the hippocampal CA1 subfield was calculated as value of ND.Neuropathological evaluation showed that in the control rats, pyramidal neurons in the CA1 hippocampus were arranged in order with 2 to 3 cell layers, the outline of the neurons was intact, nucleus was full and nucleolus was clear. The HG was 0 and ND was 210±5.7 mm-1. No significant neuronal damage was observed in the CA1 subfield at all time points observed after VAO. Neither the HG nor the ND was different from that of the control group. No change in ND or HG was found at all time points after CIP compared with that of the control or VAO group. During the first two days after the lethal ischemic insult for 8 min, no significant pyramidal neuronal damage was observed in the hippocampal CA1 subfield. However, obvious DND was observed from the third day after the ischemic insult, such as decrease in ND and increase in HG. The damage deteriorated with time manifested as pyknosis of cell bodies, karyopyknosis of the nucleus, disappearance of the nucleolus. Almost complete neurons died on the fifth and seventh day after the lethal ischemic insult, represented by more significant decrease in ND and increase in HG compared with that of the third day after the lethal ischemic insult. When the animals were pretreated with the CIP 2 days before the lethal ischemic insult, the above injured changes were prevented clearly, which indicated that the CIP protected the pyramidal neurons in the CA1 hippocampus against the DND induced normally by the lethal ischemic insult.1.2 Immunohistochemistry assayAnimals, grouping and protocols of the experiment were the same as those in neuropathological evaluation.There were very weak, but diffuse immunoparticles distributed in the peri-pyramidal neuronal structure of the hippocampal CA1 subfield in the control group. The staining pattern is consistent with other reports. Compared with the control group, the intensity of GLT-1 immunoreactivity was markedly increased at almost all time points and showed morphological characteristics of astrocytes in the VAO group. The intensity of GLT-1 immunoreactivity was further increased after the CIP compared with that of the VAO group. Very interestingly, some GLT-1 immunoreactive particles were observed in the area between the pyramidal neurons, which tightly surrounded pyramidal neurons and made the pyramidal layer looked like"shaped grade". Compared with the VAO group, the GLT-1 expression was markedly decreased at all time points after the lethal ischemic insult for 8 min, especially in the area where almost complete pyramidal neurons died, and the neighboring area of the pyramidal layer, even appeared as a sheet absence of GLT-1 immunoreactivity. But when the animals were pretreated with the CIP 2 days before the lethal ischemic insult, the decrease of the GLT-1 immunoreactivity induced by the lethal ischemic insult was prevented thoroughly. Moreover, there were more GLT-1 immunoreactive particles tightly surrounded the pyramidal neurons, which made the"shaped grid"observed in the CIP group to be more clear.In the control group, GFAP immunostaining showed that astrocytes took on star or spider-like shape with prominent processes. Very few GFAP immunoreactive particles were observed in the area between the pyramidal neurons. Compared with the control group, significant down-regulation of GFAP immunoreactivity was observed at all time points in the VAO group, and few GFAP immunoreactive particles were observed in the area tightly surrounded the pyramidal neurons. After a CIP for 3 min, The GFAP expression was significantly up-regulated at almost all time points compared with that of the VAO group. Some GFAP immunoreactive particles were observed, like GLT-1, in the area between the pyramidal neurons, which tightly surrounded the pyramidal neurons and made the pyramidal layer look like"shaped grade". The characteristics of immunostaining were similar with those of GLT-1 immunostaining mentioned above. Moreover, both the total area and average optical density of GFAP immunostaining after the CIP were significantly up-regulated compared with those of the VAO group. After the lethal brain ischemic insult for 8 min, astrocytes hypertrophied in soma with thickened processes and more intensive staining, which reached peak on 3 d after the lethal ischemia. However, no GFAP immunoreactive particles were observed in the area between the pyramidal neurons or the neighboring area of the pyramidal layer at all. From the fifth day after the lethal ischemic insult, the body of the astrocytes became more hypertrophic, whereas the processes of the hypertrophic astrocytes became collapsing and fragmenting. Although the number, total area and average optical density of immunoreactive cells were increased significantly after the lethal ischemic insult compared with those of the VAO group. When the animals were pretreated with a CIP 2 days before the lethal ischemic insult, there were many GFAP immunoreactive particles tightly surrounded the pyramidal neurons thoroughly, which made the"shaped grid"observed in the CIP group to be more clear. The phenomenon reached peak on 2 d and constantly lasted to 7 d (the end of the observed period in the experiment). However, both the number and average optical density of the GFAP immunoreactive cells were significantly decreased compared with those of the ischemic insult group.1.3 Western blotting analysisTwo hundred and five adult male Wistar rats were divided into 6 groups randomly:①control group (n=5);②sham group (n=20);③VAO group (n=45);④CIP group (n=45);⑤brain ischemic insult group (n=45);⑥CIP+brain ischemic insult group (n=45). The western blotting analysis in the sham group was performed at time 0 (immediate), 3 h, 12h and 2 d, while in the other groups was performed at time 0 (immediate), 3 h, 6h, 12h, 1 d, 2 d, 3 d, 5 d and 7 d, after the last operation or treatment (n = 5 in each time point). In our preliminary experiment, some changes in expression of GLT-1 and GFAP were observed after VAO. To clarify whether the changes were induced by surgical procedures or by VAO, the sham group was designed in western blotting analysis. Rats in the sham group were subjected to a sham operation consisting of exposing of bilateral vertebral artery and bilateral common carotid arteries, but neither vertebral arteries nor common carotid arteries were occluded. The protocols of the rats in the other groups were the same as that in part 1.1.Compared with the control group, no difference of GLT-1 levels was found at each time point observed in the sham group. However, the levels of GLT-1 expression were significantly up-regulated in CA1 subfield at almost all time points in the VAO group compared with that of the control group. The above results indicated that the GLT-1 up-regulation in the VAO group was induced by the occluding of vertebral arteries itself other than the anesthesia or surgical procedures. In the CIP group, the GLT-1 level at immediate time point was much higher than that of the control group. On the base of the up-regulated level of GLT-1 at the immediate time point, the expression of GLT-1 was further up-regulated after CIP for 3 min. In the brain ischemic insult group, the GLT-1 expression was significantly down-regulated at almost all time points compared with that of the VAO group. When the animals were pretreated with a CIP for 3 min 2 days before the lethal ischemic insult, the down-regulation of GLT-1 induced by lethal brain ischemic insult was prevented thoroughly by the CIP.No difference of GFAP levels was found at every time points in the sham group compared with that of the control group. However, compared with the control group, the GFAP levels were significantly down-regulated at all time points after the occluding of the vertebral arteries. The above indicated that the GFAP down-regulation in the VAO group was induced by the occluding of vertebral arteries itself other than anesthesia or surgical operation. Compared with the VAO group, the GFAP levels were significantly up-regulated at almost all time points except the immediate time point after CIP. Compared with the VAO group, the GFAP levels were significantly up-regulated at almost all time points except the immediate time point after lethal brain ischemic insult for 8 min. Compared with the ischemic insult group, the GFAP levels were significantly down-regulated on 5 d and 7 d, whereas up-regulated at time 0 and 6 h in CIP+brain ischemic insult group.Summary These results indicated that the surrounding of pyramidal neurons by astrocytes and up-regulation of GLT-1 induced by CIP played an important role in the acquisition of the brain ischemic tolerance induced by CIP.2 Changes in the expression of GLT-1 and GFAP in the hippocampal CA3 subfield and dentate gyrus during the induction of brain ischemic tolerance in ratsThe preparation of the model, grouping, protocols and methods were the same as those in part 1, except for that the neuropathological evaluation was performed by calculating percentages of injured neurons in the CA3 and DG.2.1 Neuropathological evaluationLethal ischemic insult for 8 min induced mild DND in the CA3 subfield and dentate gyrus (DG) in some rats. The rate of neuronal death was approximately 5% in the CA3 subfield and 19% in DG. The CIP 2 days before the lethal brain ischemic insult could also protect the neurons in the CA3 subfield and DG against the DND induced normally by the lethal ischemic insult. All the above indicated that the neurons in the CA3 subfield and DG were relatively tolerated to ischemia, and CIP could also provide protection to the neurons in the CA3 subfield and DG.2.2 The expression of GLT-1 and GFAPSome differences in the expression of GLT-1 and GFAP were found in the CA3 subfield and DG compared with the CA1 subfield in the rat hippocampus. Processes of astrocytes in the CA3 subfield and DG, especially in the CA3 subfield, extended into the area between the neurons even in the control rats, which tightly surrounded neurons and made the neuronal layer looked like"shaped grade". At the same time, many GLT-1 immunoreactive particles were observed in the same area. In the VAO group, both GFAP and GLT-1 immunoreactive particles in the area between the neurons in CA3 subfield and DG were significantly up-regulated by the VAO, which tightly surrounded neurons and made the"shaped grade"observed in the control group to be clearer. In the brain ischemic insult group, the GLT-1 expression was significantly increased and the"shaped grade"became more clearly in large area of the CA3 subfield and DG after the ischemic insult for 8 min, except for the significant decrease in the area where neurons died, and the neighboring area of the dead neurons. Compared with the CA1 subfield, the up-regulation of the expression of GLT-1 and GFAP to CIP for 3 min was also observed in CA3 subfield and DG, whereas the level of the up-regulation in CA3 subfield and DG was much higher than that in the CA1 subfield. In the CIP+ischemic insult group, similar response was observed among different subfields in the hippocampus, and the"shaped grade"existed in the CA3 subfield was very clear.Summary The tolerance of the neurons in the CA3 subfield and DG to ischemia insult maybe related to the relatively higher basal expression and stronger responsive upregulation of GLT-1 and GFAP to ischemic stimulation in the area between neurons in the both subfields. These findings further illustrated the involvement of GLT-1 in the induction of BIT. 3. Changes in concentrations of amino acids in the extracellular fluid in the rat CA1 hippocampus during the induction of BITTo investigate the role of the balance between excitatory amino acids (EAAs) and inhibitory amino acid in the induction of BIT, the concentration of glutamate, aspartate, glycine and GABA was analyzed during the induction of BIT by microdialysis combined with HPLC. Twenty four rats were divided into four groups randomly: vertebral artery occluding (VAO) group (n = 6), CIP group (n = 6), brain ischemic insult (II) group (n = 6), and CIP + brain ischemic insult group (n = 6). The preparations of the model were the same as those in part 1. The extracellular fluid in the rat CA1 region of the dorsal hippocampus was collected by intracerebral microdialysis in conscious rats. The concentration of amino acids in the dialysate was detected by HPLC. All rats were sacrificed on 7th day after the microdialysis for checking the position of the microdialysis probe and neuropathological evaluation.In all cases the microdialysis probe was correctly positioned in the CA1 subfield of the rat dorsal hippocampus. Obvious DND was observed after the brain ischemic insult for 8 min, while no obvious neuronal damage was observed in other groups. The results were consistent with our previous study.No changes in the concentration of each amino acid analyzed were observed among all samples from the VAO rats. In the CIP group, CIP for 3 min caused a significant acute increase coincident in both concentrations and time course of glutamate, aspartate, glycine and GABA, in which the peak values of the concentration were about 1.5, 2, 1.9, and 2.3 fold of their own average control level, respectively. In the brain ischemic insult group, lethal ischemic insult for 8 min evoked a significant increase in glutamate, aspartate, glycine and GABA. The acute increase of glycine and GABA appeared mono-peak at the end of the lethal ischemic insult. The peak value was about 3~4 folds of the average control level of its own. The increase of glutamate and aspartate showed a double-peak pattern. The first peaks were coincident in time course and magnitude with those in glycine and GABA. While the second peaks, which were about 5~7 fold of the control level, were higher in magnitude and appeared respectively at 7 min and 19 min after the reperfusion. In the CIP+ brain ischemic insult group, when the animals were pretreated with the CIP 2 days before the lethal ischemic insult, a significant acute increase of glutamate, aspartate, glycine and GABA coincident in both concentrations and time course was observed, in which the peak values were about 1.7, 2.5, 7, and 4 fold of their own average control level, respectively. The second higher peaks in glutamate and aspartate normally induced by brain ischemic insult were completely inhibited by the CIP. The results have shown that the lethal global brain ischemic insult for 8 min caused the imbalance between the EAAs such as glutamate as well as aspartate and GABA, the inhibitory amino acid; Whereas, when the animals were pretreated with the CIP for 3 min 2 days before the lethal ischemic insult, the increase among glutamate, aspartate, and GABA was coincident and kept balance.Summary The lethal brain ischemic insult for 8 min caused imbalance between glutamate as well as aspartate and GABA in the CA1 hippocampus, which especially manifested as a delayed but more obvious increase in the concentration of glutamate and aspartate. The imbalance could be prevented by a preceded CIP, which could protect the pyramidal neurons of the CA1 hippocampus from DND normally induced by lethal brain ischemia. These findings indicated that preventing of the imbalance between the glutamate as well as aspartate and GABA maybe one of mechanisms involved in the neuroopretection of CIP.In addition, we found that the prior occlusion of the bilateral vertebral arteries during producing 4VO global cerebral ischemic model might play a protective effect like cerebral ischemic preconditioning that can protect to some extent pyramidal neurons of the hippocampus against severe ischemic insult induced by occlusion of bilateral common carotid arteries within 48 h. The vertebral artery enters into the atlas via the transverse foramina, travels through the external aperture of the alar foramina, and then passes through the internal aperture of the alar foramina before entering into the vertebral canal. 4 Conclusions(1) The surrounding of pyramidal neurons by astrocytes and up- regulation of GLT-1 induced by CIP played an important role in the acquisition of the brain ischemic tolerance induced by CIP(2) The tolerance of the neurons in the CA3 subfield and DG to ischemia insult maybe related to the relatively higher basal expression and stronger responsive upregulation of GLT-1 and GFAP to ischemic stimulation in the area between neurons in the areas. These findings further illustrated the involvement of GLT-1 in the induction of BIT.(3) The lethal brain ischemic insult for 8 min caused imbalance between glutamate as well as aspartate and GABA in the CA1 hippocampus, which especially manifested as a delayed but more obvious increase in the concentration of glutamate and aspartate. The imbalance could be prevented by a preceded CIP, which could protect the pyramidal neurons of the CA1 hippocampus from DND normally induced by lethal brain ischemia. These findings indicated that preventing of the imbalance between the glutamate as well as aspartate and GABA maybe one of mechanisms involved in the neuroopretection of CIP.