Head Mild Hypothermia Inhibits the Increase of Extracellular Glutamate Concentration by Upregulating Glial Glutamate Transporter 1 During Ischemia-reperfusion Injury

Background Stroke is characterized by high morbidity, high mortality and disability. Therefore, it is very important to illuminate the pathological mechanism and find effective treatment strategies. Hypothermia therapy is widely used in clinical as a neuroprotective strategy practice. However, the exact mechanism is not fully understood. Methods Four-vessel occlusion (4-VO) was used to establish animal models of global cerebral ischemia-reperfusion and nasopharyngeal cavity cooling method was used for hypothermia treatment. Extracellular fluid in hippocampal CA1 area was collected by microdialysis technology, and extracellular glutamate concentration ([Glu]e) was detected by high performance liquid chromatography (HPLC). The expression levels of glial glutamate transport-1(GLT-1), Bcl-2 and Bax at the determined time points in hippocampal CA1 area were detected by Western blotting and immunohistochemistry and histopathological evaluation was performed by thionin staining. Dihydrokainate (DHK), a GLT-1 specific inhibitor, was used to confirm the function of GLT-1 in cerebral ischemia-reperfusion injury. Results Cerebral brain ischemia-reperfusion caused the downregulation of GLT-1 and Bcl-2, the upregulation of Bax, and an increase in [Glu]e, and leaded to neuron loss. Head mild hypothermia (HMH) for 2 h applied 8min after ischemic insult attenuated the abovementioned effects of ischemia-reperfusion, while pretreatment with DHK by injection into the lateral ventricle inhibited this effect. Conclusion HMH can play a neuroprotective role by upregulating GLT-1 and reducing the excitotoxicity of Glutamate during ischemia-reperfusion insult in rats.


Background
Stroke is the second leading cause of death worldwide and is the leading cause of death in China, in which a fifth of the world's population resides [1,2]. We have made great advances in the treatment of stroke, but advances in acute care also mean that more people survive with sequelae, resulting in major implications for health, the economy and society. Scholars have proposed a series of neuroprotective strategies, such as drug therapy [3], ischemic preconditioning [4], exercise preconditioning [5], and hypothermia [6]. Busto first proposed the application of hypothermia in cerebral infarction patients in 1987 [7], and many basic studies [8,9] and clinical trials [10,11] have confirmed that hypothermia can effectively improve the prognosis of patients with cerebral ischemic diseases.
Glutamate, (Glu) is a major excitatory amino acid in the mammalian brain, it acts as a neurotransmitter to transmit excitatory signals [12,13], and it is also a neurotoxin. Excessive extracellular Glu produce excitotoxicity and cause neuronal apoptosis [14]. Neurotransmission abnormalities occur in many brain diseases and are associated with abnormal concentrations of Glu [15]. GLT-1 plays a major role in maintaining a low concentration of glutamate in the extracellular space [16,17].
However, there has been little research on the relationship between GLT-1 and the neuroprotective effect of hypothermia during cerebral ischemia-reperfusion. This experiment investigated the effect of HMH on GLT-1 and the [Glu]e immediately after global cerebral ischemia-reperfusion in rats to further explore the neuroprotective mechanism of mild hypothermia.

HMH inhibits the downregulation of GLT-1 caused by ischemia-reperfusion injury
5 GLT-1 is mainly expressed in astrocytes in the hippocampal CA1 region, as determined by immunohistochemistry. A small amount of GLT-1 was uniformly distributed throughout the hippocampal CA1 region in group C. Compared with that in group C, the expression of GLT-1 in group S was higher at 0 h, and the expression peaked at 8 h. A large number of brown-stained GLT-1-immunopositive particles were observed in the hippocampal CA1 region; then, the expression decreased, increased from days 1-3, and then continued to decrease until the end of the experiment. GLT-1 expression in the timepoint subgroups were higher than that in group C at all time points (p < 0.05).. In I/R group, peak of GLT-1 in hippocampal CA1 region appeared at 0h, then decreased, and there was a large area of the hippocampal CA1 region in which GLT-1 expression was deficient at 7 d. GLT-1 expression at each time point, except for 0 h, was lower than that in group S (p < 0.05).. Compared with that in group I/R, GLT-1 expression in group HI/R was increased at 7 d (p < 0.05).. Compared with that in group HI/R, GLT-1 expression in group DHK+HI/R and group NS+HI/R was not significantly different at 7 d (p >0.05),, as shown in Fig 2. There was a small amount of GLT-1 expression in group C, as determined by western blotting. Compared with that in group C, GLT-1 expression in group S was significantly increased at all time points (p < 0.05),, and the expression peaked twice, once at 8 h and once at 3 d. The expression in group I/R peaked at 0 h and then decreased continuously. GLT-1 expression in group I/R was lower than that that in group S at each time point except 0 h and 1 d (p < 0.05).. Compared with that in group I/R, GLT-1 expression in group HI/R increased at 7 d (p <0.05).. There was no significant difference in GLT-1 expression in either group HDK+HI/R or group NS+HI/R compared with group HI/R at 7 d (p >0.05), as shown in Fig 5A- Bcl-2 was mainly distributed in the cytoplasm and neuronal synapses in the hippocampal CA1 region. A small amount of Bcl-2 was expressed in group C. Bcl-2 expression was not significantly different between group C and group S at 7 d (p > 0.05).. Compared with that in group S, Bcl-2 expression in group I/R was decreased (p < 0.05).. Compared with that in group I/R, Bcl-2 expression in group HI/R was increased (p < 0.05).. Compared with that in group HI/R, Bcl-2 expression in group DHK+HI/R was decreased (p < 0.05),, while there was no significant difference in group NS+HI/R (p > 0.05),, as shown in Fig 3. In western blotting analysis, result of Bcl-2 expression at 7d was the same with that in immunohistochemistry (Fig 5B, D) Bax was also distributed in the cytoplasm and neuronal synapses. A small amount of yellowish immunoreactivity was observed in group C. Bax expression was not different between group S and group C at 7 d (p > 0.05).. Compared with that in group S, Bax expression in group I/R was increased (p < 0.05).. Compared with that in group I/R, Bax expression in group HI/R was decreased (p < 0.05).. There was no difference between group NS+HI/R and group HI/R, (p > 0.05),, while Bax expression in group DHK+HI/R was increased compared to that in group HI/R (p < 0.05),, as shown in

HMH reduces neuronal loss in the hippocampal CA1 region caused by ischemiareperfusion injury, and DHK inhibits this effect
In the hippocampal CAl subfield in control rats, pyramidal neurons were arranged in order, there were 2 to 3 cell layers, the membranes of the neurons were intact, the 7 nucleus were full, and the nucleolus were clear (ND: 214 ±9.0). At 7 d, there was no obvious neuronal death in the hippocampal CA1 region of the rats from group S (ND: 195±14.5), and there was no significant difference compared with that in group C (p

Hypothermia and excitotoxicity
Glutamate is the main excitatory transmitter in the central nervous system [12,13].
High extracellular concentrations of Glu lead to the excessive activation of glutamate receptors and excitotoxicity in nerve cells [14]. A variety of neurological diseases, such as cerebral apoplexy, traumatic brain injury, Parkinson's disease, multiple sclerosis of the lateral cord, and Alzheimer's disease, are associated with Glu excitotoxicity [18][19][20]. The brain lacks Glu-metabolizing enzymes. The low extracellular concentration of Glu is mainly maintained by the glutamate transport system, also known as excitatory amino acid transporters (EAATs), which can uptake Glu into glial cells and neurons. The results of the present study suggested that sham operation upregulated GLT-1, while I/R injury reduced this upregulation. The occlusion of the vertebral artery, which is equivalent to a mild ischemic insult, in group S also affected cerebral blood flow. Ischemic preconditioning can upregulate GLT-1 expression by activating the p38/MAPK pathway [4] and increase the maximum binding and affinity of GLT-1 [21], thus increasing Glu reuptake and enhancing ischemic tolerance. We observed that HMH upregulated GLT-1 expression but did test the binding and affinity of GLT-1. In the present experiment, HMH increased GLT-1 expression and decreased neuronal death, which suggests that hypothermia may play a neuroprotective role by inhibiting the downregulation of GLT-1 expression. The use of the GLT-1 functional antagonist DHK to inhibit the transport of GLT-1 increased neuronal death, which suggests the accuracy of our hypothesis. In the experiment that evaluated the [Glu]e, we observed that cerebral ischemia insult caused a significant increase in the [Glu]e. High concentrations of Glu can cause intracellular Ca 2+ overload, causing nerve injury through the excessive activation of iGluRs [22]. HMH can inhibit the increase in the [Glu]e caused by ischemia. Inhibiting GLT-1 function with DHK partially reverses the effect of hypothermia. The above results indicate that mild hypothermia reduces the [Glu]e by upregulating GLT-1 and that this is one of the neuroprotective mechanisms.

GLT-1 and energy dysfunction
The role of GLT-1 in ischemia reperfusion injury is related to energy metabolism disorders. Glu reuptake is an energy-consuming process that depends on Na + .
EAATs cotransport 3 Na + and 1 H + and countertransport 1 K + in the process of extracellular Glu uptake. Glu is uptaken by EAATs, and a part of Glu generates glutamine under the action of glutamate synthase (GS), which is transferred to neurons for the synthesis of Glu and GABA. Some Glu is converted to α ketoglutaric acid by glutamate dehydrogenase in mitochondria, and it enters the TCA cycle to generate ATP, which provides energy for EAATs to uptake Glu [23]. A decrease in mitochondria can affect Glu oxidation metabolism [24]. The specific knockout of the glutamate dehydrogenase gene in the brain increases peripheral glucose transport to the brain to increase energy to compensate for the loss of glutamate-derived energy [25]. Some studies have confirmed that oxygen and glucose deprivation can cause delayed mitochondrial deficiency in astrocytic processes [26]. Mitochondrial dysfunction caused by ischemia-reperfusion injury may further affect neuronal recovery after injury by affecting Glu oxidative metabolism and the energy supply.
We are currently performing experiments to observe the effect of hypothermia on Glu oxidative metabolism.

Hypothermia and apoptosis
Bcl-2 and Bax belong to the Bcl-2 gene family and are closely related to apoptosis.
Bcl-2 can inhibit cell apoptosis and improve cell viability, mainly expresses in mitochondria, the endoplasmic reticulum and nuclear membrane. The antiapoptotic effect of Bcl-2 may be related to antioxygen free radicals, calcium overload and changes in intracellular proteases [27]. The physiological function of Bax is to promote cell apoptosis, mainly expresses in the cytoplasm. Some studies have shown that Bax functions by forming heterodimers with Bcl-2 on the mitochondrial membrane [28]. When Bax is highly expressed, Bax homodimers can be formed to promote cell apoptosis. When Bcl-2 is highly expressed, Bcl-2 and Bax form heterodimers to inhibit apoptosis [29]. Some researchers have found that the ratio of Bcl-2/Bax determines the expression level of p53 and determines whether apoptosis occurs in cells [30]. We observed that Bcl-2 decreased and Bax increased significantly after I/R insult, and the ratio of Bcl-2/Bax decreased significantly. HMH upregulated Bcl-2 expression and inhibited Bax overexpression after ischemia reperfusion. DHK treatment partially reversed this effect. High [Glu]e can cause intracellular Ca 2+ overload through the excessive activation of iGluRs [22], which further leads to mitochondrial swelling and deficiency [26]. Therefore, GLT-1 may affect the expression of Bcl-2 and Bax by reducing Glu excitotoxicity and affecting mitochondrial structure and function, thus inhibiting apoptosis caused by ischemiareperfusion injury.
There are some limitations of the present experiment. First, thionin staining cannot accurately distinguish necrosis from apoptosis. Although Bcl-2 and Bax can reflect the level of apoptosis, TUNEL and flow cytometry are more specific. Second, we only studied the protective effect of HMH on ischemia-reperfusion injury at the cellular level and molecular level, but we did not evaluate the recovery of neurological function with behavioral evaluation. Finally, we observed the Glu concentration for 60 min during ischemia reperfusion, but this reflects the early changes that result from ischemia-reperfusion but not the long-term changes.

Experiment 1: Changes in GLT-1 in rats during cerebral ischemia-reperfusion injury
One hundred and eighty male Wistar rats were randomly divided into 3 groups by random number method: (1) the control group (C, n = 12) no treatment, (2) the sham group (S, n = 84), occlusion of bilateral vertebral arteries (OBVA) were performed and then the bilateral common carotid arteries (BCCA) were exposed Six of brains were used for immunohistochemistry and thionin-staining, and 6 were used to isolate the hippocampal CA1 region, which was frozen in liquid nitrogen and stored at -80°C for Western blot analysis.

Experiment 2: Effect of HMH on GLT-1 in rats during cerebral ischemia-reperfusion injury
Our previous studies confirmed that nasopharynx cavity cooling has a neuroprotective effect [31]. To observe the effect of HMH on the downregulation of GLT-1 caused by ischemia-reperfusion injury, we added the following groups on the basis of experiment 1: (4) the hypothermia ischemia-reperfusion group (HI/R, n = 12), after OBVA for 48 h, the BCCA were exposed, and then the blood flow were blocked for 8min after the hippocampal temperature dropping to 32°C due to implementation of nasopharynx cavity cooling method, and keep hypothermia for 2 h, (5) the normal saline + hypothermia ischemia-reperfusion group (NS+HI/R, n = 12), 20μl of 0.9%NS was intracerebroventricularly injected 30min before HI/R procedure, (6) the DHK+ hypothermia ischemia-reperfusion group (DHK+HI/R group, 13 n = 12), GLT-1 functional inhibitor DHK(20μl) was injected into lateral ventricle 30min before HI/R procedure. According to the results of experiment 1, we found that the differences of GLT-1 expression levels were the most obvious at 7 d. Therefore, in this part of the experiment we only performed observations at 7 d (Fig1A).

Experiment 3: Effect of GLT-1 on the [Glu]e in rats during global cerebral ischemiareperfusion injury
HMH can reduce the Glu concentration in the brain of rats with global cerebral ischemia-reperfusion injury [31]. To clarify the neuroprotective role of GLT-1 in ischemia-reperfusion injury, we designed the following experiment to observe the changes in the [Glu]e. The treatment groups were the same as (2) -(6)

Establishment of global cerebral ischemia-reperfusion model
In this experiment, 4-vessel occlusion (4-VO) [2] was used to establish an animal model of global cerebral ischemia-reperfusion. All operational procedures were completed under isoflurane anesthesia. detailed method: oxygen flow, 0.5-0.7 L/min, isoflurane concentration: anesthesia induction with 3%-4% and maintenance with 1.5-2.5%. First, the bilateral vertebral arteries were electrocauterized under isoflurane anesthesia. After 48 h of recovery, the BCCA of the rats were exposed under isoflurane anesthesia and procaine (1%) local anesthesia. After the rats recovered from the anesthesia, the BCCA were clamped using clips for 8 min to induce brain ischemic insult. Only animals, in which the consciousness lost, then 14 righting reflex disappeared, and the pupils dilated obviously after the occlusion, were considered success of global cerebral ischemia. The wounds were sutured after each operation.

Nasopharyngeal cavity cooling method
Endotracheal intubation with ventilation was performed after isoflurane anesthesia.
The head of each rat was fixed on a stereotaxic apparatus, and a temperature probe was inserted into the right hippocampus. An 18G intravenous catheter was inserted into the nasopharynx of each rat and fixed, and a block of cotton was placed in the throat to prevent aspiration. Cool water (4°C) was pumped into to the nasopharynx at a rate of 100 ml•kg -1 •min -1 , and continuous suction was maintained simultaneously. The speed was adjusted according to the thermometer data to maintain hippocampal temperature at 33±0.5°C. Use an electric blanket to keep the body warm and maintain the anal temperature at 37±0.5°C [31].

Intracerebroventricular injection
In the DHK+HI/R group and NS+HI/R group, 20μl of DHK (Lot:064M4608V, Sigma USA) solution (200 nmol) or normal saline were slowly injected into the right ventricle (1.5 mm to the right of the anterior fontanelle, 0.8 mm posterior, and 3.9 mm below the surface of skull) with a stereotaxic instrument one hour before global cerebral ischemia [32].

Neuropathological evaluation
At the determined time points, the rats were anesthetized with chloral hydrate and perfused through the ascending aorta with normal saline followed by 4% paraformaldehyde. Coronal brain slices (3 mm thick) that included the bilateral dorsal hippocampus were excised and fixed in 4% paraformaldehyde. The brain slices were dehydrated in alcohol, and cleared with xylene, and then embedded in 15 paraffin. The paraffin embedded brain tissues were sectioned at a thickness of 5μm.
The sections were stained with thionin for neuropathological evaluation. According to the methods of Kitagawa et al. [33] and Kato et al. [34], the neuronal density

Western blotting analysis
Hippocampal CA1 tissue were frozen at -80°C, and the lysates were decontaminated and homogenized in a homogenizer. The homogenate was centrifuged at 4°C, the supernatant was collected, and the total protein was quantified by the BCA method.

Collection of extracellular fluid and measurement of the [Glu]e.
Rats were placed in a stereotaxic apparatus after anesthesia. A microdialysis probe (MAB 6.14.2.ss, MAB Sweden) was stereotaxically implanted into the parietal cortex for 2 mm (3 mm posterior to and 3 mm to the left of the bregma) and perfused with Ringer's solution at a flow rate of 2 μl/min using an infusion pump. When the electroencephalogram was stabilized, microdialysis fluid was collected every 10min for 60min. t0: 20 ~ 10 min before global cerebral ischemia, t1: 10 min 0 min before ischemia; t2: 0 min~10 min after ischemia; t3: 10 min ~ 20 min after ischemia; t4: 20 min ~ 30 min after ischemia; t5: 30 min ~ 40 min after ischemia. The samples collected should be away from light and stored at -80℃ for testing. The peak areas of Glu standard products with various concentrations were determined by high performance liquid chromatography analyzer (Thermo UltiMate 3000, Thermo USA), and the standard curve was drawn according to the functional relationship between the absorption peak area and the content. According to the regression equation of the standard curve and the recovery rate of the microdialysis probe, the actual [Glu]e in rat brain tissue was calculated. The average concentration at t1 and t2 was taken as the basic value. The ratio of [Glu]e to the basic value [Glu]r was calculated for inter group comparison [35].

Statistical analysis
Statistical analysis was performed using SPSS 21.0. All data are presented as the mean±SEMs Comparisons were performed by one-way ANOVA and the LSD method for multiple comparisons. Unmatched dates were analyzed by the nonparametric rank sum test combined with Dunn-Bonferroni test for multiple comparisons method to test the differences between the groups.