Ischemic Injury of Cortical GABAergic Neurons：Vulnerability，Mechanism and Pathological Impacts

WANG Jin-hui，HUANG Li，CHEN Na

1. University of Chinese Academy of Sciences，Beijing，100049，China

2. Department of Pathophysiology，Bengbu Medical College，Bengbu，200031，China

Cerebral ischemic stroke is the most common disease and the third reason for disability in senior individuals［1-7］. Therapeutic strategies include anticoagulation，thrombolysis，neuroprotection and antiinflammation［8-18］. Based on preclinical studies about ischemia-related molecules and biochemical reactions in brain cells，many approaches to interrupt injurious cellular and molecular processes were applied to clinical trials［19-21］. These methods have not shown the fully improvement of stroke patients［19-20，21-22］. The study of other mechanisms underlying ischemic stroke is critically needed.

Cerebral ischemia leads to the pathological changes of neurons［23-25］and the deficiency of neurovascular unit，neuron-astrocyte-endothelium［26-28］. Their ischemic malfunctions are usually ahead of their morphological impairments［29-35］. In general，the malfunction of these cells makes unable to maintain their structure，and their structural collapse leads to a complete loss of cellular functions. Therefore，the elucidation of the mechanisms underlying the dysfunctions of the brain cells in an early stage of ischemic stroke appears to be more important.

Although multiple signaling pathways in ischemia interactively initiate an impairment of nerve cells［36-38］，neuronal excitotoxicity in its early stage is presumably a major reason to drive brain cells into injury. The evidences for this assumption include an elevated glutamate action because of its high release and/or injured reuptake［35，39-42］，an elevated excitatory synaptic transmission［43-44］，an increased neuronal excitability at cortical pyramidal neurons［35］and a decreased function of inhibitory interneurons［35，41，45-47］. As GABAergic cells are active in the central nervous system［48-50］，they are vulnerable to the shortage of oxygen/glucose and the accumulation of waste products in ischemia. A study indicates that the dysfunction of cortical inhibitory neurons appears ahead of pyramidal neurons in an early phase of ischemia and that this inhibition vulnerability to ischemia facilitates neuronal overexcitation［35］.Therefore，the elucidation of the mechanisms underlying the vulnerability of inhibitory neurons during ischemia is critical for neuronal protection. The protection of inhibitory neurons from ischemic vulnerability is likely therapeutic strategy to prevent brain cells from ischemic stroke.

It is pointed out that molecular processes and signaling cascades，such as an activation of Ca2+-permeable glutamate receptors［51-53］，an increase of intracellular Ca2+［54-57］and a production of intracellular toxic molecules［58-60］that lead to ischemic neuronal excitotoxicity，are present in GABAergic neurons［61-64］.Whether these processes are involved in ischemiainduced dysfunction of GABAergic neurons remains to be studied.

These studies indicate that the impairment of GABAergic neurons is likely an important step for neuronal excitotoxicity in the early phase of ischemic stroke，subsequently leading to brain cell death. In order to facilitate the development of therapeutic strategies via securing GABAergic neurons，we focus on reviewing current knowledge related to the ischemic injury of cerebral GABAergic neurons in terms of their functional impairment，molecular mechanisms and pathological consequence.

1　The characteristics of ischemic impairment in different types of brain cells

The neurons，glia cells and vascular endothelium cells in the brain constitute a neuron-astrocyteendothelium unit［65-67］to maintain neuronal spike encoding［68-70］. Their ischemic impairments lead to clinical neuropathy［71-72］. As ischemic brain tissues are limited to be obtained from humans，the ischemiainduced injury of the brain cells is studied in animal models by bilateral carotid artery occlusion （BCAO） or middle cerebral artery occlusion （MCAO）.

1.1　Morphological changes of brain cells during ischemia

In terms of morphological impairments in brain cells，the study by magnetic resonance imaging （MRI）shows the dilation of lateral ventricles after ischemia，indicating the shrink of brain tissues［73］. Under the light microscope，the densities of cortical neurons and white matter in ischemic animals are lower than those in controls，and small vessels are dilated. These results indicate that cerebral ischemia impairs the neurons and glia cells［72-75］. Moreover，the studies by scanning and transmission electron microscopy show that the veins and arteries appear the generalized vasoconstriction 10 min after ischemia，such as the reshuffle of smooth muscle cells，the thickness of vessel base-membrane and the increase of endothelium microfilament. This vasospastic change may worsen the ischemic injury of brain cells［76-77］.

Ischemic neural impairments are characterized as apoptosis and subsequent necrosis of brain cells［22，38，78-81］.Ischemia induces the upregulation of phosphorylated c-Jun in a proportion of morphologically intact hippocampal neurons and the subsequent apoptosis-like change，such as the increases of TUNEL-positive cells and caspase-3 activation［33，82-85］. These data indicate that neuron death in hippocampus after transient forebrain ischemia may be initiated via c-Jun N-terminal kinase activation and apoptosis-mediated pathways.Apoptosis-like change contributes to the subsequent necrosis of brain cells［86］.

To temporal changes in the ischemic injuries of different types of cerebral neurons，the studies by morphology and biochemistry indicate that a generalized arrest of protein synthesis occurs shortly after ischemia.The arrest and recovery of protein synthesis appear more quickly in hippocampal interneurons than pyramidal cells［87］. An investigation by labeling GABAergic interneurons shows that transient cerebral ischemia leads to severe structural abnormality in hippocampal interneurons and delayed pyramidal cell degeneration，which is prevented by diazepam［88］. Therefore，hippocampal inhibitory neurons are likely vulnerable to cerebral ischemic damage，compared with pyramidal neurons.

1.2　Functional impairment of brain cells during ischemia

Well-organized brain functions are based on the coordinated activity of neurons and glia cells in cerebral networks. These cells and their synapses cost much energy to maintain membrane potentials，signal generation and transmission. The shortage of O2and nutrients and the accumulation of metabolic products in ischemia rapidly influence synaptic signal transmission and neuronal spike coding. Although the attenuation of neural activities lowers energy needs to maintain minimal metabolism for cell survival，this minimization can only be sustained for a few minutes. A prolonged ischemia may lead to progressive membrane depolarization and synapse dysfunction，which then elevates intracellular Ca2+via influx and internal store release. Before it reaches a critical level to trigger irreversible processes for cell death，the restore of energy supply may reactivate the membrane pumps to reestablish normal ionic gradients and membrane potentials，which facilitates the return of synaptic and neuronal functions. The protection of brain cells from immediate ischemic dysfunctions and the promotion of their reversible recovery through elevating irreversible threshold and delaying irreversible cascades should be the future efforts.

The functional injury of nerve cells occurs quickly in a range of minutes after ischemia［89］，compared with their morphological changes［90］，which is consistent with the following facts. The brain functions in humans impair immediately after ischemic stroke. Neuronal functions supported by ionic channels and transporters are quite sensitive to the changes of biochemical reactions，especially the processes for adenosine-5’-triphosphate production. Therefore，the aims to measure the functions of brain cells during ischemia and to address the mechanisms underlying their dysfunction are critically important［91］. The pharmacological treatments to such early events may be the most effective approach to control ischemia and to attenuate long-term neurological deficits［92］. Recent studies about ischemia suggest a wide spectrum of agents potentially capable of delaying or even preventing irreversible outcomes of brain ischemia［93］.

The functional measurements of brain cells in the study of ischemia include neuronal excitability，synaptic transmission and glutamate transporter by electrophysiology. In addition to permanent functional impairment by ischemic structure changes，ischemia in the minutes leads to prominent functional injuries including membrane depolarization，inhibitory synaptic depression and excitatory synaptic enhancement［30，94-98］. These processes are likely the trigger signals for neuronal excitotoxicity，irreversible cellular dysfunction and cell death［35］. This complexity in functional changes requires an early and broad spectrum of interventions at different ligand-activated receptors and voltagedependent channels for the neuroprotection before neuronal damage expands to intact regions［92］.However，these studies did not compare what kinds of cerebral neurons wee functionally vulnerable to ischemia.

2　Temporal changes among the different types of brain cells during ischemia

Neural excitotoxicity is presumably an early event to initiate neural injury during ischemic stroke［54，56，99］. Its molecular mechanisms include the malfunction of glutamate transporters to elevate glutamate actions［39-40，42，46，100-101］and the glutamate-dependent elevations of intracellular Ca2+and free radicals［57，72，81，102-104］. In terms of processes for neuronal excitotoxicity，the enhancements of signal transmission at excitatory synapses［43-44］and neuronal excitability at cortical pyramidal neurons［35］as well as the attenuation of GABAergic neuron functions［35，45-46，105］are presumably involved. That is，an excessive release of excitatory amino acids and a reduced neuronal inhibition occur in brain ischemia. However，the temporal changes of these processes have not been addressed well.

2.1　Ischemic injury of GABAergic neurons

The functions of the neurons in the central nervous system include the reception of signals from presynaptic inputs，the integration of multiple synaptic signals，the generation of sequential spikes and the output of these digital spike codes to in fl uence their downstream cells［70，106］. The analyses for the functions of GABAergic cells should include their sensitivity to excitatory synaptic input，the ability of encoding spikes and the output of inhibitory synaptic activity on the downstream neurons.

To examine ischemia-induced alternations in the output ability of GABAergic neurons，inhibitory postsynaptic potentials （IPSPs） are measured on their postsynaptic neurons. GABAergic IPSPs on cortical［35］or hippocampal［94］pyramidal cells are depressed rapidly after ischemia within a few minutes. On the other hand，in an in vivo model of transient cerebral ischemia，GABAergic IPSPs on large aspiny neurons are enhanced due to an increase of presynaptic release in 3～24 hours after ischemia，and lowered 24 hours after ischemia［107］. This early depression of GABAergic synapses may be related to the initiation of neuronal excitotoxicity，whereas an enhancement of inhibitory synapses is likely due to the in vivo intracranial anastomotic circulation or the compensatory processes in the brain. These data indicate that ischemia leads to an inability of GABAergic neurons to send their inhibitory outputs. This indication has been granted by examining ischemic changes in GABA synthesis，release，reuptake，GABAAR expression in in vitro and in vivo models of ischemia［97］.

The sensitivity of GABAergic neurons to their synaptic inputs and the ability of encoding spikes are examined on their somata. Ischemia attenuates both neuronal spike encoding and excitatory synaptic transmission on cortical GABAergic neurons［46］.This indication is consistent with other studies. For instance，ischemia reduces the ability of firing spikes at cortical GABAergic neurons through elevating spike thresholds and prolonging refractory periods［35，45］.Cerebral ischemia makes the excitability of hippocampal CA1 interneurons declined by raising action potential threshold and shifting input-output curve rightward. Such changes are associated with the ischemic inactivation of voltage-gated Na+channel （Nav） and the reduction of Nav1.1 subunit expression at parvalbumin-coexpressed GABAergic neurons［108］. These results present a diagram that ischemia leads to the dysfunction of GABAergic neurons in aspects of the reception of signals from presynaptic inputs，the ability of encoding spikes and the output of controlling their downstream cells.

The role of GABAergic dysfunction in ischemiainduced neuronal death is to drive the balance between inhibition and excitation toward neuronal excitotoxicity，and then initiate the signal pathways for cerebral ischemic stroke［35，97］. This hypothesis is also supported by seeing the neuroprotective effects，when GABAergic neurotransmission is strengthened by preventing GABA reuptake and analytic metabolism or increasing GABAAreceptor activities［47，88，97，109-110］. In addition，antiepileptic drugs by acting on the presynaptic and postsynaptic GABAergic synapses to enhance neuronal inhibition are proved to counteract abnormal brain excitability in ischemia. These experimental studies utilizing global or focal ischemia in rodents bring the insights into possible neuroprotective action of inhibition enhancers［111］.

2.2　Ischemic injury of glutamatergic neurons

The influences of ischemia on the excitability and synaptic transmission of excitatory neurons are analyzed by approaches similar to the functional studies at inhibitory interneurons. Ischemia increases the excitability of cortical pyramidal neurons，including the attenuation of spike thresholds and the elevation of firing sequential spikes［35］. In animal models of in vivo transient cerebral ischemia，hippocampal pyramidal neurons appear a raised membrane input resistance and time constant，which is associated with a decrease of hyperpolarization-activated cationic currents （Ih）. The Ih enhancement by pharmacological approach immediately after ischemia attenuates the ischemic loss of CA1 neurons［30］. The results indicate that transient cerebral ischemia increases the intrinsic excitability of pyramidal neurons by upregulating voltage-gated sodium channels（VGSCs） and downregulating Ih. The raised excitability of cortical and hippocampal pyramidal neurons in ischemia exacerbates glutamatergic excitotoxicity and subsequent neuronal loss.

Ischemia-induced changes in glutamatergic neurotransmission have received widespread attention［97］.Early indication for an enhanced glutamate release during ischemia is from the measurement of excitatory neurotransmitter by micro-dialysis. The extracellular concentration of glutamates and aspartate is elevated in the hippocampus and cortex after transient cerebral ischemia［96，112］. This indication is also examined by measuring the excitatory postsynaptic potentials（EPSP），in which spontaneous glutamate release is spike-dependent and spike-independent processes. In the progress of ischemia，the patterns of spontaneous glutamate release in hippocampus change from spikeindependent release to large amplitude spike-dependent release［35，42］. Glutamates in spike-dependent release increase may be due to neuronal hyper-excitability during ischemia，which act onto NMDA receptors and metabotropic receptors to raise neuronal excitability.Therefore，the increases of neuronal hyper-excitability and glutamate release may synergistically accelerate the development of neural excitotoxicity.

The involvement of metabotropic glutamate receptors （mGluR） in ischemic neural excitotoxicity has been received a great attention. The antagonists of mGluRs，such as MCPG for mGluR1/5，AIDC and LY367385 for mGluR1，and MPEP for mGluR5，are used to examine the effect of mGluRs on ischemic stroke.The blockade of mGluR5 activation lowers the volume of ischemic infarct［113］. The activation of mGluR5 triggers the release of free radicals for neuronal death after transient ischemia［114］. The blockade of mGluR1 activation reduces the volume of ischemic infarction by lowering PKCγ/Src signal pathway，PI3K-Atk pathway and NMDAR phosphorylation［115-117］. As we known，an activation of group I mGluR activates G-protein coupled-phospholipase C to release 1，2-diacylglycerol and activate protein kinase C. These results suggest that mGluR1/5 in the brain is involved in NMDAR phosphorylation via PKCγ/Src family kinase cascade for neural excitotoxicity after transient ischemia.

On the other hand，the agonists of mGluR1/5，such as CHPG and DHPG，are used to examine the effect of these receptors on ischemic stroke. CHPG reduces infarct volume and neurological deficit［113］.The application of DHPG leads to diverse effects on cortical neurons. DHPG potentiates NMDA-induced neuronal death in cultured cortical neurons by elevating glutamate release［118］. DHPG enhances neuronal injury under oxygen-glucose deprivation （OGD） to cortical and hippocampal cultures［119］. However，DHPG reduces OGD-induced neurotoxicity by inducing ischemic tolerance in organotypic rat hippocampal slices［120］.

The reasons for these controversies about the role of mGluR in ischemia may be followings. The preparations used for experiments are different，such as in vivo，acute slices and tissue-cell culture.The conditions to induce neuronal injury are variable，such as the applications of NMDA，OGD，lower ACSF perfusion and blood vessel occlusion，which may act onto cellular targets differently regulated by group I mGluRs. The patterns of applying reagents are variable.For example，the study demonstrates that DHPG ampli fi es NMDA-induced neural injury，but two consecutive uses of DHPG produce neuroprotection，indicating that group I mGluR is an endogenous switch for a “neuroprotective mode”，and mGluRs are subjected to activity-dependent switch in regulating excitotoxic neuronal death［121］. In addition，the study with in situ hybridization indicates that mGluR1 mRNA level is lower in ischemic animals，but mGluR5 mRNA appears a relatively high expression. The difference in mGluR expression during ischemic stroke indicates that these receptors play different functions in neuronal excitotoxicity［122］.

The dysfunction of GABAergic neurons and the enhancement of excitatory cellular functions in an early phase of cerebral ischemia may shift the balance between excitation and inhibition toward hyper-excitation for subsequent neuronal excitotoxicity. This indication remains to be systemically examined in a single experiment in terms of their spatial and temporal changes［35］to determine the time sequence of treatments to protect different kinds of brain cells.

2.3　Ischemic functional injury of glia cells

The astrocytes regulate ionic homeostasis and extracellular glutamate levels through their Na+/K+ATPase pumps that consume ATP. On the other hand，the astrocytes have substantial oxidative capacity，compared to the neurons，to upregulate glycolysis by its powerful activator，which endows the astrocytes to sustain ATP production［123］. Therefore，the astrocytes are essential for neuronal survival and function，as well as are vulnerable to ischemia and oxidative stress. It is proposed that there is an interaction between glia cells and neurons in ischemia［124］，and that the astrocytes play the diverse and important role in ischemic brain damage［125］. In addition to the ischemic injury of neuronal functions，the astrocytes undergo the ischemic dysfunction for destructions of brain tissues［126］. For instance，the astrocytes store glycogen that protects the brain against hypoglycemic damage，but aggravates brain damage during ischemia due to an enhanced lactic acidosis. The astrocytes release substances （such as glutamate，D-serine and adenosine） for ischemic brain damage. The gap junctions in astrocytic networks may spread the molecules to healthy regions leading to infarct expansion in stroke.

The GFAP-negative astrocytes in gray matter are presumably more sensitive to ischemia than the neurons in vivo. They undergo the programmed cell death in ischemia，acidosis and oxidative stress via the upregulation of caspases，since the astrocyte survival is improved by inhibiting caspase-dependent and polyADP-ribose-1 cell death pathways［79］. To the molecular targets for ischemic astrocyte dysfunction，the investigations for the ability of the astrocytes to reuptake glutamates with in situ hybridization and immunohistochemistry demonstrate that GLT1-mRNA and -proteins are decreased after cerebral ischemia.There is a positive correlation between the number of pyramidal neurons and the expression of GLT1-mRNA/protein. These data indicate that the expression of GLT1 in hippocampus is downregulated in response to ischemia［127］. In terms of the mechanisms of ischemic astrocyte injury，the disruption of Ca2+homeostasis，the generation of free radicals and the depolarization of mitochondrial membrane are involved in to apoptotic astrocyte death［123］.

The ischemic attenuation of ATP production reduces the clearance of glutamates from synaptic cleft by their transporters located in neurons and glia cells. The failure of glutamate removal results in glutamate accumulation and subsequent neuronal excitotoxicity. Such a scenario is associated with brain ischemia and hypoglycemia due to the prompt decline in ATP level. The contribution of glial vs. neuronal glutamate transporters and the involvement of distinct glutamate transporter subtypes in ischemic damage remain to be addressed［40］.

The vulnerability of oligodendrocytes to ischemia may contribute to the malfunction of central white matter.The studies with LV-MBP-GFP labeled oligodendrocytes show a progressive loss of GFP+ oligodendrocytes in white matter 24 hours after ischemia. One week after stroke，the restoration of GFP+ oligodendrocytes is seen in ischemic white matter. These results indicate transient structural deterioration in oligodendrocyte cell bodies and myelinating processes in ischemia［128］.

It is noteworthy that these studies do not indicate the time sequence of functional injury in these types of nerve cells，such that we are not able to estimate what kinds of nerve cells change early or late during ischemia and to determine what kinds of nerve cells should be protected in time windows. We have studied the functions of astrocytic glutamate transporters and GABAergic neurons during ischemia. As showed in a recent study［89］，glutamate transporter currents at the astrocytes decrease significantly one minute after ischemia，whereas synaptic currents and intrinsic spikes at GABAergic cells signi fi cantly decrease three minutes after ischemia. That is，the dysfunction of glutamate transporters at the astrocytes is ahead of the deterioration of intrinsic property and synaptic transmission on GABAergic neurons. These results indicate the importance in an early protection of astrocytes during ischemia.

3　The impairment characteristics of GABAergic neurons during ischemia

As one group of function-speci fi c neurons［48，63，70，129-133］，GABAergic inhibitory interneurons receive and integrate signals from their presynaptic inputs，encode neuronal spikes and inhibit their downstream nerve cells. The analyses of functional changes in GABAergic neurons during ischemia include their sensitivity to excitatory synaptic inputs （i.e.，excitatory postsynaptic potential），the ability of encoding spikes （spike pattern） and the output of inhibitory synaptic activity （i.e.，inhibitory postsynaptic potentials and GABA release） on their downstream neurons.

3.1　The ischemic changes of active intrinsic properties

The parameters used to present the active intrinsic properties of GABAergic neurons include the ability of converting synaptic inputs into output spikes，the intervals of inter-spike or frequency as well as the threshold and refractory periods of sequential spikes［68，134］. These active membrane properties are essentially controlled by VGSCs［135-136］. VGSC activities are regulated by intracellular Ca2+signals and protein kinases［137-138］，which are elevated during ischemic stroke［56，138-141］.

The studies with in vitro ischemia demonstrate that cerebral ischemia for five minutes impairs the ability of cortical GABAergic neurons to produce sequential spikes［35］. Cerebral ischemia makes the excitability of hippocampal interneurons declined，which is associated with shifting input-output curve toward right［108］，i.e.，reducing their ability of encoding spikes. This ischemic impairment in producing spikes is due to the elevation of spike thresholds and the prolongation of refractory periods at GABAergic neurons［35，45］，as well as the inactivation of VGSCs and the reduction of Nav1.1 subtype expression at parvalbumin-coexpressed GABAergic neurons［108］. Thus，ischemia impairs GABAergic neurons in terms of their encoding spikes，such that they are unable to inhibit their downstream nerve cells，leading to neuronal hyper-excitability.

In addition，the sensitivity of cerebral GABAergic neurons to their presynaptic inputs is impaired during ischemia. For instance，ischemia reduces excitatory synaptic transmission at cortical GABAergic cells［45-46］.The silence of GABAergic neurons in response to their synaptic inputs makes them to be initiated difficultly and slowly in maintaining the homeostatic balance of the brain.

3.2　The ischemic decrease of GABAergic synaptic transmission

The effect of ischemia on GABA release from inhibitory neurons are examined by measuring the expression of glutamate acid decarboxylase （GAD）that converts glutamate into GABA，the concentration of extracellular GABA and the signal transmission of GABAergic synapses. For instance，the studies with immunohistochemistry and western-blot show that ischemic excitotoxicity is linked with the decreases in the expression of GAD through a NMDAR-mediated process and in the processes of GABAergic neurons.These results indicate that ischemia reduces GABA synthesis and subsequent release［142］. This indication is proved by measuring GABA with micro-dialysis，in which the concentration of extracellular GABA is decreased during ischemia［107，143］.

Under the condition of stimulating GABA nerve terminals and recording inhibitory postsynaptic potentials （IPSPs），a brief in vitro ischemia invariably causes a marked depression of IPSC amplitude，and this downregulation is fully reversible after the removal of ischemic challenge［94］. A transient cerebral ischemia in vivo depresses GABAergic IPSPs on aspiny neurons following ischemia［107］. The associated changes in paired-pulse facilitation indicate presynaptic mechanism，for which endogenous adenosine appears responsible since it is prevented by using adenosine A1 receptor antagonists［94］. In addition，ischemia inhibits GABAergic inhibitory postsynaptic potentials （IPSPs）rapidly on cortical neurons［35］. These results suggest that ischemia leads to an immediate decrease of GABA release from inhibitory neurons.

In accordance with a decrease of GABA release from presynaptic terminals，ischemia appears to decline the expression of postsynaptic GABA receptors. For instance，the study by in situ hybridization reveals that the changes in the expression of the GABAARα1/β2 subunit mRNAs appear three phases，a rapid decrease within 1 hour，recovery in 4～12 hours and further decline over next three days［144］. The responses of cortical neurons to GABA appear a considerable rundown with time under anoxic condition through a process of decreasing ATP synthesis［145］. GABAergic synapses after in vivo ischemia appears reduced in their amplitudes of miniature inhibitory postsynaptic currents and GABA-evoked currents at hippocampal pyramidal neurons［146］. This can be explained by a reduced GABAAR sensitivity or clustering as well as possibly reduced GABA release. It is noteworthy that the expression of GABAAR mRNA decrease well before pyramidal cell degeneration［144］，indicating that the loss of GABAergic neurotransmission may contribute to the development of neuronal degeneration following cerebral ischemia.

3.3　The compensation of GABAergic neurons during ischemia

The ischemic decreases in the active intrinsic property and sensitivity of GABAergic neurons as well as in the signal transmission at GABAergic synapses lead to the imbalance between excitation and inhibition toward the excitotoxicity. On the other hand，this neural excitotoxicity may initiate homeostatic process to enhance inhibitory functions in the brain，a compensatory effect of GABAergic neurons，which is beneficial for the survival of the brain cells after an acute phase of cerebral ischemia. The compensation of inhibitory system in ischemia may include the increases of GABA release and postsynaptic receptors.

The compensatory presynaptic mechanisms are measured by immunohistochemistry to detect the expressions of GAD and synaptic vesicle proteins as well as by electrophysiological recording to analyze a frequency of synaptic events. For instance，cerebral ischemia leads to an increase of GAD67/65-positive neurons，in which some of GAD67-positive neurons are the phenotypic shift of pre-existing somatostatincontaining GABAergic neurons［147-148］. Cerebral ischemia by transient middle cerebral artery occlusion induces a decrease in mRNA expression of vesicular GABA transporters lasting for 3～72 hours，which is detected by immunohistochemistry and westernblot. These increased GABA release and decreased reuptake may leave the relatively high levels of GABA in synaptic cleft to strengthen inhibitory synaptic transmission after cerebral ischemia［149］.

The studies in a postsynaptic mechanism demonstrate that cerebral ischemia enhances inhibitory synaptic transmission in the late phase［148］. In addition，ischemia enhances GABAergic synaptic transmission on pyramidal neurons，mainly in the amplitudes of evoked inhibitory postsynaptic current and miniature inhibitory postsynaptic current 12 hours after ischemic procedure.This postsynaptic mechanism needs a tonic activation of adenosine A1 receptors［107，150］.

4　The mechanisms underlying ischemic injury of GABAergic neurons

The ischemic impairments of cerebral GABAergic neurons include the apoptosis and necrosis in their morphology as well as the dysfunction of their active intrinsic properties and synaptic transmissions. The mechanisms for the former case may be similar to apoptotic process for neuronal death. Here，we focus on reviewing the mechanisms underlying the functional impairment of GABAergic neurons during ischemia.Neuronal spike encoding and synaptic transmission rely on the membrane potentials［70，106］，which are basically controlled by three factors，the gradient of different ions，the functional status of ion channels and the function of ATPase-dependent ionic pumps.

Electrical and chemical gradients for ions in the neurons are maintained by ATPase-related ionic pumps on their membrane and glia cells. The functional status of ion channels is controlled by the kinetics of their channels and the properties of their gates（such as voltage-dependent and ligand-dependent）.All of these processes are regulated by intracellular signaling pathways［151-155］. In this regard，the functional impairment of GABAergic neurons may be caused by ischemic changes in their membrane，intracellular signaling and extracellular glia cells （Fig.1）.

4.1　The mechanisms mediated by intracellular signals

Fig.1 A diagram shows ischemia-induced impairment of cortical GABAergic neurons

The functional impairments of GABAergic neurons during cerebral ischemia are characterized as the attenuation in encoding spikes，releasing GABA and responding to their synaptic inputs. These pathophysiological changes are related to the membrane identities，such as ionic channels and ATPase-ion pumps［17，156］. These changes may be due to ischemic intracellular processes since there is no evidence about the direct influences of ischemia on ionic channels and pumps. Cellular pathogenesis during insufficient blood flows includes the lack of glucose/oxygen and an accumulation of metabolic toxic products，such as protons，Ca2+and free radicals，which may activate cytoplasm signaling pathways. Low ATP and high toxics subsequently influence membrane components to impair the functions of GABAergic cells. It is noteworthy that the volume of GABAergic neurons is relatively smaller and the activities of these neurons are higher，compared with pyramidal neurons. The low energy storage，high consumption and low buffer volume on the inside of GABAergic neurons make them being vulnerable to hazard situations，such as ischemia and oxidative stress.

The studies in the mechanism underlying ischemic functional impairment of GABAergic neurons are in an infant stage. The genetically labeling of GABAergic neurons in the central nervous system with green fluorescent protein confers the opportunity to accessing these cells and studying this topic，in which functional cellular imaging and electrophysiological recording plus specific treatments to GFP-labeled GABAergic neurons can be done definitely. Although the theoretical prediction in the mechanisms for the impairment of GABAergic neurons may be as many as those for other cells［57，102］，the evidences are still lack，which we summarize below.

Intracellular Ca2+elevation in cerebral ischemia leads to a functional impairment of GABAergic neurons. The evidences for this conclusion include the investigations by Ca2+imaging，intracellular Ca2+elevation and cheletor. Ca2+imaging in GFP-labeled GABAergic neurons indicate that ischemia induces an immediate increase of intracellular Ca2+，which is associated with the deterioration of neuronal spiking and receptor sensitivity［45-46］. Ischemia-induced malfunction of GABAergic cells is prevented by infusing Ca2+cheletor［105］. Moreover，an elevation of intracellular Ca2+impairs the ability of GABAergic neurons to produce spikes，and this change occludes ischemia-induced changes in the neuronal encoding each other［45，62］，indicating Ca2+elevation and ischemia share common pathways to impair GABAergic neurons. Therefore，intracellular Ca2+overload is a toxic factor for ischemiainduced dysfunction of GABAergic neurons.

Ischemic acidosis has been proposed to be a factor of impairing cerebral neurons［157-159］. The smaller volume of GABAergic neurons makes them to contain less acid-base buffers，such that these inhibitory neurons are vulnerable to intracellular acidosis. The study with proton imaging demonstrates that ischemia leads to the immediate increase of intracellular H+and the weakness of encoding spikes at cortical GABAergic cells［45］. Moreover，the impairments of GABAergic neuron functions by ischemia and cytoplasmic acidosis occlude each other［45］，indicating that cellular acidosis is involved in the functional impairment of GABAergic neurons during ischemia.

In addition，cerebral ischemia evokes an excessive release of presynaptic and astrocytic glutamate［96，112］. The accumulated glutamates in synaptic cleft may activate NMDAR and mGluR on GABAergic neurons. An activation of these receptors raises intracellular Ca2+and an elevated activity of GABAergic neurons strengthens proton production in their metabolism. Cellular Ca2+overload and acidosis may interact each other，e.g.，acidi fi cation activates acid-sensing ion channels that are highly permeable to Ca2+［158］. Finally，intracellular Ca2+overload and acidosis synergistically impair the function of cortical GABAergic neurons during ischemia［45］.

Furthermore，lowering extracellular pH signi fi cantly reduces the frequency and the amplitude of spontaneous inhibitory postsynaptic currents mediated by GABAAR.Therefore，extracellular acidosis in cerebral ischemia may impair GABAergic synaptic transmission through presynaptic and postsynaptic mechanisms［160］.

4.2　The mechanisms mediated by impairment of glia cells

One of major functions for the astrocytes in the central nervous system is to reuptake presynaptic released glutamates through glutamate transporters that require the sodium gradient established by Na+/K+ATPase pumps［65-66，161］.There is an interaction between the astrocytes and neurons during ischemia［124］. The astrocytes may undergo ischemic dysfunction［79，123，126-127］，and have diverse roles in ischemic brain injury through releasing glutamate，D-serine and adenosine［125］. Together these molecules with their attenuated ATP production，the astrocytes are unable to clear glutamates from synaptic cleft via their transporters. The failure of removing glutamates results in glutamate accumulation and subsequent neural excitotoxicity［40］. Despite its possibility，it is still not known whether this mechanism influences the function of GABAergic neurons during ischemia.

We currently observed a time sequence in the dysfunctions of astrocytes and GABAergic neurons. The function of glutamate transporters on the astrocytes appears reduced ahead of the impairments of encoding spikes and transmitting synaptic signals on GABAergic neurons.Moreover，the prevention of astrocytic dysfunction by DHPG （a speci fi c agonist of type Ⅰ/Ⅴ of mGluR） secures the function of GABAergic cells after ischemia［89］. These data indicate that an impairment of astrocytes leads to GABAergic dysfunction. Whether the activation of mGluR-I/V protects the GABAergic neurons directly remains to be studied.

Together the discussions above，we draw a diagram showing a temporal sequence of dysfunctions for glutamate transporters at the astrocytes and functions at GABAergic neurons in Figure 1. The relatively small volume of GABAergic neurons makes them to have a low level of storage in the energy and buffer systems，and their high level of activities raises their energy consume. These factors facilitate GABAergic neurons to be vulnerable to hazard conditions，such as ischemia，acid-base imbalance，oxidative stress and toxic environment. Ischemia may lead to the functional impairment of GABAergic neurons through the following steps. Ischemia impairs Glu-T function on astrocytes immediately after ischemia，since they are vulnerable to lack of energy providers. An accumulation of glutamates in synaptic cleft activates the GABAergic neurons via mGluR and NMDAR，which elevates intracellular Ca2+and cellular metabolism to facilitate the vulnerability of GABAergic neurons to ischemia.In the meantime，ischemia lowers ATP production in mitochondria of GABAergic cells，which in turn impairs molecular transportation related to ATPase-pump，e.g.，a reduction of Ca2+pumping back to mitochondria and out of cells. This process plus Ca2+release from mitochondria elevate intracellular Ca2+. In addition，intracellular protons （H+） increase in ischemic metabolism. The elevations of Ca2+and H+in cortical GABAergic neurons synergistically lead to their ischemic excitotoxicity and malfunction. The deterioration of inhibitory neurons shifts the balance between excitation and inhibition toward neuronal overexcitation through disinhibiting excitatory pyramidal neurons.

5　Pathological consequence of GABAergic neurons’impairment

The major functions of GABAergic inhibitory neurons in the central nervous system are believed to coordinate the activities of excitatory neurons，prevent neuronal over-excitation and increase neuronal sensitivity to their synaptic inputs［48，63，70，129-133，135，162-163］. The functional vulnerability of inhibitory neurons to ischemia causes the neural excitotoxicity and subsequent impairments of other brain cells. In addition to the sequential impairments of cerebral GABAergic neurons and excitatory neurons［35］，other evidences for this suggestion comes from the data that securing GABAergic function is able to prevent the ischemic impairment of the brain tissues. The strengthening of GABAergic inhibitory synaptic transmission can be realized by enhancing GABA receptor activation and by elevating GABA concentration in the synaptic cleft（prompting presynaptic GABA release and blocking GABA reuptake）.

The enhancers or agonists of GABAR to facilitate inhibitory synaptic transmission protect nerve cells from ischemic injury. For instance，GABAAR agonist muscimol increases the number of survived nerve cells and delays the process of neuronal death after cerebral ischemia［107，150］. GABABR agonists （such as baclofen and saclofen） reduce the effect of ischemia-induced glutamate release on neural damage［164］. The agonists of GABAAR and GABABR result in the neuroprotective effects against ischemia［165］. The enhancers of GABAR benzodiazepines （diazepam and imidazenil） reduce neuronal apoptosis and protect pyramidal cells from the toxic effect of transient cerebral ischemia［110，166］. In addition，the enhancement of GABAR function downregulates the excessive release of glutamates and the overactivation of GluRs triggered by cerebral ischemia［164］.

An increase of GABA concentration in synaptic cleft to facilitate inhibitory synaptic transmission protects nerve cells from ischemic injury. The inhibitors of GABA transporters to block GABA reuptake are used to strengthen inhibitory synaptic transmission.For instance，the inhibitors of GABA transporters（tiagabine and vigabatrin） prevent ischemia-induced neural malfunction［165］. GABA uptake inhibitor No-328 significantly decreases the ischemia-induced loss of cerebral neurons［166］.

These procedures for enhancing inhibitory synaptic transmission may shift an ischemic imbalance between neuronal excitation and inhibition toward the inhibitory status of brain functions，which may be beneficial for neuronal tolerance to hazard situations （such as ischemia and oxidative stress） and the brain protection itself［167］.

6　Conclusion remark

Cerebral GABAergic neurons appear vulnerable to ischemia，oxidative stress，acidosis and toxic molecules. Their vulnerability may be due to the relatively smaller volume and higher activity，compared with pyramidal neurons. The low storage and high consumption of energy as well as the low volume of intracellular buffer system make GABAergic neurons being vulnerable to the hazard internal environment.Insufficient blood flow initiates ischemic processes in brain cells，such as a lack of glucose/oxygen and an accumulation of metabolic toxic products （proton，Ca2+and free radicals），especially in GABAergic cells and astrocytes. Such changes activate cellular signaling pathways and influence membrane components in GABAergic neurons. Moreover，an ischemic failure for astrocytes to reuptake glutamates exacerbates the dysfunction of GABAergic neurons. Their dysfunctions in encoding spikes and transmitting synaptic signals may shift neural balance toward excitotoxicity，which leads to the ischemic stroke of nerve cells. Therefore，the study to understand the mechanisms underlying GABAergic neuronal vulnerability to toxic environments should provide the clues for developing therapeutic strategies in the protection of neuronal functions from ischemic injury.

［1］ Christopher R，Nagaraja D，Shankar S K. Homocysteine and cerebral stroke in developing countries［J］. Curr Med Chem，2007，14（22）：2393-2401.

［2］ Francesco Della Corte，Gian Luca Vignazia，M Cavaglia，et al. Stroke patients，what to do and what to avoid［J］.Minerva Anestesiol，2002，68（4）：273-277.

［3］ Michael Brainin，Natan M Bornstein，Gudrun Boysen，et al. Acute neurological stroke care in Europe：results of the European Stroke Care Inventory［J］. Eur J Neurol，2000，7（1）：5-10.

［4］ Kuller L H. Epidemiology and prevention of stroke，now and in the future［J］. Epidemiol Rev，2000，22（1）：14-17.

［5］ Timothy Ingall. Stroke—incidence，mortality，morbidity and risk［J］. J Insur Med，2004，36（2）：143-152.

［6］ Antonio Pinto，Antonino Tuttolomando，Domenico Di Raimondo，et al. Cerebrovascular risk factors and clinical classification of strokes［J］. Semin Vasc Med，2004，4（3）：287-303.

［7］ Gustavo Saposnik，Oscar Del Brutto. Stroke in South America：a systematic review of incidence，prevalence，and stroke subtypes［J］. Stroke，2003，34（9）：2103-2107.

［8］ Angel Chamorro. Immediate anticoagulation in acute focal brain ischemia revisited：gathering the evidence［J］.Stroke，2001，32（2）：577-578.

［9］ Dalkara，T，Michael A Moskowitz. Recent developments in the experimental stroke［J］. NeuroScience News，1999，2（5）：20-27.

［10］ Devasenapathy A，V C Hachinski. Current treatment of acute ischemic stroke ［J］. NeuroScience News，1999，2（5）：4-13.

［11］ Sebastian Jander，Michael Schroeter，Andreas Saleh.Imaging inflammation in acute brain ischemia［J］.Stroke，2007，38（2 Suppl）：642-645.

［12］ Christopher C Leonardo，Keith R Pennypacker.Neuroinflammation and MMPs：potential thera peutic targets in neonatal hypoxic-ischemic injury［J］. J Neuroinflammation，2009，6：13.

［13］ Macchi，L，Nathalie Sorel，Luc Christiaens. Aspirin resistance：definitions，mechanisms，prevalence，and clinical significance［J］. Curr Pharm Des，2006，12（2）：251-258.

［14］ Keith W Muir，M Roberts. Thrombolytic therapy for stroke：a review with particular reference to elderly patients［J］. Drugs Aging，2000，16（1）：41-54.

［15］ Alfredo Puca. Thrombolysis in cerebral ischemia. A review of clinical and experimental data［J］. J Neurosurg Sci，1993，37（2）：63-70.

［16］ Bruce D Spiess. Ischemia--a coagulation problem? ［J］ J Cardiovasc Pharmacol，1996，27（ Suppl 1）：S38-S41.

［17］ Antonino Tuttolomondo，Riccardo Di Sciacca，et al.Neuron protection as a therapeutic target in acute ischemic stroke［J］. Curr Top Med Chem，2009，9（14）：1317-1334.

［18］ Vaughan C J，N Delanty. Neuroprotective properties of statins in cerebral ischemia and stroke［J］. Stroke，1999，30（9）：1969-1973.

［19］ Myron D Ginsberg. Neuroprotection for ischemic stroke：past，present and future［J］. Neuropharmacology，2008，55（3）：26.

［20］ Myron D Ginsberg. Current status of neuroprotection for cerebral ischemia：synoptic overview［J］. Stroke，2009，40（3 Suppl）：S111-S114.

［21］ Era Taoufik，Lesley Probert. Ischemic neuronal damage［J］. Curr Pharm Des，2008，14（33）：3565-3573.

［22］ Eduardo Candelario-Jalil. Injury and repair mechanisms in ischemic stroke：considerations for the development of novel neurotherapeutics［J］. Curr Opin Investig Drugs，2009，10（7）：644-654.

［23］ Tobias Back，O G Schuler. The natural course of lesion development in brain ischemia［J］. Acta Neurochir Suppl，2004，89：55-61.

［24］ Tobias Back，Thomas Hemmen，Olaf G Schuler. Lesion evolution in cerebral ischemia ［J］. J Neurol，2004，251（4）：388-397.

［25］ Matthias Endres，Ulrich Dirnagl. Ischemia and stroke ［J］.Adv Exp Med Biol，2002，513：455-473.

［26］ Curin Y，Marie-Francoise Ritz，Ramaroson Andriantsitohaina. Andriantsitohaina，Cellular mechanisms of the protective effect of polyphenols on the neurovascular unit in strokes［J］. Cardiovasc Hematol Agents Med Chem，2006，4（4）：277-288.

［27］ Panickar K S，Norenberg M D，Astrocytes in cerebral ischemic injury：morphological and general considerations［J］. Glia，2005，50（4）：287-298.

［28］ Swanson R A，Ying W，Kauppinen T M. Astrocyte influences on ischemic neuronal death［J］. Curr Mol Med，2004，4（2）：193-205.

［29］ Brouns R，P P De Deyn. The complexity of neurobiological processes in acute ischemic stroke［J］. Clin Neurol Neurosurg，2009，111（6）：483-495.

［30］ Fan Yuan，Deng Ping，Wang Yu-Chi，et al. Transient cerebral ischemia increases CA1 pyramidal neuron excitability［J］. Exp Neurol，2008，212（2）：415-421.

［31］ Johansen F F. Interneurons in rat hippocampus after cerebral ischemia. Morphometric，functional，and therapeutic investigations［J］. Acta Neurol Scand Suppl，1993，150：1-32.

［32］ Mergenthaler，P.，U. Dirnagl，and A. Meisel，Pathophysiology of stroke：lessons from animal models［J］.Metab Brain Dis，2004，19（3-4）：151-167.

［33］ Georg Johannes Muller，Christine Stadelmann，Lone Bastholm，et al. Ischemia leads to apoptosis--and necrosislike neuron death in the ischemic rat hippocampus［J］.Brain Pathol，2004，14（4）：415-424.

［34］ Alexander G Nikonenko，Lidijia Radenovic，Pavle Andjus，et al. Structural features of ischemic damage in the hippocampus［J］. Anat Rec （Hoboken），2009，292（12）：1914-1921.

［35］ Wang Jin-Hui. Short-term cerebral ischemia causes the dysfunction of interneurons and more excitation of pyramidal neurons［J］. Brain Research Bulletin，2003，60（1-2）：53-58.

［36］ Castellanos M，Serena J. Applicability of biomarkers in ischemic stroke［J］. Cerebrovasc Dis，2007，24 （Suppl 1）：7-15.

［37］ Kristian P Doyle，Roger P Simon，Mary P Stenzel-Poore. Mechanisms of ischemic brain damage［J］.Neuropharmacology，2008，55（3）：310-318.

［38］ Hou Sheng-tao，John P MacManus. Molecular mechanisms of cerebral ischemia-induced neuronal death［J］. Int Rev Cytol，2002，221：93-148.

［39］ Alberto Camacho，Lourdes Massieu. Role of glutamate transporters in the clearance and release of glutamate during ischemia and its relation to neuronal death［J］.Arch Med Res，2006，37（1）：11-18.

［40］ Leif Hertz. Bioenergetics of cerebral ischemia：a cellular perspective［J］. Neuropharmacology，2008，55（3）：289-309.

［41］ Huang Li，Wang Chun，Shidi Zhao，et al. PKC and CaMKII inhibitions coordinately rescue ischemia-induced GABAergic neuron dysfunction［J］. Oncotarget，2017，DOI：10.18632/oncotarget.16947.

［42］ Ye Hui，Shirin Jalini，Zhang Liang，et al. Early ischemia enhances action potential-dependent，spontaneous glutamatergic responses in CA1 neurons［J］. J Cereb Blood Flow Metab，2010，30（3）：555-565.

［43］ Michelle Aarts，Mark Arundine，Michael Tymianski. Novel concepts in excitotoxic neurodegeneration after stroke［J］.Expert Rev Mol Med，2003，5（30）：1-22.

［44］ Arabadzisz D，F(xiàn)reund T F. Changes in excitatory and inhibitory circuits of the rat hippocampus 12-14 months after complete forebrain ischemia［J］. Neuroscience，1999，92（1）：27-45.

［45］ Li Huang，Chen Na，Ge Ming，et al. Ca2+and acidosis synergistically lead to the dysfunction of cortical GABAergic neurons during ischemia［J］. Biochemical and Biophysical Research Communications，2010，394：709-714.

［46］ Li Huang，Zhao Shi-di，Lu Wei，et al.，Acidosis-Induced Dysfunction of Cortical GABAergic Neurons through Astrocyte-Related Excitotoxicity［J］. PLoS One，2015，10（10）：e0140324.

［47］ Ashfaq Shuaib，Myungsook Breker-Klassen. Inhibitory mechanisms in cerebral ischemia：a brief review［J］.Neuroscience Biobehavior Review，1997，21（2）：219-226.

［48］ Freund T F，Buzsaki G，Interneurons of the hippocampus［J］. Hippocampus，1996，6：347-470.

［49］ Lu Wei，Wen Bo，Zhang Feng-yu，et al. Voltageindependent sodium channels emerge for an expression of activity-induced spontaneous spikes in GABAergic neurons［J］. Mol Brain，2014，7（1）：38.

［50］ Yu Jian-dong，Qian Hao，Wang Jin-hui. Upregulation of transmitter release probability improves a conversion of synaptic analogue signals into neuronal digital spikes［J］.Mol Brain，2012，5（1）：26.

［51］ Michelle M Aarts，Michael Tymianski. Novel treatment of excitotoxicity：targeted disruption of intracellular signalling from glutamate receptors［J］. Biochem Pharmacol，2003，66（6）：877-886.

［52］ James J P Alix. Recent biochemical advances in white matter ischaemia［J］. Eur Neurol，2006，56（2）：74-77.

［53］ Seok Joon Won，Doo Yeon Kim，Byoung Joo Gwag.Cellular and molecular pathways of ischemic neuronal death［J］. J Biochem Mol Biol，2002，35（1）：67-86.

［54］ Mark Arundine，Michael Tymianski. Molecular mechanisms of calcium-dependent neurodegeneration in excitotoxicity［J］. Cell Calcium，2003，34（4-5）：325-337.

［55］ Daniele Bano，Pierluigi Nicotera. Ca2+signals and neuronal death in brain ischemia［J］. Stroke，2007，38（2 Suppl）：674-676.

［56］ Kinga Szydlowska，Michael Tymianski. Calcium，ischemia and excitotoxicity［J］. Cell Calcium，2010，47（2）：122-129.

［57］ Blaine C White，Jonathon M Sullivan，Donald J DeGracia，et al. Brain ischemia and reperfusion：molecular mechanisms of neuronal injury［J］. J Neurological Sciences，2000，179（1-2）：1-33.

［58］ Teresa Carbonell，Ramon Rama. Iron，oxidative stress and early neurological deterioration in ischemic stroke［J］.Curr Med Chem，2007，14（8）：857-874.

［59］ Forder J P，M Tymianski. Postsynaptic mechanisms of excitotoxicity：Involvement of postsynaptic density proteins，radicals，and oxidant molecules［J］.Neuroscience，2009，158（1）：293-300.

［60］ Schmidley J W. Free radicals in central nervous system ischemia［J］. Stroke，1990，21（7）：1086-1090.

［61］ Lionel Carmant，Gavin Woodhal，Mohamed Ouardouz，et al. Interneuron-sepcific Ca2+responses linked to metabotropic and ionotropic glutamate receptors in rat hippocampal slices［J］. Eur J Neurosci，1997，9：1625-1635.

［62］ Qi Yu-long，Li Huang，Ni Hong，et al. Intracellular Ca2+regulates spike encoding at cortical GABAergic neurons and cerebellar Purkinje cells differently［J］. Biochemical and Biophysical Research Communications，2009，381（1）：129-133.

［63］ Wang Jin-hui，Paul Kelly. Ca2+/CaM signalling pathway upregulates glutamatergic synaptic function in non-pyramidal fast-spiking neurons of hippocampal CA1［J］. J Physiol（Lond），2001，533（2）：407-422.

［64］ Wang Jin-hui，Zhang Mei. Differential modulation of glutamatergic and cholinergic synapses by calcineurin in hippocampal CA1 fast-spiking interneurons［J］. Brain Research，2004，1004（1-2）：125-135.

［65］ Eduardo E Benarroch. Neuron-astrocyte interactions：partnership for normal function and disease in the central nervous system［J］. Mayo Clin Proc，2005，80（10）：1326-1338.

［66］ Paola Bezzi，Maria Domercq，Sabino Vesce，et al. Neuronastrocyte cross-talk during synaptic transmission：physiological and neuropathological implications［J］.Prog Brain Res，2001，132：255-265.

［67］ Stevens B. Neuron-astrocyte signaling in the development and plasticity of neural circuits［J］. Neurosignals，2008，16（4）：278-288.

［68］ Chen Na，Chen Shu-li，et al.，The refractory periods and threshold potentials of sequential spikes measured by whole-cell recordings［J］. Biochemical and Biophysical Research Communications，2006，340：151-157.

［69］ Chen Na，Chen Xin，Wang Jin-hui. Homeostasis established by coordination of subcellular compartment plasticity improves spike encoding ［J］. J Cell Science，2008，121（17）：2961-2971.

［70］ Wang Jin-hui，Wei Jian，Chen Xin，et al. The gain and fidelity of transmission patterns at cortical excitatory unitary synapses improve spike encoding［J］. J Cell Science，2008，121（17）：2951-2960.

［71］ Block F. Global ischemia and behavioural deficits［J］.Prog Neurobiol，1999，58（3）：279-295.

［72］ Suzuki R，Yamaguchi T，Kirino T，et al. The effects of 5-minute ischemia in Mongolian gerbils：I. Blood-brain barrier，cerebral blood flow，and local cerebral glucose utilization changes［J］. Acta Neuropathol，1983，60（3-4）：207-216.

［73］ Miller B，Nagy B L Finlay，Chance B，et al. Consequences of reduced cerebral blood flow in brain development. I.Gross morphology，histology，and callosal connectivity ［J］.Exp Neurol，1993，124（2）：326-342.

［74］ Mennel H D，Sauer D Rossberg，Bielenberg G W，et al.Morphology of tissue damage due to experimental cerebral ischemia in rats［J］. Exp Pathol，1988，35（4）：219-230.

［75］ Mennel H D，H El-Abhar，M Schilling，et al. Morphology of tissue damage caused by permanent occlusion of middle cerebral artery in mice［J］. Exp Toxicol Pathol，2000，52（5）：395-404.

［76］ del Zoppo G J，Mabuchi T. Cerebral microvessel responses to focal ischemia［J］. J Cereb Blood Flow Metab，2003，23（8）：879-894.

［77］ Wisniewski H M，Ryszard Pluta，Albert Lossinsky，et al.Ultrastructural studies of cerebral vascular spasm after cardiac arrest-related global cerebral ischemia in rats［J］.Acta Neuropathol，1995，90（5）：432-440.

［78］ Brad R S Broughton，David C Reutens，Christopher G Sobey. Apoptotic mechanisms after cerebral ischemia［J］.Stroke，2009，40（5）：e331-339.

［79］ Rona G Giffard，Raymond A Swanson. Ischemia-induced programmed cell death in astrocytes［J］. Glia，2005，50（4）：299-306.

［80］ Suresh L Mehta，Namratta Manhas，Ram Raghubir.Molecular targets in cerebral ischemia for developing novel therapeutics［J］. Brain Res Rev，2007，54（1）：34-66.

［81］ Nakka V P，Gusain A，Mehta S L，et al. Molecular mechanisms of apoptosis in cerebral ischemia：multiple neuroprotective opportunities［J］. Mol Neurobiol，2008，37（1）：7-38.

［82］ Chen Jun，Jin Kun-lin，Chen Min-zhi，et al. Early detection of DNA strand breaks in the brain after transient focal ischemia：implications for the role of DNA damage in apoptosis and neuronal cell death［J］. J Neurochem，1997，69（1）：232-245.

［83］ Robert S B Clark，Chen Jun，Simon C Watkins，et al.Apoptosis-suppressor gene bcl-2 expression after traumatic brain injury in rats［J］. J Neurosci，1997，17（23）：9172-9182.

［84］ Nakashima K. Temporal and spatial profile of apoptotic cell death in transient intracerebral mass lesion of the rat［J］. J Neurotrauma，1999，16（2）：143-151.

［85］ Newcomb J K，Zhao X，Pike B R，et al. Temporal profile of apoptotic-like changes in neurons and astrocytes following controlled cortical impact injury in the rat［J］. Exp Neurol，1999，158（1）：76-88.

［86］ Pang Zhen，Vinala Bondada，Tomoko Sengoku，et al. Calpain facilitates the neuron death induced by 3-nitropropionic acid and contributes to the necrotic morphology［J］. J Neuropathol Exp Neurol，2003，62（6）：633-643.

［87］ Johansen F F，Diemer N H. Temporal profile of interneuron and pyramidal cell protein synthesis in rat hippocampus following cerebral ischemia［J］. Acta Neuropathol，1990，81（1）：14-19.

［88］ Jon R Inglefield，Christina A Wilson，Rochelle D Schwartz-Bloom。 Effect of transient cerebral ischemia on gamma-aminobutyric acidA receptor alpha 1-subunitimmunoreactive interneurons in the gerbil CA1 hippocampus［J］. Hippocampus，1997，7（5）：511-523.

［89］ Liu Zhan，Huo Wei，Sun Wei，et al.，A sequential impairment of cortical astrocytes and GABAergic neurons during ischemia is improved by mGluR（1），（5） activation［J］. Neurol Sci，2013，34（7）：1189-1195.

［90］ Irshad H Chaudry. Cellular mechanisms in shock and ischemia and their correction［J］. Am J Physiol，1983，245（2）：R117-134.

［91］ Dale Corbett，Suzanne Nurse. The problem of assessing effective neuroprotection in experimental cerebral ischemia［J］. Prog Neurobiol，1998，54（5）：531-548.

［92］ Heiko J Luhmann. Ischemia and lesion induced imbalances in cortical function［J］. Prog Neurobiol，1996，48（2）：131-166.

［93］ Kresimir Krnjevic. Electrophysiology of cerebral ischemia［J］. Neuropharmacology，2008，55（3）：319-333.

［94］ Diego Centonze，Saulle E，Pisani A，et al. Adenosinemediated inhibition of striatal GABAergic synaptic transmission during in vitro ischemia［J］. Brain，2001，124（Pt9）：1855-1865.

［95］ Martine Hamann，David J Rossi，Claudia Mohr，et al.，The electrical response of cerebellar Purkinje neurons to simulated ischemia［J］. Brain，2005，128：2408-2420.

［96］ Roettger V，Lipton P. Mechanism of glutamate release from rat hippocampal slices during in vitro ischemia［J］.Neuroscinece，1996，75（3）：677-685.

［97］ Schwartz-Bloom R D，Renu Sah. gamma-Aminobutyric acid（A） neurotransmission and cerebral ischemia［J］. J Neurochem，2001，77（2）：353-371.

［98］ Kortaro Tanaka. Alteration of second messengers during acute cerebral ischemia：adenylate cyclase，cyclic AMP-dependent protein kinase and cyclic AMP response element binding protein ［J］. Progress in Neurobiology，2001，65（2）：173-207.

［99］ Dennis W Choi. Calcium-mediated neurotoxicity：relationship to specific channel types and role in ischemic damage［J］. Trends Neurosci，1988，11（10）：465-469.

［100］ Inage Y W，Itoh M，Wada K，et al. Expression of two glutamate transporters，GLAST and EAAT4，in the human cerebellum：their correlation in develoment and neonatal hypoxia-ischemic damage［J］. J Neuropathology and Experimental Neurology，1998，57（6）：554-562.

［101］ Akihide Yamashita，Koshi Makita，Toshihiko Kuroiwa，et al. Glutamate transporters GLAST and EAAT4 regulate postischemic Purkinje cell death：an in vivo study using a cardiac arrest model in mice lacking GLAST or EAAT4 ［J］.Neuroscience Research，2006，55（3）：264-270.

［102］ Peter Lipton. Ischemic cell death in brain neurons［J］.Physiological Review，1999，79（4）：1431-1568.

［103］ Rochelle D Schwartz-Bloom，Renu Sah. r-aminobutyric acid A neurotransmission and cerebral ischemia［J］. J Neurochemistry，2001，77：353-371.

［104］ John P Welsh，Genevieve S Yuen，Dimitris Placantonakis，et al. Why do Purkinje cells die so easily after global brain ischemia? Aldolase C，EAAT4，and the cerebellar contribution to posthypoxic myoclonus［J］. Advanced Neurology，2002，89：331-359.

［105］ Zhao Shi-di，Chen Na，Yang Zhi-lai，et al. Ischemia deteriorates the spike encoding of rat cerebellar Purkinje cells by raising intracellular Ca2+［J］. Biochemical and Biophysical Research Communications，2008，366（2）：401-407.

［106］ Shepherd G M. Electronic properties of axons and dendrites［M］. From Molecular to Networks：An Introduction to Cellular and Molecular Neuroscience，J H Byrne，J L Roberts，2004，New York：Elsevier Science （USA）：91-113.

［107］ Li Y，Z Lei，Xu Z C. Enhancement of inhibitory synaptic transmission in large aspiny neurons after transient cerebral ischemia［J］. Neuroscience，2009，159（2）：670-681.

［108］ Zhan Ren-zhi，J Victor Nadler，Rochelle D Schwartz-Bloom. Impaired firing and sodium channel function in CA1 hippocampal interneurons after transient cerebral ischemia［J］. J Cereb Blood Flow Metab，2007，27（8）：1444-1452.

［109］ Makoto Saji，Melissa Cohen，Alan D Blau，et al. Transient forebrain ischemia induces delayed injury in the substantia nigra reticulata：degeneration of GABA neurons，compensatory expression of GAD mRNA［J］. Brain Res，1994，643（1-2）：234-244.

［110］ Schwartz-Bloom R D，Miller K A，Evenson D A，et al. Benzodiazepines protect hippocampal neurons from degeneration after transient cerebral ischemia：an ultrastructural study［J］. Neuroscience，2000，98（3）：471-484.

［111］ Paolo Calabresi，Letizia M Cupini，Diego Centonze，et al.Antiepileptic drugs as a possible neuroprotective strategy in brain ischemia［J］. Ann Neurol，2003，53（6）：693-702.

［112］ Helene Benveniste，Jorgen Drejer，Arne Schousboe，et al.，Elevation of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis［J］. J Neurochemistry，1984，43：1369-1374.

［113］ Bao W L，Williams A J，F(xiàn)aden A I，et al. Selective mGluR5 receptor antagonist or agonist provides neuroprotection in a rat model of focal cerebral ischemia［J］. Brain Res，2001，922（2）：173-179.

［114］ Rao Muralikrishna Adibhatla，James F Hatcher，R J Dempsey. Neuroprotection by group I metabotropic glutamate receptor antagonists in forebrain ischemia of gerbil［J］. Neurosci Lett，2000，293（1）：1-4.

［115］ Kazutoshi Murotomi，Norio Takagi，Gen Takayanagi，et al.mGluR1 antagonist decreases tyrosine phosphorylation of NMDA receptor and attenuates infarct size after transient focal cerebral ischemia［J］. J Neurochem，2008，105（5）：1625-1634.

［116］ Tania Scartabelli，Elisabetta Gerace，Elisa Landucci，et al. Neuroprotection by group I mGlu receptors in a rat hippocampal slice model of cerebral ischemia is associated with the PI3K-Akt signaling pathway：a novel postconditioning strategy? ［J］ Neuropharmacology，2008，55（4）：509-516.

［117］ Norio Takagi，Shintaro Besshoh，Hirotsugu Morita，et al. Metabotropic glutamate mGlu5 receptor-mediated serine phosphorylation of NMDA receptor subunit NR1 in hippocampal CA1 region after transient global ischemia in rats［J］. Eur J Pharmacol，2010，644（1-3）：96-100.

［118］ Uta Strasser，Doug Lobner，M Margarita Behrens，et al.Antagonists for group I mGluRs attenuate excitotoxic neuronal death in cortical cultures［J］. Eur J Neurosci，1998，10（9）：2848-2855.

［119］ Pellegrini-Giampietro D E，Cozzi A，Peruginelli F，et al. 1-Aminoindan-1，5-dicarboxylic acid and （S）-（+）-2-（3’-carboxybicyclo［1.1.1］ pentyl）-glycine，two mGlu1 receptor-preferring antagonists，reduce neuronal death in in vitro and in vivo models of cerebral ischaemia［J］. Eur J Neurosci，1999，11（10）：3637-3647.

［120］ Claudia G Werner，Tania Scartabelli，Tristano Pancani，et al. Differential role of mGlu1 and mGlu5 receptors in rat hippocampal slice models of ischemic tolerance［J］. Eur J Neurosci，2007，25（12）：3597-3604.

［121］ Valeria Bruno，Giuseppe Battaglia，Agatta Copani，et al.An activity-dependent switch from facilitation to inhibition in the control of excitotoxicity by group I metabotropic glutamate receptors［J］. Eur J Neurosci，2001，13（8）：1469-1478.

［122］ Agnes Simonyi，J P Zhang，Grace Sun，Changes in mRNA levels for group I metabotropic glutamate receptors following in utero hypoxia-ischemia［J］. Brain Res Dev Brain Res，1999，112（1）：31-37.

［123］ Gerald A Dienel，Leif Hertz. Astrocytic contributions to bioenergetics of cerebral ischemia［J］. Glia，2005，50（4）：362-388.

［124］ Michele Zoli，Giuseppe Biagini，Rosaria Ferrari，et al.Neuron-glia cross talk in rat striatum after transient forebrain ischemia［J］. Adv Exp Med Biol，1997，429：55-68.

［125］ David J Rossi，James D Brady，Claudia Mohr. Astrocyte metabolism and signaling during brain ischemia［J］. Nat Neurosci，2007，10（11）：1377-1386.

［126］ Sonia Villapol，Antoinette Gelot，Sylvain Renolleau，et al.Astrocyte responses after neonatal ischemia：the yin and the yang［J］. Neuroscientist，2008，14（4）：339-344.

［127］ Torben Bruhn，Line M Levy，Mogens Nielsen，et al.Ischemia induced changes in expression of the astrocyte glutamate transporter GLT1 in hippocampus of the rat［J］.Neurochem Int，2000，37（2-3）：277-285.

［128］ McIver S R，Muccigrosso M，Gonzales E R，et al.Oligodendrocyte degeneration and recovery after focal cerebral ischemia［J］. Neuroscience，2010，169（3）：1364-1375.

［129］ Bruce E McKay，Ray W Turner. Physiological and morphological development of the rat cerebellar Purkinje cell［J］. J Physiology （London），2005，567（Pt3）：829-850.

［130］ Chris J McBain，Andre Fisahn. Interneurons unbound［J］.Nature Reviews Neuroscience，2001，2（1）：11-23.

［131］ Ni Hong，Huang Li，Chen Na，et al. Upregulation of barrel GABAergic neurons is associated with cross-modal plasticity in olfactory deficit［J］. PLoS ONE，2010，5（10）：e13736.

［132］ Peter Somogyi，Thomas Klausberger. Defined types of cortical interneurone structure space and spike timing in the hippocampus ［J］. J Physiology （London），2005，562（1）：9-29.

［133］ Michael Wehr，Anthony M Zador. Balanced inhibition underlies tuning and sharpens spike timing in auditory cortex［J］. Nature，2003，426：442-446.

［134］ Chen Na，Zhu Yan，Gao Xin，et al. Sodium channelmediated intrinsic mechanisms underlying the differences of spike programming among GABAergic neurons［J］.Biochemical and Biophysical Research Communications，2006，346（1）：281-287.

［135］ Chen Na，Chen Xin，Yu Jian-dong，et al. Afterhyperpolarization improves spike programming through lowering threshold potentials and refractory periods mediated by voltage-gated sodium channels［J］.Biochemical and Biophysical Research Communications，2006，346（3）：938-945.

［136］ Chen Na，Yu Jian-dong，Qian Hao，et al. Axons amplify somatic incomplete spikes into uniform amplitudes in mouse cortical pyramidal neurons［J］. PLoS ONE，2010，5（7）：e11868

［137］ Raphael Hourez，Karima Azdad，Gilles Vanwalleghem，et al. Activation of protein kinase C and inositol 1，4，5-triphosphate receptors antagonistically modulate voltagegated sodium channels in striatal neurons［J］. Brain Res，2005，1059（2）：189-196.

［138］ Li Ming，James W West，Yvonne Lai，et al. Functional modulation of brain sodium channels by cAMP-dependent phosphorylation［J］. Neuron，1992，8（6）：1151-1159.

［139］ Hao Zhi-bin，Pei Dong-sheng，Guan Qiu-hua，et al.Calcium/calmodulin-dependent protein kinase II（CaMKII），through NMDA receptors and L-Voltage-gated channels，modulates the serine phosphorylation of GluR6 during cerebral ischemia and early reperfusion period in rat hippocampus［J］. Brain Res Mol Brain Res，2005，140（1-2）：55-62.

［140］ Hiroshi Onodera，Yasundo Yamasaki，Kyuya Kogure，et al. Calcium/calmodulin-dependent protein kinase II and protein phosphatase 2B （calcineurin） immunoreactivity in the rat hippocampus long after ischemia［J］. Brain Res，1995，684（1）：95-98.

［141］ Teresa Zalewska，B Zablocka，Krystyna Domanska-Janik.Changes of Ca2+/calmodulin-dependent protein kinase-II after transient ischemia in gerbil hippocampus［J］. Acta Neurobiol Exp （Wars），1996，56（1）：41-48.

［142］ Hubert Monnerie，Peter D Le Roux. Reduced dendrite growth and altered glutamic acid decarboxylase （GAD）65- and 67-kDa isoform protein expression from mouse cortical GABAergic neurons following excitotoxic injury in vitro［J］. Exp Neurol，2007，205（2）：367-382.

［143］ Gabriella Nyitrai，Katalin A Kekesi，Gabor Juhasz.Extracellular level of GABA and Glu：in vivo microdialysis-HPLC measurements［J］. Curr Top Med Chem，2006，6（10）：935-940.

［144］ Li Hhi-ling，Ruth E Siegel，Rochelle D Schwartz. Rapid decline of GABAA receptor subunit mRNA expression in hippocampus following transient cerebral ischemia in the gerbil［J］. Hippocampus，1993，3（4）：527-537.

［145］ Norio Akaike. Time-dependent rundown of GABA response in mammalian cns neuron during experimental anoxia［J］.Obes Res，1995，3（Suppl 5）：769S-777S.

［146］ Zhan Ren-zhi，J Victor Nadler，Rochelle D Schwartz-Bloom. Depressed responses to applied and synapticallyreleased GABA in CA1 pyramidal cells，but not in CA1 interneurons，after transient forebrain ischemia［J］. J Cereb Blood Flow Metab，2006，26（1）：112-124.

［147］ Tae-Cheon Kang，Seung-Kook Park，In-koo Hwang，et al.Spatial and temporal alterations in the GABA shunt in the gerbil hippocampus following transient ischemia［J］.Brain Res，2002，944（1-2）：10-18.

［148］ Li Yan，Glenn Dave Blanco，Lei Zhi-gang，et al. Increased GAD expression in the striatum after transient cerebral ischemia［J］. Mol Cell Neurosci，2010，45（4）：370-377.

［149］ Raghu Vemuganti. Decreased expression of vesicular GABA transporter，but not vesicular glutamate，acetylcholine and monoamine transporters in rat brain following focal ischemia［J］. Neurochem Int，2005，47（1-2）：136-142.

［150］ Liang R，Pang Z P，Deng P，et al. Transient enhancement of inhibitory synaptic transmission in hippocampal CA1 pyramidal neurons after cerebral ischemia［J］.Neuroscience，2009，160（2）：412-418.

［151］ Maxim Dobretsov，Joseph R Stimers. Neuronal function and alpha3 isoform of the Na/K-ATPase［J］. Front Biosci，2005，10（1-3）：2373-2396.

［152］ Jeffrey Magee，Dax Hoffman，Costa Colbert，et al.Electrical and calcium signaling in dendrites of hippocampal pyramidal neurons［J］. Annu Rev Physiol，1998，60：327-346.

［153］ Pierre J Magistretti. Neuron-glia metabolic coupling and plasticity［J］. J Exp Biol，2006，209（Pt 12）：2304-2311.

［154］ Paul A Rutecki. Neuronal excitability：voltagedependent currents and synaptic transmission［J］. J Clin Neurophysiol，1992，9（2）：195-211.

［155］ Sheridan L Swope，Stephen J Moss，Craig D Blackstone，et al. Phosphorylation of ligand-gated ion channels：a possible mode of synaptic plasticity［J］. FASEB J，1992，6（8）：2514-2523.

［156］ Tanaka E，Yamamoto S，Yoshihisa Kudo，et al. Mechanisms underlying the rapid depolarization produced by deprivation of oxygen and glucose in rat hippocampal CA1 neurons in vitro［J］. J Neurophysiol，1997，78（2）：891-902.

［157］ Siesjo B K，Anders Ekholm，Ken-ichiro Katsura，et al.Acid-base changes during complete brain ischemia［J］.Stroke，1990，21（11 Suppl）：III194-III199.

［158］ Simon R，Xiong Z. Acidotoxicity in brain ischemia［J］.Biochem Soc Trans，2006，34（Pt6）：1356-1361.

［159］ Shono Yuji，Masahiro Kamouchi，Takanari Kitazono，et al. Change in intracellular pH causes the toxic Ca2+ entry via NCX1 in neuron- and glia-derived cells［J］. Cell Mol Neurobiol，2010，30（3）：453-460.

［160］ Zhou Chun-yi，Xiao Cheng，Deng Chun-yu，et al.Extracellular proton modulates GABAergic synaptic transmission in rat hippocampal CA3 neurons［J］. Brain Res，2007，1145：213-220.

［161］ Dwight E Bergles，Craig E Jahr. Synaptic activation of glutamate transporters in hippocampal astrocytes［J］.Neuron，1997，19：1297-1308.

［162］ Giorgio A Ascoli，Lidia Alonso-Nanclares，Stewart A Anderson，et al. Petilla terminology：nomenclature of features of GABAergic interneurons of the cerebral cortex［J］. Nat Rev Neurosci，2008，9（7）：557-568.

［163］ Thomas Klausberger，Peter Somogyi. Neuronal diversity and temporal dynamics：the unity of hippocampal circuit operations［J］. Science，2008，321（5885）：53-57.

［164］ Changhan Ouyang，Guo Lian-jun，Lu Qing，et al.，Enhanced activity of GABA receptors inhibits glutamate release induced by focal cerebral ischemia in rat striatum［J］. Neurosci Lett，2007，420（2）：174-178.

［165］ Cinzia Costa，Giorgia Leone，Emilia Saulle，et al.，Coactivation of GABA（A） and GABA（B） receptor results in neuroprotection during in vitro ischemia［J］. Stroke，2004，35（2）：596-600.

［166］ Johansen F F，N H Diemer，Enhancement of GABA neurotransmission after cerebral ischemia in the rat reduces loss of hippocampal CA1 pyramidal cells［J］. Acta Neurol Scand，1991，84（1）：1-6.

［167］ Obrenovitch T P. Molecular physiology of preconditioninginduced brain tolerance to ischemia［J］. Physiol Rev，2008，88（1）：211-247.