**3. Discussion**

The degree of hippocampal neuronal death following IR injury varies according to the ischemic duration. It has been demonstrated that the selective death of pyramidal cells in hippocampal CA1 occurs four to five days after five minutes of transient ischemia in the forebrain of gerbils [4,26]. We reported that the death of pyramidal cells in CA1-3 of gerbils subjected to 15 min of transient ischemia was observed from two days post-ischemia, indicating that a longer period of transient ischemia could evoke the acceleration and aggravation of pyramidal cell death in the hippocampus [7,25]. In this study, we examined the neuroprotective effects of ATX in CA1-3 of gerbils subjected to transient ischemia for 15 min using Nissl staining, NeuN immunohistochemistry, and FJB histofluorescence staining, and found that pretreatment with ATX significantly alleviated ischemia-induced pyramidal cell death in CA1-3. We believe that this finding was the first showing that ATX elicited strong neuroprotection against severe IR injury in the gerbil hippocampus. In rats, global cerebral ischemia for 10 min induced pyramidal cell death in CA1 alone, which might be a phenomenon of mild IR injury in the hippocampus compared to severe IR injury in the hippocampus, and the administration of ATX significantly reduced the selective death [27]. To the best of our knowledge, it is impossible to develop a rat model of severe IR injury in the hippocampus by occluding four vessels (two common carotid arteries and two vertebral arteries) for 15 min or more because rats subjected to this ischemic injury die a few days later due to failure of the brainstem which receives arterial blood supply from the vertebral arteries [28,29]. However, as described in the Materials and Methods Section, a gerbil model of severe IR injury in the hippocampus can be developed, because

gerbils with severe IR injury induced by the ligation of two common carotid arteries (not two vertebral arteries) survive a long time without brainstem failure [6,7]. Regarding the anatomy of the circle of Willis in rats and gerbils, rats have a perfect circle of Willis, whereas gerbils have an incomplete circle of Willis so that there is a lack of posterior communicating arteries [7,30]. These anatomical characteristics helped us to develop a gerbil model of severe IR injury in the forebrain. Many years ago, Jarrott and Dommer (1980) reported that a mortality rate of 100% was obtained if the occlusion of bilateral common carotid arteries was maintained for 60 min in gerbils weighing 45–55 g [31].

Oxidative stress is a phenomenon induced by an imbalance between the production and accumulation of ROS in cells and/or tissues and the ability of a system to detoxify the reactive products [32]. When the production of ROS increases, harmful effects are exerted on important structures in cells such as proteins, lipids, and nucleic acids [33]. In ischemic insults to the brain, oxidative stress has also been considered to be a major factor to overcome following ischemic brain injury. Many preclinical studies have demonstrated that oxidative stress following ROS accumulation during IR irreversibly damaged cellular biomolecules, including lipids, proteins, and DNA, and led to neuronal death [8,10]. In particular, early during IR, due to an elevated oxygen supply, large quantities of oxygen radicals increase rapidly in the ischemic lesions and DNA damage occurs [34]. Lipid peroxidation can be described as a process under which oxidants including ROS attack lipids containing carbon–carbon double bond(s) [35]. In this respect, various substances reducing oxidative stress have been demonstrated to attenuate ischemic brain injury following IR injury [36–38]. In this study, the immunoreactivity of 8OHdG (an oxidative DNA adduct) in CA1-3 pyramidal cells was significantly enhanced on day 1 after IR (before the time of IR-induced loss of CA1-3 pyramidal cells), but it was significantly reduced by ATX treatment. Resveratrol, which is a naturally occurring polyphenolic compound and elicits a variety of biological and pharmacological functions, including anti-oxidant and anti-inflammatory properties, protected HT22 cells from oxygen-glucose deprivation and reoxygenation-induced injuries by reducing DNA damage (decreases in 8OHdG levels) [39]. In addition, we found that the immunoreactivity of 4HNE (an end-product of lipid peroxidation) in CA1-3 pyramidal cells was significantly increased on day 1 after IR. However, ATX significantly decreased 4HNE immunoreactivity. These results indicate that pretreatment with ATX mitigated severe IR-induced oxidative stress in CA1-3 pyramidal cells, which may be closely associated with ATX-mediated neuroprotection against severe IR injury.

The antioxidant defense system neutralizes or scavenges ROS and plays a protective role against ischemic brain injury [11,40]. One of the key enzymes in the antioxidant defense system is SOD, which acts first to defend against the harmful effects of oxidative stress. Targeting SOD expression has shown substantial success in animal models of transient brain ischemia. Sugawara et al. (2002) showed that IR-induced superoxide production and the selective death of CA1 pyramidal cells were significantly decreased in transgenic rats overexpressing SOD1 [41]. Keller et al. (1998) also found that lipid peroxidation, protein nitration, and brain infarction following transient focal cerebral ischemia were significantly reduced in transgenic mice overexpressing SOD2 [42]. Moreover, we reported that the increased expression of SOD1 and SOD2 in CA1 pyramidal cells following pretreatment with natural antioxidants derived from plant materials, such as quercetin and chlorogenic acid, contributed to the protection of CA1 pyramidal cells against IR injury in gerbils [43,44]. In this study, SOD1 and SOD2 immunoreactivity in the CA1-3 pyramidal cells of both the vehicle-IR and ATX-IR groups were gradually decreased with time after severe IR, but, in the ATX-IR group, SOD1 and SOD2 immunoreactivity were significantly high compared to the vehicle-IR group. These results suggest that SOD1 and SOD2 in CA1-3 pyramidal cells were increased by ATX pretreatment and these increases could contribute to the neuroprotection against oxidative stress following severe IR.

Collectively, this study showed that pretreatment with ATX attenuated the massive loss (death) of CA1-3 pyramidal cells in gerbil hippocampi following severe IR injury induced by 15-min transient forebrain ischemia, and the pretreatment with ATX significantly enhanced the expressions of SOD1 and 2 in the CA1-3 pyramidal cells before IR and reduced oxidative stress in pyramidal cells following severe IR. These findings suggest that ATX can be used as a potential dietary supplement to prevent the progression of severe IR injury in brains. However, we need an in-depth study on the mechanism of the neuroprotective effect of ATX. For instance, examination of changes in antioxidant markers such as nuclear factor erythroid 2-related factor 2 (Nrf2) should be conducted to demonstrate the antioxidant defense of ATX against IR. Moreover, since IR injury in gerbils triggers delayed neuronal death in the hippocampus, examination of behavioral changes needs to be carried out to demonstrate the amelioration of IR-induced cognitive decline by treatment with ATX.
