In Vivo EPR Estimation of Bilateral Hippocampal Antioxidant Ability of Rats with Epileptogenesis Induced by Amygdalar FeCl3 Microinjection


Address correspondence and reprint requests to Dr. Yuto Ueda, Section of Psychiatry, Department of Clinical Neuroscience, University of Miyazaki, 5200 Kihara, Kiyotake-cho, Miyazaki 889-1692, Japan. E-mail:


Summary: Purpose: To measure the neural antioxidant function in the hippocampus of rats with epileptogenesis induced by microinjection of FeCl3 into the amygdala using the decay rate of the nitroxide radical as estimated by L-band electron paramagnetic resonance (EPR) spectroscopy.

Materials and Methods: Region-selected intensity determination (RSID) was used for the estimation of the nitroxide decay ratio. It is possible to estimate the in vivo hippocampal antioxidant ability using the half-life of the EPR signal of the blood–brain barrier-permeable nitroxide radical. Rats were microinjected with aqueous FeCl3 into the right amygdaloid body. Recording from chronically implanted depth electrodes showed the development of spike discharges with recurrent seizures arising from amygdalar regions with propagation into both hippocampi. Rats with unilateral aqueous FeCl3 lesions were injected systemically with the nitroxide radical and then had EPR for RSID estimation at 5, 15, and 30 days after the iron salt injection.

Results: The in vivo antioxidant ability of the dorsal hippocampus was significantly decreased bilaterally in animals with FeCl3-induced seizures when compared to the control.

Conclusion: Neural antioxidant function in the hippocampi of rats with chronic seizures induced by amygdalar FeCl3was decreased early and both ipsilaterally and bilaterally.

Redox balance in neural tissue has an important role in the pathophysiology of neurotoxicity from free radical generation, but also in the pathophysiology of neurotoxicity from free radical generation, as well as in the function of transmembrane proteins. Both N-methyl-d-aspartate receptors (NMDA-R) and excitatory amino acid transporters (EAATs), which are associated with glutamate neurotoxicity and seizures, possess redox regulatory sites (Rondouin, et al., 1992; Lipton et al., 1993; Trotti et al., 1995, 1996). Brain redox balance is maintained between oxidative and reductive conditions. Normally, there is a steady-state balance between the production of reactive oxygen species and their destruction by in vivo antioxidants.

The generation of reactive oxygen species (ROS), or free radicals, can exceed the scavenging ability of endogenous antioxidants, resulting in a shift of the redox state of the brain to the oxidative state. To study the redox state in epileptogenesis, reductants associated with in vivo antioxidant efficacy in the brain have been evaluated. However, quantitative analysis of reductants in extracts of brain tissue is problematic because in vivo antioxidants, such as ascorbic acid, α-tocopherol, and glutathione (GSH), are working synergistically to exert themselves as effective free radical scavengers (Bendich et al., 1984; McCay, 1985; Niki, 1987).

Recently, we developed two methods for the direct analysis of antioxidant ability in vivo (Ueda et al., 1998b; Yokoyama et al., 1999; Tokumaru et al., 2000) using X-band and L-band electron paramagnetic resonance (EPR) spectroscopy. The methodological principles for X-band and L-band EPR spectroscopy are based on the observation that nitroxide radicals applied exogenously in the brain are reduced and lose their paramagnetism reductants within biological systems. Thus, the decay rate or the half-life of EPR signal intensities appears to reflect the antioxidant ability in biological systems (Schallreuter et al., 1990; Zhang and Fung, 1994; Fuchs et al., 1997; Ueda et al., 1998b; Yokoyama et al., 1999; Tokumaru et al., 2000).

We have developed an advanced L-band EPR spectroscopy method to estimate the in vivo decay rates of EPR signals for a nitroxide radical in rat brain using the region-selected intensity determination (RSID) method (Yokoyama et al., 2005). The superiority and advantages of the RSID method make it possible to analyze in vivo antioxidant ability acting synergistically at two positions in the brain while performing one experiment on a living animal.

The seizure induction by amygdalar FeCl3 injection used in this study is thought to mimic posttraumatic epilepsy (Willmore et al., 1978; Ueda et al., 1998). The mechanisms resulting in chronic recurrent focal seizures could be related to the initiation of free radical reactions and lipid peroxidation at the injection site (Willmore and Rubin, 1982; Willmore et al., 1983; Triggs and Willmore, 1984; Willmore and Triggs, 1991). Hydrogen abstraction from polyunsaturated lipid acid by free radicals is hypothesized to play a causative role in lipid peroxidation (Halliwell and Gutteridge, 1986). We wondered if the antioxidant ability changes unilaterally or bilaterally following unilateral amygdalar FeCl3 injection. The objective of this study, then, was to qualitatively analyze the time-dependent bilateral hippocampal in vivo antioxidant ability using the RSID method with a single EPR measurement to estimate the brain biochemical redox state during epileptogenesis.


All animal experiments were conducted in accordance with the criteria outlined in the 1987 guidelines prepared by the Japanese Association for Laboratory Animal Science. The Committee for the Ethics on Animal Experiments of the Faculty of Medicine, University of Miyazaki, reviewed and approved the experimental design (approval number: 1998-158-2).


A water-soluble blood–brain barrier (BBB)-permeable nitroxide radical, 3-hydroxymethyl-2,2,5,5-tetramethylprrolidine-1-oxyl (hydroxymethyl-PROXYL), was synthesized in laboratory as follows (Yokoyama, et al., 2002b). Two grams of 3-carboxy-2,2,5,5-tetramethylpyrrolidine-1-oxyl (Aldrich Chemical, St. Louis, Mo, U.S.A.) was added to a suspension of 0.6 g of lithium aluminum hydride (Kanto Chemical, Tokyo, Japan) in anhydrous ether (Kanto) while being stirred. Stirring continued for 24 h. To this mixture, an 18% aqueous solution of hydrogen chloride (Kanto) was added until the emulsion disappeared. The mixture was alkalized by adding a 10% aqueous solution of sodium hydroxide (Kanto). The ether phase was separated and dried with magnesium sulfate (Wako Pure Chemical, Osaka, Japan). The solvent was evaporated, leaving yellow-colored crystals (1.8 g, 76% yield) of hydroxymethyl-PROXYL. Its molecular structure was confirmed by using nuclear magnetic resonance (JNM-A620, JEOL, Tokyo, Japan) and mass spectrometry (Zabspect Q, Micromass, Manchester, UK).

Experimental epilepsy induced by a unilateral amygdalar injection of FeCl3

An epilepsy model of limbic epileptogenesis was constructed according to the methods described elsewhere (Ueda et al., 1998a). In brief, 24 male Wistar rats were either kept in hanging cages or housed separately. They had unlimited access to food and water and were exposed to 12-h light/dark cycles. The rats weighed 200–250 g at the time of surgery. They were anesthetized with sodium pentobarbital (37.5 mg/kg, i.p.). The stereotaxic coordinates were determined according to the rat brain atlas of Paxinos (Paxinos and Watson, 1986). The incisor bar was set −3.3 mm below the intraaural line. Rats were prepared by stereotaxic placement of a 22-gauge guide cannula with the lumen occluded with a stylet and the end positioned just above the right amygdaloid body. The guide cannula was attached to a stainless electrode wire for EEG recording from the right amygdaliod body. After positioning the guide cannula, the stylet was removed and replaced with a fused silica tube (0.075 mm inside diameter, 0.15 mm outside diameter) that was attached to a microinjection pump (EP-60, EICOM, Japan). The tip was positioned into the right amygdaloid body. Rats were then injected over 5 min with a total of 1.5 μL of 100 mM FeCl3 (Fe group, n = 18) or with an equal volume of 0.9% NaCl with the pH adjusted to 2.2 to equal that of the ferric chloride solution (control group, n = 6). During injection and for the following 30 min, all rats underwent continuous EEG recording from the injection site (Model EEG-5414, NIHON-KODEN, Tokyo, Japan). The cannula was removed after the injection, and the animals were allowed to recover from anesthesia. Each rat in the Fe group had epileptiform spike discharges from the amygdaloid body and the behavior of limbic seizures, which developed into secondarily generalized seizures, as reported elsewhere (Ueda et al., 1998a). Rats injected with FeCl3 underwent the measurement of antioxidant ability at 5 days ((Fe-5d); n = 6), 15 days ((Fe-15d); n = 6), and 30 days ((Fe-30d); n = 6) after unilateral amygdalar FeCl3 injection. Rats were postnatal (PN) 49–56 days in all groups. Amygdalar saline-injected control animals at PN 49 days were used as well. The bilateral hippocampal antioxidant ability in the injection control and FeCl3-induced seizure rats was estimated by the RSID method using L-band EPR spectroscopy.

Estimation of in vivo antioxidant ability by the RSID method

EPR spectrometer

A 700-MHz radio frequency (RF) EPR spectrometer constructed in our laboratory has already been described in detail (Yokoyama et al., 1999, 2001, 2002a; Ueda et al., 2002b; Yokoyama et al., 2003). It consists of an EPR resonator, a main electromagnet, a pair of field scan coils, a pair of field gradient coils, a pair of field modulation coils, power supplies, a personal computer, 700-MHz RF circuits for homodyne detection, and intermediate frequency circuits for lock-in detection at a field modulation frequency of 100 kHz.A bridged loop-gap resonator (BLGR) (Ono et al., 1986) was used as an EPR resonator. The BLGR used in this study is a two gap type with bridge shields that are located inside the resonator. Its axial length is 10 mm with an inner diameter of 41 mm. The resonator was driven at a frequency of approximately 700 MHz. The unloaded Q of the resonator was about 700. Linear magnetic field gradients along the x-, y-, and z-axes are produced by the gradient coils (up to 1 mT/cm in a 20 mm range from the center, Yonezawa Electric Wire). The z- and x-axes are defined as the directions of the static and RF magnetic fields, respectively. The y-axis is perpendicular to the z–x plane. The coil for the z-gradient, as well as those for the x- and the y-gradients, was an anti-Helmholtz type. Fixing animals in the apparatus yielded a loaded Q of the BLGR of about 50.

Nitroxide radical injection

Under pentobarbital anesthesia, control experimental animals received an i.v. injection of 0.4 mmol of hydroxymethyl-PROXYL that had been dissolved in physiological saline. Each animal was stabilized in the BLGR by stereotaxic means fixing the incisor teeth and the external auditory canals with the intraaural line aligned 9 mm posterior to the center.

Estimation of in vivo antioxidants

Temporal EPR measurements of rat brain were performed just after injection of the hydroxymethyl-PROXYL. Selected regions were set at the bilateral dorsal hippocampi in accordance with the rat brain atlas (Paxinos and Watson, 1986). These selected regions were alternately changed, and the measurements were repeated. The distance between the two regions was sufficiently greater than the spatial resolution (2.7 mm) that was obtained by a phantom study (Yokoyama et al., 2005). The EPR conditions were: RF power, 50 mW at 684 MHz; static magnetic field, 24.1 mT; field sweep speed, 10 mT/s at a width of 10 mT; time constant, 1 ms; field modulation, 0.2 mT at 100 kHz; field gradient, 1 mT/cm; direction number of gradient, 49; accumulation number with gradient, 49; and accumulation number without gradient, 4.

Statistical analysis

Statistical analysis was carried out using two-way ANOVA (TIME × SIDE effect) with repeated measures followed by Fisher's Protected least significant difference. Data were considered significant at the level of p < 0.05.


Fig. 1 shows a typical example of a semilogarithmic plot of the temporal changes in signal intensity, as previously reported (Yokoyama et al., 2005). The half-life obtained from the RSID method in control and amygdalar FeCl3-injected rats (5, 15, and 30 days) is summarized in Fig. 2. The in vivo antioxidant ability of the dorsal hippocampus was significantly decreased bilaterally in the FeCl3-induced rats when compared to the control, especially at 5 and 30 days. The difference of laterality was observed at 30 days after FeCl3 injection, but not in the acute phase of injection.

Figure 1.

Representative eliminating decay ratio of the EPR signal intensities of nitroxide radical in the bilateral hippocampus of FeCl3-injected rats at 5 days. Squares indicate the decay of the nitroxide radical in the rat hippocampus during epileptogenesis, while circles are from control rats. Closed marks indicate ipsilateral data, while open marks show contralateral data. The decay ratio is slower bilaterally in the rats with epileptogenesis (square) than it is in the control (circle).

Figure 2.

Statistical analysis of changes in bilateral hippocampal half-life of the nitroxide radical, which was performed using two-way ANOVA (TIME × SIDE effect) with repeated measures followed by Fisher's Protected LSD (F1,3 = 4.247, p = 0.0196). Data represent mean +/− S.E.M. *p < 0.05, **p < 0.01 vs. control on each side. #p < 0.01, half-life of the ipsilateral side vs. the contralateral side.


Nitroxide radicals have been reported to react with hydroxyl radicals and superoxide anion radicals in the presence of cysteine or nicotinamide adenine dinucleotide NADH. Takeshita et al. (2002) reported that the EPR signal intensity of nitroxide radicals decays linearly in reactions with other radicals. However, the EPR signal intensity of the nitroxide radical decays exponentially in reactions with reductants. Thus, the reaction dynamics of nitroxide radicals with reductants differs from that with other radicals. In our study of rats with chronic seizures, the EPR signal intensity of hydroxymethyl-PROXYL decayed exponentially. We concluded that the prolonged half-life in the experimental group was related to decreased endogenous antioxidant ability bilaterally and in regions remote from the site of direct injection. Therefore, the antioxidant ability decreased bilaterally in the seizure group, but not in the control.

Several mechanisms associated with epileptogenesis may be causal in these observed changes in antioxidant ability. Recurrent seizures are thought to generate ROS and free radicals with resultant hippocampal sclerosis (Chung and Han, 2003, Freitas, et al., 2004, 2005; Rajasekaran, 2005). Thus, depletion of neural antioxidants could lead to lipid peroxidation and hippocampal sclerosis. Mitochondrial damage also occurs with the depletion of in vivo antioxidants (Ueda et al., 2002a). Recurrent seizures during epileptogenesis induce mitochondrial dysfunction (Ekdahl et al., 2003; Gupta and Dettbarn, 2003; Sullivan et al., 2003; Chuang, et al., 2004; Kann et al., 2005). We question, therefore, whether the depletion of in vivo antioxidants in our seizure model is related to mitochondrial damage.

Numerous studies have shown that antioxidants, such as melatonin, inhibit the promotion of epileptogenesis (Ekdahl, et al., 2003; Gupta and Dettbarn, 2003; Sullivan, et al., 2003; Chuang et al., 2004; Kann et al., 2005). Thus, depletion of antioxidants may well be one of the causative factors promoting epileptogenesis. The in vivo antioxidant ability of the dorsal hippocampus was significantly decreased bilaterally in the FeCl3-induced seizure animals, especially at 5 and 30 days after injection. Difference in laterality was observed at 30 days after FeCl3 injection, but not in the acute phase following injection. Recurrent seizures at 30 days propagating from the primary seizure focus to the other regions of the brain appear to cause depletion in ipsilateral hippocampus (see Fig. 2, p < 0.01). We suggest that impairment of hippocampal antioxidant function is important biological evidence showing that oxidative stresses, such as hypoxia or ischemia, which accompany convulsions in patients with epilepsy, are important in the mechanisms of neural injury from seizures.


Acknowledgment:  This study was supported by a Grant-in-Aid for Scientific Research (C)(2) (16591146 and 18591297) from the Ministry of Education, Science, Sport, and Culture, Japan (to Y.U.).