Possible Role of the Glycogen Synthase Kinase-3 Signaling Pathway in Trimethyltin-Induced Hippocampal Neurodegeneration in Mice

Trimethyltin (TMT) is an organotin compound with potent neurotoxic effects characterized by neuronal destruction in selective regions, including the hippocampus. Glycogen synthase kinase-3 (GSK-3) regulates many cellular processes, and is implicated in several neurodegenerative disorders. In this study, we evaluated the therapeutic effect of lithium, a selective GSK-3 inhibitor, on the hippocampus of adult C57BL/6 mice with TMT treatment (2.6 mg/kg, intraperitoneal [i.p.]) and on cultured hippocampal neurons (12 days in vitro) with TMT treatment (5 µM). Lithium (50 mg/kg, i.p., 0 and 24 h after TMT injection) significantly attenuated TMT-induced hippocampal cell degeneration, seizure, and memory deficits in mice. In cultured hippocampal neurons, lithium treatment (0–10 mM; 1 h before TMT application) significantly reduced TMT-induced cytotoxicity in a dose-dependent manner. Additionally, the dynamic changes in GSK-3/β-catenin signaling were observed in the mouse hippocampus and cultured hippocampal neurons after TMT treatment with or without lithium. Therefore, lithium inhibited the detrimental effects of TMT on the hippocampal neurons in vivo and in vitro, suggesting involvement of the GSK-3/β-catenin signaling pathway in TMT-induced hippocampal cell degeneration and dysfunction.


Introduction
Trimethyltin (TMT), an organotin compound, has potent neurotoxic effects characterized by neuronal destruction in selective regions such as the limbic system [1][2][3]. Humans accidentally exposed to TMT develop a syndrome characterized by seizures, disorientation, confusion, memory deficits and aggressiveness [1,4]. In experimental animals, exposure to TMT induces neurotoxicity due to the initial oxidative burden in select regions [5,6]. TMT-induced selective neuronal cell death and neuroinflammation contribute to neurodegeneration [7]. Several in vivo and in vitro studies demonstrated that c-Jun N-terminal kinase signaling, cyclooxygense-2, caspases-3/-8 and diverse proinflammatory cytokines are activated by TMT treatment and might be involved in the pathological mechanism of TMTinduced brain injury [8][9][10][11]. In mice, TMT treatment causes seizures, hyperactivity, memory deficits, and neuronal cell loss, especially in the hippocampal dentate gyrus (DG) [12,13]. Recently, several studies suggested the phosphoinositol 3-kinase (PI3K)/Akt pathway to be a target for neuroprotection in TMTinduced central nervous system (CNS) injury [6,14,15]. Thus, TMT-induced neurotoxicity is regarded as a useful model for the study of neurodegenerative diseases and hippocampal dysfunction, such as Alzheimer's disease (AD) [7]. However, the precise mechanism underlying TMT-induced neuronal cell death remains unclear.
In the present study, we elucidated the changes in the GSK-3/ b-catenin signaling pathway in TMT-induced hippocampal degeneration and the neuroprotective effects of lithium on TMT-induced neurotoxicity in vivo and in vitro to elucidate the possible role of GSK-3 signaling in chemical-induced neurodegeneration.

Results
Figure l shows a schematic diagram of the procedures used for in vivo tests evaluating the effect of lithium treatment on TMTinduced neurodegeneration and behavioral disability.

TMT Induced the Change of GSK-3/b-catenin Signaling in the Hippocampus
To determine the effect of TMT treatment on the GSK-3 pathway, the inhibitory serine phosphorylation of GSK-3 and the b-catenin expression levels in hippocampal extracts prepared 2, 4 and 7 days post-treatment (n = 3 mice in each time-point) were assessed by Western blot analysis ( Fig. 2 and Fig. S1). TMT treatment led to significant increases in the inhibitory phosphorylation of GSK-3a (Ser 21 ) 4 days post-treatment (p,0.01 vs. controls; Fig. 2A), and GSK-3b (Ser 9 ) 4 and 7 days post-treatment (p,0.05 vs. controls; Fig. 2B). The treatment also markedly increased the level of b-catenin expression 2 (p,0.05 vs. controls), 4 (p,0.05 vs. controls) and 7 days post-treatment (p,0.001 vs. controls) (Fig. 2C).
Consistent with the Western blotting results, the phosphorylated GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and b-catenin expression levels, measured by immunohistochemistry, were localized primarily in Cornu Ammonis (CA) 1 pyramidal and dentate gyrus (DG) granule neurons in the hippocampus, and markedly increased in the granular cell layer (GCL) of the DGs 4 days after TMT treatment (Fig. S2).

Lithium Treatment Rescued TMT-induced Seizure
TMT exposure causes symptoms such as tremor, seizure and aggressive behavior in mice (Fig. 3). However, the TMT-induced seizure score in lithium-treated mice was significantly lower than that in TMT-treated controls (n = 25 mice per group; Fig. 3). The seizure behaviors in TMT-treated controls and TMT+lithiumtreated mice had disappeared on day 6 after TMT treatment. Table 1 summarizes the effect of lithium on the clinical symptoms of TMT-treated mice.

Lithium Treatment Ameliorated TMT-induced Memory Deficits in Mice
We first assessed mouse basal locomotor activity 7 days after TMT treatment in a novel environment by open-field analysis (n = 10 mice per group). The open-field analysis quantified the overall activity that reflects the motivation and performance of the mice. The control, lithium-, TMT-, and TMT+lithium-treated mice showed comparable ambulatory movement counts, moving distances, ambulatory movement times, and resting times, with no significant differences observed in any group (Fig. S3).
Next, we assessed recognition memory in mice (n = 9 mice per group) by a sensitive hippocampus-dependent paradigm, object recognition memory [31,32]. Control, lithium-, TMT-, and TMT+lithium-treated mice displayed equal preference for the two objects during training 7 days after TMT treatment (Fig.  S4A). Additionally, the total number of interactions during training was 13.7860.55 in vehicle-treated controls, 14.1061.40 in lithium-treated mice, 13.3360.78 in TMT-treated mice and 14.1160.63 in TMT+lithium-treated mice (Fig. S4B). There were no significant differences among groups, suggesting that mice had comparable attention, motivation and visual perception.
During the test 24 h after training, one conditioned old object was replaced with a novel object. If mice retained memory for the old objects, they would show preference for the novel object. The preferences (mean6SEM) toward a novel object were 74.4462.49% in controls, 70.2460.97% in lithium-treated mice, 55.9861.61% in TMT-treated mice, and 70.5963.60% in TMT+lithium-treated mice (Fig. 4A). Thus, TMT-treated mice showed significant deficits in novel object recognition (p,0.01 vs. controls), which were ameliorated by lithium treatment (p,0.01 vs. TMT-treated mice).
We further examined hippocampus-dependent spatial memory in mice (n = 10 mice per group) using the Morris water maze test [32]. In the visible platform training, mice learned to find the escape platform which had an attached visual cue. There was no significant difference between groups in escape latency during visible platform training (Fig. 5A). However, TMT-treated mice showed significantly longer escape latency compared to controls during the hidden platform training (p,0.05; Fig. 5A).
During the probe test 24 h after training, the percentages of time spent in the target quadrant were 50.8264.92% in controls, 54.9365.57% in lithium-treated mice, 36.2363.14% in TMTtreated mice and 49.1364.17% in TMT+lithium-treated mice. Thus, TMT-treated mice showed significant deficits in spatial memory (p,0.05 vs. controls), which was rescued by lithium treatment (p,0.05 vs. TMT-treated mice) (Fig. 5B). Lithium Treatment Ameliorated TMT-induced Neuronal Cell Death in the Hippocampus According to behavioral data, lithium treatment ameliorates TMT-induced hippocampal dysfunction, suggesting that lithium decreases TMT-induced neuronal cell death in the hippocampal DG. Therefore, we first performed hematoxylin and eosin staining 2, 4, and 7 days post-treatment. There were no significant differences in hippocampal structure among the groups under low magnification (Fig. S5). However, remarkable granular cell death, characterized by eosinophilic cytoplasm, nuclear pyknosis, nuclear karyolysis and cell loss, was evident in the hippocampal DG under high magnification at each time-point after TMT treatment. Lithium treatment reduced the TMT-induced granular cell death in the hippocampal DG (Fig. S5).
We additionally performed Fluoro-jade B (FJB) staining and NeuN immunostaining to detect neuronal degeneration and survival, respectively, to clarify the protective effects of lithium on TMT-induced neuronal cell death in the hippocampal DG (n = 3 mice in each time point). Semi-quantitative analysis of the FJB-positive intensity showed that lithium treatment resulted in a significant decrease in degenerative neurons in TMT-treated mice 2 (p,0.05 vs. TMT-treated mice), 4 (p,0.001 vs. TMT-treated mice) and 7 days (p,0.01 vs. TMT-treated mice) after TMT administration (Fig. 6A). Semi-quantitative analysis of NeuNpositive intensity showed that TMT treatment resulted in significant neuronal cell loss in mice 2, 4 and 7 days after TMT administration (p,0.001 vs. controls). However, lithium treatment significantly reduced TMT-induced neuronal cell loss 2, 4 and 7 days (p,0.01 vs. TMT-treated mice) after TMT administration (Fig. 6B).

Lithium Treatment Inhibited the TMT-induced Change of GSK-3/b-catenin Signaling in the Mouse Hippocampus
To confirm that lithium treatment inhibits GSK-3 activity in the hippocampus of TMT-treated mice, we evaluated the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression in the hippocampus by Western blotting (n = 3 mice in each time-point; Fig. S1). Lithium treatment significantly increased the inhibitory phosphorylation of GSK-3a (Ser 21 ) 2 and 4 days after TMT treatment (p,0.05 vs. TMTtreated mice; Fig. 7A), and GSK-3b (Ser 9 ) 2 days after TMT treatment (p,0.05 vs. TMT-treated mice; Fig. 7B) in the hippocampus. Lithium treatment also increased the level of bcatenin expression 4 days after TMT treatment (p,0.05 vs. TMTtreated mice; Fig. 7C).

Lithium Treatment Protected Hippocampal Cultured Neurons against TMT-induced Neurotoxicity
Based on our previous study [33], we tested whether lithium treatment rescued TMT-induced cytotoxicity in mature hippocampal cells at 12 days in vitro (DIV) using lactate dehydrogenase (LDH) release assays. TMT (5 mM) increased LDH release from cultured hippocampal neurons 24 h post-treatment (n = 6 cultures per condition; p,0.001 vs. controls; Fig. 8A). However, TMTinduced cytotoxicity was significantly inhibited by lithium (1-10 mM) in a dose-dependent manner (n = 6 cultures per condition; Fig. 8A).
To confirm the protective effect of lithium on TMT-induced neurotoxicity, we performed NeuN immunostaining in primary hippocampal cultures 24 h post-treatment. Lithium treatment remarkably reduced TMT-induced neuronal cell death (n = 3 cultures per condition; Fig. 8B). Thus, consistent with the in vivo data, lithium significantly rescued neuronal cell death induced by TMT treatment in mature hippocampal cells in vitro.

Lithium Treatment Inhibited the Effect of TMT on GSK-3/ b-catenin Signaling in Hippocampal Cultured Neurons
To determine if GSK-3 activity is altered by TMT exposure and inhibited by lithium preconditioning in hippocampal cultured neurons, we assessed the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression in mature hippocampal cells at 12 DIV by Western blotting. TMT treatment significantly decreased the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression 24 h post-treatment. However, lithium treatment markedly increased the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression (Fig. 9).

Discussion
Our data demonstrate that inhibition of GSK-3 signaling attenuates TMT-induced neurodegeneration in adult mice, as reflected by neuronal cell death in the hippocampal DG, clinical symptoms characterized by tremor/seizure and memory deficit, and TMT-induced cytotoxicity in primary hippocampal cultured neurons. Furthermore, in present study, TMT and/or lithium modulated the GSK-3/b-catenin signaling in hippocampal neurons in vivo and in vitro.
GSK-3, a multifunctional Ser/Thr kinase that regulates many cellular processes, is regulated by the Wnt and/or PI3K/Akt signaling pathways through distinct mechanisms [34,35]. In the Wnt signaling pathway, a multi-protein destruction complex consisting of the Wnt receptor, Frizzled low-density lipoprotein receptor-related protein 5/6, axin, adenomatous polyposis coli, bcatenin and casein kinase-1, regulates GSK-3 activity by means of protein-protein interactions [36,37]. Moreover, GSK-3 activity is regulated by direct phosphorylation of Akt, cyclic-AMP-dependent protein kinase, p70 ribosomal S6 kinase, and p90 ribosomal S6 kinase [38][39][40][41]. GSK-3-mediated Ser33/37 phosphorylation of bcatenin is reduced by inactivation of either Wnt or PI3K/Akt signaling, leading to its accumulation in the cytosol and transcriptional activation [42][43][44]. Furthermore, since b-catenin is a key mediator of the GSK-3 signaling pathway and plays an important role in neuroprotection, it is regarded as a potent survival factor [45,46]. Therefore, b-catenin accumulation may be used as an indicator of inhibition of GSK-3 activity. In the present study, TMT exposure dramatically altered the inhibitory phosphorylation of GSK-3 and accumulation of b-catenin in the hippocampus, which is a selective TMT target region in mice, suggesting the GSK-3/b-catenin signal pathway to be associated with TMT-induced neurodegeneration.
Several studies have suggested dysregulation of GSK-3 activity to be involved in neurodegeneration by neurotoxic agents in vivo Mice were treated with lithium (50 mg/kg, i.p.) 0 and 24 h after TMT (2.6 mg/kg, i.p.) administration, and evaluated using the novel object recognition memory test (n = 9 mice per group). During the test, TMTtreated mice showed a significantly decreased preference for the novel object; however, TMT+lithium-treated mice showed a preference for the novel object comparable to control mice. The data are reported as the means6SEM. **p,0.01 vs. controls. {{ p,0.01 vs. TMT-treated mice. Cont, controls; Li, lithium-treated mice; TMT, TMT-treated mice; TMT+Li, TMT+lithium-treated mice. doi:10.1371/journal.pone.0070356.g004 and in vitro [34,47,48]. Recently, a microarray study revealed that GSK-3a gene expression increased in the hippocampus of rats 3 days after TMT treatment [49]. We hypothesized that TMT treatment might induce upregulation of GSK-3 activity in the hippocampus, because the present study showed TMT-induced neuronal cell death in the mouse hippocampus. However, the in vivo data showed that TMT treatment significantly increased the inhibitory serine phosphorylations of GSK-3a (Ser 21 ) 2 days posttreatment and GSK-3b (Ser 9 ) 4 and 7 days post-treatment in the adult mouse hippocampus. We also observed accumulation of bcatenin in the mouse hippocampus 2, 4 and 7 days post-treatment. These results are consistent with the kainate-induced neurodegeneration model [29], in which the kainate induced an increase of inhibitory phosphorylation of GSK-3 in hippocampal neurons as a compensatory survival response against kainate toxicity. Consequently, we suggest that the increases in the inhibitory serine phosphorylation of GSK-3 and the accumulation of b-catenin in the mouse hippocampus elicited by TMT treatment might mediate neuronal cell survival against TMT-induced neurotoxicity.
Lithium was recently shown to directly inhibit GSK-3 activity by competition with Mg 2+ for binding to its catalytic site and indirectly by inhibitory serine phosphorylation of GSK-3 via Akt pathways [50,51]. Previous studies revealed that lithium provides neuroprotective effects by inhibition of GSK-3 activity in various animal models of neurodegenerative disorders including TBI, AD, and ischemic stroke [26,52,53]. Lithium has also been shown to have an anticonvulsant effect on chemical-induced seizure behavior [54,55]. Furthermore, lithium improves the cognitive impairment induced by streptozotocin, mild TBI, and brain  ischemia [26,[56][57][58][59]. In the present study, lithium treatment significantly ameliorated TMT-induced seizure behavior and cognitive impairments in the object recognition memory and Morris water maze paradigms.
Accumulating in vivo and in vitro evidence suggests that inhibition of GSK-3 activity by lithium may contribute to its therapeutic effects [26,60,61]. Previous studies revealed that lithium treatment reduces neuronal cell death, inflammation and oxidative stress, and inhibits apoptotic pathways dependent on GSK-3 inhibition [57,[62][63][64]. Pre-or post-treatment with lithium reduces apoptotic cell death in an AD model and protects primary-cultured neurons from glutamate-induced excitotoxicity via inhibition of GSK-3 [65,66]. In the present study, lithium treatment significantly reduced the histopathological lesions-as shown by hematoxylin and eosin staining-the density of FJB-positive degenerating neurons, and the loss of NeuN-positive neurons in the hippocampal DG in TMT-treated mice. Therefore, inhibition of GSK-3 in the mouse hippocampus by lithium treatment may reduce neuronal cell death following TMT treatment. Furthermore, lithium treatment significantly increased the inhibitory serine phosphorylation of GSK-3 and the level of b-catenin expression in the mouse hippocampus post-treatment. These findings suggest that inhibition of GSK-3 is crucial for the therapeutic effects of lithium on TMT-induced neurotoxicity.
In addition, our in vitro data showed that inhibition of GSK-3 activity by lithium treatment markedly protected hippocampal neurons against TMT-induced cytotoxicity. Pre-treatment of lithium attenuates chemical-induced cytotoxicity through the GSK-3 signaling pathway [30,67], and increases the inhibitory phosphorylation of GSK-3 induced by neurotoxic agents in cultured cells [67,68]. In the present in vitro study, Western blot analysis showed that TMT treatment significantly decreased the level of the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression 24 h posttreatment. Conversely, lithium treatment markedly increased the inhibitory phosphorylation of GSK-3a (Ser 21 ) and GSK-3b (Ser 9 ) and the level of b-catenin expression 24 h post-treatment, suggesting that lithium-induced accumulation of b-catenin may be related to its crucial neuroprotective role against TMT-induced neurodegeneration. Therefore, consistent with our initial hypothesis, the in vitro data suggest that GSK-3 activity is closely related to TMT-induced neurotoxicity in mouse hippocampal neurons.
In conclusion, TMT exposure dramatically altered the GSK-3/ b-catenin signaling pathway in the mouse hippocampus, and GSK-3 inhibition significantly reduced TMT-induced behavioral and histological changes in the hippocampus, as well as cytotoxicity in hippocampal cultured neurons. These in vivo and in vitro findings suggest that the GSK-3 signal pathway is associated with TMT-induced hippocampal neurodegeneration. Furthermore, the use of a selective GSK-3 inhibitor, lithium, may have therapeutic effects on chemical-induced neurodegeneration, including TMT-induced neurotoxicity.

Animals
Male C57BL/6 mice, 8-to 9-weeks-old, were obtained from a specific-pathogen-free colony at Orient Bio, Inc. (Seoul, Korea). The Institutional Animal Care and Use Committee of Chonnam National University approved the protocols used in this study (CNU IACUC-YB-2012-18) and the animals were cared for in accordance with the Chonnam National University Guide for the Care and Use of Laboratory Animals.

Drug Treatment and Tissue Sampling
TMT (Wako, Osaka, Japan) and lithium chloride (Sigma-Aldrich, St. Louis, MO, USA) were dissolved in sterilized 0.9% saline. To assess the time-dependent effects of TMT on the GSK-3 signal pathway in mouse hippocampus, mice were sacrificed and the brains were dissected at 2, 4 and 7 days after a single intraperitoneal (i.p.) injection of vehicle (0.9% saline) or 2.6 mg/kg TMT (n = 6 mice per group). All the animals were killed by decapitation. The hippocampal samples and brains were stored at -70uC until used for Western blot analysis and stored in 30% sucrose after fixation in 4% paraformaldehyde in phosphatebuffered saline (PBS, pH 7.4) for immunohistochemistry, respectively.
To assess the effect of lithium on TMT-induced seizures, lithium chloride (50 mg/kg, i.p.) or vehicle (0.9% saline, i.p.) was administered to mice at 0 and 24 h after an administration of 2.6 mg/kg TMT (n = 25 mice per group), and seizure behaviors were observed for 5 consecutive days after TMT treatment. To assess the effects of lithium on TMT-induced memory deficits, hippocampus-dependent memory tests (the object recognition memory test [n = 9 mice/group] and the Morris water maze test [n = 10 mice per group]) were performed 7 days post-treatment, at which time seizure behavior had disappeared.
To assess the effects of lithium treatment on inhibitory phosphorylation of GSK-3 in the hippocampus, mice were sacrificed 2, 4, and 7 days post-treatment (n = 6 mice per group). The samples were embedded in paraffin wax after fixation in 4% paraformaldehyde in PBS (pH 7.4) using routine protocols (n = 3 mice per group) and stored at -70uC for biochemical analysis (n = 3 mice per group). Antibodies Polyclonal rabbit anti-phospho-GSK-3a (Ser 21 ), polyclonal rabbit anti-phospho-GSK-3b (Ser 9 ), monoclonal rabbit anti-active b-catenin, monoclonal rabbit anti-b-catenin, monoclonal rabbit anti-GSK-3a, and monoclonal rabbit anti-GSK-3b antibodies were purchased from Cell Signaling Technology (Beverly, MA, USA). Monoclonal mouse anti-neuronal nuclei (NeuN) antibody (Millipore, Temecula, CA, USA) was used to detect neurons. Monoclonal mouse anti-beta actin was purchased from Sigma-Aldrich. For immunoblot analysis, horseradish peroxidase (HRP)conjugated anti-rabbit IgG and anti-mouse IgG were obtained from Vector Laboratories (Burlingame, CA, USA). Immunofluorescent staining was performed using a fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Vector).
SDS sample buffer (64) was added to each homogenized sample, and the samples were heated at 100uC for 10 min. The samples were then separated by 10% SDS-PAGE (Bio-Rad, Hercules, CA, USA). Medium was completely removed from hippocampal cell culture by aspiration and SDS sample buffer (64) was added to each culture. Cells from each culture were scraped and sonicated for 4 s. The samples were heated at 100uC for 10 min, and then separated by 10% SDS-PAGE.

Immunohistochemistry
Detailed in the Supporting Information (Text S1).

Open-field Test
Detailed in the Supporting Information (Text S1).

Object Recognition Memory Test
The object-recognition memory test was used to examine hippocampus-dependent memory [31,32]. The test was similar to a test described previously [70]. Briefly, two randomly selected, different-shaped objects were presented to each mouse for 10 min during training. Next, 24 h after training, another pair of objects (one old object and one novel object) was presented to the trained mice. If, for example, cube-and pyramid-shaped objects were presented during training, then a cylinder-shaped object was used as a novel object during testing. The interactions of the mouse with each object, including approaches and sniffing, were scored. If the mouse remembered an old object, preference toward the novel object was demonstrated during testing. The preference percentage was defined as the number of interactions for a specific object divided by the total number of interactions for both objects.

Morris Water Maze Test
The Morris water maze test was used to assess hippocampusdependent spatial memory [71]. Mice were individually trained in a circular pool (100 cm diameter, 30 cm height) filled with water maintained at 25uC and made opaque using a non-toxic washable white paint. The maze was located in a lit room with extramaze cues. The escape platform (10 cm diameter) was placed in the center of a designated quadrant of the pool with its top positioned 1 cm below the water surface. Mice were not allowed to swim in the pool before training. During the visible platform training, the platform was marked by a flag (5 cm tall). Mice were subjected to six trials daily for 2 days. The six trials were divided into two blocks with 1-h intervals. Three trials per block were conducted at 10-min intervals. Each trial lasted for 60 s unless mice reached the platform. The time elapsed until the mouse reached and land on the platform was scored as the escape latency. If the mouse failed to find the platform within 60 s, it was gently navigated to the platform by hand. Whether mice found or failed to find the platform within 60 s, they were allowed to stay on the platform for 30 s. After the visible platform training, the mice were further subjected to the hidden platform training, during which the platform was placed 1 cm below the opaque water. The position of the platform was fixed and starting positions were used pseudorandomly between trials. Mice were subjected to four trials at 1-h intervals daily for 4 consecutive days. Probe trials were performed 24 h after the hidden platform training. The platform was removed from its previous location and mice were allowed to swim in the pool for 1 min. The time spent in quadrant, number of crossings for the location of the hidden platform and swim speed were measured by video-based tracking system (SMART VIDEO-TRACKING; Panlab, Barcelona, Spain).

FJB Staining
FJB (a high-affinity fluorescent marker for the localization of neuronal degeneration) histofluorescent staining was performed according to a method described previously [72]. In brief, the sections were first transferred to a solution of 0.06% potassium permanganate and then to 0.0004% FJB (Millipore) staining solution. After washing, the sections were counterstained with DAPI before being mounted. The FJB-stained sections were examined by immunofluorescence microscopy using a BX-40 apparatus with an eXcope X3 digital camera.

Immunofluorescence
NeuN immunoreactivity was examined using immunofluorescence labeling of the same section 2, 4 and 7 days post-treatment and on primary hippocampal culture 24 h post-treatment. In brief, the sections and hippocampal cultures were blocked with 10% normal goat serum (Vector) in PBS-T, and then incubated overnight at 4uC with mouse anti-NeuN antibody (1:50 dilution) in PBS-T for neuron detection. After incubation with primary antibody, sections were exposed to FITC-labeled anti-mouse IgG (1:50 dilution) for 1 h at RT. After washing, immunofluorescence-stained sections and hippocampal cultures were examined by immunofluorescence microscopy using a BX-40 apparatus with an eXcope X3 digital camera and a Leica DM IRB apparatus (Leica Microsystems, Wetzlar, Germany) with a ProgResH CFscan digital camera (Jenoptik, Jena, Germany), respectively.

Semi-quantitative Analyses of FJB Intensity and NeuNpositive Reactions
The FJB intensity and NeuN immunoreactivity in the hippocampus were quantified using the ImageJ software (NIH, Bethesda, MD, USA). Mouse brains were sampled at approximately 2.12 mm behind the bregma. A standardized analysis area that contained 5-mm-thick coronal sections in a 1-in-10 series of sections representing the rostral/mid-hippocampus was used. For each mouse, three non-overlapping sections were analyzed, one from each of the three regions of the hippocampus (,50 mm apart). All positively labeled cells in the DG were quantified. The mean value of positive intensity in the three sections of each mouse was taken as n = 1. Intensities were expressed as means6SEM (n = 3).

Primary Hippocampal Cell Culture and Drug Treatment
The primary hippocampal cell culture method has been described previously [73]. Briefly, hippocampi were dissected from C57BL/6 mice pups at 17-18 gestational days, and prepared for culturing. After dissection, tissues were chopped and digested with 10 units/mL papain (Worthington, Freehold, NJ, USA) and 100 units/mL DNase I (Roche, Basel, Switzerland) in dissociation buffer at 37uC for 30 min. The digestion was triturated with Neurobasal A medium (Invitrogen, Carlsbad, CA, USA). The cells were seeded at a density of 0.3610 6 cells/well on poly-D-lysine hydrobromide (150 mg/mL; Sigma-Aldrich)-coated 24-well plates (NUNC TM ; Thermo Fisher Scientific, Inc.). Neurobasal A was replaced 1 h after plating with growth medium including Neurobasal A, 16 B27 supplement (Invitrogen), 100 units/mL penicillin, 0.1 mg/mL streptomycin, and 0.5 mM glutamine (Invitrogen). All cultures were maintained at 37uC and 5% CO 2 . Primary hippocampal cultured neurons at 12 DIV were treated with 0, 1, 5, or 10 mM of TMT and assayed 24 h post-treatment. To evaluate the cytoprotective effects of lithium on mature hippocampal neurons with TMT, lithium (0-10 mM) was added 1 h before TMT treatment (n = 6 cultures per condition).

Cytotoxicity Evaluation
Cytotoxicity in hippocampal cultured neurons was evaluated using a LDH release assay. A commercially available LDHcytotoxicity assay kit from Biovision (Mountain View, CA, USA) was used as recommended by the manufacturer. The optical densities at 490 nm were determined using a microplate reader (Emax, Molecular Devices).

Statistical Analysis
The data are reported as means6SEM and were analyzed by one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls post hoc test for multiple comparisons. In all analyses, p,0.05 was taken to indicate statistical significance. Figure S1 Immunoblot images for phospho-GSK-3a (Ser 21 ), total GSK-3a (,51 kDa), phospho-GSK-3b (Ser 9 ), total GSK-3b (,46 kDa), b-catenin (,92 kDa) and b-actin (,45 kDa) in the mouse hippocampus 2 (A), 4 (B) and 7 days (C) after TMT (2.6 mg/kg, i.p.) treatment. Cont, controls; Li, lithium-treated mice; TMT, TMT-treated mice; TMT+Li, TMT+lithium-treated mice. (TIF) Figure S2 Immunohistochemical assays of phospho-GSK-3a (Ser 21 ), phospho-GSK-3b (Ser 9 ), and active bcatenin expression levels in the adult mouse hippocampus after TMT treatment. The expression levels of phospho-GSK-3a (Ser 21 ), phospho-GSK-3b (Ser 9 ), and active b-catenin in the mouse hippocampus, especially the dentate granule cell blades, increased significantly at day 4 post-treatment. Cont, controls; TMT, TMT-treated mice. Scale bars represent 50 mm.  Figure S4 The preference for the two objects and the total number of interactions during training 7 days after TMT treatment. (A) The control, lithium-treated, TMTtreated, and TMT+lithium-treated mice showed equal preference for the two objects during training. (B) There was no significant difference in the interaction with the two training objects during training. The data are reported as the means6SEM. Cont, controls; Li, lithium-treated mice; TMT, TMT-treated mice; TMT+Li, TMT+lithium-treated mice. (TIF) Figure S5 Histopathological findings in the hippocampus of mice after TMT and TMT+lithium treatment. Low magnification images of adult mouse hippocampus (upper panels at each day post-treatment) and high magnification images of the DG in the hippocampus (lower panels at each day post-treatment) at 2, 4 and 7 days post-treatment. CA, cornu amonis; GCL, granular cell layer; DG, dentate gyrus. The sections were stained with hematoxylin and eosin. Scale bars represent 300 mm (upper panels at each day post-treatment) and 30 mm (lower panels at each day post-treatment).

(TIF)
Text S1 Supporting information for the Materials and Methods. (DOC)