Diacylglycerol Kinase β Knockout Mice Exhibit Lithium-Sensitive Behavioral Abnormalities

Background Diacylglycerol kinase (DGK) is an enzyme that phosphorylates diacylglycerol (DG) to produce phosphatidic acid (PA). DGKβ is widely distributed in the central nervous system, such as the olfactory bulb, cerebral cortex, striatum, and hippocampus. Recent studies reported that the splice variant at the COOH-terminal of DGKβ was related to bipolar disorder, but its detailed mechanism is still unknown. Methodology/Principal Findings In the present study, we performed behavioral tests using DGKβ knockout (KO) mice to investigate the effects of DGKβ deficits on psychomotor behavior. DGKβ KO mice exhibited some behavioral abnormalities, such as hyperactivity, reduced anxiety, and reduced depression. Additionally, hyperactivity and reduced anxiety were attenuated by the administration of the mood stabilizer, lithium, but not haloperidol, diazepam, or imipramine. Moreover, DGKβ KO mice showed impairment in Akt-glycogen synthesis kinase (GSK) 3β signaling and cortical spine formation. Conclusions/Significance These findings suggest that DGKβ KO mice exhibit lithium-sensitive behavioral abnormalities that are, at least in part, due to the impairment of Akt-GSK3β signaling and cortical spine formation.

DGKb, a member of the DGK family, is widely distributed in the brain, particularly in the olfactory bulb, cerebral cortex, striatum, and hippocampus [5]. DGKb is mainly expressed in neurons, and its expression rapidly increases after 14 days of age in the rat, which is coincident with synapse formation in the brain [6]. In addition, we and other researchers previously demonstrated that DGKb regulates spine formation by regulating lipids, and DGKb knockout (KO) mice show impaired memory [7,8]. Furthermore, it was reported that the splice variant at the COOH-terminal of DGKb was related to bipolar disorder [9]. These results indicate that DGKb plays an essential role in neurons, and its functions are closely related to neurodegenerative disease and mental diseases.
Bipolar disorder is mental disorder characterized by unusual shifts in mood from the heights of mania to the depths of depression. The lifetime prevalence of this disease is approximately 1%, and epidemiological surveys suggested that some genes such as Ankyrin 3 and DGKg contribute to the onset of this disorder [10,11]. A common treatment for bipolar disorder is lithium. This drug is comparatively effective, but it has a high incidence of adverse effects, in particular because lithium has a narrow therapeutic window. Therefore, new drugs which have both efficacy and fewer side effects are needed. The purpose of our present work was to investigate the involvement of DGKb in bipolar disorder. For this purpose, we have developed DGKb KO mice [8] and performed behavioral, pharmacological and morphological tests using these mice.

Animals
DGKb KO (C57BL6/N) mice were generated using the Sleeping Beauty transposon system as described previously [8,12,13]. This study was approved by the Animal Experiment Committee of Gifu Pharmaceutical University (permission number; 2008-114, 2009-038, 2009-062, and 2009-331), the Institutional Animal Care and Use committee of Kobe University (permission number; 19-5-02) and the committee for Safe Handling of Living Modified organisms in Kobe University (H19-2). All procedures relating to animal care and treatment conformed to animal care guidelines of these committees. All efforts were made to minimize both suffering and the number of animals used. The animals (male, 10-24 weeks old) were housed at 2462uC under a 12 hr light-dark cycle (lights on from 8:00 to 20:00) and had ad libitum access to food and water. In all experiments, we used wild-type (WT) littermates as a control group for DGKb KO mice. Behavioral experiments were performed between 10:00 a.m. and 6:00 p.m. except for the 24 hr home cage locomotor activity test.

Locomotor activity in the home cage
To measure locomotor activity in a novel environment, a mouse was placed in a transparent plastic cage (length 24.56width 17.56 height 12.5 cm) with sawdust bedding on the floor. ''Home cage'' in this draft means the same cage as they are usually housed, i.e. the same size and same color. Animals were placed in the cages at 12:00 p.m. and left there for a 48 hr period. Locomotion was measured every hour for 1 day after 24 hr accumulation using a digital counter with an infrared sensor (NS-ASS01; Neuroscience, Inc., Tokyo, Japan). Movement of the mice was detected by the infrared ray sensor on the basis of released infrared rays associated with their temperature. When objects emitting an infrared ray comes across the sensor, the sensor detected the action as locomotor counts.

Open field test
To assess the effect of a single-dose drug treatment, LiCl (100 and 200 mg/kg), haloperidol (0.1 mg/kg), diazepam (1.0 mg/kg), or imipramine (20 mg/kg) was dissolved in saline or carboxymethyl cellulose, and injected intraperitoneally. After drug treatment, each mouse was placed in the periphery of the open field apparatus (length 306 width 306 height 30 cm). Thirty minutes later, the total distance moved in the arena and the time spent in the center area (length 15 6 width 15 cm) were recorded for the 1 hr test session using a computer-operated EthoVision XT system (Noldus, Wageningen, The Netherlands). The number of scratching behaviors was manually counted for the first 10 min of each test session in a blind manner by a single observer (M.I.). For assessment of chronic lithium treatment, LiCl was mixed into the drinking water at 600 mg/L and given for 10 days. Subsequently, each mouse was placed in the periphery of the open field apparatus and recorded the total distance moved in the arena and the time spent in the center area for 1 hr using EthoVision XT.

Elevated plus maze test
Elevated plus maze test apparatus consisted of two open arms (length 30 6 width 5 cm) and two closed arms of the same size, along with a semi-transparent wall (height 15 cm) and central platform (length 5 6 width 5 cm). These arms and central platform were elevated 50 cm above the floor. Each mouse was administered diazepam (3 mg/kg, suspended in 0.5% carboxymethylcellulose solution) or vehicle i.p. Thirty-minutes after the drug administration, each mouse was placed in the central platform, facing one of the open arms. During a 10 min test session, mouse behavior was recorded using EthoVision XT. The number of entries into the open and closed arms, the time spent in the open arms, and the number of times falling from the apparatus were scored. This assessment session was conducted two times (pre or post test) before and after the 10 days of LiCl treatment.

Forced swim test
Mice were placed in a glass cylinder (diameter 10 cm) filled with 10 cm of water (2561uC) for a period of 6 min; only for the last 5 min was immobility time measured. Each mouse was administered imipramine hydrochloride (20 mg/kg, dissolved in saline) or vehicle i.p. 30 min before the trial. Mice were judged to be immobile when they remained floating passively in the water, making only small movements to keep their heads above the water.

Tail suspension test
Each mouse was administered imipramine hydrochloride (20 mg/kg, dissolved in saline) or vehicle i.p. Thirty min later, mouse was tail -suspended with an adhesive tape 50 cm above the floor, and their behavior recorded for 7 min. Immobility time was measured for the last 6 min of the test automatically with the aid of the EthoVision XT system. Mice were judged to be immobile when the mobility score of the system was less than 10%.

Histological staining
Mice were anesthetized with sodium pentobarbital (50 mg/kg) and perfused with phosphate buffered saline (PBS) (pH 7.4) until the outflow became clear, and was immediately followed by 0.1 M PBS (pH 7.4) containing 4% paraformaldehyde (Wako) for 15 min. Brains were removed and postfixed in the same fixative for 24 hr at 4uC. For cresyl violet staining, brain sections were equilibrated in 25% sucrose solution and quickly frozen in Tissue-Tek O.C.T. (Sakura Finetek, Torrance, CA, USA). Ten mm thick coronal sections (Bregma + 0.5 mm) were stained with cresyl violet (Sigma, St. Louis, MO, USA). For golgi staining, brain sections were immersed in 30% sucrose for 2 to 3 days. The tissue block was placed in 2% potassium dichromate for 2 days at 4uC and then in 2% silver nitrate solution for 2 days at 4uC in the dark. The block was cut into 60 mm thick and placed into distilled water. Finally, the sections were mounted onto slides, dried for 10 min, and dehydrated through 95% alcohol, 100% alcohol, and clear in xylene.

Statistical analysis
Data are presented as the means 6 S.E.M. Statistical comparisons were made by t-test or one-way ANOVA followed by Dunnett's test or Tukey's test using Statview version 5.0 (SAS Institute Inc., Cary, NC, USA), with p,0.05 being considered to indicate a statistical significance.

DGKb KO mice exhibited hyperactivity
In order to know the effects of DGKb deficit on the general behavior and circadian rhythm in mice, we first performed locomotor activity tests in the home cage on DGKb KO mice and their WT littermates. In this test, the activity of DGKb KO mice was relatively low during the light-phase, and markedly increased just as the dark-phase began, indicating that DGKb KO mice displayed a normal circadian rhythm (Figure 1a). However, the total (24 hr) and dark-phase locomotor activity counts were larger in KO mice than in WT mice (Figures 1a and b).
Next, DGKb KO mice were subjected to an open field test, and their activity and stereotyped behavior in the novel environment was compared. Total travel distance and the number of scratching behaviors were significantly greater in KO mice than in WT mice (Figures 2a-d). Moreover, administration of a mood stabilizer, lithium revealed that the horizontal activity of lithium-treated DGKb KO mice was decreased (Figure 2c). In comparison, this concentration of lithium did not affect WT mice locomotor (Figure 2c). On the other hand, other drugs, such as haloperidol (0.1 mg/kg, i.p.), diazepam (1.0 mg/kg, i.p.), or imipramine (20 mg/kg, i.p.), did not affect the lcomotoer activity of DGKb KO mice ( Figure S1a and d).

DGKb KO mice exhibited less anxiety
In open field test, we also assessed the anxiety level of DGKb KO mice by measuring their stay time in the center area of open field apparatus (anxiety-provoking area for mice). In this assessment, DGKb KO mice spent more time in the center area of the apparatus than WT mice (Figure 2e). To investigate the effect of lithium on fearless behavior of DGKb KO mice, mice were given a low-dose of lithium for 10 days, and their anxiety level in an open field test was measured. Low-dose lithium treatment did not reduce the activity of DGKb KO mice (Figure 2f). This paradigm of LiCl treatment produces a stable serum Li + concentration [14] which is at the low end of the therapeutic range for human patients [15]. On the other hand, fearless behavior of lithium-treated DGKb KO mice was significantly decreased when compared with vehicle-treated DGKb KO mice (Figures 2g and h). No significant changes were observed in total distance moved, duration in center zone, or frequency to center zone between vehicle-and LiCl-treated WT mice ( Figure S2a-c), which was consistent with previous reports [14].
To further assess anxiety levels of DGKb KO mice, we conducted another anxiety measuring behavioral test, an elevated plus maze test. In this test, total distance moved on the apparatus was not different between genotypes (WT vs. KO

DGKb KO mice decreased immobility in forced swim test and tail suspension test
The mouse model for mood disorders generally exhibits changes in their depressive states, i.e. model mouse of mania exhibits lower depressive states [14]; we evaluated these changes in DGKb KO mice using forced swim test and tail suspension test. In each test, both anti-depressant (imipramine)-treated WT mice and DGKb KO mice showed a reduction in the despair state, as assessed by immobility time in these tests (Figures 4a and b). DGKb KO mice did not exhibit impairment in sensorimotor gating and social interaction To further analyze the effects of DGKb deficits on psychomotor behavior, we evaluated schizophrenic-like behaviors in DGKb KO mice using a prepulse inhibition (PPI) test and a social interaction test (Method S1). In the PPI test, sensorimotor gating can be assessed following PPI of the startle reflex, which is the modulation of the startle response, following a weak prepulse. Sensorimotor gating is the neural process which allows attention to be focused on one stimulus. DGKb KO mice showed normal responses to both startle amplitude for the pulse-only trial ( Figure  S3a) and PPI ( Figure S3b). In a social interaction test, we assessed social affiliative behavior in DGKb KO mice, and the results revealed that DGKb KO mice also exhibited normal social interaction ( Figure S3c).

Phosphorylation levels of Akt and GSK3b
Using Western blot analysis, we evaluated the phosohorylation levels of Akt and GSK3b, which are the downstream of the proteins PKC and PA. Moreover, these proteins were reported as one of the pharmacological action mechanisms of lithium [16,17]. The total and phosohorylation levels of Akt and GSK3b proteins in the hippocampus and striatum were not different between WT mice and DGKb KO mice (data not shown). However, the phosphorylation levels of Akt (Ser473) and GSK3b were decreased in the cortex of DGKb KO mice despite the normal levels of total Akt and GSK3b proteins (Figures 5a-d). Furthermore, the treatment of lithium attenuated these reductions in the phosphorylated proteins (Figures 5a-d). The phosphorylation level of Akt (Thr308) was also decreased in the cortex of DGKb KO mice, compared with WT mice [WT;

Morphological changes in spine formation of the cortical neuron
To investigate the morphological factors relating to behavioral changes and the decrease in the phoshorylation levels of Akt and GSK3b in cortex, we first analyzed brain sections stained with cresyl violet. In this analysis, DGKb KO mice showed normal cortical laminar structure (Figures 6a-d). On the other hand, the results of Golgi staining revealed that the cortical spine density of DGKb KO mice was significantly decreased in comparison with that of WT mice (Figures 6 e-i).

Discussion
In the present study, we performed a comprehensive behavioral analysis of DGKb KO mice in order to investigate the role of DGKb in higher brain functions and the relationship between DGKb and bipolar disorder.
In the locomotor activity and open field tests, DGKb KO mice exhibited hyperactivity in their home cage and in a novel environment. Increased locomotor activities were observed in other rodent models of mental disease, such as during the mania state of bipolar disorder [18], attention deficit hyperactive disorder (ADHD) [16], and positive symptoms of schizophrenia [19,20]. In the open field test, LiCl at 100 mg/kg or more significantly attenuated the hyperactivity of DGKb KO mice. It is known that LiCl at these doses does not cause toxicity to mice [21] and, in the present study, LiCl did not affect WT mice locomotor behavior. On the other hand, haloperidol, diazepam, or imipramine did not improve locomotor behaviors in DGKb KO mice. These results indicate that the hyperactivity of DGKb KO mice is a lithiumsensitive behavioral abnormality, similar to the mania exhibited in some animal models. DGKb KO mice also displayed more frequent scratching behavior than WT mice. Stereotypical behaviors such as scratching and head twitching are induced by the administration of psychostimulants [22], and it is also observed in mouse model of psychiatric disorders [23].
Using an open field test or an elevated plus maze test, one is able to evaluate anxiety levels in rodent models. In the open field test, DGKb KO mice spent more time in the center of the apparatus than WT mice, indicating that DGKb KO mice showed less-anxiety. Moreover, chronic treatment with a low-dose (600 mg/L) of lithium decreased fearless behavior in KO mice without decreasing their locomotor activity. With this dose of LiCl treatment, a serum concentration of Li + has reported to be stable at approximately 0.41 mmol/L [14], a concentration which is typically used to treat mania patients [15]. When the effects of the drug on animal anxiety level are measured, a low concentration is recommended because high concentration of LiCl causes a decrease in mice locomotor activity. In an elevated plus maze test, DGKb KO mice displayed enhanced open arm selectivity, which also indicates lowered anxiety of DGKb KO mice. Chronic LiCl treatments in DGKb KO mice also attenuated these behavioral changes. During each session, the activity of each group of mice was not different significantly, and therefore these results verify that lithium specifically inhibits fearless behaviors of DGKb KO mice in a manner not due to change in their locomotor activity. These lithium-sensitive changes of anxiety in mice are observed especially in the animal model of mania [14].
Forced swim test and tail suspension test are commonly used to evaluate depression-like behaviors of rodent models. Because existing antidepressants specifically inhibit the immobile time in these tests, the utility of these experiments is recognized. In the present study, DGKb KO mice exhibited antidepressant-like behavioral changes in both tests.
In addition to the variety of behavioral changes shown above, we previously investigated DGKb KO mice exhibiting impaired memory [8]. The cognitive functioning of DGKb KO mice were analyzed, and these were less affected by locomotor activity or swimming ability [8]. Therefore, the relationship between hyperactivity and memory impairment in DGKb KO mice is weak. On the other hand, DGKb KO mice did not show schizophrenia-like behavioral changes, as typified by PPI and social interaction deficits ( Figure S3). Taken together, DGKb KO mice present lithium-sensitive excitatory psychomotor effects related to their mood state and cognitive impairment. These characteristics of behavior are similar to those observed in mouse models of mania [14] and ADHD [16].
Next, to analyze the mechanisms of abnormal behavior generation in DGKb KO mice and the lithium sensitive mechanisms in DGKb KO mice, we assessed the phosophorylation levels of Akt and GSK3b, which lie downstream of PKC and PA. The activity of DGK to catalyze the phosphorylation of DG into PA is an intrinsic component of the phosphatidylinositol cycle. Additionally, PA contributes to the activation of Akt/phosphatidylinositol 3-kinase (PI3K) signaling by stimulating PI4P5-kinase [24]. Indeed, using DGKf-deficient mice, DGKf has been reported to be involved in the PI3K/Akt pathway [25] and phosphorylated Akt inhibits the activity of GSK3b by phosphorylating at Ser9 [26,27]. It was reported that Akt and GSK3b are one of target proteins of lithium; lithium increases the phosphorylation levels of Akt and GSK3b [16,28,29,30,31]. In the present study, the phosophorylation levels of Akt (Ser473) and GSK3b were decreased in the cortex of DGKb KO mice. In addition, lithium significantly attenuated the alteration of these phosophorylation levels in DGKb KO mice. The phosphorylation level of Akt (Thr308) was also decreased in the cortex of DGKb KO mice, compared with WT mice. These decreases in levels of phosphorylated Akt and GSK3b were observed in the striatum of dopamine transporter (DAT) KO mice [16]. The effects of lithium on phosphorylated Akt and GSK3b in DGKb KO mice are also similar to those observed in DAT KO mice [16]. Transgenic mice overexpressing GSK3b also display mania-like behaviors [32]. Additionally, b-catenin, the downstream protein of GSK3b over expression mice exhibit lithium-sensitive behaviors [33]. On the other hand, lithium attenuates the behavior of GSK3b heterogeneous mice in locomoter activity and tail suspension test [34]. We have also investigated the changes of phospho-GSK3a and tau hyperphosphorylation (using the AT8 antibody) in the cortex of WT and DGKb KO mice. The protein level phospho-GSK3a was decreased in DGKb KO mice compared with WT mice ( Figure  S5c). On the other hand, the level of tau hyperphosphorylation in DGKb KO mice trended to increase, but did not reach a significant level ( Figure S4e). Although we can not fully explain the reason that the level of tau hyperphosphorylation was not significant, it is possible that the phosphorylation of tau may be regulated by multiple kinases. Especially, tau pshosphorylation at Ser 202 (detected by AT8) is regulated not only by GSK3b, but also by cyclin-dependent protein kinase 5 (cdk5) [35]. On the other hand, GSK activity reflected the ratio of phosphor-Ser9 GSK3b to total GSK3b levels, as others reported [36]. These results suggest that the alteration of GSK3b function affected mood-behavior in mice, and the lithium-sensitive excitatory mood behaviors of DGKb KO mice are partly caused by the reduced phosphorylation levels of Akt and GSK3b. Some reports have shown that lithium affects both Akt and GSK3 phosphorylation in the striatum of WT animals [36]. On the other hand, in the present study we focused solely on the cortical changes of Akt and GSK3 phosphorylation. Effects of lithium on Akt and GSK3 phosphorylation may vary depending upon the brain region or the mouse strain. Additionally, our results showing that LiCl (200 mg/kg i.p.) did not affect locomotor activity in WT mice (Figure 2c) may support the results of Akt and GSK3 phosphorylation in WT cortex.
Previously, we and other groups have reported the contribution of DGKb to hippocampal spine formation [7,8]. Our present study revealed that impaired spine formation also occurred in the cortex of DGKb KO mice without changing cortical laminar structure. The decrease of spine density in cortex was observed in many animal models expressing abnormal psychomotor behaviors [20,37,38]. In addition, postmortem brain analysis revealed that spine density on the primary apical dendrites of the cortex layer III pyramidal neurons were also decreased in psychiatric diseased patients [39,40]. As the reports implied, dendrite spine dysplasia led to the expression of abnormal behavior, and in DGKb KO mice the decrease of cortical spine density, in turn, may have led to the observed abnormalities in behavior.
In conclusion, DGKb KO mice exhibited lithium-sensitive excitatory psychomotor behaviors related to their mood state. The phenotypes of DGKb KO mice may be caused, at least in part, by the impairment of Akt-GSK3b signaling and cortical spine dysplasia.

Supporting Information
Method S1 Prepulse inhibition and social interaction test.  and KO mice (n = 9). (a) In the 120 dB-pulse-only trials, startle amplitude did not differ significantly between DGKb KO and WT mice. (b) The PPI is expressed as a percentage of the startle response to a 120 dB-pulse. DGKb KO mice showed normal PPI at each prepulse intensity. (c) Social interaction test in a novel environment in WT (n = 8) and KO (n = 8) mice. Two genetically identical mice that had been housed separately were placed in the same cage. Their social interaction was then monitored for 10 min. There was no significant difference in duration per contact between WT and DGKb KO mice. Found at: doi:10.1371/journal.pone.0013447.s004 (1.03 MB TIF) Figure S4 Western blot analysis in Akt-GSK3b signaling. Effects of LiCl on Akt-GSK3b signaling in the cortex of WT mice were measured (a-e). LiCl (200 mg/kg, i.p.) or vehicle was administrated at 30 min before the western blotting. (a) Representative images of immunoblottin showing p-GSK3a/b, total GSK3a/b, p-Akt (Ser473), total Akt, and b-actin. Quantitative analysis of (b) p-GSK3a/GSK3a, (c) p-GSK3b/GSK3b, and (d) p-Akt (Ser473)/Akt (n = 7). (e) Tau phosphorylation (using the AT8 antibody) in the cortex of WT and DGKb KO mice (n = 5 and 6).