Lithium reverses increased rates of cerebral protein synthesis in a mouse model of fragile X syndrome

Abstract

Individuals with fragile X syndrome (FXS), an inherited form of cognitive disability, have a wide range of symptoms including hyperactivity, autistic behavior, seizures and learning deficits. FXS is caused by silencing of FMR1 and the consequent absence of fragile X mental retardation protein (FMRP). FMRP is an RNA-binding protein that associates with polyribosomes and negatively regulates translation. In a previous study of a mouse model of FXS (Fmr1 knockout (KO)) we demonstrated that in vivo rates of cerebral protein synthesis (rCPS) were elevated in selective brain regions suggesting that the absence of FMRP in FXS may result in dysregulation of cerebral protein synthesis. Lithium, a drug used clinically to treat bipolar disorder, has been used to improve mood dysregulation in individuals with FXS. We reported previously that in the Fmr1 KO mouse chronic dietary lithium treatment reversed or ameliorated both behavioral and morphological abnormalities. Herein we report that chronic dietary lithium treatment reversed the increased rCPS in Fmr1 KO mice with little effect on wild type mice. We also report our results of analyses of key signaling molecules involved in regulation of mRNA translation. Our analyses indicate that neither effects on the PI3K/Akt nor the MAPK/ERK 1/2 pathway fully account for the effects of lithium treatment on rCPS. Collectively our findings and those from other laboratories on the efficacy of lithium treatment in animal models support further studies in patients with FXS.

Introduction

Individuals with fragile X syndrome (FXS), an inherited form of intellectual disability, show a broad spectrum of morphologic, cognitive, behavioral, neurological, and psychological features (Chonchaiya et al., 2009). FXS is caused by the silencing of the FMR1 gene and the consequent absence of its protein product, fragile X mental retardation protein (FMRP) (Brown, 2002). FMRP is a RNA-binding protein that associates with polyribosomes and negatively regulates translation of certain mRNAs (Penagarikano et al., 2007). It is thought that absence of FMRP results in a dysregulation of protein synthesis and that it is this dysregulation that has such profound consequences for development and function of the nervous system. Protein synthesis in the nervous system is required for normal nervous system development and maintenance. It is also essential for implementation of long lasting changes in synaptic strength such as occur in long term potentiation (LTP) and long term depression (LTD). Some forms of both LTD and LTP are impaired in Fmr1 knockout (KO) mice (Huber et al., 2004; Larson et al., 2005; Li et al., 2002; Suvrathan et al., 2010; Wilson and Cox, 2007; Zhao et al., 2005), and it has been proposed that a dysregulation of protein synthesis may underlie these impairments.

In accord with the idea that a dysregulation of protein synthesis may be at the heart of the defect in FXS, we have shown that adult male Fmr1 KO mice have elevated regional rates of cerebral protein synthesis (rCPS) compared with wild type (WT) littermates (Qin et al., 2005). We measured rCPS in vivo with a quantitative autoradiographic method (Smith et al., 1988). Effects were regionally selective occurring mainly in hippocampus, thalamus and hypothalamus. Other investigators have monitored the incorporation of ³⁵S-methionine/cysteine into new protein in vitro. Their findings also indicate that incorporation is higher in hippocampal slices from Fmr1 KO mice (Dölen et al, 2007; Osterweil et al, 2010).

Treatment for FXS is largely symptom-based and multidisciplinary. Current treatments include special education, behavioral interventions, physical therapy and symptom-directed pharmacotherapy. There are several classes of pharmacotherapeutic agents currently under investigation in FXS that are aimed at the biochemical processes thought to underlie the disease. Among them are mGluR antagonists, antibiotics, and GABA agonists (Levenga et al., 2010). Lithium, an effective mood stabilizer for the treatment of bipolar disorder, has been used successfully in individuals with FXS to stabilize mood dysregulation (Al-Semaan et al., 1999). Results of a pilot add-on trial demonstrated that lithium may improve behavior, adaptive skills, and verbal memory in patients with FXS (Berry-Kravis et al., 2008). We and other groups have demonstrated that chronic dietary lithium treatment ameliorates many of the behavioral deficits seen in Fmr1 KO mice (Liu et al., 2011; Mines et al., 2010; Yuskaitis et al., 2010). Moreover, lithium treatment partially normalized the immature dendritic spine morphology in medial prefrontal cortex (Liu et al., 2011). Lithium has also been shown to modify abnormal behavior and morphology in a Drosophila model of FXS (McBride et al., 2005). Taken together results from these studies suggest that lithium could provide significant therapeutic benefits in FXS.

How lithium treatment may effect these therapeutic outcomes is not understood. In the current study, we measure rCPS in WT and Fmr1 KO mice to investigate whether lithium may act through an effect on the control of translation. Our results indicate that, in addition to its effects on behavior and morphology, lithium also normalizes the elevated rCPS in Fmr1 KO mice. We also present in this paper effects of lithium treatment on some of the steps in signaling pathways that control cap-dependent translation.

Materials and methods

Animals

Male WT and Fmr1 KO mice (n=79), generated by FVB/NJ-fmr1^tm1Cgr breeding pairs (heterozygous females and homozygous males), were used. The generation of Fmr1 KO mice and their genotyping by PCR amplification of tail DNA were described previously (Qin et al., 2005). All mice were group housed in a central facility and maintained under controlled conditions of normal humidity and temperature with standard alternating 12-hr periods of light and darkness. All procedures were carried out in accordance with the National Institutes of Health Guidelines on the Care and Use of Animals and approved by the National Institute of Mental Health Animal Care and Use Committee.

Four groups of mice were studied: 1) WT fed control diet (WT-C) (n=23); 2) KO fed control diet (KO-C) (n=22); 3) WT fed lithium-supplemented diet (WT-Li) (n=17); 4) KO fed lithium-supplemented diet (KO-Li) (n=17). The base diet was NIH-31 and the lithium-supplemented diet was NIH-31 to which 0.3% (w/w) lithium carbonate had been added (Harlan Teklad, Madison, WI). Feeding commenced at weaning (3 weeks of age) and was provided ad libitum until mice were studied at 12 weeks of age. Mice fed lithium-supplemented chow were given drinking water with 1.5% (w/v) sodium chloride to counteract potential toxicity of lithium. Mice fed control chow were provided tap water. As we reported previously (Liu et al, 2011), whole blood lithium concentrations were 0.38 ± 0.03 (n = 4) and 0.38 ± 0.02 mEq/L (n = 4) in WT and Fmr1 KO mice, respectively. No toxicity of the lithium treatment was observed. This concentration is equivalent to a plasma lithium level that is at the low end of the therapeutic range of plasma lithium for treatment of bipolar disorder in human subjects (0.4-1.2 mEq/L).

2.2. Determination of rCPS

Mice, under light isoflurane anesthesia, were prepared for studies of rCPS by insertion of polyethylene catheters (PE-10) into a femoral artery and vein. Mice recovered from the surgery overnight and food and water were available ad libitum. Body temperature was maintained by warming the bottom of the cage with hand warmers (Grabber, Grand Rapids, MI). Mice were permitted to move freely throughout the recovery and experimental periods. To monitor the physiological state of each mouse, mean arterial blood pressure, hematocrit and arterial plasma glucose concentrations were measured at 0 and 45 min during the rCPS experiment.

The experimental period was initiated by an intravenous pulse injection of 100 μCi/kg of L-[1-¹⁴C]leucine (specific activity, 60 mCi/mmol, Moravek Biochemicals and Radiochemicals, Brea, CA). Timed arterial samples were collected during the following 60 min for determination of the time courses of plasma concentrations of leucine and [¹⁴C]leucine. Labeled and unlabeled leucine concentrations in the acid-soluble fractions of arterial plasma were assayed by liquid scintillation counting and amino acid analysis, respectively. At the end of the experimental interval, brains were removed and frozen, and serial sections, 20 μm in thickness, were prepared for quantitative autoradiography by means of a Leica cryostat (Leica Microsystems, Inc, Buffalo Grove, IL). Sections were mounted on gelatin-coated slides, fixed and washed in 10% formalin, and exposed to Ektascan B/RA film (Eastman Kodak, Rochester, NY) along with calibrated [¹⁴C]methylmethacrylate standards as described previously (Qin et al., 2005). Autoradiograms were digitized (MCID Analysis, Interfocus Imaging Ltd, Linton, Cambridge, UK), the concentration of ¹⁴C in each region of interest was determined, and rCPS was calculated by means of the operational equation of the method (Smith et al, 1988). The value of lambda in the equation was 0.603 (Qin et al, 2005). Brain regions were identified by reference to a mouse brain atlas (Paxinos and Franklin, 2001).

2.3. Western blotting

At 12 weeks of age, 32 mice (n = 8 per group) were decapitated and brains were removed quickly and placed on ice. Hippocampi were dissected immediately, and snap frozen in dry ice. To prepare lysates, frozen tissues were thawed on ice and homogenized in 5% (w/v) ice-cold T-PER tissue protein extraction reagent (Thermo Scientific, Rockford, IL) with 1% Halt protease inhibitor (Thermo Scientific) and 1% phosphatase inhibitor cocktails (Sigma-Aldrich, St. Louis, MO). Homogenates were centrifuged (12,000 × g, 4°C, 15 min) and supernatant fractions were collected. Protein concentrations were determined by BCA protein assay kit (Thermo Scientific). Protein samples were subjected to electrophoresis on NuPAGE 10% Bis–Tris or 3-8% Tris-Acetate gels (Invitrogen Life Technologies, Grand Island, NY), and transferred electrophoretically to nitrocellulose membranes. Membranes were incubated with phospho-Akt (Ser 473) monoclonal antibody (1:4,000), Akt antibody (1:5,000), phospho-mTOR (Ser2448) antibody (1:1,000), mTOR antibody (1:2,000), phospho-p70 S6K1 (Thr389) antibody (1:500), p70 S6K1 (Thr389) antibody (1:1000), phospho-p44/42 ERK1/2 (Thr202/Tyr204) antibody (1:2000), p44/42 ERK1/2 antibody (1:5000), β-tubulin monoclonal antibody (1:8,000). β-tubulin was used as a loading control. All antibodies were obtained from Cell Signaling Technology, Inc (Danvers, MA). Immunoblots were developed by Western Breeze Chemiluminescent kit (Invitrogen). Signals of protein bands were quantified by densitometry with the MCID image processing system.

2.4. Immunohistochemistry

Mice (5/group, WT-C & KO-C; 2/group, WT-Li & KO-Li) were deeply anesthetized and perfused transcardially with normal saline containing 2% sodium nitrite followed by 4% paraformaldehyde containing 2.5% acrolein. Brains were removed and immersed in 30% sucrose at 4 °C. Cryostat sections, 30 μm in thickness, were prepared. Floating sections were incubated with p-mTOR (Ser 2448) rabbit monoclonal antibody (Cell signaling) (1:3,000) for 1 h at room temperature followed by 48 h at 4 °C, and then secondary biotinylated anti-rabbit IgG (Vector Laboratories, Burlingame, CA) (1:600). Sections were stained with nickel-3,3′-diaminobenzidine chromogen (Sigma-Aldrich, St. Louis, MO) (Hoffman et al, 2008).

2.5. Statistical analysis

The rCPS data were analyzed by 3-way repeated measures analysis of variance (RM ANOVA) with genotype (WT, KO) and treatment (untreated, lithium-treated) as between-subjects factors and brain region as a within-subjects factor. The Western blotting data were analyzed by 2-way ANOVA with genotype and treatment as factors. Post hoc pair-wise comparisons were further conducted with Bonferroni t-tests. The significance level was set at p ≤ 0.05. We used the SPSS program (IBM, Armonk, NY) for statistical computations.

3. Results

3.1. Cerebral protein synthesis

We studied WT and Fmr1 KO mice under two conditions, control with normal rodent chow and experimental with lithium-supplemented chow. Physiological variables were measured at the time of determination of rCPS. The four groups were well matched with regard to age, body weight, and physiological status (Table 1). Arterial plasma leucine concentration was affected by the lithium treatment; leucine concentrations were lower in the lithium-treated mice regardless of genotype (F_{(1, 29)} = 30.51, p < 0.001).

Table 1.

Physiological variables

	WT-C (10)	KO-C (9)	WT-Li (7)	KO-Li (7)
Age (days)	87.4 ± 0.9	87.0 ± 0.9	87.0 ± 1.0	87.1 ± 0.9
Body weight (g)	28.0 ± 0.5	29.4 ± 0.7	28.5 ± 0.9	27.7 ± 0.7
Hematocrit (%), 0 min	43.3 ± 0.4	42.4 ± 0.5	42.1 ± 0.7	42.2 ± 0.8
45 min	39.0 ± 0.4	38.6 ± 0.5	38.5 ± 0.6	38.4 ± 0.9
Arterial plasma glucose concentration (mg/dl), 0 min	9.1 0± 0.25	7.94 ± 0.38	8.29 ± 0.64	8.57 ± 0.72
45 min	9.12 ± 0.18	8.33 ± 0.48	8.73 ± 0.71	8.69 ± 0.42
Mean arterial blood pressure (mmHg), 0 min	116.6 ±2.4	112.0 ± 1.6	126.5 ± 3.4	125.0 ± 2.3
45 min	113.8 ± 3.8	107.3 ± 2.6	117.7 ± 4.4	115.2 ± 2.1
Arterial plasma leucine concentration (nmol/ml)	135.5 ± 4.2	142.7 ± 4.5	116.7 ± 7.5	104.8 ± 4.1

Values were means ± SEM for the number of mice indicated in parentheses. Arterial plasma leucine concentrations are the means for each animal of samples drawn over the 60 min experimental interval weighted by the integrated [¹⁴C]leucine activity during the sampling interval. There were no statistically significant interactions between genotype and treatment (2-way ANOVA) for any of the variables measured. The main effect of treatment, but not genotype, was statistically significant for arterial plasma leucine concentration (P<0.001).

We determined rCPS in 27 regions of brain (Fig. 1&2). In general, rCPS values were higher in KO-C mice compared with WT-C mice (3-20%). With lithium treatment rCPS values were decreased in both genotypes, but the effects were consistent and significant only in Fmr1 KO mice. Data were analyzed by means of 3-way RM ANOVA with genotype, treatment and region as factors and repeated measures on region. The 3-way interaction was statistically significant (F_{(4.8, 139.6)} = 4.50, p < 0.01). Post hoc pair-wise comparisons revealed that compared with WT-C mice, rCPS values were higher (p ≤ 0.05) in KO-C mice in 6 out of 27 regions including basolateral amygdala, paraventricular nucleus, supraoptic nucleus, arcuate nucleus, CA1 pyramidal cell layer of hippocampus and granule cell layer of dentate gyrus (Fig.1&2). Increases ranged from 11-20% in these regions. In posterior parietal cortex rCPS was 9% higher in KO-C compared with WT-C (p = 0.055). Values of rCPS in other regions analyzed were 3-10% higher in KO-C than WT-C mice, but these changes were not statistically significant. In Fmr1 KO mice lithium treatment resulted in statistically significant 9-20% decreases in rCPS in 23 of the 27 regions analyzed. In the other four regions decreases approached statistical significance (0.05 ≤ p ≤ 0.10). In WT mice, decreases (1-11%) in rCPS after lithium treatment were smaller than in Fmr1 KO mice and none reached or approached statistical significance. These data indicate that chronic lithium treatment may have very minor effects on rCPS in WT mice, but lithium reverses increased rCPS in Fmr1 KO mice.

Fig. 1.

Effects of chronic lithium treatment on rCPS in WT and Fmr1 KO mice. Bars represent the means ± SEM for WT-C (n-10), KO-C (n=9), WT-Li (n=7) and KO-Li (n=7) mice. Abbreviations are as follows: FrA, frontal association cortex; mPFr, medial prefrontal cortex (infralimbic and prelimbic); Cg, anterior cingulate cortex; M1, primary motor cortex; S1, somatosensory cortex; PPtA, posterior parietal cortex; VHi, ventral hippocampus; CA1 pyr, CA1 sector of the pyramidal cell layer of ventral hippocampus; Rad, stratum radiatum of ventral hippocampus; GrDG, granule cell layer of the dentate gyrus of ventral hippocampus; BLA, basal lateral amygdala; CeA, central amygdaloid nucleus; DTh, dorsal thalamus; Vth, ventral thalamus; LHb, lateral habenula; MHb, medial habenula; LH, lateral hypothalamic area; PVN, paraventricular nucleus of the hypothalamus; SON, supraoptic nucleus, Arc, arcuate nucleus; SuMM, Supramammillary nucleus; S, septal nucleus; CPu, caudate putamen; DR, dorsal raphe; MR, median raphe; Ce, cerebellum; CC, corpus callosum. Data were analyzed by means of 3-way RM ANOVA, followed by pair-wise comparisons with Bonferroni t-tests. The 3-way interaction (genotype x treatment x region) was statistically significant (F_(4.8,139.6)=4.50, p<0.01).

Post-hoc comparisons:

Statistically significantly different from WT-C, *, 0.01 ≤ P ≤ 0.05; **, 0.001 ≤ P ≤ 0.01; ^#, P ≤ 0.06 Statistically significantly different from KO-C, ^†, 0.01 ≤ P ≤ 0.05; ^††, 0.001 ≤ P ≤ 0.01

Fig. 2.

Digitized autoradiographic images color-coded for rCPS at the level of dorsal hippocampus (A-D) and at the level of the anterior hypothalamus (E-H). Representative images at the level of dorsal hippocampus from each experimental group are as follows: (A) WT-C, (B) KO-C, (C) WT-Li, (D) KO-Li. Upper color bar and the scale bar in A (2 mm) apply to all four images (A-D). Representative images from each experimental group at the level of anterior hypothalamus are as follows: (E) WT-C, (F) KO-C, (G) WT-Li, (H) KO-Li. Lower color bar and the scalebar in E (1 mm) apply to all four images (E-H). The paraventricular nuclei (PVN) are indicated in E.

3.2. Akt/mTOR and MAPK/ERK1/2 signaling

It has been proposed that the therapeutic actions of lithium likely involve alterations in signaling pathways and gene expression (Chiu and Chuang, 2010). One of the principal targets of lithium may be the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt) pathway. We examined the phosphorylation state of Akt in hippocampus of control and lithium-treated mice and found that lithium decreased phospho-Akt (p-Akt) (Ser 473) in both WT (20% decrease, p=0.02) and Fmr1 KO (31% decrease, p<0.001) mice and had no effects on total Akt (Fig. 3A&B). In untreated mice, average p-Akt was 16% higher in Fmr1 KO mice compared with WT (p<0.01), whereas in lithium-treated mice there was no difference between the two genotypes. Neither the interaction between genotype and treatment (F_{(1, 28)} = 2.81, p = 0.105) nor the main effect of genotype (F_{(1, 28)} = 3.189, p = 0.08) reached statistical significance. The main effect of treatment was statistically significant (F_{(1, 28)} = 35.0, p < 0.001), indicating that chronic lithium decreased the level of p-Akt regardless of genotype. With respect to the level of total Akt, neither the interaction between genotype and treatment (F_{(1, 28)} = 0.06, NS) nor the main effect of treatment (F_{(1, 28)} = 1.28, NS) was statistically significant. The main effect of genotype was statistically significant (F_{(1, 28)} = 7.969, p < 0.01), but the effect was very small (4% higher in KO compared with WT).

Fig. 3.

Western blot analysis and representative Western blots of (A) p-Akt (ser473), (B) total Akt, (C) p-mTOR (ser2448), (D) total mTOR levels (E) p-p70 S6K1 (Thr389), and F p70 S6K1 in hippocampal lysates. Optical densities were normalized to β-tubulin and values were expressed as percent of WT-C. A. p-Akt, both the genotype x treatment interaction (F_{(1, 28)} = 2.81, p = 0.11) and the main effect of genotype (F_{(1, 28)} = 3.19, p = 0.09) approached statistical significance, and the main effect of treatment was statistically significant (F_{(1, 28)} = 35.00, p < 0.001). Because the interaction only approached statistical significance pairwise comparisons are not strictly valid. B. Total Akt, neither the genotype x treatment interaction (F_{(1, 28)} = 0.06, NS) nor the main effect of treatment (F_{(1, 28)} = 1.28, NS) was statistically significant, but the main effect of genotype (F_{(1, 28)} = 7.97, p < 0.01) was. C. p-mTOR, the genotype x treatment interaction (F_{(1, 28)} = 8.29, p < 0.01) was statistically significant. D. Total mTOR, the genotype x treatment interaction (F_{(1, 28)} = 0.06, NS) was not statistically significant, neither was the main effect of treatment (F_{(1, 28)} = 0.29, NS) nor genotype (F_{(1, 28)} = 0.02, NS). E. p-p70 S6K1, the genotype x treatment interaction (F_{(1, 28)} = 1.79, NS) was not statistically significant, neither was the main effect of treatment (F_{(1, 28)} = 0.57, NS) nor genotype (F_{(1, 28)} = 0.66, NS). F. Total p70 S6K1, the genotype x treatment interaction (F_{(1, 28)} = 0.41, NS) was not statistically significant, neither was the main effect of treatment (F_{(1, 28)} = 0.91, NS) nor genotype (F_{(1, 28)} = 0.20, NS). Results of post-hoc Bonferroni t-tests are shown on the figure; *, p ≤ 0.05; **, p ≤ 0.001.

We also examined the phosphorylation state of mammalian target of rapamycin (mTOR) in hippocampus from treated and untreated WT and Fmr1 KO mice (Fig. 3C&D). The signaling kinase, mTOR, is downstream from PI3K/Akt, and is a pivotal regulator of protein synthesis. mTOR has been shown to participate in control of cell growth and proliferation and is thought to be involved in synaptic plasticity (Jaworski and Sheng, 2006). In untreated mice, the average level of phospho-mTOR (p-mTOR) (Ser2448) was higher in Fmr1 KO mice by 11% (p=0.05). Treatment with lithium lowered p-mTOR by 11% in Fmr1 KO mice (p=0.03) and increased it by 9% in WT (NS). For p-mTOR the interaction between genotype and treatment was statistically significant (F_{(1, 28)} = 8.29, p < 0.01) indicating a difference in the response to lithium treatment between the two genotypes. Total mTOR levels did not differ significantly among groups (Interaction: F_{(1, 28)} = 0.06, NS; genotype: F_{(1, 28)} = 0.02, NS; treatment: F_{(1, 28)} = 0.29, NS).

We further examined the hippocampus for changes in p-mTOR by means of immunohistochemistry in a parallel series of mice (Fig. 4). Expression of p-mTOR in WT-C was highest in the fasciola cinereum and in the polymorph layer of the dentate gyrus. Staining was also intense in layers just above and below the pyramidal cell layer of CA1. The heavier band of staining is the innermost portion of the stratum oriens, the locus of the basal dendrites of the pyramidal cells. The layer of staining below the pyramidal cell layer is the innermost portion of stratum radiatum and the locus of the portions of apical dendrites closest to the soma. There is also discernable staining in the granule cell layer of the dentate gyrus and at the hippocampal fissure. Comparison of the staining pattern in KO-C with WT-C suggests that staining in the layers above and below the pyramidal cells and in stratum radiatum is higher in the KO-C mice. Following lithium treatment this enhanced staining in the KO mice appears to be abated.

Fig. 4.

Immunohistochemical localization of p-mTOR (Ser 2448) in the dorsal hippocampus from (A & B) WT-C, (C & D) WT-Li, (E & F) KO-C, and (G & H) KO-Li mice. Scale bars in A and B are 250 and 25 μm, respectively. Boxes in the low power images (A, C, E, & G) indicate the area from which the high power images were obtained (B, D, F, & H). Staining is particularly intense in fasciola cinereum (visible in C & D), CA2, and in the polymorph layer of the dentate gyrus. Staining is also noteworthy in layers just above and below the pyramidal cell layer of CA1. The upper band of staining is the innermost portion of the stratum oriens, the locus of the basal dendrites of the pyramidal cells. The layer of staining below the pyramidal cell layer is the innermost portion of stratum radiatum and the locus of the portions of apical dendrites closest to the soma. In the higher power images (B, D, F, & H) there is clear activity in the cytoplasm of the pyramidal cells and in the pyramidal cell dendrites. There is also discernable staining in the granule cell layer of the dentate gyrus and at the hippocampal fissure. High p-mTOR activity in pyramidal cell dendrites (particularly proximal segments) (F) appears to be normalized in lithium-treated mice (H).

One of the downstream targets of mTOR is a ribosomal protein p70 S6 kinase (p70 S6K1). We found no differences in the relative levels of either the total or phosphorylated forms of p70 S6K1 in hippocampal extracts between WT and Fmr1 KO mice with or without chronic lithium treatment (Fig. 3E&F).

We also considered the possibility that lithium may act through effects on the mitogen-activated protein kinase (MAPK)/ extracellular-regulated kinase (ERK1/2) pathway. We analyzed the effects of lithium treatment on phosphorylated (Thr202/Tyr204) and total ERK1/2 in hippocampal homogenates. Similar to our results on p70 S6K1 we found no statistically significant effects, although there was a tendency for p-ERK1/2 levels to be higher in untreated Fmr1 KO mice (Fig. 5).

Fig. 5.

Western blot analysis and representative Western blots of (A) phospho-p44/42 ERK1/2 (Thr202/Tyr204), (B) total p44/42 ERK1/2 in hippocampal lysates. Optical densities were normalized to β-tubulin and values were expressed as percent of WT-C. A. p-p44/42 ERK1/2 (Thr202/Tyr204), the genotype x treatment interaction (F_{(1, 28)} = 0.56, NS) was not statistically significant, neither was the main effect of treatment (F_{(1, 28)} = 0.35, NS) nor genotype (F_{(1, 28)} = 0.80, NS). B. Total p44/42 ERK1/2, the genotype x treatment interaction (F_{(1, 28)} = 0.13, NS) was not statistically significant, neither was the main effect of treatment (F_{(1, 28)} = 2.30, p=0.14) nor genotype (F_{(1, 28)} = 1.83, NS).

4. Discussion

The central finding of our study is that chronic treatment with lithium reversed the increased rCPS found in Fmr1 KO mice and had little, if any, effect on rCPS in WT mice. Increased rCPS in the untreated Fmr1 KO mice were found primarily in the limbic system and hypothalamus, whereas the effects of lithium occurred throughout the brain. We extended our studies to examine in hippocampus the effects of lithium treatment on some of the signaling pathways that influence translation. Our results indicate that neither effects on the PI3K/Akt nor the ERK 1/2 pathway can fully account for the effects on rCPS under basal conditions.

The protein absent in Fmr1 KO mice is FMRP, an RNA-binding protein that negatively regulates translation of specific mRNAs (Laggerbauer et al., 2001; Li et al, 2001; Zhang et al., 2001). We posit that a dysregulation of general translation is the primary defect of FXS that underlies the wide range of signs and symptoms of the disease. Protein synthesis is an integral part of the plasticity response in nervous tissue. As such, its tight regulation is vital for this critical brain function. We have shown previously that adult Fmr1 KO mice have elevated rCPS in selective regions of the brain including the hippocampus, parts of the hypothalamus and thalamus, amygdala, and frontal and parietal cortex (Qin et al., 2005). We used the in vivo autoradiographic leucine method (Smith et al., 1988) to quantify protein synthesis rates in our studies. This methodology made it possible to measure steady state rates of ongoing protein synthesis in conscious functioning mice. We took care not to disturb the metabolic state of the animals. During the study we administered a tracer amount of a radiolabeled naturally occurring amino acid, and we sampled arterial blood to monitor the clearance of the tracer and the concentrations of amino acids in the arterial plasma while the animals were awake and freely moving. Since we did nothing to perturb the metabolic state by our procedures, measurements should reflect ongoing processes in the intact nervous system. Our measurements are average rates of leucine incorporation into all tissue proteins; they do not target specific proteins. The spatial resolution of the autoradiographic method is 50 μm (Smith, 1983). This permits us to distinguish specific nuclei and cell layers, including dendritic-rich and cell body layers. We cannot, however distinguish specific cell types within a layer.

After chronic dietary lithium treatment, the increased rCPS in Fmr1 KO mice was reversed. Because the effects on protein synthesis are so close to the genetic abnormality, we think that this may underlie the amelioration of lithium on behavioral deficits and immature dendritic spine morphology. We and others have previously discovered that the same lithium treatment strategy leads to multiple beneficial effects on behavioral phenotypes in Fmr1 KO mice such as hyperactivity, low generalized anxiety, impaired social interaction, impaired performance on the passive avoidance test (Liu et al., 2011; Mines et al., 2010; Yuskaitis et al., 2010) and increased incidence of audiogenic seizures (Min et al., 2009). We have also found that lithium corrected the abnormally increased spine length and density in medial frontal cortex of fragile X mice (Liu et al., 2011).

In the literature there are reports of some indirect measures that suggest a negative effect of lithium on mRNA translation in brain (Berg, 1968; De Bernardi et al., 1969; Wolcott, 1981; Wolcott, 1982; Kuznetsov, 1989). Studies employing microarray analysis of rodent brain showed that 20-50 transcripts were significantly down-regulated after chronic lithium treatment; none of the transcripts was up-regulated (Bosetti et al., 2002a; Chetcuti et al., 2008). The concentration of a key regulator of translation, eIF-2Bε, was decreased in rat brain extracts by 29% following chronic treatment with lithium (Bosetti et al., 2002b). In our study, negative effects of lithium treatment on rCPS were clear in Fmr1 KO mice.

The process of protein synthesis is primarily regulated by translation factors that transiently associate with ribosomes. Numerous components involved in the initiation and elongation stages of translation are regulated by mTOR, a protein kinase that phosphorylates 4E-BP (eIF4E-binding proteins) and the ribosomal protein p70 S6 kinase 1 (p70 S6K1). The PI3K/Akt signaling pathway is an important upstream regulator of mTOR. At the synapse, this pathway is coupled to membrane receptors. Results of recent studies indicate that PI3K/Akt signaling may be dysregulated in FXS. Basal PI3K activity in cortical synaptoneurosomes from developing Fmr1 KO mice was found to be excessive likely due to increased synthesis and synaptic localization of its catalytic subunit, p110β, both of which are normally regulated by FMRP (Gross et al., 2010). In hippocampal slices from Fmr1 KO mice, mTOR signaling was also found to be elevated possibly due to an increase in PI3K enhancer (PIKE) (Sharma et al., 2010), which is also normally regulated by FMRP. It has been suggested that these effects lead to the exaggerated protein synthesis in the Fmr1 KO mice.

In the present study, we looked at the phosphorylation status of some components of the PI3K/Akt pathway to try to understand the effects of lithium treatment. Herein we investigated the expression of total and phosphorylated forms of Akt, mTOR, and p70 S6K1 in homogenates of hippocampus. Relative levels of p-Akt (p=0.01) and p-mTOR (p-0.05) were elevated in untreated Fmr1 KO mice, but not in the downstream component, p-p70 S6K1. Our results in the untreated mice agree with other measurements of p-Akt and p-mTOR in lysates of freshly dissected hippocampus (Osterweil et al., 2010; Sharma et al., 2010), although the percent changes in our study were less than previously reported. Our result on p-p70 S6K1 contrasts with that of a previous report (Sharma et al., 2010) in which p-S6K was increased by 32% in Fmr1 KO mice, but in the previous study a 60 kDa band was analyzed whereas we analyzed a 70 kDa band (Shima et al., 1998; Ruvinsky and Meyuhas, 2006). It is also possible that quantitative differences between our results and those of others are due to differences in the ages of animals studied. In both of the other studies, mice were 28-42 days of age (Osterweil et al., 2010; Sharma et al., 2010), whereas in our study mice were 80-90 days of age. Another factor that may play a role is the effect of postmortem ischemia which may vary with preparation time and conditions.

Chronic lithium treatment resulted in substantial decreases in p-Akt in both genotypes, but effects of lithium on p-mTOR and p-p70 S6K1 were slight and were only seen in Fmr1 KO mice. The effects of lithium on GSK-3β activity may play a role in the diminished effects further down the PI3K/Akt pathway. In our previous study, we showed that whole brain extracts from adult Fmr1 KO mice had a decreased (-30%) level of the phosphorylated form of glycogen synthase kinase-3β (Ser9) (GSK-3β) indicating an increased activity compared to WT; this was reversed by lithium treatment (Liu et al., 2011). Decreases in both Akt and GSK-3β activities may have opposing effects on mTOR via their opposite effects on tuberosclerosis protein 2 (TSC2). GSK-3β activates TSC2 and consequently further suppresses mTOR, whereas Akt inhibits TSC2 thus decreasing TSC2 suppression of mTOR. Assuming that the effects of lithium on GSK-3β activity that we found in whole brain homogenates (Liu et al, 2011) also occur in the hippocampus, these opposing effects on TSC2 may explain the diminished effects of the lack of FMRP and of lithium treatment on p-mTOR and p-p70 S6K1.

We also considered the possibility that effects of lithium on translation may occur through the MAPK/ERK1/2 pathway. Our analysis of the relative level of the phosphorylated form of ERK1/2 in the hippocampus under basal conditions indicates no difference between WT and Fmr1 KO mice in agreement with others (Osterweil et al, 2010; Gross et al, 2010). Moreover, we found no effect of lithium treatment.

Collectively our results clearly indicate that lithium treatment ameliorates increased protein synthesis in Fmr1 KO mice but they do not support an effect of lithium treatment on protein synthesis by means of either the PI3K/Akt or the MAPK/ERK1/2 pathway. The mechanisms underlying the up-regulation of protein synthesis and the effects of lithium treatment in fragile X are likely more complicated than originally thought. Multiple pathways may be involved. These will be the focus of future studies.

These findings are of interest in light of a model proposed for the bimodal mechanism of action of lithium (Jope, 1999). Lithium, a drug used primarily in the treatment of bipolar disorder, has been shown to attenuate both hyperactivity (Jope, 1999) and hypoactivity (Lerer et al., 1980; Smith, 1988). Lithium elevates basal adenylyl cyclase activity, but it decreases receptor-coupled stimulation of adenylyl cyclase (Jope, 1999). In mouse brain slices, lithium acutely inhibits and chronically upregulates and stabilizes glutamate uptake (Dixon and Hokin, 1999). In animal studies lithium has been shown to activate Akt (Chalecka-Franaszek and Chuang, 1998) and to downregulate Akt (Erdal et al., 2005; Nemoto et al., 2008). In the bimodal model of lithium action, lithium reduces the low and high extremes, resulting in a stabilization of fluctuations in activity (Jope, 1999).

Our data add credence to the idea that dysregulated protein synthesis in the nervous system is central to FXS. We have confirmed our original finding that mice lacking FMRP have elevated in vivo rCPS (Qin et al., 2005). We further show that chronic lithium treatment of these mice reverses the excessive translation, but we do not find evidence that this effect occurs through PI3K/Akt or MAPK/ERK1/2 signaling. Moreover, chronic lithium treatment ameliorates the behavioral and morphological characteristics of the Fmr1 KO mouse (Liu et al., 2011). These results in the mouse model coupled with the results from other laboratories studying the mouse model (Min et al., 2009), the Drosophila model (McBride et al., 2005), and human subjects (Berry-Kravis et al., 2008) make a strong case for instituting a placebo-controlled trial of lithium in subjects with FXS.

Research Highlights.

• Chronic lithium reverses increased cerebral protein synthesis in Fmr1 KO mice

• Chronic lithium has little effect on cerebral protein synthesis in WT mice

• PI3K/Akt pathway signaling does not fully account for effects on hippocampal rCPS

• MAPK/ERK 1/2 pathway signaling does not fully account for effects on hippocampal rCPS

Abbreviations Footnote

FXS

fragile X syndrome

FMRP

fragile X mental retardation protein

rCPS

regional rate of cerebral protein synthesis

WT

wild type

KO

knock out

Akt

protein kinase B

mTOR

mammalian target of rapamycin

PI3K

phosphatidylinositol 3-kinase

ERK

extracellular-signal-regulated kinase

p70 S6K

p70 S6 kinase

PIKE

PI3K enhancer

4E-BP1

eIF4E-binding protein