Address correspondence and reprint requests to Mary C. McKenna, PhD, Department of Pediatrics, University of Maryland School of Medicine, 655 W. Baltimore St., Room 13-019, Baltimore, MD 21201, USA. E-mail: firstname.lastname@example.org
Fragile X syndrome (FXS) is the most common form of inherited mental retardation and is studied in the Fmr1 knockout (KO) mouse, which models both the anatomical and behavioral changes observed in FXS patients. In vitro studies have shown many alterations in synaptic plasticity and increased density of immature dendritic spines in the hippocampus, a region involved in learning and memory. In this study, magnetic resonance imaging (MRI) and 1H magnetic resonance spectroscopy (MRS) were used to determine in vivo longitudinal changes in volume and metabolites in the hippocampus during the critical period of early myelination and synaptogenesis at post-natal days (PND) 18, 21, and 30 in Fmr1 KO mice compared with wild-type (WT) controls. MRI demonstrated an increase in volume of the hippocampus in the Fmr1 KO mouse compared with controls. MRS revealed significant developmental changes in the ratios of hippocampal metabolites N-acetylaspartate (NAA), myo-inositol (Ins), and taurine to total creatine (tCr) in Fmr1 KO mice compared with WT controls. Ins was decreased at PND 30, and taurine was increased at all ages studied in Fmr1 KO mice compared with controls. An imbalance of brain metabolites in the hippocampus of Fmr1 KO mice during the critical developmental period of synaptogenesis and early myelination could have long-lasting effects that adversely affect brain development and contribute to ongoing alterations in brain function.
Fragile X syndrome (FXS) is the most commonly inherited form of X-linked mental retardation (Garber et al. 2008), and is caused by mutational insertion of CGG trinucleotide repeats that leads to hypermethylation and silencing of the FMR1 gene (Oostra et al. 1993; Garber et al. 2008). The rodent Fmr1 gene and the human FMR1 gene code for the fragile X mental retardation protein (FMRP), which acts as a translational repressor and binds specific mRNAs (Brown et al. 2001), including group 1 metabotropic glutamate receptor (mGluR) mRNA (Darnell et al. 2011). Silencing of the Fmr1 gene can result in structural, biochemical, and physiological changes in brain that impact development, and contribute to the pathology exhibited in the Fmr1 knockout (KO) mouse (Greenough et al. 2001; Bassell and Warren 2008; Dolen and Bear 2008). The mGluR5-dependent increase in long-term depression (LTD) involves dynamic translation and degradation of FMRP, which serves as a regulatory system controlling synaptic plasticity (Hou et al. 2006). The increased protein synthesis that occurs in the absence of FMRP can lead to enhanced LTD (Huber et al. 2002; Qin et al. 2005).
Fmr1 KO mice exhibit behavioral and cognitive alterations similar to patients with FXS such as hyperactivity, anxiety, and deficits in learning and memory (Bakker et al. 1994; Paradee et al. 1999; Spencer et al. 2005; Brennan et al. 2006; Smith et al. 2012). The hippocampus has a crucial role in learning and memory formation (Eichenbaum 1996; Scoville and Milner 2000), and this region shows increased volume in young FXS patients (Reiss et al. 1994; Kates et al. 1997). An increased density of morphologically immature dendrites has been observed in the dentate gyrus (Grossman et al. 2010), and the CA1 region of the hippocampus (Levenga et al. 2011) of Fmr1 KO mice compared with controls.
Most studies on the Fmr1 KO mouse have focused on alterations in dendritic spines, electrophysiological properties, signaling and RNA processing pathways, and behavior (Bardoni et al. 2006; Bassell and Warren 2008; Dolen and Bear 2008; Garber et al. 2008; Narayanan et al. 2008; Liu-Yesucevitz et al. 2011). Little emphasis has been placed on metabolic and biochemical changes. Metabolites offer a glimpse into functional processes of FMRP regulation (beyond the well-studied protein and mRNA interactions), which can offer additional insight into regulatory pathways of the brain (Davidovic et al. 2011). Many alterations in metabolism can contribute to impaired brain development and dysfunction (Yudkoff 1999; Moffett et al. 2007); however, few groups have studied metabolic or biochemical alterations in Fmr1 KO mouse brain (Gruss and Braun 2001; Qin et al. 2002; Davidovic et al. 2011). Both structural and biochemical alterations have been reported in brain regions of children with autism syndrome disorders (ASD) (Chugani 2012), which shares many features with FXS (Garber et al. 2008). Evidence of alterations in cholinergic pathways have been reported in patients with FXS (Kesler et al. 2009).
Alterations in metabolites in the hippocampus of the FVB strain of Fmr1 KO mice have been reported at post-natal days 12 (Davidovic et al. 2011) and 28–32 (Gruss and Braun 2001). However, no studies have determined in vivo longitudinal changes in the developing hippocampus of Fmr1 KO mice. Non-invasive methods such as in vivo magnetic resonance imaging (MRI) and 1H magnetic resonance spectroscopy (MRS) allow monitoring of brain structure and metabolic integrity longitudinally during the brain development and provide information on changes in neurotransmitters and neuromodulators such as glutamate, GABA, and taurine, as well as metabolites including N-acetylaspartate, myo-inositol, and choline-containing compounds.
We hypothesized that longitudinal assessment of the hippocampus in Fmr1 KO mice may reveal a characteristic pattern of metabolic alterations that contribute to abnormal development of this brain region. This study was performed during the critical period of myelination and synaptogenesis at PND 18, 21, and 30, in the hippocampus where developmental alterations in metabolites could be affected by a lack of FMRP. We measured volumetric and metabolic changes over time in the hippocampus of Fmr1 KO mice compared to wild-type C57Bl/6J (WT) mice using in vivo MRI and MRS, respectively. To our knowledge, this study is the first report on in vivo1H-MRS assessment of longitudinal metabolic changes in the developing Fmr1 KO mouse brain.
Methods and materials
Male Fmr1 KO mice (JAX B6.129P2-fmr1tmICgr mice; Jackson Laboratory, Bar Harbor, ME, USA) from a C57Bl/6J background were obtained from a breeding colony at the University of Maryland, Baltimore. All protocols were approved by the Institutional Animal Care and Use Committee at University of Maryland, Baltimore. Twelve Fmr1 KO mice and 12 wild-type (JAX C57Bl/6J) mice were used for each group; the mice used in this study came from three different litters, respectively. Food and water were given ad libitum. All mice were weaned at PND 21. Serial MRI and 1H-MRS data sets were acquired at PND 18, 21, and 30 for each mouse.
In vivo MRI/MRS
All experiments were performed on a Bruker Biospec 7.0 Tesla 30-cm horizontal bore scanner using Paravision 5.1 software (Bruker Biospin MRI, Ettlingen, Germany). A four-channel Bruker 1H surface array coil was used as the receiver and a Bruker 72-mm linear-volume coil as the transmitter. Mice were anesthetized using 2% isoflurane during the set up and maintained between 1 and 1.5% isoflurane at 1 L/min oxygen administration via a nose cone for the rest of the experiment. Fmr1 KO and WT control mice were carefully exposed to the same duration and level of isoflurane at each age to control for the potential effect of isoflurane exposure on developing brain metabolism (Kulak et al. 2010). The anesthetized mice were placed prone on a specialized holder with warm circulating water flowing through a pad that was placed over the body. Pain and discomfort were minimized by limiting scanning time to less than 2 h, maintaining body temperature at 36–37°C, and monitoring respiration throughout the experiment (SA Instruments, Stony Brook, NY, USA).
Proton density-weighted MR images covering the entire brain were obtained using a 2D rapid acquisition with relaxation enhancement (RARE) sequence (TR/TE = 5500/9.5 ms; slice thickness 0.5 mm; 16 slices) that served as anatomical reference for 1H MRS (10 min). Shimming over the hippocampus on average yielded a water line of 11 ± 2 Hz in both the WT and Fmr1 KO mice. A point-resolved spectroscopy (PRESS) pulse sequence (TR/TE = 2500/20 ms) with water suppression was used for data acquisition from a 2 × 4 × 2 mm3 voxel that covered the hippocampus area (Fig. 2). The 1H MRS voxel used was larger than the actual mouse hippocampus and included part of the cortex and the corpus callosum, but the majority of the scanned volume was over the hippocampus. A larger voxel was used to improve signal-to-noise ratio and to minimize MRS scanning time. For each spectrum, 300 acquisitions were averaged for a total of 13 min. Shimming was performed using the Fastmap technique (Gruetter 1993) over the voxel of interest. 1H MRS data were fitted using the LCModel software package (Provencher 2001). Only those metabolite ratios relative to total creatine (tCr) that passed the Cramer–Rao lower bound (CRLB) of 25% were used for further analysis. The CRLB values for tCr were all less than 10 percent.
Hippocampal volume was measured using Medical Image Processing, Analysis and Visualization tool (MIPAV v5.3.1, CIT; NIH, Bethesda, MD, USA) from the proton density-weighted images. The region of interest (ROI) was manually outlined in the proton density-weighted images for each image from around interaural 2.74 mm to interaural 0.28 mm [regions referenced from Franklin and Paxinos (2007)], then these ROI were grouped and volume was calculated.
A repeated-measures analysis of variance (anova; SPSS, Armonk, NY, USA) was used to determine differences between Fmr1 KO and WT mice across the three ages for the hippocampal volume and each metabolite ratio separately, with genotype being the between-subject factor and age being the within-subject factor. Significant differences in the genotype or the interaction of age by genotype were further examined using independent two-tailed Student t-tests on the volume or metabolite ratio at each time point to determine significance between Fmr1 KO and WT control mice. Metabolites that had a CRLB value higher than 25% were excluded from analysis. The decreased sample numbers for creatine and phosphocreatine are because of incomplete peak separation in some samples.
Fmr1 KO and WT control mice were studied at post-natal days (PND) 18, 21, and 30, as this time frame encompasses the critical period of synaptogenesis and rapid myelination in mice (McKenna et al. 1994; Gruss and Braun 2004; Massa et al. 2004) and while FMRP is still expressed above adult levels in the brain of WT mice (Lu et al. 2004; Davidovic et al. 2011).
The hippocampi were manually outlined in the proton density-weighted images for each mouse at each age and the volumes were calculated. Analysis of hippocampal volume revealed a significant effect of age, F(2,34) = 8.28, p < 0.005 and of genotype F(1,17) = 24.35, p < 0.001 (Fig. 1); however, no age by genotype interaction was observed. The hippocampal volume in the Fmr1 KO mice was larger than WT control mice at all ages: 16.6% average increase at PND 18, 19.3% average increase at PND 21, and 11.0% average increase at PND 30 (Fig. 1). Movement artifacts prevented analysis of three images from PND 21 and one from PND 30. The results agree with clinical reports of increased hippocampal volume in children with FXS (Reiss et al. 1994; Kates et al. 1997).
Neurochemical profile of Fmr1 KO mice compared to wild-type mice
High-quality 1H spectra of the hippocampal region were obtained at PND 18, 21, and 30 from both WT control and Fmr1 KO mice. Typical in vivo1H MR spectra from the hippocampus at each age are shown in Fig. 2. Data from 12 metabolites were compared between the hippocampus of WT and Fmr1 KO mice. The peak from N-acetylaspartate (NAA) methyl singlet at 2.01 ppm was assigned as the reference from which the rest of the metabolites were identified. The ratio of metabolites to total creatine (tCr; creatine plus phosphocreatine) in the hippocampus was determined longitudinally in the WT control and Fmr1 KO mice at each age (Table S1 and Fig. 3). There was no difference in total creatine content between WT controls and Fmr1 KO mice at any of the ages studied. Metabolite data from many in vivo studies of human and rodent brain are reported relative to the value for tCr (Mukonoweshuro et al. 2001; Xu et al. 2011).
Longitudinal changes in metabolites in the developing hippocampus
N-acetylaspartate (NAA) is a well-accepted marker of neuronal health (Moffett et al. 2007). Analysis of NAA/tCr revealed a significant effect of age, F(2,40) = 19.64, p < 0.001 as well as a significant effect of genotype, F(1,40) = 4.78, p < 0.05 (Fig. 3a). The interaction of age by genotype was not significant.
Analysis of taurine/tCr confirmed a significant effect of age, F(2,38) = 140.4, p < 0.001, genotype, F(1,19) = 43.5, p < 0.001, as well as a significant age by genotype interaction, F(2,38) = 4.83, p < 0.05 (Fig. 3b). The ratio of taurine/tCr was significantly increased in the hippocampus of Fmr1 KO mice compared with the WT controls (Fig. 3b) at PND 18 t(19) = 3.60, p < 0.005; PND 21 t(21) = 6.22, p < 0.005 and PND 30 t(21) = 2.16, p < 0.05.
Analysis of myo-inositol (Ins/tCr) confirmed a significant effect of age, F(2,38) = 13.87, p < 0.001 and a significant age by genotype interaction, F(2,38) = 7.06, p < 0.004 (Fig. 3c). The effect of genotype was not significant. A significant decrease was detected in the ratio of myo-inositol (Ins/tCr) in Fmr1 KO mice compared with the WT controls (Fig. 3c) at PND 30 t(21) = 3.06, (p < 0.01), and trend toward decrease was observed at PND 21 t(20) = 1.95, (p < 0.07).
Analysis of total choline (Cho/tCr) showed a significant effect of age, F(2,36) = 15.53, p < 0.0001 (Fig. 3d). There was no significant age by genotype interaction between WT and Fmr1 KO mice, and the main effect of genotype was not significant. Analysis of glutamate (Glu/tCr) showed a significant effect of age, F(2,38) = 5.28, p < 0.01. There were no significant effects of genotype or age by genotype interaction (Fig. 3e) in Fmr1 KO mice compared to WT controls. Analysis of glutamine (Gln/tCr) (Fig. 3f) and GABA/tCr (Fig. 3g) did not show within and between-group effects.
Analysis of creatine/tCr and phosphocreatine/tCr, the sum of which equals tCr, confirmed no significant interaction of age by genotype and no significant difference in genotype (Table S1). A significant effect of age was observed in both creatine, F(2,20) = 5.29, p < 0.05 and phosphocreatine, F(2,24) = 8.51, p < 0.01 in both Fmr1 KO mice and WT controls.
To our knowledge, this study is the first to characterize the longitudinal changes of metabolites in vivo in the developing hippocampus of the Fmr1 KO mouse compared to WT controls. We determined the metabolites on post-natal days 18, 21, and 30, during the critical developmental period of synaptogenesis and early myelination. Our study found increased hippocampal volume as well as alterations in NAA, taurine, and inositol in the hippocampus of Fmr1 KO mice. An imbalance of these metabolites that are involved in osmoregulation, signal transduction, and neuromodulation during this important period of synaptogenesis and early myelination could have long-lasting effects that profoundly alter brain development.
A recent study by Kulak et al. (2010) used high-field in vivo1H MRS to determine longitudinal changes in the neurochemical profile of developing anterior and posterior cortex in C57Bl/6J mice from PND 10–60 and provides useful reference data for the WT controls in this study. These authors (Kulak et al. 2010) reported significant changes over time in the concentration of many of the metabolites studied. In this study, values are reported relative to total creatine as is frequently done in 1H-MRS studies (Mukonoweshuro et al. 2001; Kumar et al. 2003; Xu et al. 2011). The developmental profile of the data for our WT mice can be compared to the data from Kulak et al. (Kulak et al. 2010) (when their data are expressed as metabolite/tCr). Kulak et al. (2010) reported a gradual increase in NAA and more striking increases in Ins and choline in anterior and posterior cortex from PND 20 to 30 in C57Bl/6J mice. The developmental patterns between PND 18 and 30 for metabolites in the hippocampus in the C57Bl/6J WT controls in this study parallel the developmental patterns in cortex found by Kulak et al. (2010) in that we observed a small increase in NAA and much larger increases in choline and inositol. The increase in choline reported by Kulak et al. (2010) between PND 20 and 30 was much greater in the lipid-rich cortex (~100%) than we observed in the hippocampal region (~31%) in this study. However, it should be noted that an ex vivo study by Yao et al. (1999) reported a decrease in phosphatidylcholine and total choline from PND 21 to adult in whole brain of C57Bl/6J mice. In addition, Tkac et al. (2003) reported a small decrease in choline in the hippocampus from PND 21 to 28 in Sprague–Dawley rat brain.
Kulak et al. (2010) observed a striking decrease in taurine concentration in the cortex from PND 20 to 30, and this study found a 26% decrease in taurine in hippocampus of WT mice from PND 18 to 30. The same group also found a small increase in glutamate and region-specific changes in GABA and glutamine between PND 20 and 30 (Kulak et al. 2010). They found no change in GABA in anterior cortex and a slight decrease in posterior cortex, and an increase in glutamine in anterior cortex, but no change in posterior cortex (Kulak et al. 2010). In this study, we found a small age-related increase in glutamate (~13%) in hippocampus in the WT mice from PND 18 to 30, and no change in either GABA or glutamine. Overall, the longitudinal developmental changes in metabolites in the hippocampus of WT mice in this study are remarkably similar to the longitudinal changes from day 20–30 reported by Kulak et al. (2010) for cortex of C57Bl/6J mice.
There is relatively little information about the metabolite changes in the hippocampus of Fmr1 KO mice, and the two studies in the literature report ex vivo values for hippocampal metabolites in male FVB strain of Fmr1 KO mice. An HPLC study by Gruss and Braun (2001) reported values for hippocampal metabolites in PND 28–32 and adult brain, and an ex vivo study by Davidovic et al. (2011) that reported metabolite values from PND 12 mouse brain. The recent study by Davidovic et al. (2011) using 1H high-resolution magic angle spinning nuclear magnetic resonance spectroscopy reported alterations in metabolites in the cortex, striatum, hippocampus, and cerebellum in the FVB Fmr1 KO mice at PND 12, an age at which FMRP concentration is high in the brains of WT mice (Davidovic et al. 2011). Metabolite values showed regional variation in the WT mouse brain, reflecting maturational differences of region-specific function at this early developmental time point (Davidovic et al. 2011). The FVB Fmr1 KO mice, however, displayed a poorly differentiated, more homogenous metabolic state across the same regions (Davidovic et al. 2011). The authors suggested that this less diverse metabolic profile in the brain regions studied in FVB Fmr1 KO mice was because of delayed brain maturation and delayed functional anatomical differentiation (Davidovic et al. 2011).
In this study, the in vivo longitudinal developmental pattern of several important metabolites including the neurotransmitters glutamate and GABA, and also Gln and choline were not different in the hippocampus of Fmr1 KO mice compared with controls. In contrast, several key metabolites had different developmental profiles in Fmr1 KO mice compared with controls. Ins/tCr was not different on PND 18, but was significantly decreased in Fmr1 KO mice at PND 30 compared with controls. Taurine/tCr was significantly higher in the hippocampus of Fmr1 KO mice than controls at all ages studied, and decreased 32% from PND 18 to 30 in Fmr1 KO mice, compared with a 25% decrease in WT mice. The longitudinal data in the present study are consistent with the findings of an HPLC study of brain amino acids (Gruss and Braun 2001) that found increased taurine in the hippocampus of 28–32-day-old FVB Fmr1 KO mice, and no differences in Glu, Gln, and GABA between Fmr1 KO mice and WT controls. Interestingly, the same study (Gruss and Braun 2001) did not find any differences in the metabolite levels in the hippocampus or in the other brain regions studied between the adult FVB Fmr1 KO mice and the adult WT mice (Gruss and Braun 2001). The differences observed in metabolite levels in vivo in this study, as well as data from the HPLC study of Gruss and Braun (Gruss and Braun 2001) suggest that some of the alterations in hippocampal metabolites that occur in the young Fmr1 KO mice from PND 18 to 30 can be transient. Indeed, the genotype difference in NAA/tCr was most apparent at PND 18 as the ratios were similar to controls at the older ages. The increased ratio of taurine/tCr at PND 18 and 21 in Fmr1 KO mice were closer to, but still significantly higher than the values for WT controls at PND 30. However, this was not the case for all metabolites as the decrease of Ins/tCr ratio in Fmr1 KO mice was more pronounced at 30 days compared with controls, than at PND 21 where it only approached significance.
Change in N-acetylaspartate (NAA)
NAA increases significantly in brain from birth to adulthood (Birken and Oldendorf 1989), and is a widely used clinical marker for neuronal health and/or neuronal mitochondrial integrity (Moffett et al. 2007). NAA is synthesized in neurons, released and taken up by oligodendrocytes where it is cleaved by the glial enzyme aspartoacylase to aspartate and acetyl CoA, which serves as an important precursor for the synthesis of myelin lipids in the developing brain (Burri et al. 1991; Urenjak et al. 1992; Moffett et al. 2007). Decreased formation of NAA in neurons has been shown to impair myelination in developing brain (Degaonkar et al. 2002). Decreased NAA/tCr has been observed in hippocampus of children with ASD (Kleinhans et al. 2007; Gabis et al. 2008; Chugani 2012) and in other related disorders characterized by psychomotor delay (Gabis et al. 2008; Chugani 2012). Such observations are particularly important as Kleinhans (Kleinhans et al. 2007) demonstrated that the NAA concentration was correlated to brain activation determined by functional MRI (fMRI), and lower NAA was related to under connectivity in prefrontal cortex. The effect of genotype on NAA/tCr in Fmr1 KO mice compared to WT suggests impairment in neuronal development, and/or in neuronal mitochondrial function in the hippocampus in this model of FXS.
NAA acts as an agonist to N-methyl-d-aspartic acid (NMDA) receptors in neurons (Rubin et al. 1995); therefore, the decrease in NAA at PND 18 in the Fmr1 KO mice could lead to decreased activation of NMDA receptors in the hippocampus. This hypothesis is consistent with reports of decreased long-term potentiation and decreased amplitude of excitatory post-synaptic current on activation of NMDA receptors in the hippocampus of Fmr1 KO mice (Gibson et al. 2008; Yun and Trommer 2011). The increased mGluR-mediated LTD in the hippocampus of Fmr1 KO mice compared with WT mice (Huber et al. 2002) can result in weakened synaptic connections, and is a putative mechanism leading to altered memory and cognition characteristic of the syndrome (Pfeiffer and Huber 2009). The relationship between NAA levels and hippocampal connectivity has not been studied in Fmr1 KO mice.
Alterations in osmoregulatory molecules in the hippocampus of Fmr1 KO mice
NAA, taurine, and myo-inositol (Ins) can all function as osmolytes in brain cells (Heilig et al. 1989; Estevez et al. 1999; Baslow 2010). Baslow (Baslow 2010) has proposed that release of NAA and NAAG from neurons has a key role in regulating neuronal water and in overall brain function. Extracellular taurine has been shown to increase in response to hypoosmotic swelling (Davies et al. 1998). Decreased Ins/tCr has been reported in all regions in ASD patients (Gabis et al. 2008), and in temporoparietal junction in high-functioning ASD patients (Bernardi et al. 2011) compared with age-matched controls.
The significant effect of age in this study shows that the ratios of taurine/tCr and Ins/tCr are also changing during development in both WT control and Fmr1 KO mice. However, the increase of taurine and decrease of Ins could cause alterations in brain osmoregulation. An increase in both taurine and Ins was found in the hippocampus of the FVB Fmr1 KO mice at PND 12 and such an increase could affect osmoregulation and phosphatidylinositol (PI) metabolism (Davidovic et al. 2011). The increase of Ins at PND 12 in the hippocampus reported by Davidovic et al. (2011) could be an age-specific change or may reflect differences between the FVB model and the C57Bl/6J strain of Fmr1 KO mice. However, as noted below, alterations of NAA, taurine, and Ins can have other effects beyond osmoregulation.
The enzyme phosphoinositide 3-kinase (PI3K) phosphorylates PI, and PI3K signaling plays an important role in differentiation and cell growth (Sanchez et al. 2004; Peltier et al. 2007). PI signaling involves continuous breakdown to Ins by phosphatases and lipid resynthesis of PI from Ins (Shears 1989). The decreased Ins/tCr in hippocampus of Fmr1 KO mice at PND 30 could lead to alterations in PI level and potential alterations in the PI3K signaling pathways during brain development. Increased PI3K activity in the hippocampus of Fmr1 KO mice was shown in two different studies (Gross et al. 2010; Sharma et al. 2010) and a PI3K antagonist was able to rescue neurons from the Fmr1 KO phenotype in vitro (Gross et al. 2010). PI increases steadily from gestational day 15 to PND 21 in WT mice (Yao et al. 1999). In this study, we found Ins/tCr increased from PND 21 to 30 consistent with the developmental profile reported by Kulak et al. (2010) for C57Bl/6J mice. However, at PND 30 values for Ins/tCr in hippocampus of Fmr1 KO mice in this study were only 65% of the WT mice. Phosphatidylinositol 4,5-bisphosphate has a crucial role in the attachment of myelin basic protein to oligodendrocyte membranes (Nawaz et al. 2009). Thus, the decrease of Ins/tCr at PND 30 in Fmr1 KO mice compared with WT controls could potentially contribute to alterations in both myelination and signal transduction in the hippocampus. The importance of inositol in brain is underscored by the report that the Ins/tCr level was directly associated with performance IQ scores in ASD patients (Gabis et al. 2008).
The non-protein inhibitory amino acid taurine is present in high concentration in immature brain and decreases with age (Lima et al. 2004; Kulak et al. 2010). Taurine is a key osmolyte and also protects against the effects of glutamate excitotoxicity in vitro by stabilizing calcium concentration in the cytoplasm to basal levels (El Idrissi and Trenkner 1999). Thus, an increased level of taurine could potentially protect mitochondria from the potentially harmful increase in intracellular calcium (Nicholls 2009) resulting from mGluR5 activation. In this study, the increased taurine/tCr in Fmr1 KO mice at all ages may reflect delayed or altered maturation of the hippocampus compared with WT mice.
El Idrissi et al. (2010) showed that the taurine-dependent decrease in paired-pulse inhibition observed in hippocampal slices from WT mice was not seen in hippocampal slices from Fmr1 KO mice. The same group showed that supplementation with taurine improved performance in inhibitory avoidance in Fmr1 KO mice compared with the non-treated group (El Idrissi et al. 2009). During the developmental window in this study, taurine can have an inhibitory function in the brain (Curtis and Watkins 1960) as the high concentration of taurine found in the developing brain can activate glycine and GABAA receptors of pyramidal cells in the immature rat hippocampus (Wu and Xu 2003). The increased levels of taurine at PND 18, 21, and 30 in Fmr1 KO compared with WT mice could contribute to neuroprotection by modulating the intracellular calcium levels, and by contributing to the inhibitory activity in hippocampal neurons.
Neurotransmitters and related metabolites
Choline and choline-containing lipids are important constituents of plasma membranes, which are increased with cell density; thus the total choline peak is used as a marker of proliferation (Usenius et al. 1994; Warren et al. 2000). The total choline peak at 3.21 ppm contains both the excitatory neurotransmitter acetylcholine (Fan 1996), which is involved in learning and plasticity (Broide and Leslie 1999; Shinoe et al. 2005; Hasselmo 2006), and phosphatidylcholine, a necessary component of myelin which increases during brain development (Diamond 1971; Loffelholz 1998; Martinez and Mougan 1998). In vivo MRS in FXS patients showed a decrease in relative choline concentration in the dorsolateral prefrontal cortex relative to control subjects (Kesler et al. 2009). Altered activity of muscarinic acetylcholine receptors in the Fmr1 KO mice suggested a decrease in the cholinergic driven plasticity in the subiculum (D'Antuono et al. 2003) and in the CA1 region of the hippocampus (Volk et al. 2007). Our study found no differences in the developmental profile of Cho/tCr in the hippocampus of Fmr1 KO mice compared to WT controls. However, acetylcholine released by cholinergic terminals enervating the hippocampus is likely to be synthesized within the lateral septum of brain where these projections originate (McKinney et al. 1983; Parent and Baxter 2004).
Glutamate (Glu) is the major excitatory neurotransmitter in the brain, synthesized from α-ketoglutarate that is derived from glucose metabolism via the TCA cycle, or by deamination of the metabolite glutamine (Gln), which is synthesized in astrocytes (Schousboe et al. 1997; Danbolt 2001; Sonnewald and Kondziella 2003; McKenna 2007). Consistent with the findings of Gruss and Braun (Gruss and Braun 2001), we did not find any alteration in the ratios of Glu, Gln, or GABA to tCr levels in the hippocampus between Fmr1 KO and WT mice during the critical development period of myelination and synaptogenesis. It is important to note that Glu and GABA are both highly compartmentalized in brain (Waagepetersen et al. 2003), thus the overall ratio in hippocampus may mask alterations in specific hippocampal regions or changes in the distribution of these neurotransmitters. Indeed, reports from other groups demonstrate that the GABAergic system is affected by the absence of FMRP in Fmr1 KO mice without an overall alteration of the GABA neurotransmitter concentration (El Idrissi et al. 2005; Gibson et al. 2008). There is clear evidence of an imbalance in excitatory versus inhibitory neurotransmission in Fmr1 KO brain (Gibson et al. 2008). Decreased excitation of inhibitory neurons in the neocortex of FVB Fmr1 KO mice at PND 14, 21, and 28 has been reported (Gibson et al. 2008), as well as alterations in GABAA receptor activation (El Idrissi et al. 2005), receptor subunit expression in the cortex (D'Hulst et al. 2006), and decreased expression of glutamate decarboxylase (D'Hulst et al. 2009).
Hippocampal volume differences
The most striking difference observed was the significantly increased hippocampal volume in the brains of the Fmr1 KO mice at all ages compared with controls. Increased hippocampal volume has been reported in children with autism and boys with fragile X syndrome (Reiss et al. 1994; Kates et al. 1997; Geuze et al. 2005). The hippocampus is rich in FMRP in normal brain, and it has been suggested that the altered growth of the hippocampus may be related to lack of this protein in fragile X syndrome (Bray et al. 2011). Increased density of dendritic spines and delayed morphological development of spines were found in the dentate gyrus (Grossman et al. 2010) and the CA1 region of the hippocampus (Levenga et al. 2011) of Fmr1 KO mice compared with C57Bl/6J controls. Although differences in FMRP have not been reported for hippocampal subregions, the increased activation of mGluR5 is thought to contribute to the alterations in dendritic spines in the hippocampus of Fmr1 KO mice (Grossman et al. 2010; Levenga et al. 2011). Reports that LTD is altered in the CA1 (Huber et al. 2002) and dentate gyrus (Yun and Trommer 2011), but spines are not altered in the CA3 region (Levenga et al. 2011), suggest that alteration in the mGluR5 signaling pathway may be restricted to the CA1 region and may contribute to the regional differences in altered spine phenotype in the hippocampal of Fmr1 KO mice (Grossman et al. 2010; Levenga et al. 2011). The increase of hippocampal volume we observed in vivo at PND 18, 21, and 30 in Fmr1 KO mice (Fig. 1) may reflect an increased number of immature dendritic spines (Greenough et al. 2001; Grossman et al. 2010), or may be related to alterations in the osmolyte profile, or altered connectivity in this region.
During the critical developmental period of synaptogenesis and early myelination included in this study, a number of alterations were observed, particularly in metabolites involved in osmoregulation, signal transduction, and neuromodulation in the hippocampus of Fmr1 KO mice when compared with age-matched controls. Although some alterations may be normalized in adulthood (Gruss and Braun 2001), such profound metabolic perturbations during synap-togenesis and modulation of neuronal connections could lead to lifelong alterations in the hippocampus and can affect both neuronal and astrocytic functions. The alterations observed in these metabolites may contribute to the FXS phenotype observed in the Fmr1 KO mouse.
This study was supported by the Core for Translational Imaging @ Maryland (C-TRIM), the FRAXA Research Foundation, and NIH P01 grant HD016596. The authors do not have any conflicts of interests.