Altered differentiation of neural stem cells in fragile X syndrome

Abstract

Fragile X syndrome, a common form of inherited mental retardation, is caused by the absence of the fragile X mental retardation protein (FMRP) due to a mutation in the FMR1 gene. We investigated the differentiation of neural stem cells generated from the brains of fmr1-knockout (KO) mice and from postmortem tissue of a fragile X fetus. Mouse and human FMRP-deficient neurospheres generated more TuJ1-positive cells (3-fold and 5-fold, respectively) than the control neurospheres generated from normal mouse and human brains, and these cells showed morphological alterations with fewer and shorter neurites and a smaller cell body volume. The number of cells expressing glial fibrillary acidic protein and generated by these neurospheres was reduced because of increased apoptotic cell death. Furthermore, there was an increase in a population of cells with intense oscillatory Ca²⁺ responses to neurotransmitters in differentiated cells lacking FMRP. In addition, the number of cells in a cohort of bromodeoxyuridine-labeled newborn cells was increased in the subventricular zone of the telencephalon of the fmr1-KO mouse in vivo. These results demonstrate substantial alterations in the early maturation of FMRP-deficient neural stem cells in fragile X syndrome and in the fmr1-KO mice.

Fragile X syndrome is a common form of inherited mental retardation with an incidence of one in every 4,000 males (1). Many patients exhibit attention deficits, hyperactivity, autistic-like behavior, unusual responses to sensory stimuli, and epileptic seizures (for a review, see ref. 2). The syndrome is caused by the absence or dysfunction of the fragile X mental retardation protein (FMRP), most often due to a mutation in the FMR1 gene leading to transcriptional silencing of the gene (2). FMRP is an RNA-binding protein that associates with polyribosomes and acts as a translational repressor of specific mRNAs at synaptic sites, and in that way it can regulate synapse growth and function (3-8). FMRP is highly expressed in the human central nervous system (9, 10). The absence of FMRP results in abnormalities of dendritic spines in fragile X patients (11-14). The phenotype of the fmr1-knockout (KO) mice exhibits similarities with human fragile X syndrome (15, 16). Abnormalities in synaptogenesis, synaptic structures and functions have been observed in the KO mice in vivo (17-21). One particularly interesting finding has been the augmentation of metabotropic glutamate receptor (mGluR)-dependent long-term depression in the fmr1-KO mice (22).

Neural stem cells (NSCs) are multipotent, self-renewing cells that can be propagated in culture. Human NSCs provide a source of cells for in vitro studies attempting to elucidate neural mechanisms in the pathogenesis of human neurological disorders (23, 24). In the present study, we exploited mouse and human NSCs to study the pathogenesis of fragile X syndrome and the cellular mechanisms leading to cognitive impairment and epilepsy. We compared the differentiation of FMRP-deficient mouse NSCs to the differentiation of WT cells both in vitro and in vivo and demonstrated that histology and Ca²⁺ signaling are altered in NSCs generated from FMRP-deficient mice and human fragile X NSCs.

Materials and Methods

Brain Tissue. Mouse NSCs were generated from the brains of FVB fmr1-KO mice at embryonic day 13 (E13) and postnatal day 6 (P6) and their WT littermates (15). Postmortem human fetal tissue was obtained in accordance with the guidelines of National Institutes of Health, the government of Finland, and the local ethics committee of the Kuopio University Hospital. All investigations were conducted according to the principles of the Declaration of Helsinki. Full informed consents were obtained. A mutation in the FMR1 gene was detected by PCR and Southern analysis. Primary cultures of human brain cells were initiated from an 18-week-old fragile X fetus with a methylated repeat expansion of 276-300 trinucleotide repeats in the 5′ untranslated region of the FMR1 gene and from control fetuses at 7, 12, and 18 weeks of gestation as determined by intrauterine ultrasound examinations.

Cell Cultures. NSCs were generated from the wall of the lateral ventricles as described by Clarke et al. (25) and Supporting Materials and Methods, which is published as supporting information on the PNAS web site.

Differentiation of Cells. Neurospheres and papain-treated single cells (10⁵ cells per ml) were plated on cover glasses coated with 10 μg/ml poly-d-lysine (Sigma) and 5 μg/ml laminin (GIBCO/BRL) in culture medium without mitogens and differentiated for the indicated periods of time (see Figs. 1, 4, and 5).

Fig. 1.

Differentiation of NSCs generated from the brains of fmr1-KO mice. (A) FMRP expression in the periventricular region in coronal brain sections of WT and fmr1-KO mouse. (B) The average number of TuJ1-positive cells generated from neurospheres derived from E13 and P6 fmr1-KO mice was significantly higher than the number of TuJ1-positive cells generated from the WT neurospheres. (C) Representative images of the TuJ1-positive cells generated from NSCs derived from WT and fmr1-KO mice. (D) The percentage of neurons with more than two primary neurites, the length of the longest neurite, and the mean cell-body volume of WT and FMRP-deficient TuJ1-positive cells after differentiation for 1 day (1d) and 5 days (5d). (E) The average number of GFAP-positive cells generated from WT neurospheres was significantly higher than the number of GFAP-positive cells generated from FMRP-deficient neurospheres. (F and G) The number of TUNEL-positive and TuJ1-negative cells (F) and [³H]Thymidine incorporation (G) obtained from four separate experiments in WT and FMRP-deficient neurosphere cultures. (H) The mean size of the neurospheres derived from fmr1-KO brains (n = 31) did not differ significantly from controls 3 days after cell dissociation. TuJ1-positive and GFAP-positive cells were counted from 6-10 cover glasses at different developmental stages, and the cell numbers are shown per 10⁴ cells. (Scale bars, 100 μm.) Data are represented as mean ± SEM. Asterisks indicate statistically significant differences (P < 0.05) versus respective controls.

Fig. 4.

Differentiation of normal and FMRP-deficient human NSCs. (A) NSCs were propagated as neurospheres in the presence of mitogens. After withdrawal of mitogens, cells started to differentiate into neuronal cells and acquired the morphological properties of neurons and glia. (B) FMRP expression (red) was detectable in undifferentiated clustered human NSCs derived from control fetuses but not in proliferating NSCs derived from fetal brain with fragile X syndrome during the first 3 weeks in culture. (C) FMRP expression (red, arrows) in normal neurosphere-derived cells 7 days after differentiation. FMRP staining was mainly in GFAP-negative cells (blue, arrowhead). (D) The differentiation of fragile X neurospheres into neuronal and glial cells differed from the differentiation of control neurospheres. (Insets) The difference can be appreciated at the edge of the neurospheres during the first day of differentiation. Photographs represent TuJ1-positive neurons (red) and GFAP-positive astrocytes (green) generated from neurospheres. (Scale bars, 100 μm.)

Fig. 5.

The number of TuJ1-positive and GFAP-positive cells generated from human fragile X neurospheres differed from the controls. (A) Photographs represent the TuJ1-positive cells produced from human control and fragile X neurospheres. (B) The human fragile X neurospheres gave rise to significantly more TuJ1-positive cells and less GFAP-positive cells than the age-matched control neurospheres after differentiation for 2 weeks. Cells were counted from 10 single neurospheres in control and fragile X cover glasses. (C) The average neurite length and the mean cell-body volume of differentiated TuJ1-positive cells derived from the human control and the fragile X neurospheres are shown. Data were collected from two different staining experiments; 246 fragile X neurites and 69 control neurites were measured. The cell-body volume was measured from 69 human fragile X neural cells and from 36 control cells. (D) The average number of oscillating cells in eight cover glasses of fragile X cells (236 cells measured individually) and 14 cover glasses of control cells (450 cells measured individually) and shown as percentages of the total number of measured cells (±SD). (E) Fura-2 recordings from representative single cells in response to ACh (100 μM) from normal control and fragile X neurosphere-derived cells. Bars indicate the application of ACh and time. Data are represented as means ± SEM. Asterisks indicate statistically significant differences (P < 0.05) versus respective controls.

Cell Proliferation Analyses. The incorporation of 0.5 μCi/ml (1 Ci = 37 GBq) [³H]thymidine (Amersham Biosciences, which is now GE Healthcare) was investigated in cell cultures seeded at a density of 10⁵ cells per ml in DMEM/F12 culture medium supplemented with mitogens. After incubation for 24 h, cells were precipitated with 50 μl/ml trichloroacetic acid for 20 min at 4°C, solubilized in 80 μl of 1 M NaOH for 20 min at room temperature, and neutralized with 0.25 ml of 1 M HCl. The radioactivity of the lysates was counted with scintillation fluid (OptiPhase HiSafe 3) in a liquid scintillation counter (1450 MicroBeta, Wallac). See Supporting Materials and Methods for detailed descriptions of BrdUrd administration in vivo and stereological assessment of newborn cells.

Immunohistochemistry. Cultured cells on the cover glasses were fixed with paraformaldehyde in 4% PFA in PBS, pH 7.4, for 10 min and permeabilized with ice-cold methanol for 20 min. After blocking with PBS containing 20% normal goat serum for 20 min, the cells were incubated with antibodies. Primary antibodies were monoclonal anti-FMRP (clone 1C3-1a, 1:1,000, Euromedex, Mundolsheim, France), mouse anti-β-III-tubulin (clone TuJ1, 1:500, Babco, Richmond, CA), rabbit anti-glial fibrillary acidic protein (GFAP, 1:250, Sigma), and anti-BrdUrd (clone BU-1, Amersham Biosciences). Secondary antibodies (Jackson ImmunoResearch) were Cy3-conjugated anti-mouse IgG (1:500), 7-amino-4-methylcoumarin-3-acetic acid-conjugated anti-rabbit IgG (1:250), and Alexa Fluor 488-neutravidin-conjugated anti-mouse IgG (1:1,000, Molecular Probes). DAPI and hematoxylin staining were used to label nuclei. Cover glasses were rinsed and mounted with SlowFade antifading reagent (Molecular Probes).

Analysis of Apoptosis, Quantitative Analysis, and Ca²⁺ Imaging. Apoptotic cells were detected by using the DeadEnd Colorimetric TUNEL System (Promega) according to the instructions of the manufacturer. See Supporting Materials and Methods for a detailed description of the cell counting procedures and the morphological analysis of neurons. Statistical comparisons were performed by Student's t test, ANOVA, and the Mann-Whitney test. Ca²⁺ recordings were performed with fura-2 in neurosphere-derived cells after differentiation overnight as described in detail in Supporting Materials and Methods.

Results

FMRP Expression in NSCs. FMRP is expressed at its highest expression levels during early murine development (2, 9). In the fetal brain, FMRP expression was particularly high in the walls of the lateral ventricles (Fig. 1A) (9). Interestingly, the adult mouse brain samples also exhibited appreciable levels of FMR1 mRNA in the periventricular region, where the growth factor-responsive stem cells are known to reside throughout the life span (Fig. 6, which is published as supporting information on the PNAS web site).

Neuronal Differentiation of Mouse NSCs Lacking FMRP. We compared the differentiation of NSCs generated from the fmr1-KO mice to the differentiation of WT mouse NSCs. We found that after differentiation for 5 days the neurospheres derived from embryonic brains (embryonic day 13) of fmr1-KO mice produced 3-fold (n = 5, P < 0.05) more cells positive for neuronal marker TuJ1 than the WT neurospheres did (Fig. 1B). As shown in Fig. 1B, a significant increase was also seen in the number of TuJ1-positive cells generated from neurospheres derived from P6 KO mice when compared with the respective controls (4.5-fold, P < 0.05).

The morphological analysis of the differentiated TuJ1-positive cells revealed fewer primary neurites in FMRP-deficient cells than in WT cells (Fig. 1C). The number of cells with more than two primary neurites was significantly smaller in dissociated cell cultures derived from NSCs generated from fmr1-KO embryos (18% from the total cell number) than from WT embryos (38% from the total cell number) after differentiation for 1 day (P = 0.028) (Fig. 1D). Furthermore, the length of the longest neurite was significantly shorter in the FMRP-deficient cells than in the WT cells after differentiation for 1 day (64% of control, P = 0.038) or 5 days (51% of control, P = 0.046) (Fig. 1D and Fig. 7, which is published as supporting information on the PNAS web site). In addition, the mean cell-body volume was significantly smaller in FMRP-deficient cells than in controls after differentiation for 1 or 5 days (83% and 85% of control and P = 0.040 and 0.038, respectively) (Fig. 1D). The data suggest that the absence of FMRP altered the neuronal lineage differentiation of NSCs.

The number of GFAP-expressing cells was significantly lower in the differentiated cells derived from E13 or P6 fmr1-KO mice compared with the respective WT cells (86% and 87% of control, respectively; P < 0.05 in both) (Fig. 1E). The number of TUNEL-positive cells was increased in differentiating TuJ1-negative FMRP-deficient cells (Fig. 1F), and the quantitative data indicated that the cell death accounted for the decrease in the number of GFAP-positive cells in cell cultures derived from the KO mice. There were no significant differences in [³H]thymidine incorporation or in the mean size of the neurospheres between the neurosphere cultures derived from WT and fmr1-KO (Fig. 1 G and H).

Ca²⁺ Responses of Differentiating NSCs Generated from fmr1-KO Mice. An intracellular Ca²⁺ ([Ca²⁺]_i) response to acetylcholine (ACh) during the first day of NSC differentiation was shown in ref. 26. The majority (≈90%) of the cells measured in this study responded to ACh (Fig. 2A). The cells were also challenged with glutamate in the presence and the absence of Ca²⁺ to assess the presence of mGluR and ionotropic glutamate receptors and with elevated K⁺ (70 mM) to demonstrate the presence of voltage-gated Ca²⁺ channels in these cells. Two clearly defined cell populations could be distinguished based on their functional response. Approximately 30% (29 ± 19% in control and 36 ± 18% in FMRP-deficient cells; type I cells) of the cells responded to glutamate in the absence of Ca²⁺ with a fast, spiky response (metabotropic response) but showed a relatively small response to K⁺ (Fig. 2 A). A smaller proportion of the cells (22 ± 10% in control and 22 ± 18% in FMRP-deficient cells; type II cells) responded to elevated K⁺ with a robust Ca²⁺ elevation (Ca²⁺ elevation at >300 nM) (Fig. 2B). None of these cells showed a metabotropic Ca²⁺ response to glutamate (Fig. 2B). A third type of response could be seen in the remaining cells, which had a minor or even absent metabotropic response to glutamate and K⁺ (type III cells) (data not shown).

Fig. 2.

Effects of different concentrations of ACh, glutamate, and elevated K⁺ on [Ca²⁺]_i in differentiating neurosphere-derived mouse cells. Recordings were made from the edge of neurospheres, where a single layer of cells could be identified. (A and B) Fura-2 recordings of responses to different concentrations of ACh (100, 10, and 1 μM), 100 μM glutamate in the absence (-Ca²⁺) and presence of extracellular Ca²⁺, and elevated K⁺ from a cell with a metabotropic Ca²⁺ response to glutamate (A) and from a cell with a Ca²⁺ response to high K⁺ (B). Bars indicate the application and time. (C) The average number of oscillating cells (±SD) in cell populations derived from single neurospheres generated from WT (n = 7; 809 cells) and fmr1-KO mice (n = 6; 802 cells) at different concentrations (μM) of ACh. The asterisk indicates a statistically significant difference (P < 0.05) versus respective control. (D) The correlation between the number of cells responding to glutamate in the absence of Ca²⁺ and the number of oscillatory cells shown as measurements in the different cover glasses. (E) Effect of the selective mGluR antagonist on responses to glutamate. Cells were sequentially challenged with 1 μM glutamate in the absence of extracellular Ca²⁺ together with 10 μM MPEP or LY367385 (LY) where indicated. (F) The average data (±SD) from similar experiments (30, 24, and 40 cells) are shown as bar diagrams. The data include only cells for which a response to glutamate was obtained after washout of the antagonists.

One interesting feature was the frequent appearance of oscillations with a frequency of 4-6 spikes per min in response to ACh (and glutamate) in the type I cells. The oscillatory response was significantly accentuated in the FMRP-deficient mouse cells treated with 100 μM ACh when compared with WT cells (3.2-fold, P = 0.01) (Fig. 2C). There was a clear correlation between the number of cells displaying the metabotropic response to glutamate and the number of cells with an oscillatory response to 100 μM ACh (Fig. 2D). To investigate which mGluR subtype was responsible for the Ca²⁺ elevation in type I cells, they were challenged with different concentrations of glutamate (1, 10, and 30 μM) in the presence and absence of antagonists for the mGluR1 and mGluR5 subtypes. Addition of 1 μM glutamate resulted in a typical Ca²⁺ peak response. Subsequent washing and renewed challenge with glutamate in the presence of 10 μM 2-methyl-6-(phenylethynyl)pyridine hydrochloride (MPEP) resulted in no response. Challenge with glutamate in the presence of LY367385 showed a partially attenuated response compared with the control. A further challenge with glutamate after the wash was of a magnitude similar to the initial response (Fig. 2E). In the presence of 30 μM glutamate, no inhibitory responses were seen (Fig. 2F). An average of measurements from several cells (± SD) assayed in a similar manner demonstrated the higher sensitivity of the glutamate response to MPEP compared with that to LY367385 (Fig. 2F).

Production of New Cells in the Absence of FMRP in Vivo. To investigate the correlation of the in vitro findings to the in vivo situation, we compared the production of new cells in the embryonic brains of WT and fmr1-KO mice after BrdUrd injections in vivo. The total number of BrdUrd-positive cells differentiated for 4 days after multiple injections of BrdUrd was not significantly different in the radial cross sections of developing cortex in fmr1-KO brains when compared with WT brains (Fig. 3A). However, a substantial number of BrdUrd-positive cells was observed in the subventricular zone (SVZ) of the telencephalon of KO mice, whereas significantly fewer BrdUrd-labeled cells could be found in this region in the brains of WT mice (4.5-fold, P = 0.011) (Fig. 3 B and C).

Fig. 3.

Production of new cells in the brains of WT and fmr1-KO mice in vivo.(A) The cohort of newborn cells differentiated for 4 days was detected by BrdUrd antibody in coronal brain sections of WT and fmr1-KO mice. CP, cortical plate; IZ, intermediate zone; LV, lateral ventricle; NE, neuroepithelium. (B) Schematic representation of the developing (embryonic day 17) mouse cortex. (C) The increase in the number of BrdUrd-positive cells in the SVZ of the fmr1-KO mice (n = 4) when compared with WT mice (n = 3). The error bars indicate means ± SEM. The asterisk indicates a statistically significant difference (P < 0.05) versus respective control.

Neuronal Differentiation of Human Fragile X NSCs. We also generated human NSCs from postmortem tissue of a fragile X fetus and normal fetal brains and propagated them as neurospheres in the presence of mitogens (Fig. 4A). Removal of mitogens induced the differentiation of cells in neurospheres into neuronal cells (27, 28) (Fig. 4A). FMRP labeling was shown in neurospheres generated from the normal brains (Fig. 4B), whereas no staining was found in the proliferating NSCs derived from the fragile X fetus (Fig. 4B). Double staining with GFAP revealed that FMRP labeling was mainly present in GFAP-negative neuronal cells (Fig. 4C) in accordance with the pre-dominant neuronal expression of FMRP in brain (10).

We compared the differentiation of NSCs generated from the 18-week-old fetal brain with fragile X to control cells, which were prepared from the same brain region of a normal 18-week-old fetal brain. The differentiation of fragile X neurospheres differed distinctly from the differentiation of the age-matched control neurospheres. During the first day of differentiation, fragile X cells generated shorter radial processes and more TuJ1-positive cells than controls did (Fig. 4D). FMRP-deficient neurospheres generated 5.3-fold more TuJ1-positive cells and 70% less GFAP-positive cells than the age-matched controls after differentiation for 2 weeks when the cells were clearly separated (Fig. 5 A and B). The TuJ1-positive cells accounted for 4.4% of the total cells in control neurospheres and 23.1% of the cells in fragile X neurospheres (Fig. 5B). The increased numbers of Tuj1-positive cells and the decrease in GFAP-positive cells in fragile X neurospheres were also significant when the differentiation of fragile X cells was compared with that of controls generated from human brains at different developmental stages (Fig. 8, which is published as supporting information on the PNAS web site).

In agreement with the morphological changes of newborn TuJ1-positive cells derived from FMRP-deficient mouse neurospheres, the average neurite length of the TuJ1-positive cells derived from human fragile X neurospheres was significantly shorter than the average neurite length of cells derived from the age-matched fetal control (71.91 ± 3.76 μm in fragile X and 86.07 ± 5.42 μm in controls, P < 0.01) (Fig. 5C). The decrease in fragile X neurite length was also significant when compared with the neurite length of TuJ1-positive cells derived from the control brain of a fetus that was younger by 3 weeks (89.33 ± 13.36 μm). The mean cell-body volume was also significantly smaller in the FMRP-deficient human TuJ1-positive cells than in control cells (45.38 ± 1.55 μm² in fragile X and 76.31 ± 4.46 μm² in controls, P < 0.01) (Fig. 5C).

The Absence of FMRP Alters Ca²⁺ Responses of Differentiating Human NSCs. An ACh-induced [Ca²⁺]_i response was also seen in the majority of the neurosphere-derived human cells (486 of 552; 88%) during the first day of differentiation. The typical response consisted of a peak and stable phase similar to that seen in murine cells. In the fragile X cells, 190 of 256 (74%) reacted to ACh. A significant fraction of cells derived from fragile X neurospheres responded to ACh with [Ca²⁺]_i oscillations during the stable phase (Fig. 5 D and E). Only a very small number of human control cells treated with 100 μM ACh responded with [Ca²⁺]_i oscillations (Fig. 5D). The magnitude of the initial peak and the stable phase of [Ca²⁺]_i elevation in fragile X cells was not significantly different from that seen in control cells.

Discussion

The present study provides evidence that the differentiation of FMRP-deficient NCSs is different from that seen in control NSCs. The accumulation of newborn cells in the SVZ and the increase in the numbers of immature cells of neuronal lineage as well as cells showing oscillatory responses to neurotransmitters appear to be typical features in the differentiation of FMRP-deficient rodent and human NSCs. These features may reflect abnormal responsiveness of at least one developing cell type or an abnormally dominating cell type. The alterations are expected to have effects on the function of the brain, and they may be related to the previously shown abnormalities in the brains of fragile X patients and fmr1-KO mice.

We show here that the number of neural cells generated from neurospheres was increased in the absence of FMRP. A marked increase in the production of TuJ1-positive cells with morphological alterations from NSCs derived from fmr1-KO embryos and a fragile X fetus indicates that a prominent population of neurosphere-derived neurons was affected. The reduced number of primary neurites and the short neurite length of these differentiating neurons may represent the effects of the absence of FMRP on the dominating neuronal population at the early stage of neuronal development. Human and mouse FMRP-deficient neurospheres generated less GFAP-positive cells, a phenomenon that was apparently accounted for by an increase in the programmed cell death of nonneuronal cells and that suggested disturbances in mechanisms regulating the survival of differentiating cells. The Ca²⁺ imaging data revealed functional alterations in a population of cells, which responded with a metabotropic response to glutamate and showed a relatively small response to high K⁺ depolarization. In the FMRP-deficient neurosphere-derived cells, a major proportion of cells in this cell population responded with an intense oscillatory response to ACh. An oscillatory response was also seen with other stimuli, such as glutamate, ATP, and norepinephrine in FMRP-deficient cells, suggesting that the alteration was associated with Ca²⁺ signaling rather than with a specific ACh receptor.

Abnormalities in the dendritic spines have been detected in multiple cerebral cortical regions of fragile X patients and fmr1-KO mice (17, 19-21). Evidence for altered synaptic structure and activity has also been observed in cultured hippocampal neurons from fmr1-KO mice (18). Alterations in Ca²⁺ signaling have been shown to correlate with changes in the neuronal morphology and neurotransmitter expression at an early stage of neuronal maturation (29). Furthermore, an increase in Ca²⁺ oscillation frequency can dramatically increase the efficiency of gene expression (30). Thus, these findings suggest a possibility that altered Ca²⁺ signaling in FMRP-deficient, mGluR-responsive cells may in part be responsible for the disturbed neurogenesis, synapse formation, and maturation in fragile X syndrome. Stimulation of group I mGluRs results in a translation-dependent elongation of dendritic spines resembling the abnormal spines seen in neurons without FMRP expression (31). Because the translation-dependent form of long-term depression induced by mGluR is also augmented in fmr1-KO mice, it is assumed that mGluR-dependent translation is exaggerated in the absence of FMRP. A recent study by Weiler et al. (32) showed that the overall group I mGluR-stimulated translation was significantly depressed in synaptoneurosomes of the fmr1-KO mouse.

The increase in the number of cells with an oscillatory response in neurosphere-derived cells generated from a fragile X patient or fmr1-KO mice indicates that the absence of FMRP alters cellular plasticity of a particular cell type responding to mGluR activation. The higher sensitivity of the responses to the mGluR5 antagonist MPEP as compared with the mGluR1-selective antagonist LY367385 suggests that these cells express primarily mGluR5. Interestingly, systemic administration of mGluR5 antagonist was recently reported to prevent the audiogenic seizure phenotype in fmr1-KO mice (33). It has been suggested that overactive or inappropriate group 1 mGluR signaling may contribute to the pathophysiological changes found in fragile X syndrome, including epilepsy, cognitive impairment, and abnormal dendritic spines. Furthermore, it has been proposed that group 1 mGluR antagonists might be of use as potential pharmacological treatments of the neurological and psychiatric symptoms in fragile X syndrome (34). The present study provides additional evidence that disturbances in signaling through group 1 mGluRs are associated with the pathogenesis of fragile X syndrome.

At present, it is uncertain which mechanisms are involved in the oscillatory responses to transmitters generated in fragile X cells. The generators of oscillatory signals are complex and involve different messengers (35). Regulators of G protein signaling (RGS) proteins play a critical role in the generation of [Ca²⁺]i oscillations and their frequency (36). We have recently shown that mRNA levels for RGS4 are specifically altered in the brain of the fmr1-KO mice (37) and that the changes in RGS4 expression might be involved in the oscillatory responses. The oscillations in fragile X cells may imply that FMRP, acting as a translational regulator (4-8), can regulate the expression of the proteins involved in the generation of Ca²⁺ oscillations, which in turn modify gene expression and direct cells along specific developmental pathways (30, 35).

One remarkable difference was found in the number of newborn cells in the SVZ of the telencephalon between the fmr1-KO and WT brains. The data suggest that the absence of FMRP alters differentiation of a particular subpopulation of cells in vivo, which is in line with our in vitro findings. There was no difference in the total number of newborn cells in the radial cross sections of WT and KO cortex. This finding is in agreement with observations in imaging studies, which indicate that the gross morphology of the fmr1-KO mouse brain is normal (38). Structural brain abnormalities with different characteristics depending on the specific brain location, however, have been reported in MRI studies of fragile X patients (39, 40). It is possible that such subtle abnormalities might result from alterations of NSC differentiation, a possibility that remains to be investigated.