Figures
Abstract
Mutations in MECP2 are responsible for the majority of Rett syndrome cases. MECP2 is a regulator of transcription, and has two isoforms, MECP2_e1 and MECP2_e2. There is accumulating evidence that MECP2_e1 is the etiologically relevant variant for Rett. In this study we aim to detect genes that are differentially transcribed in neuronal cells over-expressing either of these two MECP2 isoforms. The human neuroblastoma cell line SK-N-SH was stably infected by lentiviral vectors over-expressing MECP2_e1, MECP2_e2, or eGFP, and were then differentiated into neurons. The same lentiviral constructs were also used to infect mouse Mecp2 knockout (Mecp2tm1.1Bird) fibroblasts. RNA from these cells was used for microarray gene expression analysis. For the human neuronal cells, ∼800 genes showed >three-fold change in expression level with the MECP2_e1 construct, and ∼230 with MECP2_e2 (unpaired t-test, uncorrected p value <0.05). We used quantitative RT-PCR to verify microarray results for 41 of these genes. We found significant up-regulation of several genes resulting from over-expression of MECP2_e1 including SRPX2, NAV3, NPY1R, SYN3, and SEMA3D. DOCK8 was shown via microarray and qRT-PCR to be upregulated in both SK-N-SH cells and mouse fibroblasts. Both isoforms up-regulated GABRA2, KCNA1, FOXG1 and FOXP2. Down-regulation of expression in the presence of MECP2_e1 was seen with UNC5C and RPH3A. Understanding the biology of these differentially transcribed genes and their role in neurodevelopment may help us to understand the relative functions of the two MECP2 isoforms, and ultimately develop a better understanding of RTT etiology and determine the clinical relevance of isoform-specific mutations.
Citation: Orlic-Milacic M, Kaufman L, Mikhailov A, Cheung AYL, Mahmood H, Ellis J, et al. (2014) Over-Expression of Either MECP2_e1 or MECP2_e2 in Neuronally Differentiated Cells Results in Different Patterns of Gene Expression. PLoS ONE 9(4): e91742. https://doi.org/10.1371/journal.pone.0091742
Editor: Namik Kaya, King Faisal Specialist Hospital and Research center, Saudi Arabia
Received: October 9, 2013; Accepted: February 14, 2014; Published: April 3, 2014
Copyright: © 2014 Orlic-Milacic et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by funding from International Rett Syndrome Foundation, Ontario Mental Health Foundation, and Cure Autism Now to JBV and BAM. JBV is a National Alliance for Research on Schizophrenia and Depression Independent Investigator. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Rett syndrome (RTT; MIM#312750) is a progressive neurodevelopmental disorder that affects females almost exclusively, with a frequency of ∼1∶10000 births. The overwhelming majority of Rett syndrome cases are caused by mutations in the Methyl-CpG-binding protein 2 gene, MECP2 [1]. The MECP2 protein has two isoforms, MECP2_e1 and MECP2_e2, which differ in exon usage and amino terminal amino acid sequences [2], [3]. There is much evidence to suggest that MECP2_e1 may be the etiologically relevant isoform in Rett Syndrome based on its expression profile in the brain and because, to-date, no mutations in Rett individuals have been discovered that affect MECP2_e2 exclusively, while mutations have been found in exon 1 that exclusively impact on MECP2_e1 [4], [5]. Overall, this would suggest that MECP2_e1 may be more relevant to Rett syndrome.
The two MECP2 isoforms exhibit different regional and temporal expression patterns in mouse brain [6], with MECP2_e1 more abundant throughout the brain, and MECP2_e2 less abundant and more restricted. It is frequently assumed that there are no functional differences between the two isoforms, although the evidence to support this assumption is limited. A recent study using complementation with isoform-specific MECP2 transgenes in MECP2-null mice suggests that both isoforms can prevent development of the Rett-like phenotype in the mice, and thus may substitute for each other, and are likely to perform similar cellular functions [7]. In silico analysis predicts that MECP2_e1 is a more stable (and N-acetylated) protein and has a longer half-life (220 hrs) than MECP2_e2 (unacetylated; 65 hrs; [8]). However, the addition of the N-terminal eGFP tag to the MECP2_e2 constructs by Kerr et al. [7] may lead to N-acetylation (MECP2_e1 transgene was tagged at C-terminus), and double the predicted half-life of the protein. Also, the degree of phenotypic rescue for the two isoforms in the Kerr et al study was not fully concordant, particularly for motor phenotypes, with the MECP2_e1 transgene more effective. Thus, while very similar, with >95% sequence identity, there still may be important functional differences between the two isoforms.
Also, in contrast to the Kerr et al study, recent studies of isoform-specific MECP2 knockouts in mice appear to show that knockout of MECP2_e2 has no detrimental effect on normal neurodevelopment and behavior [9], whereas knockdown of MECP2_e1 fully recapitulates the neurological symptoms seen in Rett syndrome [10]. Thus, there is now convincing data to support the hypothesis that disruption of MECP2_e1 rather than MECP2_e2 leads to RTT.
We hypothesize that the two isoforms regulate the transcription of different sets of genes, and for MECP2_e1 these will be enriched for genes relevant to the neurodevelopmental phenotype. To investigate this hypothesis we have infected human SK-N-SH cell lines with lentiviral constructs encoding either MECP2_e1 or MECP2_e2 or a control construct, and differentiated stably infected cells into neurons [11]. One of the major advantages of this approach is that we are able to look selectively at neuronal response, rather than at a heterogeneous group of cells including glia and astrocytes, which may greatly dilute the response from neurons. We have used gene expression microarrays supported with quantitative real time RT-PCR to catalogue genes that are either up or down-regulated under the effects of over-expression of either isoform in the differentiated cells. In addition, in order to investigate the effect of over-expression of MECP2_e1 or MECP2_e2 in MECP2 null background, we used these lentiviral vectors to stably infect primary mouse fibroblasts derived from male MECP2 knockout mice. Although human neuronal cells with no functional MECP2 expressed would have been desirable, such cells were not available for the study. As our experiments studied the effects of over-expression of MECP2, the results are also relevant to MECP2 duplication syndrome (MIM #300260).
Materials and Methods
Research Ethics Statement
Human brain tissue was obtained and used following informed written consent from next of kin and approval from the Research Ethics Board at the Centre for Addiction and Mental Health (http://www.camh.ca/en/research/research_ethics/research_ethics_board_general_information/Pages/reb_general_information.aspx).
Cell Lines
Human neuroblastoma cell line SK-N-SH (ATCC, Manassas, VA, USA) was grown in Eagle’s Minimal Essential Medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FBS (Wisent, St. Bruno, QC, Canada) and Penicillin-Streptomycin (Invitrogen, Carlsbad, CA, USA) Primary mouse fibroblasts isolated from MECP2 knockout mice Mecp2tm1.1Bird (Jackson laboratories, ME, USA) were generated in the Ellis lab by standard procedures, and were grown in Minimal Essential Medium (Invitrogen, Carlsbad, CA, USA) supplemented with 20% FBS and Penicillin-Streptomycin. Cells were kept at 37°C in humid atmosphere with 5% CO2.
Infections
The recombinant lentiviruses that we used to infect subconfluent SK-N-SH cells were described previously by Rastegar et al. [12]. These are self-inactivating lentiviruses encoding either MECP2_e1, MECP2_e2 or eGFP under the control of the EF1α promoter. The optimal multiplicity of infection (MOI) for infection of SK-N-SH cells was determined by infecting SK-N-SH cells and primary mouse MECP2 knockout fibroblasts by the eGFP-encoding lentiviral vector. The MOI of 20 resulted in over 90% of infected SK-N-SH cells, while the optimal MOI for primary mouse fibroblasts was determined to be 100. The optimal MOI for each cell line was used for infection of these cell lines with all lentivirus constructs. The infections were performed in the presence of 100 μg/ml polybrene (Sigma-Aldrich). Three independent infections with each lentiviral construct were performed for each cell line, providing three biological replicates for each treatment. Three days after infection the virus-containing medium was removed and replaced with fresh growth medium.
Neuronal Differentiation
Neuronal differentiation of infected SK-N-SH cells was achieved by treatment of cells with retinoic acid (Cedarlane Laboratories, Burlington, ON, Canada) and herbimycin A (Cedarlane Laboratories, Burlington, ON, Canada), using a modification of the protocol previously described [11]. Retinoic acid was added to the medium at 10 μM concentration seven days post-infection, and cells were grown in the presence of retinoic acid for 96 h, with medium being replaced with fresh retinoic-acid containing medium every two days. Retinoic acid treatment increases the number of neuronally committed cells in the SK-N-SH culture (Preis et al. 1988). After 96 hours, herbimycin A was added to the growth medium at 236 nM concentration, and cells were grown in the presence of both herbimycin A and retinoic acid for 72 hours, with medium being replaced with fresh medium after 48 hours. Herbimycin A was used to clear the undifferentiated cells from the culture [11].
RNA Isolation and Reverse Transcription
RNA was isolated from stably infected neuronally differentiated cells and from stably infected primary mouse fibroblasts using TRIzol (Invitrogen, Carlsbad, CA, USA), in accordance with the manufacturer’s protocol. The concentration of total RNA was determined by NanoDrop spectrophotometry. Isolated RNA was stored at −80°C. The reverse transcription was performed using 2 μg of total RNA, SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and random primers (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer’s instructions.
Microarray Gene Expression Analysis
RNA isolated from stably infected neuronally differentiated SK-N-SH cells and primary mouse fibroblasts was used for microarray gene expression analysis on human and mouse Agilent arrays, respectively (Agilent Technologies, Santa Clara, CA, USA). For SK-N-SH cells, the Agilent Whole Human Genome Oligo44K (#14850) two-color microarrays were used, and for mouse fibroblasts the Agilent Whole Mouse Gene Expression 44K (#14868) two-color arrays were used. Microarray hybridizations, data collection, and analysis of results were performed as a service by the University Health Network (UHN) Microarray Centre (http://www.microarrays.ca/). Each hybridization was performed in triplicate. Therefore, there were three technical replicates for each biological replicate. Hybridizations were two-color, and human and mouse reference RNAs (Agilent, Santa Clara, CA, USA) were used as controls for SK-N-SH and mouse primary fibroblasts, respectively. Fold changes in gene expression were calculated relative to eGFP-lentivirus-treated cells. Both ANOVA analysis and unpaired t-tests were used for statistical analysis. BioVenn (http://www.cmbi.ru.nl/cdd/biovenn/index.php [13]) was used to identify the overlap in differentially expressed genes between different treatments. We used DAVID Gene Ontology (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of differentially expressed genes.
Microarray data from this study has been deposited in the NCBI GEO, and has been assigned accession numbers GSE54258 (human/SK-N-SH) and GSE54261 (mouse/ko fibroblast).
Real-Time RT-PCR
Real-time RT-PCR was performed on ABI 7500 or Viia7™ machines (Life Technologies, Foster City, CA, USA). Expression of NAV3, FOXP2 and PTCHD1 was examined using TaqMan Gene Expression Assays (Hs00372108_m1, Hs00362817_m1, and Hs00288486_m1, respectively) and TaqMan PCR MasterMix (Life Technologies, Foster City, CA, USA). Primers with probes for analysis of MECP2_e1 and MECP2_e2 expression were custom ordered from IDT and their sequences are provided in Table 1. Expression of other genes of interest was validated by using custom-made primers and SYBRgreen PCR master mix (Life Technologies, Foster City, CA, USA). Primer sequences are also provided in Table 1. Two housekeeping genes, HPRT and GAPDH, were initially used as endogenous controls, but after it was established that the expression of these two genes did not vary between samples, only one endogenous control gene, HPRT, was used. TaqMan Gene Expression Assays for HPRT and GAPDH (Hs99999909_m1 and Hs02758991_g1, respectively) were used to quantify HPRT and GAPDH transcripts used as endogenous control for relative quantification of NAV3, FOXP2 and PTCHD1 expression. TaqMan Gene Expression Assay for Tbp (Mm00446973_m1) was used as an endogenous control to determine expression changes in MECP2_e1 and MECP2_e2 after lentiviral infection. Sequences of HPRT primers used with SYBRgreen chemistry are provided in Table 1. Each real-time PCR reaction was done in triplicate. Real-time RT-PCR results were analyzed using ABI 7500 Software v2.0.1 or ViiA7 RUO software. Microsoft Excel was used for statistical analysis of real-time RT-PCR data by F-test and to plot the data. RNA isolated from an adult human brain frontal cortex sample, kindly provided by Dr. Stephen Kish (following approval from the Research Ethics Board at the Centre for Addiction and Mental Health (http://www.camh.ca/en/research/research_ethics/research_ethics_board_general_information/Pages/reb_general_information.aspx), and written consent from next of kin) was used as a positive control for expression of genes of interest in the brain.
Microscopy
To determine the percentage of eGFP positive cells after infection, sterile glass cover slips coated with Poly-D-Lysine (Sigma-Aldrich) were put in tissue culture flasks prior to cell plating and infection. Seven days after infection, the medium was removed, and cells were washed with PBS and fixed with 4% PFA. Glass cover slips were recovered, placed on microscope slides and, after cell nuclei were counterstained with DAPI, examined for percentage of eGFP positive cells by fluorescence microscope (Eclipse E600, Nikon) in the green fluorescent field.
To examine the percentage of cells with neuronal morphology during the course of retinoic acid and herbimycinA treatment, SK-N-SH cells were grown and treated with differentiation medium in microscope slide chambers. When the treatment procedure was completed, cells were fixed with methanol, stained with Coomassie blue, and examined in the bright light field using Nikon microscope (Eclipse E600, Nikon).
Results
Stably Infected Cells Over-Express MECP2 Isoforms
Cell lines of neuronal origin, like SK-N-SH cells, as well as primary cell cultures are known to be refractory to commonly used transfection reagents. In addition, our experiment required a longer time frame than the one provided by transient transfection. Therefore, we used recombinant lentiviruses to achieve stable over-expression of human MECP2 isoforms e1 or e2, or stable expression of control gene eGFP in human neuroblastoma cell line SK-N-SH and primary mouse MECP2 knockout fibroblasts. The empirically determined optimal MOI for these cells resulted in >90% of infected cells, based on eGFP expression (Figure S1).
Stably infected SK-N-SH cells were differentiated into neurons by retinoic acid and herbimycinA treatment [11]. After being grown in differentiation medium for seven days, as described in Materials and Methods, more than 90% of cells in culture exhibited neuronal morphology (Figure S2).
Infected cells were harvested and RNA was isolated from primary mouse MECP2 knockout fibroblasts seven days after infection, and from SK-N-SH cells after the differentiation treatment was completed. Expression of MECP2 isoforms in primary mouse MECP2 knockout fibroblasts and overexpression of MECP2 isoforms in neuronally differentiated SK-N-SH cells was confirmed by real-time RT-PCR.
Neuronally differentiated SK-N-SH cells infected with the control eGFP lentiviral vector express both transcript variants of MECP2. The level of MECP2_e1 is approximately 3-fold lower than in the frontal cortex of the human brain, while MECP2_e2 levels are similar between brain and neuronally differentiated neuroblastoma cells. Stable infection with MECP2-encoding recombinant lentiviruses results in overexpression of MECP2_e1 or MECP2_e2. In SK-N-SH cells infected with the MECP2_e1 vector, MECP2_e1 transcript is 35 times more abundant than in the brain frontal cortex (∼100-fold increase relative to SK-N-SH cells infected with the control eGFP-expressing lentiviral vector), and 5 times more abundant than the MECP2_e2 transcript. In SK-N-SH cells infected with the MECP2_e2 vector, MECP2_e2 transcript is 11 times more abundant than in the brain frontal cortex and SK-N-SH cells infected with the eGFP control vector, and approximately 100 times more abundant than the MECP2_e1 transcript (Figure 1A and 1B).
A) Infection with the MECP2_e1 vector increases MECP2_e1 mRNA level ∼35-fold, compared with the brain frontal cortex and ∼100-fold compared with the eGFP control infected cells. Infection with the MECP2_e2 vector increases MECP2_e2 mRNA level 11-fold compared with the brain frontal cortex and the eGFP control. Relative quantities of MECP2_e1 and MECP2_e2 transcripts for each treatment/sample are indicated in data labels, below sample names; B) In cells infected with the MECP2_e1 vector, MECP2_e1 transcript is ∼5 times more abundant than the MECP2_e2 transcript. In cells infected with the MECP2_e2 vector, the MECP2_e2 transcript is ∼100 times more abundant than the MECP2_e1 transcript. In cells infected with the eGFP control and in the brain frontal cortex, the MECP2_e2 transcript is ∼9 times more abundant than the MECP2_e1 transcript.
Infection of MECP2 knockout primary mouse fibroblasts at MOI = 100 results in ∼80% of infected cells (Figure S1). MECP2 lentivirus vectors induce expression of MECP2 isoforms in mouse MECP2 knockout fibroblasts (Figure S3).
Predominant Expression of MECP2_e1 or MECP2_e2 Results in Distinct Gene Expression Profiles
An uncorrected analysis using one way ANOVA (p<0.01) identified 1098 genes differentially expressed between SK-N-SH (neuronally differentiated) infected with MECP2_e1, MECP2_e2 or eGFP lentiviral vectors (Table S1) and 489 genes differentially expressed between mouse Mecp2 knockout fibroblasts infected with MECP2_e1, MECP2_e2 or eGFP lentiviruses (Table S2). However, correction for multiple testing using the Benjamini-Hochberg false discovery rate (FDR) test showed no significant differences (p<0.05). Using a t-test (p<0.05) ∼800 genes were identified with at least three-fold change in expression level (up- or down-regulation) when MECP2_e1 was over-expressed in SK-N-SH cells, in comparison with the eGFP control. When MECP2_e2 was over-expressed in SK-N-SH cells, levels of ∼230 genes changed at least 3-fold. Correction for FDR, however, showed no significant differences. For microarray t-test analysis results, please refer to Table S3 and Table S4.
Six genes are at least two-fold up-regulated by MECP2_e1 in both SK-N-SH cells and primary mouse MECP2-null fibroblasts: TSHR, BMX, DOCK8, COQ10B, RNF2 and LPHN3. TMC4 is ∼2-fold up-regulated by MECP2_e2 expression in both SK-N-SH cells and primary mouse fibroblasts. Four genes are at least two-fold down-regulated by MECP2_e1 in both SK-N-SH cells and primary mouse fibroblasts: ITGBL1, GPR85, KCNJ9 and ZBTB10. SMCR8 is ∼three-fold downregulated by MECP2_e2 in both SK-N-SH cells and primary mouse fibroblasts.
Gene ontology analysis of SK-N-SH expression data.
For the purpose of gene ontology analysis using DAVID, we used a cut-off of two-fold change and p<0.05. Over-expression of MECP2_e1 in neuronally differentiated SK-N-SH cells resulted in at least two-fold down-regulation of 797 genes, while over-expression of MECP2_e2 resulted in at least two-fold down-regulation of 253 genes. Of these, 122 genes were downregulated by both MECP2_e1 and MECP2_e2, and according to DAVID functional clustering, these genes are mainly involved in transcriptional control (see Table S5 and SKNSH S1). The 675 genes that were down-regulated only by MECP2_e1 are involved in cell adhesion, cell migration and motility, synaptic transmission, axonogenesis, G-protein coupled receptor signaling, regulation of behavior, immune response, regeneration, and angiogenesis (Table S5 and SKNSH S2). The 131 genes that were down-regulated only by MECP2_e2 function in chromatin organization, regulation of transcription and apoptosis (Figure 2A; Tables S5 and SKNSH S3). Over-expression of MECP2_e1 in neuronally differentiated SK-N-SH cells up-regulated, at least two fold, 934 genes, while over-expression of MECP2_e2 up-regulated, at least two-fold, 550 genes. Of these, 286 genes are up-regulated by both MECP2_e1 and MECP2_e2 and are involved in cell adhesion, cell motility, cell migration, and embryonic development (Tables S5 and SKNSH S4). The 648 genes up-regulated at least two-fold only by MECP2_e1 were significantly enriched for synaptic transmission, activity of voltage-gated channels, G-protein coupled receptor signaling, cell adhesion, and lipid transport (Tables S5 and SKNSH S5). The 264 genes up-regulated at least two-fold by MECP2_e2 alone were significantly enriched for negative regulation of transcription, G-protein coupled receptor signaling, signal transduction, and skeletal system development (Figure 2B; Table S5 and file SKNSH S6).
Biological processes that indicated gene sets are enriched for, as identified by DAVID Functional Clustering, are indicated on the right hand side of Venn diagram.
Gene ontology analysis of mouse knockout fibroblast expression data.
Using the uncorrected t-test data (p<0.05), as FDR correction showed no significant results, the expression of MECP2_e1 in primary mouse MECP2 knockout fibroblasts resulted in >two-fold down-regulation of 236 genes (for microarray t-test results, please refer to Table S6 and Table S7). By DAVID functional clustering, these genes are mainly involved in the regulation of signaling by receptor tyrosine kinases, tissue morphogenesis, regulation of transcription, neuronal differentiation, and regulation of activity of voltage gated channels (Table S8 and Mouse S1). Expression of MECP2_e2 in primary mouse MECP2 knockout fibroblasts resulted in >two-fold down-regulation of 140 genes, mainly involved in regulation of transcription, neuronal differentiation, G-protein coupled receptor signaling, regulation of small GTPases and apoptosis (Table S8 and Mouse S2). Forty three genes are>two-fold down-regulated by both MECP2_e1 and MECP2_e2 expression and are mainly involved in transcriptional regulation (Figure 3A; Table S8 and Mouse S3). At least two-fold up-regulation of 807 genes was detected when MECP2_e1 was expressed in primary mouse MECP2 knockout fibroblasts. These genes are involved in transcriptional regulation, neuronal development, sensory organ development, DNA repair, lipid metabolism, G-protein coupled receptor signaling and vesicle-mediate transport (Table S8 and Mouse S4). Expression of MECP2_e2 in primary mouse MECP2 knockout fibroblasts resulted in >2-fold up-regulation of 94 genes involved in transcriptional regulation and cell motility and migration (Table S8 and Mouse S5). Forty three genes are at least two-fold up-regulated by both MECP2_e1 and MECP2_e2 expression and are mainly implicated in G-protein coupled receptor signaling and regulation of small GTPases (Figure 3B; Table S8 and Mouse S6).
Biological processes that indicated gene sets are enriched for, as identified by DAVID Gene Ontology Functional Clustering, are indicated on the right hand side of Venn diagrams.
Quantitative Validation of Microarray Results by RT-PCR
We used quantitative RT-PCR to verify microarray results for a total of 41 genes, including 29 that, in addition to individually significant change in expression level (uncorrected for FDR), are known to be involved in cognitive function or neuronal development, or are highly expressed in CNS tissues, or are implicated in Rett syndrome, intellectual disability or autism (Tables 2, 3 and 4, Figures 4, and 5). An additional 12 genes were identified as differently expressed in both the transfected SK-N-SH cells and in the mouse knockout fibroblasts (Table 5). As uncorrected statistical tests were used to identify differentially expressed genes by microarray, a higher rate of false-positives and false-negatives was expected. We confirmed significant up-regulation of several genes resulting from over-expression of the e1 isoform in SK-N-SH cells, including NAV3 (5-fold level increase), FOXP2 (induced by MECP2_e1 over-expression, and weakly induced by MECP2_e2 (“induced”-indicating that no expression was detectable in control eGFP-infected cells, thus no fold-change calculation was possible), Figure 5a), NPY1R (six-fold increase), SYN3 (two-fold increase), and SEMA3D (six-fold increase). Both MECP2_e1 and MECP2_e2 up-regulated GABRA2 (18-fold and three-fold increase, respectively), KCNA1 (seven-fold and three-fold increase, respectively) and FOXG1 (six-fold and three-fold increase, respectively). Down-regulation of expression in the presence of e1 was seen with UNC5C(40% down-regulation) and RPH3A (30% down-regulation). Expression levels of four genes determined by the qPCR contradicted microarray results: SPRX2, ITGA3 and NDN were reported as down-regulated by microarray analysis but found to be up-regulated by qPCR, and CNTN4, which was identified as up-regulated by microarray analysis, was indicated as down-regulated by qPCR (Table 2). Several of the genes, while potentially relevant to RTT, had very low levels of expression and could not be validated by this method, including DLGAP2, CRH, GABRB, GRIN2A and PTCHD1 (Table 2).
Expression of target genes was normalized to HPRT endogenous control and calculated relative to eGFP control treatment.
A weak signal is also visible when cells are infected with MECP2_e2 lentivirus vector. Frontal cortex was used as a positive control. Replicates represent biological replicates i.e. repeated infections. For FOXP2, real-time RT-PCR products were run on an agarose gel. For DOCK8, end point PCR was performed using KAPA2G Fast HotStartReadyMix (KapaBiosystems, KK5601. PCR consisted of 35 cycles, and the annealing temperature was 60°C. C) GAPDH was used as an endogenous control for RT-PCR.
Of the twelve genes identified to be differentially regulated by MECP2 isoforms in both human and mouse cells, we were able to confirm induction of DOCK8 expression by MECP2_e1 in neuronally differentiated SK-N-SH cells (Figure 5b). A slight increase in Dock8 expression in primary mouse MeCP2 knockout fibroblasts infected with MECP2_e1 lentivirus vector was not statistically significant. Five genes (BMX, ZBTB10, RNF2, COQ10B and LPHN3) showed expression trends by qRT-PCR, in accordance with microarray results, in either human or mouse cells, or both, but the expression level changes were not statistically significant. Two of the genes, TSHR and KCNJ9, had very low expression levels that could not be quantified. Four genes (ITGBL1, GPR85, TMC4 and SMCR8) showed expression level changes, by quantitative RT-PCR, that contradicted microarray findings. Results of microarray and quantitative RT-PCR for genes differentially regulated by MECP2 isoforms in both human and mouse cells are summarized in Table 5.
Discussion
We have used an over-expression system to identify downstream targets of both the main isoforms of MECP2. This allows us some insight into the differences and similarities between the two isoforms. The use of an over-expression system obviously does not mimic the situation in Rett patients, and, as mentioned earlier, is more etiologically relevant to the MECP2 duplication syndrome. However, the purpose of this approach was chiefly to produce a strong and measurable effect from the addition of the exogenous MECP2 isoforms. Another caveat is that direct comparison of levels of up- or down-regulation of genes for the two isoforms would be inappropriate, as the translational efficiencies of the two isoforms are different [3], and stability and longevity are also predicted to be different (see in silico analysis in Introduction). Another restriction of this approach is that it cannot distinguish whether MECP2 is a direct regulator, or whether it is acting through other effector molecules. Chromatin-immunoprecipitation followed by sequencing (ChIP-seq) may help show which of these genes are direct targets, however, recent studies using this approach suggest that MECP2 does not show discreet binding targets, but rather shows a much broader genome-wide distribution that generally tracks the density of CpG methylation [16]. It is thus likely that the gene-specificity for MECP2 depends very much upon the presence of co-repressors or activators, which may in turn be tissue and temporally specific.
Our quantification of e1 and e2 mRNA in eGFP-infected SK-N-SH cells and post mortem human frontal cortex using the ΔΔCt method of relative quantification shows ∼9-fold higher levels of e2 than e1. This contrasts with several previous studies which used quantification of end-point RT-PCR products and showed higher levels of e1 than e2 [2], [3]. Our previous study looking at e1 and e2 mRNA levels in lymphocytes using the ΔΔCt method also showed very similar levels of e1 and e2 [17], also differing from the previous reports [2], [3]. This may not, however, correspond directly to relative amount of the protein isoforms, as e1 is translated with much higher efficiency than e2 [3], and has been shown to be a more stable protein than e2 [10], and as predicted by in silico analysis of the different N-terminal sequences for likelihood of N-methionine excision and N-terminal acetylation and effect on protein stability and longevity [10].
Of the isoform-specific targets that were validated using qPCR, NAV3 is one of the most interesting, and showed approximately five-fold increase in expression with over-expression of MECP2_e1. NAV3 encodes neuron navigator 3- so called based on sequence similarity to the C-terminal end of the C. elegans protein unc-53, which is involved in axon guidance. Affymetrix gene chip data shows NAV3 to be strongly expressed in brain, and in particular fetal brain, prefrontal cortex and amygdala (microarray data for probe 204823_AT from GeneAtlas U133A; biogps.org). Its expression correlates with that of MEF2C (probe 209200_AT)- a gene linked with severe intellectual disability (ID), stereotypic movements and epilepsy [18]. Zweier et al. [19] reported several patients with de novo truncating or missense mutations in MEF2C, and showed decreased levels of MECP2 and CDKL5 expression in patient lymphocytes, and speculated that the phenotypic overlap between RTT and MEF2C mutation patients is due to a common pathway.
SEMA3D showed a six-fold increase in expression with MECP2_e1. SEMA3D encodes a semaphorin protein. Semaphorins are secreted molecules that function primarily through neuropilin receptors to guide the axonal growth cone. Liu and Halloran [20]showed that SEMA3D protein expression at the dorsal spinal cord midline in zebrafish repelled growth of peripheral axons from the spinal cord, and knockdown reduced numbers of peripheral axons. SEMA3D repels axons through receptors containing neuropilin-1A, and attracts axons through receptors containing neuropilin-1A and 2B [21].
NPY1R expression was also increased approximately six-fold under MECP2_e1. NPY1R encodes the neuropeptide Y receptor 1. Neuropeptide Y, or NPY, is a 36-amino-acid neurotransmitter, involved in neurophysiological functions such as cardiovascular homeostasis [22] and hypothalamic regulation of food intake [23]. NPY has several receptors, including NPY1R [24]. NPY1R also shows strong expression almost exclusively in fetal brain (microarray data for probe 205440_s_at from GeneAtlas U133A; biogps.org). Interestingly, Rett syndrome includes dysregulation of cardiovascular functioning, described as “an increase in sympathetic activity that is not counter balanced by adequate vagal tone” [25]. It should also be noted that the isoform-specific localization for MECP2 in mouse brains, found isoform e1 but not e2 in the hypothalamus [6].
The gene SYN3, which encodes Synapsin 3, a vesicle-associated phosphoprotein, located at the synapse, is also significantly over-expressed (∼two-fold) in SK-N-SH cells under regulation of MECP2_e1. Its expression is strongly correlated with that of KCNH6- a potassium voltage-gated channel, subfamily H (microarray data for probe 206322_at (SYN3) and 211045_s_at (KCNH6) from GeneAtlas U133A; biogps.org).
We validated two genes that were down-regulated with the over-expression of MEPC2_e1. These were UNC5C (decreased by 40%) and RPH3A (decreased by 30%). UNC5C belongs to a family of netrin-1 (MIM 601614) receptor genes and is thought to mediate the chemorepulsive effect of netrin-1 on specific axons. RPH3A (MIM *612159) encodes Rabphilin 3A, an effector of RAB3A (MIM 179390), a G-protein known to be involved in exocytosis during synaptic neurotransmission, interacting with the SNARE synaptic complex [26], [27]. RPH3A is strongly expressed in whole brain, prefrontal cortex, cingulate cortex, subthalamic nucleus, and thalamus in particular (microarray data for probe 205230_AT from GeneAtlas U133A; biogps.org).
Up-regulation of FOXP2 (MIM*605317) by MECP2 (both isoforms) is of significance because of the association between this gene and severe speech and language disorder known as developmental verbal dyspraxia (SPCH1; MIM 602081; [28]). Partial or complete loss of acquired spoken language is one of the main diagnostic criteria for Rett syndrome [29], however SPCH1 is associated with articulatory disturbance and severe orofacial dyspraxia [30] – a very different clinical manifestation. Microarray analyses on COS-7 cells transfected with a plasmid (non-lentiviral) MECP2_e1 construct also showed a >2-fold up-regulation of FOXP2 (data not shown). One of the genes showing down-regulation under MECP2_e1 by microarray, SRPX2, actually showed significant up-regulation by RT-PCR analysis (17-fold; p = 3×10−4). This gene, which encodes a sushi repeat-containing protein, was also inferred as down-regulated by MECP2 in microarray analysis in several studies of human cell lines [31], [32], but up-regulated in the study of hypothalami from null mice [33]. Disruption of SRPX2 is associated with Rolandic epilepsy, speech dyspraxia and ID [34]. It has also been shown that the SRPX2/uPAR complex is regulated by FOXP2, thus suggesting SRPX2 may play a role in speech-related disorders [35]. It is possible that the down-regulation of SRPX2 seen in our study is mediated by the up-regulation we see for FOXP2, and that this together may play a significant role in the phenotypic features of RTT.
Both isoforms appear to up-regulate expression of GABRA2, a gamma-aminobutyric acid A receptor gene located on chromosome 4p. The GABA receptor gene cluster at this locus has previously been implicated in autism [36]–[38], and GABA(A) receptors including GABRA2 have been reported to be significantly down-regulated in brains of autism subjects [39]. Dysfunction of GABA signaling is believed to underlie various Rett syndrome-like phenotypes seen in Mecp2-deficient mice, in particular forepaw stereotypic movements, increased sociability, impaired motor coordination, cognitive deficits, and respiratory dysrhythmia [40]. Our data suggests that disruption of either MECP2 isoform, would down-regulate GABRA2 expression.KCNA1 is also upregulated by both MECP2 isoforms. KCNA1 encodes a voltage-gated potassium channel, which binds to postsynaptic density proteins SAP97, PSD-93 and PSD-95. Mutations in KCNA1 may result in seizures, episodic ataxia and myokymia syndrome (MIM 160120).
Mutations in FOXG1 have recently been shown to be the cause of a syndrome frequently referred to as a congenital form of Rett (MIM 613454), typically with de novo heterozygous truncating mutations associated with infantile onset of microcephaly, ID, no speech development, poor motor development and stereotypic movements [41]–[44]. Interestingly, it has recently been shown that MECP2_e2, but not e1, promotes neuronal death, and that this activity is limited through interaction with FOXG1. Thus, it is of great interest that FOXG1 should be validated as a downstream target of MECP2, and for both isoforms (MECP2_e1: six-fold increase; MECP2_e2: three-fold increase). Both isoforms of MECP2 have recently been shown to interact directly with the FOXG1 gene, with higher level of interaction of MECP2_e2 than for _e1 [45].
A number of other genes of interest were identified through analysis of the microarray data as being over- or under-expressed, but because expression levels were at the low end in SK-N-SH cells, it was impossible to validate the findings with any reliability. These genes included several previously identified in relation to Rett, autism or ID, such as CRH (MIM*122560; Corticotropin-releasing hormone; dysregulated CRH expression shown in Mecp2 knockout mice; [33], [46]), PTCHD1 (MIM*300828; disruptions may result in autism or ID; [47], [48]), GRIN2A (MIM*138253; a component of synaptic glutamate receptors; mutations have been associated with seizures and ID, [49]), DLGAP2 (MIM*605438; CNVs disrupting this gene have been identified in autism; [50]).
The up-regulation of genes such as DOCK8, confirmed by qRT-PCR, in both SK-N-SH cells and primary mouse MeCP2 knockout fibroblast is also of much interest. Defects in DOCK8 have been reported as the cause of autosomal dominant ID type 2 (MRD2; MIM 614113 [51]). However, somewhat in contradiction to this, recessive mutations in this gene have been reported as the cause of Hyper IgE recurrent infection syndrome (MIM 243700 [52]), for which there is no cognitive deficit reported. Down-regulation of genes such as SMCR8 was observed in both SK-N-SH cells and Mecp2-null mouse fibroblasts. SMCR8 is a candidate gene for Smith–Magenis Syndrome (SMS), a developmental disorder that affects many parts of the body [53]. The major features of this condition include ID, facial features, sleep disturbances, and behavioral problems.
The identification and validation of various synapse-related genes showing altered transcription after MECP2 over-expression supports the hypothesis that mutations in MECP2 cause Rett syndrome through disruption of synaptic plasticity [54]. The results indicating that both isoforms up-regulate some of the same genes, e.g. KCNA1, FOXG1, and GABRA2, suggests that genetic disruption of the MECP2_e2 isoform could potentially have a phenotypic effect, however one would expect it to be milder than for disruption involving MECP2_e1, given the differences in location of the two isoforms in the brain [6]. Differences in the gene targets between the two isoforms may reflect actual differences in gene targets, however it cannot be ruled out that some of the differences may be downstream effects of the over-expression, or may be due in part to differences in protein stability and longevity within the infected cells. The fact that some qPCR results contradict the microarray data suggests that there are limitations to the conclusions that may be drawn on unvalidated microarray gene expression data. Moves towards transcriptome analysis using next generation sequencing approaches should improve the accuracy for determining gene targets for MECP2.
The gene ontology analysis reported here seems to show a more diverse set of functions for e1, compared to e2, and in particular an effect of e1 on regulation of genes involved in synaptic transmission, axonogenesis and activity of voltage-gated channels (SK-N-SH cells), and neuronal differentiation and development (fibroblasts). In contrast, many genes up- or down-regulated by e2 appear to be involved in chromatin organization and transcriptional regulation. In both SK-N-SH and fibroblasts, the genes down-regulated by both isoforms are categorized as being involved in transcriptional regulation.
Differences at the N-termini of the two MECP2 isoforms are predicted to generate an N-terminal N-acetylated alanine for MeCP2_e1, but a non-acetylated valine for MeCP2_e2. This is likely to result in a much shorter lifespan for the e2 isoform [8]. Whether this alone is enough to influence the gene regulatory effects of MeCP2, or whether it may also have a direct effect on DNA-binding specificity, remains to be seen.
In summary, we provide evidence that the two isoforms of MECP2 show distinct but overlapping profiles of gene transcription under their regulation, and propose that a number of these transcriptional effects translate to the clinical features of Rett syndrome.
Supporting Information
Figure S1.
Infection efficiency determined by eGFP expression in recombinant lentivirus treated cells. A) DAPI was used to visualize nuclei and determine the total number of cells in the visual field. Infection efficiency was calculated as the percentage of eGFP positive cells. SK-N-SH cells are shown. B) Comparison of infection efficiency between SK-N-SH cells and primary mouse Mecp2 knockout fibroblasts. MOI used is indicated above the bars. Calculation was done based on three independent infections for each sample.
https://doi.org/10.1371/journal.pone.0091742.s001
(TIF)
Figure S2.
Neuronal differentiation of neuroblastoma cell line SK-N-SH. A) Untreated cells. B) Cells treated with retinoic acid only. C) Cells treated with retinoic acid and herbimycin A.
https://doi.org/10.1371/journal.pone.0091742.s002
(TIF)
Figure S3.
Induction of MECP2 expression in mouse Mecp2 knockout fibroblast by stable infection with recombinant lentiviruses expressing MECP2 isoforms e1 and e2. Real-time RT-PCR amplification curves for MECP2_e1 (first row), MECP2_e2 (second row) and endogenous control Tbp (third row) are shown. Mecp2 knockout mouse fibroblasts were stably infected by recombinant lentiviruses encoding either MECP2_e1 (first column), MECP2_e2 (second column) or eGFP control (third column).
https://doi.org/10.1371/journal.pone.0091742.s003
(TIF)
Table S1.
Uncorrected one way ANOVA (p<0.05) results of gene expression microarray on neuronally differentiated SK-N-SH cells infected with MECP2_e1, MECP2_e2 or eGFP lentiviral vectors.
https://doi.org/10.1371/journal.pone.0091742.s004
(XLSX)
Table S2.
Uncorrected one way ANOVA (p<0.05) results of gene expression microarray on mouse Mecp2 knockout fibroblasts infected with MECP2_e1, MECP2_e2 or eGFP lentiviral vectors.
https://doi.org/10.1371/journal.pone.0091742.s005
(XLSX)
Table S3.
Gene expression microarray results for differentiated SK-N-SH cells with over-expression of MECP2_e1 compared to cells infected with eGFP (uncorrected t-test).
https://doi.org/10.1371/journal.pone.0091742.s006
(XLS)
Table S4.
Gene expression microarray results for differentiated SK-N-SH cells with over-expression of MECP2_e2 compared to cells infected with eGFP (uncorrected t-test).
https://doi.org/10.1371/journal.pone.0091742.s007
(XLS)
Table S5.
List of genes showing altered transcription in differentiated SK-N-SH cells under e1 or e2 over-expression (or both), categorized as either down-regulation or up-regulation. Overlap with previous transcriptome studies for MECP2 is also indicated, with genes highlighted in yellow reported as up-regulated or activated, and in green as down-regulated or repressed.
https://doi.org/10.1371/journal.pone.0091742.s008
(XLSX)
Table S6.
Gene expression microarray results for mouse Mecp2 knockout fibroblasts with over-expression of MECP2_e1 compared to cells infected with eGFP (uncorrected t-test).
https://doi.org/10.1371/journal.pone.0091742.s009
(XLS)
Table S7.
Gene expression microarray results for mouse Mecp2 knockout fibroblasts with over-expression of MECP2_e2 compared to cells infected with eGFP (uncorrected t-test).
https://doi.org/10.1371/journal.pone.0091742.s010
(XLS)
Table S8.
List of genes showing altered transcription in mouse Mecp2 knockout fibroblasts under e1 or e2 over-expression (or both), categorized as either down-regulation or up-regulation.
https://doi.org/10.1371/journal.pone.0091742.s011
(XLSX)
SKNSH S1.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in SKNSH cells by both e1 and e2.
https://doi.org/10.1371/journal.pone.0091742.s012
(TXT)
SKNSH S2.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in SKNSH cells by e1.
https://doi.org/10.1371/journal.pone.0091742.s013
(TXT)
SKNSH S3.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in SKNSH cells by e2.
https://doi.org/10.1371/journal.pone.0091742.s014
(TXT)
SKNSH S4.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in SKNSH cells by both e1 and e2.
https://doi.org/10.1371/journal.pone.0091742.s015
(TXT)
SKNSH S5.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in SKNSH cells by e1.
https://doi.org/10.1371/journal.pone.0091742.s016
(TXT)
SKNSH S6.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in SKNSH cells by e2.
https://doi.org/10.1371/journal.pone.0091742.s017
(TXT)
Mouse S1.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in mouse knockout fibroblasts by e1.
https://doi.org/10.1371/journal.pone.0091742.s018
(TXT)
Mouse S2.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in mouse knockout fibroblasts by e2.
https://doi.org/10.1371/journal.pone.0091742.s019
(TXT)
Mouse S3.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes down regulated in mouse knockout fibroblasts by both e1 and e2.
https://doi.org/10.1371/journal.pone.0091742.s020
(TXT)
Mouse S4.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in mouse knockout fibroblasts by e1.
https://doi.org/10.1371/journal.pone.0091742.s021
(TXT)
Mouse S5.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in mouse knockout fibroblasts by e2.
https://doi.org/10.1371/journal.pone.0091742.s022
(TXT)
Mouse S6.
DAVID Gene Ontology analysis (http://david.abcc.ncifcrf.gov/[14], [15]) for functional clustering of genes upregulated in mouse knockout fibroblasts by both e1 and e2.
https://doi.org/10.1371/journal.pone.0091742.s023
(TXT)
Author Contributions
Conceived and designed the experiments: JBV BAM JE PJG MOM. Performed the experiments: PJG MOM LK AM AYC HM. Analyzed the data: MOM JBV. Contributed reagents/materials/analysis tools: JBV JE. Wrote the paper: MOM JBV.
References
- 1. Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U, et al. (1999) Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat Genet 23: 185–188.
- 2. Mnatzakanian GN, Lohi H, Munteanu I, Alfred SE, Yamada T, et al. (2004) A previously unidentified MECP2 open reading frame defines a new protein isoform relevant to Rett syndrome. Nat Genet 36: 339–341.
- 3. Kriaucionis S, Bird A (2004) The major form of MeCP2 has a novel N-terminus generated by alternative splicing. Nucleic Acids Res 32: 1818–1823.
- 4. Saunders CJ, Minassian BE, Chow EW, Zhao W, Vincent JB (2009) Novel exon 1 mutations in MECP2 implicate isoform MeCP2_e1 in classical Rett syndrome. Am J Med Genet A 149A: 1019–1023.
- 5. Fichou Y, Nectoux J, Bahi-Buisson N, Rosas-Vargas H, Girard B, et al. (2009) The first missense mutation causing Rett syndrome specifically affecting the MeCP2_e1 isoform. Neurogenetics 10: 127–133.
- 6. Dragich JM, Kim YH, Arnold AP, Schanen NC (2007) Differential distribution of the MeCP2 splice variants in the postnatal mouse brain. J Comp Neurol 501: 526–542.
- 7. Kerr B, Soto CJ, Saez M, Abrams A, Walz K, et al. (2012) Transgenic complementation of MeCP2 deficiency: phenotypic rescue of Mecp2-null mice by isoform-specific transgenes. Eur J Hum Genet 20: 69–76.
- 8. Meinnel T, Peynot P, Giglione C (2005) Processed N-termini of mature proteins in higher eukaryotes and their major contribution to dynamic proteomics. Biochimie 87: 701–712.
- 9. Itoh M, Tahimic CG, Ide S, Otsuki A, Sasaoka T, et al. (2012) Methyl CpG-binding protein isoform MeCP2_e2 is dispensable for Rett syndrome phenotypes but essential for embryo viability and placenta development. J Biol Chem 287: 13859–13867.
- 10.
Yasui DH, Gonzales ML, Aflatooni JO, Crary FK, Hu DJ, et al.. (2014) Mice with an isoform-ablating Mecp2 exon 1 mutation recapitulate the neurologic deficits of Rett syndrome. Hum Mol Genet.
- 11. Preis PN, Saya H, Nadasdi L, Hochhaus G, Levin V, et al. (1988) Neuronal cell differentiation of human neuroblastoma cells by retinoic acid plus herbimycin A. Cancer Res. 48: 6530–6534.
- 12. Rastegar M, Hotta A, Pasceri P, Makarem M, Cheung AY, et al. (2009) MECP2 isoform-specific vectors with regulated expression for Rett syndrome gene therapy. PLoS One 4: e6810.
- 13. Hulsen T, de Vlieg J, Alkema W (2008) BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics 9: 488.
- 14. Huang da W, Sherman BT, Lempicki RA (2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4: 44–57.
- 15. Huang da W, Sherman BT, Lempicki RA (2009) Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37: 1–13.
- 16. Skene PJ, Illingworth RS, Webb S, Kerr AR, James KD, et al. (2010) Neuronal MeCP2 is expressed at near histone-octamer levels and globally alters the chromatin state. Mol Cell 37: 457–468.
- 17. Gianakopoulos PJ, Zhang Y, Pencea N, Orlic-Milacic M, Mittal K, et al. (2012) Mutations in MECP2 exon 1 in classical Rett patients disrupt MECP2_e1 transcription, but not transcription of MECP2_e2. Am J Med Genet B Neuropsychiatr Genet 159B: 210–216.
- 18. Le Meur N, Holder-Espinasse M, Jaillard S, Goldenberg A, Joriot S, et al. (2010) MEF2C haploinsufficiency caused by either microdeletion of the 5q14.3 region or mutation is responsible for severe mental retardation with stereotypic movements, epilepsy and/or cerebral malformations. J Med Genet 47: 22–29.
- 19. Zweier M, Gregor A, Zweier C, Engels H, Sticht H, et al. (2010) Mutations in MEF2C from the 5q14.3q15 microdeletion syndrome region are a frequent cause of severe mental retardation and diminish MECP2 and CDKL5 expression. Hum Mutat 31: 722–733.
- 20. Liu Y, Halloran MC (2005) Central and peripheral axon branches from one neuron are guided differentially by Semaphorin3D and transient axonal glycoprotein-1. J Neurosci 25: 10556–10563.
- 21. Wolman MA, Liu Y, Tawarayama H, Shoji W, Halloran MC (2004) Repulsion and attraction of axons by semaphorin3D are mediated by different neuropilins in vivo. J Neurosci 24: 8428–8435.
- 22. Grundemar L, Hakanson R (1993) Multiple neuropeptide Y receptors are involved in cardiovascular regulation. Peripheral and central mechanisms. Gen Pharmacol 24: 785–796.
- 23. Stanley BG, Anderson KC, Grayson MH, Leibowitz SF (1989) Repeated hypothalamic stimulation with neuropeptide Y increases daily carbohydrate and fat intake and body weight gain in female rats. Physiol Behav 46: 173–177.
- 24. Herzog H, Hort YJ, Ball HJ, Hayes G, Shine J, et al. (1992) Cloned human neuropeptide Y receptor couples to two different second messenger systems. Proc Natl Acad Sci U S A 89: 5794–5798.
- 25. Bissonnette JM, Knopp SJ, Maylie J, Thong T (2007) Autonomic cardiovascular control in methyl-CpG-binding protein 2 (Mecp2) deficient mice. Auton Neurosci 136: 82–89.
- 26. Lin CC, Huang CC, Lin KH, Cheng KH, Yang DM, et al. (2007) Visualization of Rab3A dissociation during exocytosis: a study by total internal reflection microscopy. J Cell Physiol 211: 316–326.
- 27. Smith R, Klein P, Koc-Schmitz Y, Waldvogel HJ, Faull RL, et al. (2007) Loss of SNAP-25 and rabphilin 3a in sensory-motor cortex in Huntington’s disease. J Neurochem 103: 115–123.
- 28. Lai CS, Fisher SE, Hurst JA, Vargha-Khadem F, Monaco AP (2001) A forkhead-domain gene is mutated in a severe speech and language disorder. Nature 413: 519–523.
- 29. Neul JL, Kaufmann WE, Glaze DG, Christodoulou J, Clarke AJ, et al. (2010) Rett syndrome: revised diagnostic criteria and nomenclature. Ann Neurol 68: 944–950.
- 30. Vargha-Khadem F, Watkins KE, Price CJ, Ashburner J, Alcock KJ, et al. (1998) Neural basis of an inherited speech and language disorder. Proc Natl Acad Sci U S A 95: 12695–12700.
- 31. Yakabe S, Soejima H, Yatsuki H, Tominaga H, Zhao W, et al. (2008) MeCP2 knockdown reveals DNA methylation-independent gene repression of target genes in living cells and a bias in the cellular location of target gene products. Genes Genet Syst 83: 199–208.
- 32. Nectoux J, Fichou Y, Rosas-Vargas H, Cagnard N, Bahi-Buisson N, et al. (2010) Cell cloning-based transcriptome analysis in Rett patients: relevance to the pathogenesis of Rett syndrome of new human MeCP2 target genes. J Cell Mol Med 14: 1962–1974.
- 33. Chahrour M, Jung SY, Shaw C, Zhou X, Wong ST, et al. (2008) MeCP2, a key contributor to neurological disease, activates and represses transcription. Science 320: 1224–1229.
- 34. Roll P, Rudolf G, Pereira S, Royer B, Scheffer IE, et al. (2006) SRPX2 mutations in disorders of language cortex and cognition. Hum Mol Genet 15: 1195–1207.
- 35. Roll P, Vernes SC, Bruneau N, Cillario J, Ponsole-Lenfant M, et al. (2010) Molecular networks implicated in speech-related disorders: FOXP2 regulates the SRPX2/uPAR complex. Hum Mol Genet 19: 4848–4860.
- 36. Vincent JB, Horike SI, Choufani S, Paterson AD, Roberts W, et al. (2006) An inversion inv(4)(p12-p15.3) in autistic siblings implicates the 4p GABA receptor gene cluster. J Med Genet 43: 429–434.
- 37. Ma DQ, Whitehead PL, Menold MM, Martin ER, Ashley-Koch AE, et al. (2005) Identification of significant association and gene-gene interaction of GABA receptor subunit genes in autism. Am J Hum Genet 77: 377–388.
- 38. Kakinuma H, Ozaki M, Sato H, Takahashi H (2008) Variation in GABA-A subunit gene copy number in an autistic patient with mosaic 4 p duplication (p12p16). Am J Med Genet B Neuropsychiatr Genet 147B: 973–975.
- 39. Fatemi SH, Reutiman TJ, Folsom TD, Thuras PD (2009) GABA(A) receptor downregulation in brains of subjects with autism. J Autism Dev Disord 39: 223–230.
- 40. Chao HT, Chen H, Samaco RC, Xue M, Chahrour M, et al. (2010) Dysfunction in GABA signalling mediates autism-like stereotypies and Rett syndrome phenotypes. Nature 468: 263–269.
- 41. Ariani F, Hayek G, Rondinella D, Artuso R, Mencarelli MA, et al. (2008) FOXG1 is responsible for the congenital variant of Rett syndrome. Am J Hum Genet 83: 89–93.
- 42. Mencarelli MA, Spanhol-Rosseto A, Artuso R, Rondinella D, De Filippis R, et al. (2010) Novel FOXG1 mutations associated with the congenital variant of Rett syndrome. J Med Genet 47: 49–53.
- 43. Philippe C, Amsallem D, Francannet C, Lambert L, Saunier A, et al. (2010) Phenotypic variability in Rett syndrome associated with FOXG1 mutations in females. J Med Genet 47: 59–65.
- 44. Kortum F, Das S, Flindt M, Morris-Rosendahl DJ, Stefanova I, et al. (2011) The core FOXG1 syndrome phenotype consists of postnatal microcephaly, severe mental retardation, absent language, dyskinesia, and corpus callosum hypogenesis. J Med Genet 48: 396–406.
- 45. Dastidar SG, Bardai FH, Ma C, Price V, Rawat V, et al. (2012) Isoform-specific toxicity of Mecp2 in postmitotic neurons: suppression of neurotoxicity by FoxG1. J Neurosci 32: 2846–2855.
- 46. McGill BE, Bundle SF, Yaylaoglu MB, Carson JP, Thaller C, et al. (2006) Enhanced anxiety and stress-induced corticosterone release are associated with increased Crh expression in a mouse model of Rett syndrome. Proc Natl Acad Sci U S A 103: 18267–18272.
- 47. Noor A, Whibley A, Marshall CR, Gianakopoulos PJ, Piton A, et al. (2010) Disruption at the PTCHD1 Locus on Xp22.11 in Autism spectrum disorder and intellectual disability. Sci Transl Med 2: 49ra68.
- 48. Filges I, Rothlisberger B, Blattner A, Boesch N, Demougin P, et al. (2011) Deletion in Xp22.11: PTCHD1 is a candidate gene for X-linked intellectual disability with or without autism. Clin Genet 79: 79–85.
- 49. Endele S, Rosenberger G, Geider K, Popp B, Tamer C, et al. (2010) Mutations in GRIN2A and GRIN2B encoding regulatory subunits of NMDA receptors cause variable neurodevelopmental phenotypes. Nat Genet 42: 1021–1026.
- 50. Pinto D, Pagnamenta AT, Klei L, Anney R, Merico D, et al. (2010) Functional impact of global rare copy number variation in autism spectrum disorders. Nature 466: 368–372.
- 51. Griggs BL, Ladd S, Saul RA, DuPont BR, Srivastava AK (2008) Dedicator of cytokinesis 8 is disrupted in two patients with mental retardation and developmental disabilities. Genomics 91: 195–202.
- 52. Zhang Q, Davis JC, Lamborn IT, Freeman AF, Jing H, et al. (2009) Combined immunodeficiency associated with DOCK8 mutations. N Engl J Med 361: 2046–2055.
- 53. Bi W, Yan J, Stankiewicz P, Park SS, Walz K, et al. (2002) Genes in a refined Smith-Magenis syndrome critical deletion interval on chromosome 17p11.2 and the syntenic region of the mouse. Genome Res 12: 713–728.
- 54. Zoghbi HY (2003) Postnatal neurodevelopmental disorders: meeting at the synapse? Science 302: 826–830.