Reduced expression of MECP2 affects cell commitment and maintenance in neurons by triggering senescence: new perspective for Rett syndrome

MECP2 mutation affects MSC biology

An enrolled RTT patient presented the clinical manifestations of classic Rett syndrome. She carried a de novo mutation (R270X) in the MECP2 gene (Supplemental File 1). We obtained MSCs from this patient and from two healthy controls as described in Materials and Methods. Following an initial expansion for 7–10 d, the MSCs reached confluence and were grown for an additional 10 d. At the end of this period, the MSCs were collected for biological assays or were further cultivated for neural differentiation.

Proliferating MSCs from the RTT patient and controls were determined by fluorescence-activated cell sorting (FACS) analysis. In several experiments, we observed a significant reduction (p < 0.05) in S-phase cells and an increase in G₁/G₀ (p < 0.05) and G₂/M cells in the RTT sample compared with controls (Figure 1A).

The MSCs were obtained from a RTT patient who had a mosaic expression of mutated and wild-type MECP2 due to X chromosome inactivation. Because the patient had a mutation that produced C-terminal–truncated proteins, we used a C-terminal–specific anti-MeCP2 antibody, combined with biological assays, to specifically analyze the MECP2-mutated cells (MECP2 negative by immunofluorescence). The analysis of somatic mosaicism in MSCs from the RTT patient evidenced 61.5% (±5.0%) of MECP2-positive cells (wild-type cells) and 38.5% (±5.9%) of MECP2-negative cells (mutated cells).

The bromodeoxyuridine (BrdU) assay showed that reduction in proliferation, as detected with FACS analysis, affected only cells with inactivated MECP2, as evidenced by the double-staining assay (Figure 1B).

The annexin assays showed a significant reduction (p < 0.05) of apoptotic cells in MECP2 cultures obtained from the RTT sample as compared with those from controls (Figure 1C). Moreover, signs of senescence were detected more in cells from the RTT patient than in those from healthy donors, as found by the in situ acid β-galactosidase assay (Figure 1D). In addition, in this case, the impairment of cell function occurred in cells with the mutated form of MECP2 and was statistically significant (p < 0.05).

Senescent cells show several changes in gene expression, including changes in known cell-cycle inhibitors or activators. Two cell-cycle inhibitors that are often expressed by senescent cells are the cyclin-dependent kinase inhibitors (CDKIs) p21^Cip1 and p16^INK4a (Campisi and d'Adda di Fagagna, 2007). These CDKIs are components of tumor-suppressor pathways that are governed by the p53 and retinoblastoma family proteins, respectively (Kapic et al., 2006; Campisi and d'Adda di Fagagna, 2007; Helmbold et al., 2009).

In proliferating cells from the RTT patient, we detected a significant up-regulation (p < 0.05) of the p53 protein level as compared with the control sample (Figure 1E). On the contrary, however, we did not find significant modifications of p21^Cip1, pRb, and p16^INK4a protein expression (Figure 1E). Note that Western blots were carried out on the whole population of MSCs from the RTT patient containing mutated and wild-type MECP2 cells; hence, subtle changes in gene expression may might been overlooked.

Overall, these data confirm our previous results obtained in patients showing the clinical manifestations of classic and “fruste” Rett syndrome (Squillaro et al., 2008a, 2008b, 2008c).

In vitro neural differentiation of MSCs from patient with classic RTT syndrome

A complete study of neural cell commitment and differentiation should be carried out with neural stem cells. Obviously, it is not possible to obtain neural stem cells from living RTT patients. We decided to use an alternative experimental model to study in vitro neurogenesis with biological samples obtained from RTT patients. In our opinion, MSCs could be a good alternative. Indeed, several groups demonstrated that MSCs can be induced to differentiate into neurons (Pittenger et al., 1999; Woodbury et al., 2000; Jori et al., 2004, 2005a, 2005b).

Neural differentiation of MSCs occurred mainly in vitro, and their in vivo contribution to neurogenesis is negligible. Nevertheless, they might represent a tool with which to study the early phases of neural commitment and differentiation.

We investigated the process of in vitro neural differentiation by defining the expression patterns of genes representing “generic” neural differentiation and those associated with specific neurotransmission. To achieve this goal, we induced neural differentiation in MSCs using an advanced neural differentiation medium (see Materials and Methods) that contains a proprietary formulation of growth factors. Morphological and molecular analyses of MSCs from the RTT patient and controls that displayed a neuron-like phenotype were performed.

We identified differentiating neurons by anti-NeuN immunostaining. In the proliferation medium, cells from healthy donors showed a lack of NeuN expression, but after 4 d of incubation in the neural induction medium, 41.7% (± 6.1%) of cells became positive for NeuN. In the RTT patient, this percentage did not change significantly in cells expressing wild-type MECP2. Conversely, the percentage of NeuN-positive cells was lower (p < 0.05) in the population bearing mutated MECP2 (Figure 2B and magnification of Figure 2B in Supplemental File 5). These data suggest that the mutation of MeCP2 results in a cell autonomous effect, in agreement with other reports showing that MeCP2 functions are largely cell autonomous in neuronal maturation (Kishi and Macklis, 2010).

As expected, differentiating cells were mainly in the G₁/G₀ cell cycle phase, and no significant differences were detected between neuron-like cultures from controls and those from the RTT patient (Figure 2A). FACS analysis was carried out on the whole population of MSCs from the RTT patient, and this would not us allow to detect minimal differences in cell cycle profile between wild-type and mutated-MECP2 subpopulations.

In progenitor cells (NeuN negative) from RTT patients, the percentage of apoptosis did not change significantly in cells expressing wild-type MECP2 as compared with the control (Figure 2C and magnification of Figure 2C in Supplemental File 6). In contrast, the percentage of apoptosis was lower (p < 0.05) in the population with mutated MECP2 (Figure 2C).

In neuron-like (NeuN-positive) cells from RTT patients, irrespective of MECP2 status, the apoptosis level was similar to that observed in the control sample (Figure 2C).

Differentiating MSCs from mutated MECP2s showed an increased percentage (p < 0.05) of senescent cells as compared with those from controls (Figure 2D). Increased senescence was observed both in the undifferentiated (NeuN-negative) and the differentiated (neuron-like) cell populations (Figure 2D).

In neuron-like cells from RTT patients, we detected a highly significant up-regulation (p < 0.01) of the pRb protein level as compared with the control sample (Figure 1E). In contrast, we did not find a significant modification of p21^Cip1, p53, and p16^INK4a protein expression (Figure 1E).

Reverse transcription (RT)-PCR was used to analyze the expression of generic neuron differentiation markers such as neuron-specific enolase (NSE) and microtubule-associated protein 2 (MAP2; Pittenger et al., 1999; Woodbury et al., 2000; Jori et al., 2004, 2005a, 2005b). MSC differentiation in neuron-like cells was associated with increased expression levels of both markers. MSCs differentiated from the RTT patient displayed a significant reduction (p < 0.05) of the expression of NSE and MAP2 compared with those differentiated from controls (Figure 2E). The observed decrease should be considered to be highly significant because RNA analysis was carried out on the whole population of MSCs from the RTT patient containing wild-type MECP2 cells, which differentiated normally.

Differentiated MSCs acquired a generic neuronal phenotype; we did not detect the expression of markers specific for cholinergic (acetylcholinesterase [ACHE]), dopaminergic (dopamine transporter [DAT]), noradrenergic (dopamine β-hydroxylase [DBH]), or serotonergic (tryptophan hydroxylase [TH]) neurons.

The senescence and reduced expression levels of neural markers in proliferating and differentiating MSCs from the RTT patient suggest that reduced activity of MECP2 protein, as observed in RTT patients, might impair neural commitment and differentiation.

In vitro neural differentiation of MSCs from patient with “fruste” RTT syndrome

To confirm and extend our data, we analyzed neural differentiation of MSCs obtained from a patient with the clinical manifestations of “fruste” Rett syndrome. She carried a de novo mutation (c. 1164–1207del) in the MECP2 gene (Supplemental File 1).

In addition, in this patient, neural differentiation of MSCs was impaired as compared with the control. The percentage of neuron-like cells (NeuN-positive) was lower (p < 0.05) in the population bearing the mutated MECP2 (Supplemental File 2A).

This result was in agreement with RT-PCR experiments. MSCs differentiated from the RTT patient displayed a highly significant reduction (p < 0.01) of the expression of NSE as compared with those differentiated from controls (Supplemental File 2B).

MSCs with a mutated MECP2 showed an increased percentage of senescent cells (p < 0.05) as compared with those from controls (Supplemental File 2C). Increased senescence was observed both in the undifferentiated (NeuN-negative) and the differentiated (neuron-like) cell populations (Supplemental File 2C).

MECP2 silencing

The paucity of biological samples available from RTT patients permits limited analyses. To extend our findings, we studied neural differentiation in a human SK-N-BE(2)-C neuroblastoma cell line in which MECP2 was silenced. This model is convenient for studies investigating neural progenitor cells. The neuroblastoma cell line was derived from human malignant neural crest cells (de Bernardi et al., 1987) and shows an intermediate (I) phenotype that has been demonstrated to represent a multipotent embryonic stem cell of the peripheral nervous system. Therefore neuroblastoma cells can be induced to differentiate toward the committed neuronal N-type and glial S-type cells (Ciccarone et al., 1989; Ross et al., 1995; Jori et al., 2001). To assess whether MECP2 inactivation affected cell fate decisions of neural progenitors and/or neuronal differentiation and maintenance, to study different cell populations, we used the RheoSwitch Mammalian Inducible Expression System to silence MECP2 in neuroblastoma cells before and after the induction of neural differentiation via retinoic acid treatment. In the first case, we evaluated the effects on multipotent embryonic precursor cells, and in the latter case, we performed the analysis using committed and neuronal-like cells.

The MECP2 gene is composed of four exons. Two alternatively spliced transcripts have been characterized: MECP2A (also called MECP2_E2) and MECP2B (also called MECP2_E1). The E1 isoform is composed of exons 1, 3, and 4, whereas the E2 isoform is composed of exons 1–4 (Singh et al., 2008). To design hairpin small interfering RNAs (siRNAs), we selected several target regions that were present in both isoforms. All experiments were carried out with pNEBR-X1-MECP2-1706, which targets nucleotides (nt) 1706–1724 of MECP2 mRNA (GenBank accession number NM_004992.3). In a previous study, we demonstrated the effectiveness of such hairpin siRNA, which we transduced in cell cultures by adenoviral vectors (Squillaro et al., 2010). Confirmatory experiments were carried out with pNEBR-X1-MECP2-4398 and pNEBR-X1-MECP2-1229, which target nt 4398–4416 and 1229–1247, respectively (see later discussion).

Proliferating neuroblastoma cells were cotransfected with pNEBR-R1 (regulator plasmid) and either pNEBR-X1-MECP2-1706 or pNEBR-X1-CTRL carrying short hairpin RNAs (shRNAs). The R1 plasmids contained the hygromycin B gene, whereas the X1 plasmid had the G418 gene. These antibiotic resistance genes were used for generation of stable cell lines.

On addition of the synthetic RSL1 ligand to the culture medium, the RheoReceptor and the RheoActivator proteins stably dimerized and bound to the response element of the pNEBR-R1 plasmids. This process, in turn, activated shRNA transcription and silenced MECP2 expression, as detected by RT-PCR using a primer pair that identified both isoforms (Figure 3A). The specific down-regulation of each isoform was evaluated with a semiquantitative RT-PCR, since primer pairs that distinguish the two isoforms were not suitable for real-time RT-PCR (Figure 3A). The silencing of MECP2 was observed in almost all cells in the culture because the antibiotic treatment selected only cells stably expressing the hairpin siRNAs (unpublished data).

Next, in cells with silenced MECP2 and in controls, we induced neural differentiation by incubating cultures for 7 d with retinoic acid. Of note, the percentage of neurons was greatly reduced (p < 0.05) in cells that demonstrated decreased MECP2 expression compared with control cultures, as detected by NeuN immunostaining (18.6 vs. 37.9%; Figure 3B). This reduction was associated with a decreased number of BrdU-positive cells (Figure 3C). Moreover, MECP2 down-regulation augmented (p < 0.05) the percentage of senescent cells both in uncommitted/undifferentiated (NeuN-negative) and neuronal-like (NeuN-positive) populations (Figure 3D).

Senescence was associated with a decrease in the percentage (p < 0.05) of apoptotic cells in uncommitted/undifferentiated cells (Figure 3E). This finding suggests that the reduced percentage of neuronal-like cells observed in MECP2-silenced cultures may be attributed to the senescence of neural precursors rather than to cell death either in uncommitted cells or in neuronal-like cells (Figure 3E).

In a second group of experiments, we induced neural differentiation in neuroblastoma cells that were stably cotransfected with pNEBR-R1 (regulator plasmid) and either pNEBR-X1-MECP2-1706 or pNEBR-X1-CTRL carrying shRNAs. Three days after retinoic acid treatment, RSL1 ligand was added to the cultures, and the cultures were incubated for an additional 4 d in differentiation medium.

At the end of this period, we assessed MECP2 mRNA levels and found that they were decreased (p < 0.01) in cells treated with pNEBR-X1-MECP2-1706 compared with those treated with the control plasmid, as expected (Figure 4A). Partial silencing of MECP2 in cells following the induction of neural differentiation did not affect the percentage of NeuN-positive cells (Figure 4B). These results indicate that MECP2 silencing after cell commitment does not modify the number of neurons.

MECP2 silencing enhanced the cell proliferation arrest observed in neural differentiation medium. MECP2 down-regulation induced an increase (p < 0.05) in senescent cells (Figure 4C). Of great interest, the percentage of senescent differentiated neurons more than doubled (38 vs. 16%) in cells that carried partially silenced MECP2 compared with control cells (Figure 4C). The down-regulation of MECP2 following neural induction also increased (p < 0.05) senescence in uncommitted/undifferentiated cells (Figure 4C). These data are consistent with results obtained when MECP2 was silenced before the induction of neural differentiation.

Senescence was associated with a decrease (p < 0.05) in the percentage of apoptotic cells in uncommitted/undifferentiated cells and in neuronal-like cells (Figure 4D).

The biological effects we detected in neuroblastoma cells with a silenced MECP2 through treatment with hairpin siRNA MECP2-1706 might be not specifically related to MECP2 down-regulation.

To rule out the off-target effects of hairpin siRNA-1706, we also transfected neuroblastoma cells with pNEBR-X1-MECP2-4398 and pNEBR-X1-MECP2-1229. Both hairpin siRNAs induced silencing of the MECP2 gene (Supplemental File 3A).

We selected pNEBR-X1-MECP2-4398 for further experiments. In cells with a silenced MECP2 through pNEBR-X1-MECP2-4398 and in controls, we induced neural differentiation by incubating cultures for 7 d with retinoic acid. It is worth noting that as detected with hairpin siRNA MECP2-1706, the percentage of neurons was greatly reduced (p < 0.05) in cells with decreased MECP2 expression as compared with control cultures (Supplemental File 3B).

In addition, siRNA-4398 augmented the percentage of senescent cells both in uncommitted/undifferentiated (p < 0.05) and neuronal-like cell populations and decreased the percentage of apoptotic cells in uncommitted/undifferentiated cells (Supplemental File 3C).

Molecular analysis of the identified patient

Blood samples were collected after informed consent was obtained. DNA was extracted from peripheral blood using a QIAamp DNA Blood Kit (Qiagen, Milan, Italy). DNA samples were screened for mutations in the four exons coding for MECP2 using transgenomic WAVE denaturing high-performance liquid chromatography (DHPLC). Analysis of the MECP2 gene for deletions/duplications was performed as previously described (Ariani et al., 2004) and as indicated in Supplemental File 1. PCR products that provided abnormal DHPLC profiles were sequenced on both strands using PCR primers with fluorescent dye terminators on an ABI PRISM 310 genetic analyzer (PE Applied Biosystems, Foster City, CA).

MSC cultures

Bone marrow was collected from a girl with RTT syndrome and two healthy children after obtaining informed consent. The cells were separated on a Ficoll density gradient (GE Healthcare, Milan, Italy), and MSC cultures were obtained as described previously (Squillaro et al., 2008c, 2010).

To induce neural differentiation, cells were plated on tissue culture dishes at ∼2500 cells/cm² in proliferating medium. When the cells reached 30% confluence, the medium was discarded, the cell plates were washed with phosphate-buffered saline (PBS), and complete neural differentiation medium was added (Thermo Scientific, Waltham, MA). The formation of neuron-like cells was observed within 24 h and peaked at 72 h. To maintain cells in a differentiated state, we added additional neural differentiation medium every 48 h.

Neuroblastoma cell cultures

SK-N-BE(2)-C neuroblastoma cells were grown in monolayers and maintained at 37°C under 5% CO₂ in RPMI 1640 containing 10% fetal bovine serum, 2 mM l-glutamine, 50 U/ml penicillin, and 100 μg/ml streptomycin. To induce neural differentiation, 10⁻⁶ M all-trans retinoic acid (Sigma Aldrich, St. Louis, MO) was added to the culture medium for 7 d.

Silencing

Hairpin siRNAs targeted against human MECP2 mRNA (GenBank accession number NM_004992.3) were selected using the design algorithm developed by Cenix Bioscience (Dresden, Germany) on the basis of rules and criteria as already described (Pittenger et al., 1999; Reynolds et al., 2004).

We selected three target regions of MECP2 mRNA to design hairpin siRNAs:

5′-GCAAAGCAAACCAACAAGA-3′ (nt 1706–1724)

5′-GGTTGTCACTGAGAAGATG-3′ (nt 4398–4416)

5′-GGACTGAAGACCTGTAAGA-3′ (nt 1229–1247)

The siRNA negative control was as described (Harborth et al., 2001) and targets the firefly (Photinus pyralis) luciferase gene (GenBank accession number X65324). The negative control siRNA exhibits no target mRNA in the human transcriptome.

For each hairpin siRNA, we designed two complementary oligonucleotides (55-mers); each included the 19-mer siRNA in sense orientation, followed by the loop sequence 5′-TTCAAGAGA-3′ and by the 19-mer siRNA in antisense orientation. The 5′ ends of the two oligonucleotides had XhoI and BamHI restriction site overhangs to facilitate the efficient directional cloning into the pNEBR-X1 plasmid (see later discussion).

The complementary hairpin siRNA template oligonucleotides were synthesized (MWG Biotech, Ebersberg, Germany) and then were annealed to form a double-stranded DNA. The annealed siRNA template inserts were ligated into the pNEBR-X1 plasmid according the manufacturer's protocol and routine procedures for DNA cloning (Sambrook and Russell, 2001).

We cloned and produced four recombinant plasmids: pNEBR-X1-MECP2-1706; pNEBR-X1-MECP2-4398; pNEBR-X1-MECP2-1229, which targeted MECP2; and pNEBR-X1-CTRL, a control shRNA, which targeted the firefly luciferase gene.

Neuroblastoma cell cultures were cotransfected with pNEBR-R1 (regulator plasmid) and either pNEBR-X1-MECP2 plasmids or pNEBR-X1-CTRL.

The neomycin resistance gene in pNEBR1-R1 and the hygromycin resistance gene in the pNEBR-X1 plasmid were used to select stable cells growing in media containing 500 μg/ml G418 and 50 μg/ml hygromycin. Stable cell lines expressing the RheoReceptor and the RheoActivator proteins via pNEBR1-R1 were generated. In these cell lines, shRNA transcription was turned off under basal conditions. By adding the synthetic RSL1 ligand to culture medium, the RheoReceptor and the RheoActivator proteins were dimerized stably and bound to the response element of pNEBR-R1 plasmids. This in turn activated shRNA transcription.

Following the silencing of MECP2, the rescue of its expression was obtained via the removal of RSL1 ligand from the culture medium after cultivating the cells for 48 h.

Cell cycle analysis

For each assay, 3 × 10⁵ cells were collected and resuspended in a hypotonic buffer containing propidium iodide. Cells were incubated in the dark and then analyzed. Samples were acquired using a FACSCalibur flow cytometer with CellQuest software (BD Biosciences, Franklin Lakes, NJ) and analyzed according to standard procedures with CellQuest and ModFitLT software (BD Biosciences ).

Detection of MECP2 by immunocytochemistry

MECP2 was detected using the anti-MECP2 (clone Mec-168) primary antibody (Abcam, Cambridge, United Kingdom) according to the manufacturer's protocol. Cells were analyzed using either fluorescence or a light microscope. In the first case, the cells were incubated with goat anti-mouse secondary antibodies and conjugated to Texas red (Jackson ImmunoResearch Laboratories, West Grove, PA), and the nuclei were counterstained with Hoechst 33342. Alternatively, the cells were incubated with goat anti-mouse secondary antibodies, conjugated to peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA), and then treated with the DAB substrate (Roche, Mannheim, Germany) or conjugated to alkaline phosphatase (Santa Cruz Biotechnology) and then treated with the Vector Red Alkaline Phosphatase Substrate Kit (Vector Laboratories, Burlingame, CA).

The percentage of MECP2-positive cells was calculated by counting at least 500 cells in different microscopic fields (see also Supplemental File 11).

BrdU assay

For BrdU immunostaining, the cells were grown on glass coverslips and incubated for 10 h with 10 μM BrdU (Sigma-Aldrich). Briefly, the cells were rinsed with PBS, fixed with 100% methanol at 4°C for 10 min, air dried, and then incubated with 2 N HCl for 60 min at 37°C. HCl was neutralized with 0.1 M borate buffer (pH 8.5). Following additional PBS washes, the slides were incubated with an anti-BrdU rabbit polyclonal antibody (1:200; Bioss, Woburn, MA). After 60 min at room temperature, the cells were washed with PBS and incubated with goat anti-rabbit secondary antibodies conjugated to peroxidase (Santa Cruz Biotechnology) for 60 min at room temperature. Finally, after additional washes in PBS, the slides were treated with DAB substrate (see also Supplemental File 11).

Detection of apoptotic cells

Apoptotic cells were detected using fluorescein-conjugated annexin V (Roche, Milan, Italy) according to the manufacturer's instructions. Apoptotic cells were observed using a fluorescence microscope (Leica Italia, Milan, Italy). In each experiment, at least 1000 cells were counted in different fields to calculate the percentage of dead cells in each culture (see also Supplemental File 11).

Senescence-associated β-galactosidase assay

The cells were fixed with a solution of 2% formaldehyde and 0.2% glutaraldehyde. The cells were then washed with PBS and incubated at 37°C for at least 2 h with a staining solution according to the manufacturer's protocol (Roche, Italy). The percentage of senescent cells was calculated based on the number of β-galactosidase–positive cells among at least 500 cells in different microscopic fields; see also Supplemental File 11).

Neuronal nuclei detection

NeuN was detected by immunocytochemistry using the anti-NeuN primary antibody (Millipore, Milan, Italy) according to the manufacturer's protocol. Cells were analyzed using either a fluorescence or a light microscope. In the first case, the cells were incubated with goat anti-rabbit secondary antibodies conjugated to tetramethylrhodamine isothiocyanate or aminomethylcoumarin acetate (Jackson Immuno­Research Laboratories), and the nuclei were counterstained with Hoechst 33342. Alternatively, the cells were incubated with goat anti-rabbit secondary antibodies conjugated to peroxidase (Santa Cruz Biotechnology) and then treated with the appropriate substrate.

The percentage of NeuN-positive cells was calculated by counting at least 500 cells in different microscopic fields (see also Supplemental File 11).

RNA extraction, RT-PCR, and real-time PCR

Total RNA was extracted from cell cultures using Omnizol (Euroclone, Milan, Italy) according to the manufacturer's protocol. mRNA levels were measured by RT-PCR amplification as previously reported (Galderisi et al., 1999). The primers used are listed in Supplemental File 4.

Real-time PCR assays were performed using an Opticon-4 machine (Bio-Rad, Hercules, CA). The reactions were performed according to the manufacturer's instructions using a SYBR green PCR Master Mix.

Western blotting

Cells were lysed in a buffer containing 0.1% Triton for 30 min at 4°C. The lysates were then centrifuged for 10 min at 10,000 × g at 4°C. After centrifugation, 10–40 μg of each sample was loaded electrophoresed in a polyacrylamide gel and electroblotted onto a nitrocellulose membrane. All the primary antibodies were used according to the manufacturers' instructions. Immunoreactive signals were detected with a horseradish peroxidase–conjugated secondary antibody (Santa Cruz Biotechnology) and reacted with ECL Plus reagent (GE Healthcare, Italy).

Statistical analysis

Statistical significance was evaluated by analysis of variance followed by Student's t test or Bonferroni test.