SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • High throughput screening;
  • Repressor element-1 silencing transcription factor;
  • Neuron restrictive silencer factor;
  • Neural stem cells

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Decreased expression of neuronal genes such as brain-derived neurotrophic factor (BDNF) is associated with several neurological disorders. One molecular mechanism associated with Huntington disease (HD) is a discrete increase in the nuclear activity of the transcriptional repressor REST/NRSF binding to repressor element-1 (RE1) sequences. High-throughput screening of a library of 6,984 compounds with luciferase-assay measuring REST activity in neural derivatives of human embryonic stem cells led to identify two benzoimidazole-5-carboxamide derivatives that inhibited REST silencing in a RE1-dependent manner. The most potent compound, X5050, targeted REST degradation, but neither REST expression, RNA splicing nor binding to RE1 sequence. Differential transcriptomic analysis revealed the upregulation of neuronal genes targeted by REST in wild-type neural cells treated with X5050. This activity was confirmed in neural cells produced from human induced pluripotent stem cells derived from a HD patient. Acute intraventricular delivery of X5050 increased the expressions of BDNF and several other REST-regulated genes in the prefrontal cortex of mice with quinolinate-induced striatal lesions. This study demonstrates that the use of pluripotent stem cell derivatives can represent a crucial step toward the identification of pharmacological compounds with therapeutic potential in neurological affections involving decreased expression of neuronal genes associated to increased REST activity, such as Huntington disease. Stem Cells 2013;31:1816-1828


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Several pathological states affecting the central nervous system are associated with perturbations of the expression of neuronal genes before actual neural cell loss or transformation occurs. This, for instance, is the case for Huntington disease (HD) [1-3]. One identified mechanism that leads to such phenomena is an increased activity of the repressor element-1 silencing transcription factor (REST), also known as neuron restrictive silencer factor (NRSF). Accordingly, a therapeutic avenue for those pathologies may be to interfere pharmacologically with REST inhibition of its target genes (RE1-genes) [4, 5]. Indeed, REST inhibition achieved in vitro via the overexpression of a dominant-negative form of REST (D/N-REST) lifts REST-mediated repression or silencing of several hundreds of neuron-specific genes, among which, notably, the brain-derived neurotrophic factor (BDNF) [6-8]. While D/N-REST-mediated inhibition has limited clinical relevance, pharmacological intervention that would similarly increase neuroprotective gene expression through REST inhibition would open promising therapeutic perspectives. This study has accordingly been undertaken in a search for compounds that could meet that goal.

REST is a Krüppel-type zinc finger transcription factor that exerts its repressive cis-activity in the nucleus upon binding to a 21-nucleotide DNA sequence called repressor element-1 (RE1). REST is the key component of a nuclear complex consisting of associated core factors such as SIN3A, SIN3B, and RCOR1 (a.k.a Co-Rest), and epigenetic regulators such as histone-deacetylases (HDACs), histone-methyltransferase (EHMT2), and histone-demethylase (KDM1A) that mediate chromatin compaction [9]. The composition of the nuclear complex and, as a consequence, the exact role of REST is dependent on the cell type and developmental stage [9, 10]. REST has been initially described as an inhibitor of neuronal genes in non-neuronal cells [11, 12], and its activity has more recently extended to several aspects of development, in particular pluripotent and neural stem cell (NSC) maintenance and commitment [13-16].

Levels of REST transcript and protein decrease during transition from mouse embryonic stem cells (ESCs) to neurons [17, 18]. REST degradation leads to neural commitment and further differentiation of neural cells via the expression of critical proneural and neuronal activators containing RE1 site(s) in their promoters [13, 17, 19, 20]. Different levels of REST may also control neuronal and glial lineage diversification [21-23] and the maintenance of the quiescent state characteristic of postmitotic neurons [24-26].

This study has been undertaken within the framework of a program seeking therapeutics for HD, using chemical compounds that may decrease the activity of REST in the human brain. Human pluripotent stem cells (hPSC: embryonic: hESC or induced: hiPSC) were used as they give access to an unlimited supply of human neural cells [27, 28]. Harnessing this biological resource, we developed a cell-based reporter system to monitor REST activity in hPSC neural derivatives and carried out high throughput screening (HTS) that revealed hit compounds. The mechanism of action of the most potent of these hit compounds was subsequently determined. The activity of this compound was measured in vitro in human NSC carrying the Huntington mutation and in vivo in a phenotypic model of HD-like striatal degeneration. This work opens a potential path for the development of therapeutic agents against neurological diseases that involve loss of expression of neural genes controlled by REST.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Cell Culture

SA-01 (Cellartis) and RC9 (RoslinCells) hESC repeats lines (WT, XY) and HD1-iPS4 [29] hiPSC line (HD 72 CAG, XY) were differentiated into NSCs as described previously [27]. NSCs were grown on polyornithine/laminin-coated tissue culture plates in NSC medium containing Neurobasal, Dulbecco's modified Eagle's medium (DMEM)/F-12, N-2, and B-27 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 0.55 mM 2-mercaptoethanol, 10 ng/mL epidermal growth factor (EGF) (R&D Systems, Minneapolis, MN, http://www.rndsystems.com), and 10 ng/mL fibroblast growth factor 2 (FGF2) (Invitrogen). NSCs were passaged every 5–7 days up to 20 passages. To obtain neurons, confluent NSCs were grown during 7 days in NSC medium without BDNF, EGF, and FGF2. Cells were then plated at 400,000 cells per centimeter square in NSC medium supplemented with 10 μM DAPT (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Cells were harvested after 14 days. Human embryonic kidney 293 (HEK) cells were passaged every 3–4 days and grown in DMEM, high glucose (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen).

Vectors

A 1,200 nucleotide (nt) fragment of the elongation factor 1 (EF1α) promoter containing the transcription start site (TSS) was blunted and cloned into HindIII-digested pGL4.82 [hRluc/Puro] luciferase reporter vector (Promega, Madison, WI, http://www.promega.com), upstream of renilla luciferase. Alternatively, a 540-nt fragment of phosphoglycerate kinase (PGK) containing the TSS was blunted and cloned into the NheI-digested pGL4.82 upstream of renilla luciferase. Sense and antisense 90 nt DNA fragment (see sequences in Supporting Information Table 3) containing three 17-nt-long RE1 or mutant RE1 sites, each one separated by 9 nt spacer, were synthesized by Invitrogen. These fragments were then annealed and cloned upstream of PGK or EF1α. Subcloning strategies for 6, 12, and 24 RE1 sites or mutantRE1 sequences were based on polymerase chain reaction (PCR) amplification of inserts with adapter primers or digestion/ligation of DNA fragments containing RE1 or mutant RE1 sequences. The D/N-REST vector was kindly provided by Dr. Noel Buckley (King's College, London). The D/N REST cassette corresponding to 234–437 amino-acid residues of mouse sequence coding for REST [30] was subcloned downstream of EF1α promoter in pIRES backbone (Clontech, Mountain View, CA, http://www.clontech.com) (pEF1α-D/N-REST: D/N-REST and pIRES: backbone vector). pMission small hairpin RNA (shRNA) plasmids (TRCN0000014785, Sigma-Aldrich) are directed against human REST gene (exon IV). Myc-DDK-tagged open reading frame (ORF) clone of Homo sapiens REST, transcript variant 1 was purchased from Origene (Rockville, MD, http://www.origene.com).

Bioluminescence Studies

Cells were transiently transfected using the Nucleofector Technology (Lonza, Basel, Switzerland, http://www.lonza.com). One and five million cells were transfected per nucleofection for HEK cells and NSCs, respectively. One microgram of reporter plasmids was used per transfection. Four microgram of D/N-REST or backbone plasmids were cotransfected with 1 μg reporter plasmids. Si-REST (Hs_REST_5) and si-RCOR1 (Hs_RCOR1_6) were purchased from Qiagen (Hilden, Germany, http://www1.qiagen.com) and were cotransfected at 100 and 10 nM, respectively, with 1 μg reporter plasmids. HEK cells and NSCs were plated in 96-well plates in 100 μL media at 350,000 and 500,000 cells per centimeter square, respectively. Cells were treated 5 hours after seeding with chemical compound or dimethyl sulfoxide (DMSO) only. Plates were then incubated for 24 hours at 37°C, with 95% humidity, and 5% CO2. Enduren substrate (Promega) was then added in each well and the bioluminescent signal was read 90 minutes later on an AnalystGT microplate reader (Molecular devices, Union City, CA, http://www.moleculardevices.com). This measure was immediately followed by addition of CellTiter-Glo reagent (Promega) and the second bioluminescence signal (viability) was read 40 minutes later. Dual-glo luciferase assay, CellTiter-Glo, and Enduren live cell substrate experiments were done according to the manufacturer's protocols (Promega). The percentage of repression, de-repression, activity and the specificity index were calculated as follows:

  • % repression = [E/CTG (mutant RE1) – E/CTG (RE1)]/[E/CTG (mutant RE1)] where E is the Enduren substrate signal, CTG is the CellTiter-Glo viability signal, and mutant RE1 or RE1 is the reporter plasmid.
  • % de-repression by D/N-REST = [E/CTG(RE1 + D/N-REST) – E/CTG(RE1+backbone)]/[E/CTG(RE1+D/N-REST)].
  • % de-repression by si-RNA = [E/CTG (RE1 + si-RNA) – E/CTG(RE1+ si-CTRL)]/[E/CTG (RE1 + si-RNA)] where si-RNA is si-REST or si-RCOR1. % activity = [(Sc/MN)) – 1] where Sc is the sample treated by compound and MN is the mean of samples treated by negative control (DMSO) in the same plate
  • Specificity index = [% activity (RE1 + backbone) – % activity(RE1+D/N-REST)]/[% activity (RE1 + backbone)].

Primary HTS

Two chemical libraries were purchased: one from CHEM-X Infinity (Romainville, France, http://www.chem-x-infinity.com) and the other from Prestwick company (Illkirch, France, http://www.prestwickchemical.com); these libraries consisted of 5,864 compounds and 1,120 compounds, respectively. The primary screening was conducted on the Biocel1800 (Agilent, Palo Alto, CA, http://www.agilent.com) platform starting with 150 million NSCs that were transiently transfected with p12RE1-EF1α-Luc plasmids. Cells were then seeded in 44 384-well plates coated with poly-ornithine and laminin (15,000 cells per well in 38 μL of NSC medium). Five hours after seeding, each compound of Prestwick (5 μM final) or Chem-X library (2 μM final) was transferred in duplicate into wells. Positive control (valproic acid [VPA] 10 mM in DMSO 0.1% v/v) and negative control (DMSO 0.1% v/v) were added in columns 1 and 2 of each plate. Plates were then incubated for 24 hours. One day later, Enduren substrate CellTiter-Glo reagents were added and bioluminescence signals were measured as described above. Data analysis of the screening was done with Spotfire software (Tibco Co, Palo Alto, CA, http://spotfire.tibco.com/). The robustness of the HTS was evaluated using Z′ factor calculated as follows: Z′ = 1 – [3(SDH + SDL)/(MH – ML)] where MH and ML correspond to the means of the positive and negative controls, respectively, and SDH and SDL correspond to the standard deviation of the positive and negative controls. Z-score method was applied to normalize values.

Quantitative Reverse Transcriptase PCR

RNA from NSCs SA-01, RC9, HD1-iPS4 or HEK293 cells was extracted after 1 day of treatment with DMSO (0.1% final) or with X5050 (100 μM final) in NSC medium without cytokines or HEK medium. The NSCs were lysed directly in the culture dishes, and RNA was isolated using RNeasy Mini kit (Qiagen) with DNAse I digestion. After quantification using a NanoDrop ND-1000A spectrophotometer, reverse transcription was performed with SuperScript III reverse transcriptase (Invitrogen) and random primers (Invitrogen). Gene expression was determined by quantitative reverse transcriptase PCR (QRT-PCR) performed with LC480 SYBR Green I Master mix (Roche, Basel, Switzerland, http://www.roche-applied-science.com). Primer sequences are presented in Supporting Information Table 4. For all experiments on human samples, values were related to 18S housekeeping gene then to appropriate control. For all experiments on mouse samples, values were related to β-actin housekeeping gene then to appropriate control.

Electrophoretic Mobility Shift Assay

Sense and antisense DNA fragments containing two RE1 sites separated by 9 nt spacer were synthesized (Invitrogen) and annealed (see sequences in Supporting Information Table 4). γ-dATP was incorporated by polynucleotide kinase T4 (Promega), and probes were purified on illustra MicroSpin G-50 Columns (GE Healthcare, Chalfont St Giles, United Kingdom, http://www.gehealthcare.com). Lysate from HEK cells transiently transfected with plasmids overexpressing REST (REST lysate) was purchased from Origene. Five microgram of lysate proteins were preincubated for 30 minutes at RT with or without competitor DNA (100-fold radioactive probe approximately 300 ng) in 22 μL of solution containing 0.1% DMSO (v/v) or 100 μM X5050, and the binding mix consisting of 8% (v/v) glycerol, 0.1 mM EDTA, 25 mM Hepes (pH 7.9), 5 mM MgCl2, 34 mM KCl, 1 mM dithiothreitol (DTT), and 1 μg polydI-dC. Approximately 3 ng (≈40,000 cpm) of probes were then added, and the reaction mixture was incubated for another 20 minutes at RT. Reactions were run in 22.5 mM Tris-borate/0.5 mM EDTA buffer and electrophoretic mobility shift assay (EMSA) was performed using 5% polyacrylamide gels. Gels were then fixed, dried, and exposed to Biomax X-ray film (Kodak) for 72 hours.

Western Blot

Cells were resuspended in RIPA lysis buffer (Sigma-Aldrich) in the presence of Protease Inhibitor Cocktail (Sigma-Aldrich) and anti-phosphatases PhosphoSTOP (Roche). Protein concentration of cell extracts was determined using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc, Waltham, MA, http://www.thermofisher.com) according to the manufacturer's instructions. Proteins from each sample were mixed with NuPAGE lithium dodecyl sulfate (LDS) sample buffer 4× (Invitrogen) and DTT 1 M (Sigma-Aldrich) then heated at 70°C for 10 minutes. SDS poly-acrylamide gel electrophoresis (SDS-PAGE) was performed using NuPAGE Novex 4%–12% Bis-Tris Gels (Invitrogen) and NuPAGE 2-(N-morpholino)ethanesulfonic acid (MES) SDS running buffer (Invitrogen) with addition of NuPAGE antioxidant (Invitrogen). Twenty microgram of total proteins was loaded per well along with HiMark Pre-stained Protein Standard (Invitrogen). Protein migration was performed during 45 minutes at 200 V at RT. Proteins were transferred onto nitrocellulose membranes using the iBlot Gel Transfer Stack (Invitrogen) and the iBlot Dry Blotting System (Invitrogen). Membranes were blocked with 5% non-fat milk in phosphate-buffered saline (PBS) containing 0.1% Tween 20 (PBST) for 1 hour, then incubated overnight at 4°C with REST polyclonal antibody (Abcam, Cambridge, U.K., http://www.abcam.com). After several washes with PBST, blots were incubated for 1 hour at room temperature with rabbit horseradish peroxidase-conjugated secondary antibody. Membranes were then washed with PBST and incubated in Amersham ECL Plus Western Blotting Detection Reagents (GE Healthcare) in order to reveal immunoreactive bands by using the ImageQuant LAS 4000 mini luminescent image analyzer (GE Healthcare). Results were normalized to β-actin revealed with AC-74 antibody (Sigma-Aldrich). For myc-tagged experiments, we used anti-myc antibody (Invitrogen)

Transcriptome

RNA was extracted from six samples, corresponding to three independent cultures of NSCs SA-01, each one treated either with DMSO (0.1%) or with X5050 (100 μM). RNA was isolated using RNeasy Mini kit with DNase I digestion (Qiagen). Quality control was assessed using Agilent Bioanalyzer (Agilent Technologies, Germany) and NanoDrop spectrophotometer ND-1000A. Genome-wide gene expression profiling was performed by hybridization on oligonucleotide microarrays (in total six GeneChips human Gene 1.0 ST) according to standards supplied by the manufacturer (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). All quality controls and statistics were performed using Partek Genomic Suite. Raw data were normalized using the Robust Multichip Algorithm in Partek [32]. We first made a hierarchical clustering (Pearson's dissimilarity and average linkage) and principal component analysis for unsupervised analysis with all samples. To find differentially expressed genes, we applied a two-way ANOVA (factor treatment and factor culture) and computed the fold-change for each gene. All data obtained by microarray analysis have been submitted on gene expression omnibus (GEO) site with this accession number (GSE40695).

Gene set enrichment analysis (GSEA) was carried out using the motif database from the Broad Institute [31, 33]. Genes belonging to the enrichment core of datasets with false discovery rate (FDR) < 0.05 were selected. For core enrichment genes in each dataset, gene ontology was performed using the online database DAVID (http://david.abcc.ncifcrf.gov [34, 35]) and some of these genes were selected and confirmed by QRT-PCR. Hierarchical clustering with Ward method was done using JMP software.

Intraventricular Injection of X5050 in Quinolinic Acid Lesioned Mice

12-week-old male C57Bl6 mice (Charles River, France) were used in this study (n = 14). All experimental procedures were performed in strict accordance with the recommendations of the European Commission (86/609/EEC) concerning the care and use of laboratory animals. Mice were anesthetized with 0.1 mL/10 g of a mixture of ketamine (100 mg/mL) and 0.5 mL xylazine (20 mg/mL). Quinolinic acid was injected into the striatum, using a 34-gauge blunt-tip canula linked to a Hamilton syringe by a polyethylene catheter. A total volume of 1 μL (80 mM) was injected at 0.5 μL/minute. The stereotaxic coordinates were: anteroposterior, +1 mm; lateral, +2 mm from the bregma; and ventral, −2.7 mm from the dura, with tooth bar set at 0 mm. At the end of the injection, the needle was left in place for 5 minutes before being slowly removed. The skin was sutured and mice were allowed to recover. One week after the lesion, mice received simultaneous bilateral injection of X5050 (2× 2 μL of 20 mM in 10% DMSO in water) in the lateral ventricles (the stereotaxic coordinates were: anteroposterior, −0.46 mm; lateral, ±1 mm from the bregma; and ventral, −2.25 mm from the dura, with tooth bar set at 0 mm). An equal volume of 10% DMSO in water was injected in controls. Before injection, the needles were fully removed to allow cerebrospinal fluid (CSF) to exit from the needle tract, lower CSF pressure, and validate needle placement. At the end of the injection, the needles were left in place for 5 minutes before being slowly removed. One day after intraventricular injection, mice were killed and the brain was removed, blocked, and cut into 1-mm-thick coronal slices. On one coronal slice (+1 mm from bregma), tissue punches from the striatum were taken by using a tissue corer (1.5-mm in diameter). From the adjacent (anterior) slice, the prefrontal cortex was dissected out. RNA from all tissue punches was isolated with Trizol Reagent and RNeasy micro Kit according to the manufacturer's instructions (Qiagen).

Statistical Analysis

With the exception of the microarray analysis, all statistical analyses were performed using Graph Pad Prism5 and JMP software. For multiple comparisons we used one-way ANOVA analysis. In paired experiments, one sample t test or Student's t test were used depending on each experiment as indicated in the figure legends.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Assay Development for Measuring REST Activity

REST activity was first measured during neuronal differentiation of SA-01 hESCs, in order to control the relevance of the cell model. REST mRNA levels were quantified using QRT-PCR in three cell populations: undifferentiated hESCs, hESC-derived NSCs, and neurons differentiated for 21 days from NSCs; Nestin (NES) and SOX1 expression peaked in NSCs while synaptophysin (SYP), neural cell adhesion molecule L1 (L1CAM), synaptosomal-associated protein 25 (SNAP25), and α-synuclein (SNCA) were highest in neurons (Supporting Information Fig. S1A). In keeping with previous studies [13, 17], the expression of the predominant and longer transcript of REST was maximal in hESCs and NSCs and decreased by 25-fold in neurons, while that of the alternatively spliced and shorter transcript REST4 was minimal in hESCs and NSCs and increased by more than 100-fold in neurons (Supporting Information Fig. S1B) confirming observations made in rodent brain [36-38] and PC12 derivatives [39]. As a control, RCOR1, a core member of the REST nuclear complex, was not significantly modulated in any of the three cell populations. Western blot analyses confirmed the presence of the 122 kDa longer isoform of the REST protein in NSCs, the identity of which was further established by knocking-down REST using specific shRNA (Supporting Information Fig. S1C). The functionality of REST protein in NSCs was checked using a dominant/negative REST (D/N-REST) isoform. Expression of three RE1-genes containing two RE1 sites, L1CAM, SNAP25, and SYP, was significantly increased 24 hours after transfection (Fig. 1D). As control of specificity of the effects of D/N-REST, similar changes were not observed on mRNA levels of SOX1 or NES that are not regulated by REST or on levels of REST itself.

image

Figure 1. Luciferase-based reporter assay for REST activity in neural stem cells (NSCs). (A): Schematic of REST reporter cassette and RE1 consensus and mutated (red) sequences. (B): REST activity expressed as percentage of repression/de-repression of luciferase cassette. From left to right : Luciferase repression comparing signals with mutant 6RE1-PGK and 6RE1-PGK plasmids: luciferase de-repression by D/N-REST comparing signals with 6RE1-PGK + D/N-REST and 6RE1-PGK + backbone plasmids; luciferase de-repression by si-REST comparing 6RE1-PGK + si-REST and 6RE1-PGK + si-CTRL; luciferase de-repression by si-RCOR1 comparing 6RE1-PGK + si-RCOR1 and 6RE1-PGK + si-CTRL. Error bars: mean ± SEM (n = 4). One sample t test compared to 0. (C): Percentage of repression related to the number of RE1 sites. Experiments similar to (B) with plasmids with EF1α (continuous line) or PGK (dashed line) promoter driving luciferase. Error bars: mean ± SEM (n = 11). *, p < .05; **, p < .01 by unpaired two-tailed Student's t-test. Significant logarithmic fit of means (p = .014 for EF1α and p = .007 for PGK). (D): REST activity and REST mRNA levels in neural derivatives. Upper panel: Percentage repression quantified as in (B) but using 12RE1-EF1α and mutant 12RE1-EF1α plasmids. Lower panel: relative REST mRNA levels by quantitative reverse transcriptase polymerase chain reaction. NSCs pretreated for 7 days in medium without mitogens before being treated with one of three media (FGF2 + EGF, without cytokines, DAPT). Error bars: mean ± SEM (n = 6). *, p < .05; **, p < .01 by one-way ANOVA and Dunnett's multiple comparison test. Abbreviations: D/N REST, dominant-negative form of REST; DAPT, N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester; EGF, epidermal growth factor; EF1α, elongation factor 1α; FGF2, fibroblast growth factor; PGK, phosphoglycerate kinase; RE-1, repressor element-1; REST, repressor element-1 silencing transcription factor

Download figure to PowerPoint

Several RE1-containing reporter plasmids were designed and constructed in order to measure the activity of REST in hESC derivatives. “RE1-plasmids” included the coding sequence of Renilla luciferase under the control of the promoter of either EF1α or PGK, itself located downstream of 3, 6, 12, or 24 consensus RE1 sites (Fig. 1A). In control vectors (“mutant RE1-plasmids”), six nucleotides of the consensus sequence located at position with the highest position-scoring matrix (PSM) were mutated. The repressor activity of REST on luciferase expression was measured in SA-01 derived NSCs, 24 hours after transfection with either RE1-plasmids or mutant RE1-plasmids. Luciferase signals normalized to cell viability were lower in cells transfected with RE1-plasmids. This suggested the RE1-mediated repression of the luciferase expression cassette by endogenous REST in transfected cells (Fig. 1B). Additional experiments confirmed the specificity of the assay, that is, the relationship between luciferase signal and repressor activity of endogenous REST on RE1 sites upstream of PGK/EF1α promoter. RE1-plasmids were cotransfected with either D/N-REST plasmids or small interfering RNA (siRNA) targeting REST or RCOR1 to impair endogenous REST function (Fig. 1B). De-repression of these control conditions measured was found equivalent across the range of experiments. The assay specificity was further validated using reporter plasmids containing increasing numbers of RE1 or mutant RE1 sites. Repression of the activity of EF1α and PGK promoters under the control of RE1 sites increased in parallel to the number of RE1 sites. Repression curves were similar for both promoters and fitted a logarithmic model (Fig. 1C). The capacity of REST to inhibit luciferase expression of RE1-plasmids was independent of the orientation (sense or reverse) of the RE1 sequences and of the normalization method (Supporting Information Fig. S1A, S1B).

In order to test the correlation between REST mRNA levels and repression of the reporter gene in cells transfected with RE1-plasmids, three neural cell populations containing different levels of REST were assayed: cells cultured with FGF2 and EGF, in which REST level is the highest, cells coaxed to neuronal differentiation by the notch inhibitor DAPT (Supporting Information Fig. S2C) that have the lowest level and cells cultured without mitogens that have intermediate level. Levels of REST enzymatic reporter activity paralleled levels of REST mRNA (Fig. 1D).

HTS for REST Inhibitors

The conditions that achieved statistical robustness compatible with HTS of several thousand chemical compounds were defined in SA-01-derived NSCs. Several pharmacological inhibitors of enzymes present in the REST nuclear complex were examined in order to identify the most suitable positive control for our REST assay in NSCs. The HDAC inhibitor (VPA, 10 mM) showed the highest activity, as normalized to negative control treatment (DMSO) (Supporting Information Fig. S3A). Using these controls, the mean score of Z′ factor calculated for the REST assay in NSCs on a set of five 384-well plates was 0.5 ± 0.2 (mean ± SD), a value that is appropriate for HTS (Supporting Information Fig. S3B).

The effect of 6,984 compounds was tested in duplicate in SA-01-derived NSCs transfected with 12 RE1-plasmids. Compounds were selected as primary hits—potential REST inhibitors—when viability signal was not reduced below 70% and when REST activity was above a two-sigma threshold (mean plate signal + 2 SD) (Fig. 2A). Fifty compounds matched those first selection criteria. Retest screening and counter-screening using mutant RE1-plasmid retained 23 of them (Fig. 2B) out of which 20 showed dose-dependent activity. We then applied an arbitrary threshold for maximum activity of 30% over controls (D/N-REST plasmids) and thus selected the 11 most potent compounds. These 11 compounds mostly clustered into two chemical families, namely benzoimidazole-5-carboxamide and pyrazole propionamide derivatives (Supporting Information Fig. S2C; Table 4). Properties of the two most potent and specific members of each chemical family are shown in Figure 2C, 2D, and 2E. The benzoimidazole-5-carboxamide derivative X5050 exhibited both the highest activity and specificity. Its activity was dependent on the number of RE1 sites (Fig. 2F) and was confirmed both using an alternative normalization strategy based on a firefly luciferase control plasmid and in another NSC line derived from RC9 (WT) hESCs (Supporting Information Fig. S2D). X5050 compound was therefore selected for subsequent studies.

image

Figure 2. Identification by high throughput screening of candidate repressor element-1 silencing transcription factor (REST) inhibitors in neural stem cells (NSCs). (A): Primary screening and hit selection. Activity of each compound expressed as % increase in bioluminescence-to-viability ratio signal with compound normalized to bioluminescence-to-viability ratio signals obtained with DMSO. Primary hit compounds shown in green, toxic compounds in red, inactive compounds in blue, negative controls (DMSO) in yellow, and positive controls (10 mM VPA) in purple. (B): Attrition cascade of primary hits. (C): Dose-response activity for four most potent hit compounds. NSCs cotransfected with 12RE1-PGK and backbone plasmids (REST activity in red) or D/N-REST plasmids (nonspecific activity in blue). Error bars: mean ± SD of three wells. Nonlinear fit using the inhibitory dose-response curves with variable slope model with Graph pad Prism5. (D): Specificity index of the four selected hits. Specificity index calculated from the difference in plateau levels of dose-response REST activity related to control. Error bars: mean ± SEM (n = 6). ***, p < .001 by one sample t test compared to 0. (E): Characteristic parameters for the four selected hits. (F): Dose-dependent activity of X5050 to the number of RE1 sites. NSCs transfected with RE1 plasmids with increasing number of RE1 sites treated for 1 day with 8 μM X5050. Error bars: mean ± SEM (n = 6). *p < .05 by unpaired two-tailed Student t test. Abbreviations: DMSO, dimethyl sulfoxide; VPA, valproic acid.

Download figure to PowerPoint

Mechanism of Action of X5050 on REST Activity

Mechanism of action of X5050 was sought on REST binding to RE1 site, REST transcription, alternative RNA splicing, and protein degradation. EMSA with a radioactive oligonucleotide containing two RE1 sequences revealed a distinctive labeled band that disappeared after addition of 100-fold excess of unlabeled probes in extracts of HEK293 cells overexpressing REST (Fig. 3A). Addition of X5050 (100 μM) did not decrease the band intensity indicating that it did not affect the in vitro binding of endogenous REST to RE1 sites. Quantification of REST levels using primers recognizing all types of REST transcripts showed no significant change in X5050 treated cells, excluding a transcriptional effect (REST-all in Fig. 3C). There was no change in the titer of transcripts for the longer isoform of REST either (REST in Fig. 3C). However, levels of the REST4 shorter transcripts increased twofold, which is likely associated to the initiation of neuronal commitment in treated cultures (REST4 in Fig. 3C).

image

Figure 3. Mechanism of action of X5050. (A): Electrophoretic mobility shift assay of radiolabeled oligonucleotides with 2 RE1 sequences. Arrow corresponds to REST-RE1* complex. (B): Percentage of de-repression by D/N-REST with increasing X5050 concentrations in NSCs and HEK cells. Cells cotransfected with 12RE1-EF1α and backbone or D/N-REST plasmids. Values are mean ± SD (three wells). (C, E): Relative mRNA expression levels of REST and RCOR1 in NSCs derived from SA-01 wild-type (WT) human embryonic stem cells (C) or in HEK cells (E) treated with X5050. (D, F): Effect of X5050 on endogenous REST protein level in NSCs. Left panel: one representative immunoblot. SA-01 (WT)-derived NSCs (D) or HEK (F) treated 1 day before protein extraction, with 50 or 100 μM X5050 or with DMSO. β-Actin as loading control. Right panel: Densitometry (values are normalized to DMSO-treated cells). (G): Effect of X5050 on transgenic myc-tagged REST protein level in transfected HEK cells. Left panel: One representative immunoblot. HEK cells transfected with Myc-tagged REST plasmid 1 day before treatment. One day treatment with 50 or 100 μM X5050 or with DMSO. β-Actin as loading control. Right panel: Densitometry. One day treatment with 50 or 100 μM X5050. For (C–G): Error bars, mean ± SEM (n = 6). **, p < .01; ***, p < .001 by one sample t test compared to 1. Abbreviations: DMSO, dimethyl sulfoxide; HEK, human embryonic kidney; NSC, neural stem cell; REST, repressor element-1 silencing transcription factor; RE-1, repressor element-1.

Download figure to PowerPoint

In the absence of effects of X5050 at DNA or RNA levels, the levels of REST protein were then analyzed. Western blot analyses of NSC extracts showed that 24-hour-long treatments with increasing doses of X5050 induced a dose-dependent decrease in the 122 kDa longer REST isoform (Fig. 3D). Treatment of NSCs with Bortezomib (100 nM) or MG132 (10 μM), two cell-permeable proteasome inhibitors, increased the activity of REST as measured by the repression of the reporter cassette but did not prevent X5050 dose-dependent inhibition. This result indicated that the compound does not act directly on the proteasome activity (Supporting Information Fig. S4). Although NSCs assayed are proliferative, one possible confusing parameter in these experiments could be the concomitant stimulation of neuronal differentiation in the culture, as REST levels decrease over that process. The activity of X5050 was therefore assessed in non-neural cells expressing high levels of REST in the absence of any potential neuronal differentiation. Twenty-four hours after cotransfection of HEK cells with 12RE1-plasmids and either backbone or D/N-REST plasmids in the presence of increasing concentrations of X5050, REST activity—that is, de-repression of the reporter cassette—decreased in a dose-dependent manner, similarly to results obtained in treated NSCs (Fig. 3B). As in NSCs, REST expression was not changed by X5050 treatment, the expression of the neuron-specific and shorter transcript REST4 being below detection level (Fig. 3E). Western blot analyses of HEK extracts showed that 24-hour-long treatments with increasing doses of X5050 induced a dose-dependent decrease in the 122 kDa longer REST isoform (Fig. 3F). This was also the case for transgenic MYC-tagged-REST protein (Fig. 3G). These results altogether support the hypothesis that X5050 reverses REST repression on neuronal genes by promoting its degradation.

Functional Impact of X5050 on Gene Expression in Human NSC

The functional impact of the changes induced by X5050 was then investigated using a whole-genome differential transcriptomic approach. Changes in gene expression resulting from the treatment of NSCs with 100 μM X5050 for 24 hours were analyzed using GSEA on Affymetrix human Gene 1.0 ST array. Genes modulated by X5050 that contained a common regulatory sequence (e.g., a binding site for a transcription factor or for a micro-RNA) were identified using the analysis software and “motif” gene set database from the Broad Institute. Eight out of the 828 gene sets were over-represented in a statistically significant (FDR <0.05) manner at the top (two upregulated gene sets) or bottom (six downregulated gene sets) of the list of genes ranked according to their modulation by X5050 (Fig. 4A; Supporting Information Table 2). Gene ontology of the core enrichment group of each of these eight gene sets—that is, the list of genes that contributed most to the enrichment result—were analyzed using DAVID bioinformatics resource (Fig. 4A). Among the six downregulated gene sets, four comprised genes that encode nucleic acid-interacting proteins such as histones and proteins present in nucleosomes or spliceosomes. The last two gene sets contained genes involved in focal adhesion and cellular contractility. One of the two upregulated gene sets consisted of genes containing a consensus binding sequence for miR-380. However, the result that attracted most attention was the other upregulated gene set, V$NRSF, as it clusterized the 72 genes that display at least one validated RE1 sequence and are, accordingly, the most likely targets of REST silencing (Fig. 4B). Furthermore, gene ontology of the 26 genes that formed the core enrichment group of V$NRSF revealed a statistically significant over-representation of RE1-genes implicated in neuronal function or development (p < 10−3 to 10−6). In order to further confirm the specific impact of X5050 on RE1-genes, a larger set of 494 genes that were identified as functional RE1-genes with PSM score >0.9 (http://bioinformatics.leeds.ac.uk/RE1db_mkII/) was analyzed. Hierarchical clustering of these data segregated transcriptome observations of DMSO-treated cells from X5050-treated cultures (Fig. 4C). Analysis of upregulated genes showed the predominance of neuronal genes. Gene expression data were confirmed by QRT-PCR in another NSC line derived from RC9 wild-type (WT) hESCs on selected upregulated or downregulated transcripts. At 100 μM, X5050 upregulated by more than twofold two representative RE1 neuronal genes, SNAP25 and BDNF (Fig. 4D) while downregulating histone 1H genes (HIST1H2BM, HIST1H3J) (Supporting Information Fig. S5). Differential regulation of BDNF splice variants was measured using primers specific for upstream 5′ untranslated exons, including exon II that contains an RE1 site. Upregulation of other RE1-genes tested was less marked (SYP, HTT) or not statistically significant (L1CAM). Expression of REST and other control genes without known RE1 sites such as, RCOR1 or antisense BDNF was not notably upregulated. Altogether, the unbiased GSEA and the supervised analysis of a large number of RE1-genes indicated that X5050 upregulates neuronal RE1-genes in human NSCs.

image

Figure 4. Transcriptome analysis of wild-type NSCs treated with X5050. (A): Significant motif gene sets by GSEA of SA-01 NSC samples. NES (values for genes upregulated and downregulated by X5050 in red and blue, respectively). GO analysis of enrichment core of genes in corresponding gene set. (B): V$NRSF gene set. Upper panel: Plot of enrichment score for genes present in the gene set. Bars underneath the plot correspond to genes of the gene set. Lower panel: Heatmap of core enrichment genes of the V$NRSF gene set. (C): Hierarchical clustering of experiments using 494 functionally identified RE1 genes with position-scoring matrix >0.9. (D): Quantitative reverse transcriptase polymerase chain reaction confirmation of transcriptome data in RC9 (WT) derived NSCs. Left panel: Expression level of RE1 genes (L1CAM, SNAP25, SYP, HTT) and control genes (RCOR1, REST), Right panel: Expression level of BDNF alternative transcripts, 1 day after treatment with 50 μM (light blue bars) or 100 μM (dark blue bars) of X5050. Values normalized to values in NSCs treated with DMSO. Error bars: mean ± SEM (n = 12). *, p < .05; **, p < .01; ***, p < .001 compared to DMSO by one sample t test compared to 1. Abbreviations: BDNF, brain-derived neurotrophic factor; DMSO, dimethyl sulfoxide; FDR, false discovery rate; GO, gene ontology; GSEA, gene set enrichment analysis; NES, normalized enrichment score; NSC, neural stem cell; REST, repressor element-1 silencing transcription factor; UTR, untranslated region; WT, wild type.

Download figure to PowerPoint

Functional Impact of X5050 in HD Models

The functional impact of the changes induced by X5050 was finally investigated in HD pathological conditions. X5050 activity was measured on NSCs derived from a human induced pluripotent stem cells (HD1-iPS4 line) previously generated from somatic cells of a patient diagnosed with HD [29]. The dose-dependent effect of X5050 was tested in HD1-iPS4 derived NSCs (Fig. 5A). Dose-response curves for REST activity measured in parallel in HD-iPSC4 NSCs and RC9 (WT) NSCs showed no significant difference. Effect of X5050 on the expression levels of selected RE1- and control genes in HD-NSCs was measured by QRT-PCR. The profile of regulation by X5050 of the expression of RE1-genes and control genes was similar to that measured in WT-NSCs (Figs. 5B, 4D). The highest and most significant upregulations were those observed for SNAP25 and BDNF alternative transcript containing exon II and to a lesser degree for BDNF transcripts containing exons IV and VIII and SYP (Fig. 5B). Basal levels of BDNF II and BDNF IV were lower in HD-NSCs than in WT-NSCs (Supporting Information Fig. S6), confirming previous report using similar cells derived from HD1-iPS4, WT-iPSC, and HD1-iPS4 clones genetically corrected the for CAG expansion [40].

image

Figure 5. Functional impact of X5050 in Huntington disease models: (A): Dose-response activity for X5050 in NSC derived from RC9 (WT) and HD1-iPS4 (HD) pluripotent stem cells. NSCs were cotransfected with 12RE1-EF1a and backbone plasmids (REST-activity in red) or D/N-REST plasmids (nonspecific activity in blue). Nonlinear fit dose-response curves for WT-NSCs (dashed lines) and for HD-NSCs (solid lines). Error bars: mean ± SEM of four wells. (B): Quantitative reverse transcriptase polymerase chain reaction (QRT-PCR) analyses of X5050 activity in HD1-iPS4-derived NSCs. Left panel: Expression level of RE1 genes (L1CAM, SNAP25, SYP, HTT) and control genes (RCOR1, REST), Right panel: Expression level of BDNF alternative transcripts, 1 day after treatment with 50 μM (pink bars) or 100 μM (red bars) of X5050. Values normalized to values in HD-NSCs treated with DMSO. Error bars, mean ± SEM (n = 12). *, p < .05; **, p < .01; ***, p < .001 compared to DMSO by one sample t test compared to 1. (C, D): QRT-PCR analyses of X5050 activity in the brain of mice with QA-induced excitotoxic striatal lesions. (C) Expression level in striatal samples of striatal neurons marker (Darpp32, Snap25), 24 hours after X5050 bilateral intraventricular injection of X5050 (i.e., 1 week after QA injection). (D) Expression level in prefrontal cortex samples of selected neuronal RE1-genes [Bdnf (variant containing exon II, IV and all variants), Snap25, Trim9, and Omg] 24 hours after bilateral intraventricular injection of X5050. Values normalized to median of values of samples from same brain region of unlesioned hemisphere injected with vehicle. Values for samples from unlesioned brain hemisphere (blue bars), from QA-lesioned hemisphere (red bars), from X5050-treated mice (cross-hatched bars). Error bars, mean ± SEM (n = 7 per group). *, p < .05; **, p < .01; ***, p < .001 by unpaired two-tailed Student t test. Abbreviations: BDNF, brain-derived neurotrophic factor; DMSO, dimethyl sulfoxide; HD, Huntington disease; NSC, neural stem cell; QA, quinolinic acid; REST, repressor element-1 silencing transcription factor; WT, wild type.

Download figure to PowerPoint

In vivo activity of X5050 was finally measured in the brain of mice with Quinolinic Acid (QA)-induced excitotoxic striatal lesions. C56BL6 mice received unilateral injection of 80 nmol of QA to induce neurodegeneration of medium spiny striatal projection neurons (Darpp32 immunopositive), the major population of striatal neurons [41]. Although striatal QA does not produce direct excitotoxic neuronal death in the cortical regions anatomically connected to the striatum, a loss of BDNF levels has been reported in the cerebral cortex ipsilateral to the lesion [42]. Thus, we reasoned that the QA model would be appropriate to test the pharmacological efficacy of X5050 in vivo to increase BDNF expression. One week after lesioning, mice received bilateral injection of either X5050 or vehicle in the lateral ventricles in order to bypass the blood brain barrier. Effect of QA lesion and X5050 acute treatment on the expression levels of selected RE1 and control genes was measured 24 hours after X5050 injection, in striatum and in the prefrontal cortex, a cortical area spared by the direct excitotoxic effect of QA but anatomically connected to the striatum and localized near the anterior part of the lateral ventricles where X5050 was injected. A decrease in Darpp32 and Snap25 expression was found by QRT-PCR in the lesioned striatum as compared with the contralateral striatum. This confirmed that QA had induced significant neurodegeneration in this brain region (Fig. 5C). Expression levels in samples from prefrontal cortex from all animals were also measured by QRT-PCR (Fig. 5D). Results showed a significant (p < .05) upregulation of Bdnf, Snap25, Trim9, and Omg (the mouse homologs of the top four RE1-genes of the V$NRSF-1 core enrichment list presented in GSEA, Fig. 4B) by X5050 in the prefrontal cortex ipsilateral to the QA-lesioned striatum (Fig. 5D). Most importantly, using splice variant specific primers for BDNF, QRT-PCR revealed that, in QA-lesioned hemisphere, X5050 significantly increased the levels of Bdnf II splice variant (exon II containing variant) while Bdnf IV levels were not significantly changed. Altogether, the functional analysis of X5050 in HD-NSCs and QA-lesion mice suggested that this compound was active in an HD pathological context (Fig. 5D).

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

The main result of this study is the identification of a benzoimidazole-5-carboxamide derivative (X5050) that promotes the expression of neuronal genes including BDNF and SNAP25 via the degradation of REST in human neural stem cells. Combination of HTS and pluripotent stem cell technologies was instrumental in identifying that compound as an inhibitor of REST activity from a proprietary library of several thousand molecules and in exploring its mechanism of action in human neural stem cells. In vitro and in vivo functional analyses in HD models revealed that X5050 is active in HD pathological context. In particular, X5050 upregulated among several known REST-regulated genes, the expression of BDNF in the cortex of mice with striatal lesions. This study underlines the value of a strategy aimed at modulating REST in the attempt to restore key neuronal gene transcription in the brain. This may reveal valuable to tackle neurodegenerative conditions involving downregulation of BDNF in particular in the case of HD for which BDNF impairment results in part from increased REST activity.

REST transcriptional regulation of RE1-genes is modulated at multiple levels including transcription, protein degradation, intracellular localization, and REST-nuclear complex composition and binding to RE1-sequences [18]. We have developed a luciferase-based assay for REST activity on transgenic RE1 sites that could report all these types of regulation. This assay efficiently measured REST activity levels in NSCs as it successfully integrated a number of challenges against REST function including: (a) the inhibition of REST binding to RE1 either via the mutation of the RE1 sequences or via the competitive binding of D/N-REST, (b) REST expression knockdown by REST-targeting siRNA, (c) the combined reduction of REST mRNA and protein levels via induction of NSC neuronal differentiation, and finally (d) the impairment of the formation of REST nuclear complex via the knockdown of RCOR1, one of the main core cofactors of this complex.

With the ultimate goal of discovering chemical compounds that may decrease the activity of REST in the human brain, we took advantage of the specificity of this assay to identify by HTS two clusters of candidate REST inhibitors. None of the FDA-approved drugs (Prestwick library) we tested displayed a significant effect against REST activity, suggesting that drug repositioning may not be an option. All hits were identified from a proprietary chemical library of over 5,000 synthetic molecules. The most potent REST inhibitor was the benzoimidazole-5-carboxamide derivative X5050. Several compounds closely related to X5050 were active against REST, although with reduced potency.

The exploration of the mechanisms of action of X5050 in NSCs has evidenced drug-induced degradation of REST. Treatment of non-neural cells, HEK, with X5050, resulted in a similar reduction in endogenous or transgenic Myc-tagged REST levels, indicating that this reduction was not mediated by the induction of the neuronal differentiation of NSCs, also known to decrease REST bioavailability [17]. The actual target protein(s) of X5050 remain(s) to be identified. Since inhibition of proteasome did not reduce X5050 activity, good candidates may therefore be found among the enzymes that control REST ubiquitination by tagging REST for proteasome degradation. Several studies have indeed demonstrated that REST degradation in the proteasome via the SCF-βTRCP complex, which consists of Skp1, Cul1 and the F-box protein β-TrCP (β-transducin repeat containing protein), critically regulates REST activity and, consequently, neuronal commitment of NSCs [13, 19, 43].

BDNF deficiency in the brain has been linked to psychiatric and neurodegenerative disorders including depression, Alzheimer, Parkinson, or HD (for review see [44, 45]). In the later case, reduced BDNF level in the striatum has been reported both in HD patients and in HD genetic models [46, 47]. BDNF is mostly anterogradly transported to the striatal tissue from cortical neurons [48]. Studies in transgenic mice lacking cortical BDNF expression best illustrate that BDNF is required for the differentiation and long-term survival of the medium spiny neurons in the striatum, the neurons most affected in HD [49]. HTT mutation disrupts both the BDNF transport from the cortex to the striatum [50, 51] and the expression of cortical BDNF via HTT mutation-mediated impairment of REST [7, 52, 53]. These observations have set BDNF as a promising therapeutic target for HD [54]. Along that this line, pharmacological upregulation of BDNF level in the brain through REST inhibition should open interesting perspectives. The backbone of X5050 appears of particular interest since transcriptome analysis has shown that, among RE1-genes encompassing a wide variety of organ-specific cell types, X5050 affected specifically those related to neuronal function. This included BDNF and most significantly the BDNF splice variant known to be regulated by REST containing exon II: BDNF II. Effect of X5050 on BDNF was not limited in WT-NSCs to BDNF II but was detectable for BDNF I, IV, VII, VIII mRNA variants. This could result from a positive feedback loop induced by X5050 activity on BDNF. Increased release of BDNF that may result from X5050 upregulation of BDNF II could activate BDNF receptor TrkB and its downstream intracellular effectors such as cyclic CREB (cAMP response element binding) in turn responsible for BDNF I and IV promoters' activation (for review see [55]). Effect of X5050 on BDNF II expression was as well confirmed in HD-NSCs. An et al. [40] have linked the reduced expression of BDNF in HD1-iPS4 derived NSCs (compared with level in WT-NSCs) to the HTT mutation, genetically correcting the CAG expansion (72 CAGs) in HD1-iPS4 clones. In this work, we showed that X5050 rescued BDNF II reduced expression in NSCs derived from the same HD1-iPS4 line.

A pilot assessment of the acute activity of the X5050 compound was conducted in mice with QA lesion of the striatum that reproduce the neuronal loss observed in the striatum of HD patient (for review see [56]). Reduced BDNF protein levels in the cortex have been reported within weeks of QA injection in the ipsilateral striatum and have been monitored to assess the efficacy of new neuroprotective approach [42, 57, 58]. Using the same animal model, Rite et al. [59] have also demonstrated that the expression of BDNF in cortical areas projecting to striatum is dependent on both target integrity and neuronal activity. We consequently explored the activity of X5050 in a cortical region known to project in the striatum, the prefrontal cortex. Neurons in this cortical region should be affected by the loss of their striatal targets. Indeed, we found that X5050 treatment impacted on BDNF and other RE1-neuronal gene expression in QA-lesioned hemisphere. The identification of novel chemical compounds that could lift the repression of neuronal genes such as BDNF mediated by pathologically over-active REST may have direct therapeutic applications. Strategies to further characterize the therapeutic potential of X5050 will involve chronic administration of this hit compound or one of its chemical derivatives in a genetic model of HD in mice and should ultimately aim at exploring the capacity of the agent injected to slow neurodegeneration and dysfunction in HD animal. Overall our data show that compounds with X5050 backbone appear as potential candidates for normalizing expression level of key neuronal genes in patients with HD and even other neurodegenerative disorders featuring alteration of BDNF level.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

We thank Dr. G. Daley (Children's Hospital Boston, MA) for providing the HD1-iPS4 line; Dr. P. De Sousa (RoslinCells, U.K.) for the RC9 line; Dr. P. Charbord for helpful discussions and transcriptome analysis; Dr. J. Denis and Dr. N.J. Buckley for helpful discussions; and Dr. F. Letourneur and F. Dumont of the Genomic/transcriptomic platform (Cochin Institute, Paris, France). This work was supported by additional grants from European Union programs FP6 STEM-HD and FP7 Neurostemcell, the laboratoire d'Excellence Revive (Investissement d'Avenir; ANR-10-LABX-73) and by a FRM fellowship to J.C. J.C. is currently affiliated with Department of Cell and Molecular Biology, Karolinska Institutet, Stockholm, Sweden; P.P. is currently affiliated with CYTOO Cell Architects, Bâtiment BHT- BP 50, 7 parvis Louis Néel, Grenoble, France; F.C. is currently affiliated with Discngine-Parc Biocitech, Romainville, France; M.L. is currently affiliated with Institut de la Vision, UMRS INSERM 968, UMRS CNRS 72.10, Université Pierre et Marie Curie, Paris, France.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure of Potential Conflicts of Interest
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1430-sup-0001-suppfig1.tiff335KSupporting Information Figure 1
stem1430-sup-0002-suppfig2.tiff145KSupporting Information Figure 2
stem1430-sup-0003-suppfig3.tiff601KSupporting Information Figure 3
stem1430-sup-0004-suppfig4.tiff174KSupporting Information Figure 4
stem1430-sup-0005-suppfig5.tiff364KSupporting Information Figure 5
stem1430-sup-0006-suppfig6.tiff156KSupporting Information Figure 6
stem1430-sup-0007-suppinfo1.docx11KSupporting Information
stem1430-sup-0008-suppinfo2.docx11KSupporting Information
stem1430-sup-0009-suppinfo3.xlsx92KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.