Abstract
Differentiation of oligodendrocyte progenitor cells (OPCs) into mature oligodendrocytes is regulated by the interplay between extrinsic signals and intrinsic epigenetic determinants. In this study, we analyze the effect that the extracellular ligands sonic hedgehog (Shh) and bone morphogenetic protein 4 (BMP4), have on histone acetylation and gene expression in cultured OPCs. Shh treatment favored the progression toward oligodendrocytes by decreasing histone acetylation and inducing peripheral chromatin condensation. BMP4 treatment, in contrast, inhibited the progression toward oligodendrocytes and favored astrogliogenesis by favoring global histone acetylation and retaining euchromatin. Pharmacological treatment or silencing of histone deacetylase 1 (Hdac1) or histone deacetylase 2 (Hdac2) in OPCs did not affect BMP4-dependent astrogliogenesis, while it prevented Shh-induced oligodendrocyte differentiation and favored the expression of astrocytic genes. Transcriptional profiling of treated OPCs, revealed that BMP4-inhibition of oligodendrocyte differentiation was accompanied by increased levels of Wnt (Tbx3) and Notch-target genes (Jag1, Hes1, Hes5, Hey1, and Hey2), decreased recruitment of Hdac and increased histone acetylation at these loci. Similar upregulation of Notch-target genes and increased histone acetylation were observed in the corpus callosum of mice infused with BMP4 during cuprizone-induced demyelination. We conclude that Shh and Bmp4 differentially regulate histone acetylation and chromatin structure in OPCs and that BMP4 acts as a potent inducer of gene expression, including Notch and Wnt target genes, thereby enhancing the crosstalk among signaling pathways that are known to inhibit myelination and repair.
Introduction
Oligodendrocytes are the myelin-forming cells of the CNS. Their differentiation from oligodendrocyte progenitor cells (OPCs) results from transcriptional changes in gene expression and is favored by sonic hedgehog (Shh; Orentas et al., 1999; Nery et al., 2001), thyroid hormone (Barres et al., 1994; Orentas et al., 1999; Pombo et al., 1999; Nery et al., 2001), IGF1 (Mason et al., 2000; Back et al., 2007) and retinoic acid (Huang et al., 2011) and antagonized by mitogenic stimuli such as PDGF and FGF (Barres et al., 1994; Baron et al., 2000; Fortin et al., 2005; Zhou and Armstrong, 2007), Wnt (Fancy et al., 2009; Feigenson et al., 2011), Notch (Wang et al., 1998; Zhang et al., 2009, 2010) and BMP4 (Grinspan et al., 2000; Gomes et al., 2003; See et al., 2004; Sim et al., 2006). The progression from OPCs to mature oligodendrocytes occurs in stages and is characterized by complex cytoskeletal (Liu et al., 2003) and transcriptional changes (Dugas et al., 2006; Cahoy et al., 2008; Swiss et al., 2011) and progressive chromatin compaction (Mori and Leblond, 1970; Peters and Sethares, 2004; Menn et al., 2006). We and others (Marin-Husstege et al., 2002; Shen et al., 2005, 2008; Liu et al., 2006, 2010; Ye et al., 2009) have previously reported that changes of acetylation levels in nucleosomal histones, the basic unit of chromatin, are critical for oligodendrocyte differentiation in the developing brain and in repair after demyelination in adult mice. However it is not clear whether extracellular signals with antagonistic roles in oligodendrocyte differentiation, also differentially regulate histone acetylation. In this study we asked whether two extracellular signals that either favor (i.e., Shh) or inhibit (i.e., BMP4) OPC differentiation into oligodendrocytes, modulate changes in gene expression by differential effects on histone acetylation.
Shh belongs to the hedgehog family of factors, which promotes oligodendrogliogenesis (Orentas et al., 1999; Nery et al., 2001; Soula et al., 2001; Tekki-Kessaris et al., 2001), while BMP4 is a member of the TGFβ super family of ligands which promotes astroglial lineage determination (Gross et al., 1996) at the expense of oligodendrogliogenesis (Wada et al., 2000; Miller et al., 2004; Samanta and Kessler, 2004) and prevents differentiation of OPC (Grinspan et al., 2000; See et al., 2004). More recently, several studies have suggested that BMP4 also affects neurogenesis and gliogenesis in the adult brain (Colak et al., 2008; Jablonska et al., 2010) and modulates repair after demyelination (Cate et al., 2010; Sabo et al., 2011). Therefore, we also asked whether in vivo BMP4 would induce similar changes to those detected in vitro, in cultured oligodendrocytes and affect histone acetylation and gene expression.
Materials and Methods
Antibodies and reagents.
Acetyl- Histone 3 (06–599, Millipore); Histone 3 (ab1791, Abcam); Histone H3K9Ac (ab4441, Abcam); A2B5, O4 and O1 antibodies (conditioned medium from hybridoma cells provided by Dr. R. Bansal, University of Connecticut, Farmington, CT); HDAC1 (PA1–186, ABR); HDAC2 (ab7029 Abcam); HDAC2 (SC-7899, Santa Cruz Biotechnology); HDAC3 (SC-11417, Santa Cruz Biotechnology); HDAC8 (SC-11405, Santa Cruz Biotechnology); GFAP (Z0334, Dako); GFAP (AVES Laboratories INC.); Caspase-3 (AB3623, Millipore); Shh (461-SH, R&D Systems); BMP4 (314-BP, R&D Systems); Noggin (719-NG, R&D Systems); valproic acid (VPA) (P4543, Sigma); XAV 939 (3748, Tocris Bioscience); iCRT 14 (4299, Tocris Bioscience), SAHM1(491002, EMD Chemicals).
Cell culture.
Oligodendrocyte progenitors were isolated from the cortex of postnatal day 1 rats, and cultured according to McCarthy and de Vellis (1980). After shaking from the flasks, progenitors were immunoselected using A2B5 antibodies and secondary antibodies conjugated with magnetic beads (Miltenyi Biotech). OPCs were maintained proliferating in the presence of bFGF (20 ng/ml) and PDGF (10 ng/ml), while oligodendrocyte differentiation was induced by culturing the cells in the absence of mitogens. Treatment with VPA (1 mm), Shh (1 μg/ml), BMP4 (50 ng/ml) or Noggin (1 μg/ml) was initiated when the culture medium was replaced with mitogen-free medium.
Time-lapse video microscopy.
For the time-lapse imaging of Shh or BMP4-induced OPC differentiation, primary oligodendrocyte progenitors were plated at a density of 2 × 104 into 8-well cover glass bottom slides (Nunc) and treated with either Shh (1 μg/ml) or BMP4 (50 ng/ml). Live cell images were captured every 5 min for the first 4 h of treatment, using the AxioObserver Z1 inverted fluorescence microscope equipped with controlled humidity, temperature and CO2 incubation system (Carl Zeiss MicroImaging). Cells were then allowed to re-equilibrate pH and CO2 levels in the incubator for 23 h and the following day, additional live cell images were captured every 5 min for an additional 4 h.
HDAC activity assay.
HDAC activity was measured by HDAC Activity Assay/Drug Discovery Kit (BIOMOL Research Laboratories). Experimental procedures were designed and performed according to the protocol provided (BIOMOL Research Laboratories). Lysates from cultured cells were used as sources for HDAC activity. Sample lysates containing 20 μg of protein were added to a 96-well plate in 25 μl HDAC assay buffer (BIOMOL Research Laboratories). A fluorimetric acetylated substrate was added and the reaction was allowed to proceed at room temperature for 1 h, followed by an incubation with developer for 10–15 min. Enzymatic activity was measured using a microtiter plate-reading fluorimeter (excitation at 360 nm, detection of emitted light at 460 nm).
Immunocytochemistry.
Cells were grown on CC2-coated 8 well chambers (Lab-Tek) for all immunocytochemistry. For staining oligodendrocyte lineage markers, cells were rinsed gently with PBS and incubated live with A2B5, O4, or O1 hybridoma supernatant (1:10) for 30 min at 37°C. Cells were then fixed with 1% paraformaldehyde for 20 min at room temperature and first incubated with pageing solution (PGBA plus 10% normal goat serum) for 30 min followed by incubation with secondary antibodies for 1 h at room temperature. For costaining with antibodies against AcH3, GFAP, class I HDACs, cells were equilibrated with Triton X-100 containing pageing solution (PGBA plus 10% normal goat serum, and 0.5% TX-100) for 30 min, then processed for additional primary antibody staining followed by the appropriate secondary antibody. Cells were then counterstained with 4′, 6′-diamidino-2-phenylindole (DAPI; 1:5000; Invitrogen) to visualize cell nuclei. Labeled cells were counted in at least three fields, from two or three separate cultures performed in duplicate, using a confocal microscope (LSM710, Zeiss). The intensity of AcH3 fluorescence staining was measured in arbitrary units using NIH ImageJ software.
Induction of demyelination.
Cuprizone-mediated demyelination was induced by feeding 8-week-old female C57BL/6 mice powdered feed (Barastoc) containing 0.2% cuprizone (w/w: bis-cyclohexanone-oxaldihydrazone; Sigma-Aldrich; Alfa Aesar) daily as previously described (Sabo et al., 2011).
Immunohistochemistry.
Mice were anesthetized and perfused intracardially with PBS followed by 4% paraformaldehyde. Brains were postfixed for 30 min at 4°C and then equilibrated in 20% sucrose in PBS overnight at 4°C for cryoprotection before being embedded in Tissue-Tek O.C.T. (Sakura), frozen using isopentane (Ajax Finechem) on dry ice, and stored at −80°C. For fluorescence microscopy, 10 μm coronal cryostat sections were probed with platelet-derived growth factor α receptor (PDGRA; rat; eBioscience; 1:500) and Acetyl-Histone H3 (AcH3; rabbit; Millipore; 1:200) overnight at room temperature followed by incubation with secondary antibodies for 1 h. The secondary antibody incubation included Hoescht 33342 (1:10,000; Invitrogen) to visualize the nucleus of all cells. BrdU detection was described previously (Cate et al., 2010). Cell counts were performed blind to experimental treatment. Immunopositive cells were quantified as previously described (Cate et al., 2010). Images were taken from 3 sections 50 μm apart for each animal at locations between bregma −1.58 mm and −1.82 mm as previously described.
Intraventricular infusion.
Acute demyelination was induced after 4-weeks of cuprizone ingestion and then mice received recombinant BMP4 (R&D Systems) dissolved in artificial CSF (aCSF) at a dose of 400 ng/d or aCSF delivered by ALZET osmotic pumps (Model 1007) at a flow rate of 0.50 ml/h for 5 days during cuprizone challenge. The ALZET Brain Infusion Kit III was used with the osmotic pumps to infuse via a cannula, which was implanted at coordinates previously described (Sabo et al., 2011). Mice undergoing surgery for cannula and osmotic pump implantation were monitored as previously described (Cate et al., 2010). Mice received 1 mg/ml BrdU in their drinking water for the final 3 days of the infusion.
In situ hybridization.
DIG labeled in situ riboprobes were generated by in vitro transcription from a ∼1.1kb cDNA clone (Open Biosystems). In situ hybridizations were performed on 10 μm cryostat sections with Hey1 probe according to standard protocols. Fluorescent in situ hybridization was performed with Fast Red (Roche) and followed by Acetyl-Histone H3 immunohistochemistry. The secondary antibody incubation included Hoescht 33342 (1:10,000; Invitrogen) to visualize the nucleus of all cells.
Electron microscopy for evaluation of nuclear condensation.
After 4 days of treatment, cells were fixed for 30 min with 4% glutaraldehyde in 0.1 m sodium cacodylate buffer with 1 mm CaCl2. The location of the cell on the coverslip grid was determined using brightfield illumination. The cells were then processed for transmission electron microscopy, washed and treated with 1% osmium tetroxide, 1.5% potassium ferracyanide in 0.1 m cacodylate buffer for 1 h at 4°C. After dehydration, cells were kept in 3% uranyl acetate in 70% ethanol for 12 h at 4°C, further dehydrated and then embedded (Embed 812 kit; Electron Microscopy Sciences) and sectioned. Sections were contrasted with lead citrate and uranyl acetate, and serial sections of the nucleus for the cell of interest were documented at magnifications of ×12,000 and ×50,000.
Lentiviral shRNA infection.
Hdac1, Hdac2, Hdac3 and Hdac8 Lentiviral shRNA Transduction Particles were purchased from Sigma-Aldrich.
The sequences of the shRNAs targeting the following mouse genes are
Hdac1: TRCN0000039401: CCGGCCCTACAATGACTACTTTGAACTCGAGTTCAAAGTAGTCATTGTAGGGTTTTTG
Hdac2: TRCN0000039396: CCGGGCTGTGAAATTAAACCGGCAACTCGAGTTGCCGGTTTAATTTCACAGCTTTTTG
Hdac3: TRCN0000039389: CCGGCGTGGCTCTCTGAAACCTTAACTCGAGTTAAGGTTTCAGAGAGCCACGTTTTTG
Hdac8: TRCN0000088000: CCGGGCAGATACTATTGCCGGAGATCTCGAGATCTCCGGCAATAGTATCTGCTTTTTG
ShRNA infection in primary OPCs was performed in cell suspension following harvesting. M.O.I 5 was used for the infection. Polybrene was added at a final concentration of 4 μg/ml to increase the infection efficiency. The virus/cell mixture was incubated in 37°C for 2 h, and plated at desired density. The next day, cells were changed to complete growth medium (PDGF, FGF) for 24 h, followed by subsequent experimental treatments.
Quantitative real time-PCR.
RNA was isolated following manufacturer's instruction and cleaned using RNeasy Mini kit (Qiagen; Transcripts were amplified from samples derived from primary rat OPC, using the primers listed in Table 1 and from mouse OliNeu cells, using the primers listed in Table 2).
Chromatin immunoprecipitation assay.
Olineu cells (10 × 106) treated with Shh or BMP4 were crosslinked in 1% formaldehyde for 10 min and lysed in cell lysis buffer [HEPES 5 mm, KCl 85 mm, NP40 0.5%, protease inhibitor cocktail 1× (p8340, Sigma), phosphatase inhibitor cocktail 1× (Sigma catalog #p0044), TSA 30 nm (Sigma catalog #T8552)] on ice for 15 min. The nuclei, were lysed in buffer containing Tris-HCl 50 mm, pH 8, EDTA 10 mm, SDS 1% and chromatin was fragmented using the Bioruptor (Diagenode) sonicator to produce an average size of 250–300 base pairs. Chromatin concentration was measured by its absorbance at 260 nm in the Nanodrop (1 U/ml of chromatin corresponds to OD260 = 1) and diluted 1:10 with dilution buffer (NaCl 165 mm, SDS 0.01%, Triton X-100 1.1%, EDTA 1.2 mm, Tris-HCl 16.7 mm, pH 8.0). One unit of chromatin was immunoprecipitated by protein A magnetic beads (Dynabeads-Invitrogen 100.01D) after incubation with antibodies specific for histone H3-Ac (Millipore 06–5999), histone H3K9Ac (Abcam ab4441), HDAC1 (Thermo Scientific PA1869), HDAC2 (Abcam ab7029). A nonantibody sample was used as negative control. Following overnight immunoprecipitation, beads were washed twice consecutively with the following buffers: Lio-B (HEPES 50 mm, pH 8.0, NaCl 140 mm, Triton X-100 1%, Sodium deoxycholate 0.1%, EDTA 1 mm), Hio-B (HEPES 50 mm, pH 8.0, NaCl 500 mm, Triton X-100 1%, Sodium deoxycholate 0.1%, EDTA 1 mm), LiCl (Tris-HCl 10 mm, LiCl 250 mm, NP-40 0.5%, Sodium deoxycholate 0.5%, EDTA 1 mm) and TE (Tris-HCl 10 mm, pH 8.0, EDTA 1 mm). Immunoprecipitated chromatin and input DNA were reverse crosslinked in elution buffer (Tris-HCl 50 mm, EDTA 10 mm, SDS 1%) in the presence of proteinase K (50 μg/ml) by shaking (1300 RPM) at 68°C using a thermomixer (Eppendorf) for 2 h. DNA was purified using phenol-chloroform and precipitated overnight in ethanol at −20°C. DNA pellets were dissolved in 200 μl of ddH2O and amplified by quantitative PCR, using primers specific for selected genomic regions, as indicated in Table 3.
Microarray data analysis.
RNA samples were extracted from cultured cortical rat OPCs treated for 24 h with either Shh (1 μg/ml), BMP4 (50 ng/ml) or with Shh (1 μg/ml)+ VPA (1 mm) and the isolated RNA was hybridized to the Rat Affymetrix 230.2 Gene Chip, containing 31,100 probe sets. Affymetrix CEL files were processed with Affy and Limma packages from the Bioconductor project in R (Gautier et al., 2004; Gentleman et al., 2004; Smyth, 2004). The MAS5 algorithm from Affy was used for data normalization and background correction. The statistical assessment of differentially expressed genes was performed using Limma. A false discovery rate (FDR) of 5% was chosen as statistical significance to define the genes that were differentially expressed.
Gene ontology analysis.
The gene ontology analysis was performed using DAVID database (Database for Annotation, Visualization and Integrated Discovery; Dennis et al., 2003; Huang da et al., 2009b). The final gene list, obtained from the intersection of genes upregulated by BMP4 and Shh+VPA, but not upregulated by Shh, was analyzed for enrichment of gene ontology terms found in the “biological process” library. Only those ontology terms with a significant p value of enrichment (<0.05) were considered for this analysis.
Transcription factor binding sites.
The software MatInspector from Genomatix (Genomatix Software) was used to search the binding sites for SMAD, Gli Homeodomain- and basic-helix-loop-helix family of transcription factors on the promoter regions of selected genes.
Statistical methods.
The statistical significance between two groups was determined by Student's t test. Data were expressed as mean ± SEM. The statistical significance is either described in figure legends, or indicated as asterisks (*). A p value of <0.05 was considered to be statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001).
Results
The opposing effects of Shh and BMP4 on the decision of OPC to differentiate along the oligodendrocytic or astrocytic lineage are associated with differences in nuclear chromatin
Our experimental system consisted of a homogeneous population of A2B5+ oligodendrocyte progenitors isolated from neonatal rat cortex by immunoselection using antibodies conjugated to magnetic beads. The relative proportion of A2B5-immunoreactive cells in this experimental system was >97.8%, with 0.47% being GFAP+ astrocytes and 0.01% microglial cells (Fig. 1A,B). These cells were cultured in the presence of either Shh or BMP4 for 5 days and then analyzed by immunocytochemistry using antibodies specific for the lineage markers O4 and O1 (to test progression into oligodendrocytes) and for GFAP (to test differentiation into astrocytes). In agreement with previous reports, Shh treatment for 5 days favored the generation of O4+ late oligodendrocytes progenitor and O1+ mature oligodendrocytes compared with chemically defined mitogen-free medium. BMP4, in contrast, induced the generation of GFAP+ astrocytes at the expenses of oligodendrocytes (Fig. 1C,D). Cotreatment with BMP4 and its receptor antagonist Noggin, attenuated the extent of astrocytic differentiation and promoted the generation of O4+ and O1+ mature oligodendrocytes (Fig. 1C,D). Interestingly, cotreatment with BMP4 and Shh prevented differentiation of progenitors into either lineage and retained the majority of the cells as undifferentiated and characterized by the expression of A2B5 (Fig. 1E,F).
To ascertain whether the effects of BMP4 and Shh were due to modulation of differentiation rather than a nonspecific effect on cell survival, we treated cells with either BMP4 or Shh for 1, 2, or 3 days and assessed the presence of apoptotic cells by colabeling with lineage markers (Fig. 2A). No difference was observed between the two treatment groups. We also assessed the total number of cells during the same time period and revealed that it was not statistically significantly different in the distinct treatment groups (Fig. 2B), thereby ruling out the possibility that the differences in number of mature oligodendrocytes and astrocytes observed after treatment with Shh or BMP4 could be consequent to selective death or amplification of specific subpopulations of cells. To further confirm that Shh and BMP4 act on the same cell types and modulate lineage choices we also performed live imaging of cultured OPCs during the first 32 h of treatment with BMP4 or Shh (Fig. 2C). On the first day of treatment, OPC cultures were homogeneous and characterized by cells with the typical bipolar morphology. No major morphological changes were detected during the first 4 h of treatment. Twenty four hours later, however, it was possible to distinguish between the BMP4-treated cells, characterized by few thick cytoplasmic expansions and a slight enlargement and flattening of the cell body and the Shh-treated cells, characterized by multiple thin processes and rounding-up of the soma (Fig. 2C). Together with the cell counts, the live imaging data supported the conclusion that the effect of Shh or BMP4 treatment on OPC was on lineage progression, rather than survival.
In addition to the morphological changes and the induction of cell-specific markers, Shh and BMP4 had distinct effects on nuclear chromatin. While Shh-treated cells were characterized by electron-dense chromatin aggregates distributed along the nuclear periphery, BMP4-treated cells retained a dispersed and homogeneously distributed euchromatin (Fig. 2D). Thus, the differential effect of Shh or BMP4 treatment on OPCs was associated with clear differences in nuclear chromatin.
Opposing effects of Shh and BMP4 on histone acetylation and nuclear chromatin conformation in cultured progenitors
It is well established that changes in nuclear chromatin are consequent to post-translational modifications of nucleosomal histones and therefore we asked whether Shh and BMP4 differentially modulated histone acetylation. Confocal analysis of cells stained with acetyl-H3 revealed strong immunoreactivity in BMP4-treated cells, weak immunoreactivity in OPCs treated with Shh or with BMP4 and its antagonist Noggin and intermediate levels of acetylation in OPCs that were cotreated with BMP4 and Shh (Fig. 3A,B). These immunocytochemical data were supported by measurement of nuclear HDAC enzymatic activity, which revealed higher deacetylation in OPCs treated with Shh for 24 h, compared with BMP4-treated cells (Fig. 3C). To evaluate the functional relevance of this finding, we treated cells with the pharmacological inhibitor of HDAC, valproic acid (Göttlicher, 2004), and evaluated the consequences on Shh-induced oligodendrogliogenesis or BMP4-induced astrogliogenesis (Fig. 3D,E). Pretreatment of cells with VPA for 24 h before exposure to Shh resulted in high levels of acetyl-histone H3 and prevented oligodendrocyte differentiation, while it did not affect BMP4-dependent astrogliogenesis (Fig. 3D,E). This suggested that HDAC activity is important for Shh-induced oligodendrocyte differentiation and dispensable for BMP4-induced astrogliogenesis.
Silencing of Hdac1 or Hdac2 prevents Shh-induced oligodendrocyte differentiation
To identify the specific HDAC isoforms downstream of Shh, we used a silencing approach with shRNA lentiviral constructs. Two days after infection with lentiviruses, OPC were incubated with Shh or BMP4 for 4 days, and silencing efficiency was confirmed by immunocytochemistry using antibodies specific for each HDAC isoform. Silencing of Hdac1 or Hdac2 dramatically reduced the percentage of O4+ cells from 84.95% (in mock infected cultures) to <3% (in silenced cells), and upregulated the astrocytic marker GFAP (Fig. 4A–D). Silencing of Hdac3 and Hdac8 had no effect on cell lineage progression (Fig. 4A–D). These results indicated that Hdac1 and Hdac2 isoforms were critical for Shh-induced oligodendrocyte differentiation, and that their silencing reversed the effect of this morphogen and favored the expression of astrocyte-specific genes. Silencing of Hdac isoforms however, had no significant effect on BMP4-induced astrocytic lineage progression and did not favor the expression of oligodendrocyte-specific genes in treated OPCs (Fig. 5A–D). Therefore, Hdac1 and 2 are important for repressing astrocytic genes during Shh-induced OL differentiation, but they are dispensable for BMP4-induced astrogliogenesis.
Genome-wide transcriptomic analysis reveals important crosstalk among key developmental signaling pathways
To begin unraveling genome-wide changes induced by Shh or BMP4, we adopted an unbiased approach and conducted a genome-wide transcriptional profiling, using the Rat 230.2 Affymetrix Gene Chip and RNA isolated from OPCs that were treated with Shh, BMP4 or Shh and VPA for 24 h. A heat map representation of gene expression profiles among the different groups of treatment (we averaged the results of three independent replicates for each treatment) was initially used to visualize the overall differences (Fig. 6A). Of the 31,100 probe sets present in the array, BMP4 upregulated 8451 genes and downregulated 5497 compared with untreated OPCs. Shh treatment, in contrast downregulated only 634 genes and upregulated 604 genes, but the presence of VPA in the medium resulted in the upregulation of 6075 genes, compared with untreated OPCs. These findings were consistent with the results of HDAC enzymatic activity and revealed that the progression from OPC to oligodendrocytes requires fewer transcriptional changes than the conversion to astrocytes and that HDAC activity in OPCs is responsible for the repression of other lineage genes. To further identify the BMP4 and Shh target genes whose transcriptional regulation is mediated by HDAC activity, we computed the intersection set between the genes upregulated by BMP4 and Shh-VPA, but not upregulated by Shh (Fig. 6B). This rationale was based on the repressive role of HDAC activity on gene expression. The functional classification of the resulting set (containing a total of 1821 genes) was performed using DAVID gene ontology analysis (Huang da et al., 2009a,b). Among the most relevant groups of differentially regulated genes in BMP4 and Shh-treated cells, we identified those involved in “developmental processes,” “signaling cascade,” “synaptic transmission,” and “negative regulation of transcription” (Fig. 6C). Within these groups, we noticed several genes associated with Notch signaling (Jagged1, Hey1, Hey2, Hes1, Hes5) and others associated with Wnt signaling (Tbx3; Fig. 6D). These results were validated by RT-PCR using RNA samples isolated from cultured OPCs (Fig. 6E) and suggested the intriguing possibility that the inhibitory effect of BMP4 treatment on oligodendrocyte lineage progression, could be partially mediated by crosstalk with important developmental signals such as Wnt and Notch, due to modulation of the acetylation state of the identified target genes (Fig. 7A). To test this possibility, we performed immunoprecipitation (ChIP) on chromatin isolated from BMP4 or Shh-treated Oli-neu cells and precipitated with antibodies specific for acetyl-H3 (total or K9-specific) or for Hdac1 and Hdac2. We asked whether differences in the recruitment of Hdac or in the acetylation of histone H3 could be found in the target genes Hey1, or Hey2, or Tbx3. Interestingly, the regulatory elements of these genes include the consensus sequence for Shh-activated transcription factors (i.e., Gli, E-box and homeobox) very closely located to the consensus sequence for BMP4 activated transcription factors (i.e., SMAD; Fig. 7A). We therefore reasoned that activation of Shh or BMP4 pathways would result in differential regulation of target genes by differential recruitment of histone-modifying enzymes to the same region of chromatin. In agreement with our prediction, using Chromatin immunoprecipitation, we detected recruitment of Hdac1 and Hdac2 at the regulatory regions of Hey1, Hey2, and Tbx3 in OPCs treated with Shh, which was consistent with lower levels of these transcripts (Fig. 7A). In contrast, BMP4-treated cells were characterized by high levels of these target genes and the increased transcription from these loci, nicely correlated with high levels of histone acetylation and decreased Hdac recruitment at the same chromatin loci. The results of this experiment further support the concept that histone acetylation of target genes is differentially modulated by BMP4 and Shh. This interpretation was also supported by a competition experiment, where Shh-treated cells were treated with increasing concentrations of BMP4 (Fig. 7B). Since Id2 and Id4 are downstream genes of BMP4 pathways, we used the levels of these transcripts as positive controls for BMP4 treatment and measured the levels of Notch target genes in these conditions. The results are in agreement with the Chip data and show upregulation of Hey1 and Hey 2 in a dose-dependent fashion in cells treated with Shh and increasing concentrations of BMP4 (Fig. 7B). Interestingly, the reverse experiment (i.e., constant concentration of BMP4 and increasing concentrations of Shh) was not equally effective (data not shown), therefore indicating that the presence of large transcriptional complexes containing histone acetyltransferase (like those recruited by BMP4 signaling pathway) are harder to displace than repressive complexes containing Hdac (like those found in Shh-treated cells).
To further exclude the possibility that the BMP4-induced expression of Wnt and Notch target genes was caused by the presence of Wnt or Notch ligands in cultured OPCs, we pretreated cells either with specific pharmacological inhibitors of the Wnt or Notch pathway before Bmp4 treatment. To page Wnt signaling we used the tankyrase inhibitor XAV939 (to decrease β-catenin availability, due to axin2 stabilization) or a small molecule inhibiting the binding of β-catenin to Tcf4 (i.e., iCRT14) and then measured the levels of Tbx3 transcripts. To page the Notch pathway, we used SAHM1, an inhibitor of the assembly of the NICD activator complex, and then measured the levels of Hey1, Hey2, and Hes5 transcripts (Fig. 7C). The levels of Tbx3 induced by BMP4 treatment were not affected by the presence of Wnt inhibitors of β-catenin. With the exception of Hes5 transcript levels, which were reduced by SAHM1 treatment, the other Notch-dependent genes (Hey1, Hey2) were expressed at similar levels in cells treated with BMP4 in the absence and presence of the Notch inhibitor (Fig. 7). Together with the ChIP results, these data strongly support the interpretation of a direct effect of BMP4 signaling on Notch and Wnt target genes.
To define whether BMP4 similarly modulated histone acetylation and the expression of Notch target genes (i.e., Hey1) in vivo, we studied OPCs in the adult brain of mice during myelin repair from cuprizone-induced demyelination. We have previously reported that repair occurs from the fourth week of cuprizone treatment and therefore mice were infused with vehicle or BMP4 into the lateral ventricle, using osmotic minipumps for 5 days, starting during the fourth week of cuprizone treatment. Acetylated-histone H3 was visualized by immunohistochemistry followed by confocal analysis and Hey1 transcripts were detected using in situ hybridization in the demyelinated corpus callosum at the end of the 5 days of BMP4 infusion. In agreement with the results in cultured OPCs, we also detected increased number of PDGFRA+AcH3+ double-positive cells in BMP4 infused mice compared with vehicle controls (Fig. 8A,B). These cells were also characterized by the presence of Hey1 transcripts, as shown by in situ hybridization (Fig. 8C). To further assess the consequences of BMP4 treatment on lineage progression we also performed GFAP staining and measured the number of immunoreactive cells, 1 week following the infusion. In agreement with previous reports (Sabo et al., 2011), we detected increased numbers of astrocytes and decreased numbers of oligodendrocytes. Within the demyelinated lesion, BMP4 infusion increased the number of astrocytes that were immunoreactive for acetylated-histone H3 and BrdU (Fig. 8D,E). Together these results are in agreement with the data in cultured oligodendrocytes and indicate that the actual levels of BMP4 surrounding areas of demyelination, might significantly impact repair.
Discussion
BMP4 and Shh in OPCs differentially modulate histone acetylation
Differentiation of OPCs results from the integration of extracellular signals and the chromatin state of a cell. OPCs are characterized by euchromatin and amenable to “accept” environmental signals (i.e., Shh, BMP4, Notch, Wnt) to modulate gene expression and differentiation. Previous studies reported impaired OPC generation after Shh loss-of-function (Davies and Miller, 2001) and ectopic oligodendrogenesis after Shh gain-of-function (Nery et al., 2001; Yung et al., 2002). BMP4 was shown to increase the number of astrocytes at the expenses of oligodendrocytes (Gomes et al., 2003) and inhibit later stages of differentiation (Grinspan et al., 2000) and myelin gene expression (Wada et al., 2000; Miller et al., 2004; Samanta and Kessler, 2004; See et al., 2004). In this study we describe the effect of BMP4 and Shh on nuclear chromatin and gene expression in OPCs. We show that Shh activates Hdac activity and promotes oligodendrocyte differentiation and peripheral chromatin compaction, while BMP4 prevents oligodendrogliogenesis and favors histone acetylation on the promoter of several transcriptional inhibitors (including genes that are known down-stream effectors of Notch or Wnt).
BMP4 signals via serine-threonine receptors (BMPRIa, Ib and II) that activate Smad1/5/8 proteins (R-Smads), associate with the common mediator-Smad 4 (co-Smad; Nohe et al., 2004; Liu and Niswander, 2005) and activate gene expression by interacting with transcriptional coactivators (Feng et al., 1998; Janknecht et al., 1998). Binding of Shh to the Ptchd1 receptor, releases the inhibition of the transmembrane protein Smoothened and allows the translocation of Gli2/3 transcriptional activators into the nucleus, where they induce the expression of Shh-target genes.
A competitive mechanism between Shh and BMP4 was identified in OPCs (Samanta and Kessler, 2004) and involved the functional sequestration of Shh target molecules (i.e., Olig2) by the inhibitors of differentiation protein Id2 and Id4, which are induced by BMP4 (Rivera and Murre, 2001; Wang et al., 2001; Yokota, 2001; Kondo, 2009). In mouse cerebellar cultures, in contrast, BMPs interfere with the Shh-induced proliferation by decreasing the levels of Gli1 and Smo (Rios et al., 2004), while in the chick spinal cord, BMP has been shown to repress Shh target genes Olig2 and Nkx2.2 (Mekki-Dauriac et al., 2002). In the adult brain the inhibitory effect of BMP4 on oligodendrogliogenesis has been attributed to the effect of Smad4 on the expression of Olig2. In a series of elegant studies on mice lacking Smad4 in neural stem cells, higher numbers of Olig2+ cells and oligodendrocytes were detected, a finding that was replicated by infusion of noggin (Colak et al., 2008).
This manuscript identifies an additional layer of regulation of OPC differentiation by Shh and BMP4 by exerting opposite effect on histone acetylation and chromatin compaction, thereby inducing genome-wide changes of OPC transcriptome.
Transcriptomic analysis of BMP4-treated oligodendrocyte progenitors reveals the presence of Notch and Wnt target genes
Our analysis revealed that BMP4 treatment favored a relaxed and transcriptionally competent chromatin structure associated with the induction of thousands of genes during the first 24 h. In contrast, Shh-treatment activated Hdac activity and induced peripheral condensation of chromatin and induced repression and activation of hundred genes during the progression from OPC to oligodendrocytes. A potential interpretation of these findings is that Shh favors Hdac activity to allow oligodendrocyte lineage differentiation while BMP4 decreases Hdac activity and massively turns on the transcription of large gene programs that inhibit oligodendrocyte while favoring astrocyte generation. In support of this interpretation, we noticed that 73% of the BMP4 activated genes corresponded to transcripts that were repressed by Hdac activity in oligodendrocytes. Silencing of Hdac1 or Hdac2“released” this inhibition and resulted in the expression of astrocytic genes in Shh-treated cultures. Gene ontology analysis of the genes activated by BMP4 and repressed during the progression of OPC to oligodendrocytes, revealed genes involved in synaptic transmission, Notch and Wnt signaling pathways.
The BMP4-mediated induction of Wnt target genes was not surprising if we take into consideration the fact that BMP4 and Wnt are both expressed in dorsal regions of the developing neural tube and have both been shown to prevent oligodendrocyte differentiation in vivo (Shimizu et al., 2005; Kim et al., 2008; Fancy et al., 2009; Feigenson et al., 2009) and in vitro (Feigenson et al., 2011). In addition, OPCs lacking both BMPr1A and BMPr1b lost also the ability to respond to Wnt3a, while cells lacking β-catenin, an essential component of the Wnt pathway, retained their ability to respond to BMP4, further supporting the existence of a cross talk between BMP4 and Wnt (Feigenson et al., 2011). These data suggest that BMP4 treatment may render the cell more responsive to the inhibitory effect of Wnt, possibly by upregulating specific transcription factors, such as Tbx3 (Eblaghie et al., 2004), although the enhanced responsiveness remains to be experimentally tested. Tbx3 encodes for a T-box-containing transcription factor, whose expression has been previously shown to be differentially regulated by BMP4 and Shh during specification of the identity of the posterior digits in the chick (Suzuki et al., 2004). Modulation of Tbx3 levels by BMP4 was reported in the developing retina (Behesti et al., 2006) and in the adult brain (Eriksson and Mignot, 2009), although its role in the nervous system has not been clearly elucidated, one possibility is that Tbx factors act as repressors of Shh (Manning et al., 2006), thereby decreasing the responsiveness of progenitors to ventral-derived signals. Alternatively, Tbx3 could suppress the Arf locus, which encodes for tumor suppressor and cell cycle inhibitory molecules and thereby favor proliferation (Brummelkamp et al., 2002).
Our study also detected the BMP4-mediated induction of Notch target genes, including Jagged1 and the downstream effectors Hey1, Hey2, Hes1, Hes5. The role of Notch signaling in gliogenesis (Gaiano et al., 2000) is well established and in the oligodendrocyte lineage, Notch has been linked to maintenance of the undifferentiated progenitor state during development (Wang et al., 1998) and defective repair after demyelination in the adult brain (Zhang et al., 2009). The downstream targets belong to the Hairy Enhancer of Split family of bHLH transcriptional repressors whose role has been linked to maintenance of neural stem cells (Kageyama and Nakanishi, 1997; Wang et al., 1998; Lai, 2004), inhibition of oligodendrocyte differentiation (Wang et al., 1998) and myelin gene expression (Liu et al., 2006) and possibly increased proliferation, consequent to the repression of p27Kip1 (Havrda et al., 2006), a cell cycle inhibitor which has been previously shown to modulate cell cycle exit in OPCs (Casaccia-Bonnefil et al., 1997, 1999).
Role of BMP4 in the adult brain and potential implications of our findings for repair
Finally, it is important to mention that inhibition of BMP signaling has been implicated for oligodendrogliogenesis during human development (Sim et al., 2006) and high BMP4 levels have been reported in human brains from patients with multiple sclerosis (Deininger et al., 1995). High levels of BMP4 were also detected in animal models of spinal cord injury, and correlated with increased astrocyte generation (Wang et al., 2011) and severe glial scar formation (Fuller et al., 2007). BMP4 levels were also upregulated in mice with induced experimental allergic encephalomyelitis (Ara et al., 2008) and in the cuprizone-induced model of toxic demyelination and they correlated with a dose-dependent increase in astrocyte numbers at the expenses of oligodendrocytes (Cate et al., 2010). The inhibition of BMP4 signaling by infusion of Noggin reversed the effect and reduced the astrocyte numbers (Cate et al., 2010) while promoting oligodendrocyte regeneration and remyelination (Sabo et al., 2011). In addition to the role of BMP4 in astrogliogenesis, it is important to remember that BMP also modulates adult neurogenesis in neural stem cells residing in the subventricular zone (SVZ) and that ventricular infusion of noggin favors oligodendrocyte generation, at the expenses of neurogenesis (Colak et al., 2008). Recent studies in the SVZ after lysolecithin model of demyelination in the corpus callosum have further revealed that another inhibitor of BMP signaling, chordin, is upregulated and that this favors the transition from neurogenesis to gliogenesis, although both, astrogliogenesis and oligodendrogliogenesis were favored (Jablonska et al., 2010). Our studies contribute to this field by revealing an additional role of BMP4 in the adult brain, where it also promotes the expression of Notch target genes, presumably involved in the proliferative response to BMP4 infusion.
In conclusion, our study reveals histone acetylation and chromatin changes as points of convergence for Shh and BMP4 in OPC differentiation. We demonstrate that Hdac activity is required for the suppression of transcriptional programs activated in other lineages during Shh-induced differentiation. In contrast, the effect of BMP4 on the oligodendrocyte lineage does not rely on repressive regulation of genes; rather, it requires extensive activation of gene expression and the upregulation of downstream signaling molecules of pathways that also exert a negative modulation of oligodendrocyte differentiation. Together these results suggest that inhibitory strategies aimed at dampening BMP4 effect may be important to favor myelin formation in the developing and diseased brain.
Footnotes
This study was supported by Grant NIH-R01-NS 42925-10 to P.C. and by a Minority Fellowship Supplement to M.H. (NIH-R01-NS 42925-08S2). J.K.S. is supported by a Multiple Sclerosis Research Australia Postgraduate Scholarship, and H.S.C. is supported by the National Health and Medical Research Council of Australia.
The authors declare no competing financial interests.
- Correspondence should be addressed to Patrizia Casaccia, Mount Sinai School of Medicine, One Gustave Levy Place, Box 1065, New York, NY 10029. patrizia.casaccia{at}mssm.edu