Abstract
Oligodendrocyte progenitors respond to biophysical or mechanical signals, and it has been reported that mechanostimulation modulates cell proliferation, migration, and differentiation. Here we report the effect of three mechanical stimuli on mouse oligodendrocyte progenitor differentiation and identify the molecular components of the linker of nucleoskeleton and cytoskeleton (LINC) complex (i.e., SYNE1) as transducers of mechanical signals to the nucleus, where they modulate the deposition of repressive histone marks and heterochromatin formation. The expression levels of LINC components increased during progenitor differentiation and silencing the Syne1 gene resulted in aberrant histone marks deposition, chromatin reorganization and impaired myelination. We conclude that spatial constraints, via the actin cytoskeleton and LINC complex, mediate nuclear changes in oligodendrocyte progenitors that favor a default pathway of differentiation.
SIGNIFICANCE STATEMENT It is recognized that oligodendrocyte progenitors are mechanosensitive cells. However, the molecular mechanisms translating mechanical stimuli into oligodendrocyte differentiation remain elusive. This study identifies components of the mechanotransduction pathway in the oligodendrocyte lineage.
Introduction
Myelin is critical for the proper function of the CNS as it insulates axons, allowing rapid synaptic transmission and providing trophic support to neurons (Rinholm et al., 2011; Fünfschilling et al., 2012; Lee et al., 2012). Myelinating oligodendrocytes differentiate in response to soluble chemical signals (Hart et al., 1989; Barres et al., 1994; Gao et al., 1998; Pombo et al., 1999; Lu et al., 2000; Nery et al., 2001; Samanta and Kessler, 2004; Fortin et al., 2005; Zhou and Armstrong, 2007; Furusho et al., 2011; Dugas et al., 2012; Wu et al., 2012) as well as physical perturbations (Rosenberg et al., 2008; Jagielska et al., 2012), but the pathways connecting spatial constraints to transcriptional changes have not been identified. Previous work on epigenetics of developmental myelination highlighted the importance of heterochromatin formation during oligodendrocyte differentiation (Marin-Husstege et al., 2002; Ye et al., 2009; Magri et al., 2014; Liu et al., 2015), with an emphasis on histone deacetylation and repressive trimethylation of lysine residues (Marin-Husstege et al., 2002; Shen et al., 2005; Ye et al., 2009; Liu and Casaccia, 2010; Sher et al., 2012; Liu et al., 2015).
In this study we sought to identify the molecules responsible for converting mechanostimulation (i.e., response to spatial constraints in the presence of constant soluble signals) into changes of the epigenetic landscape of oligodendrocytes.
Materials and Methods
Antibodies and reagents.
Reagents used in this study were as follows: methylated K9 histone 3 (H3K9me3, ab8898); phalloidin (Alexa Fluor 546 phalloidin; A22283, ThermoFisher Scientific); Deoxyribonuclease I (D12371, ThermoFisher Scientific); MBP (SMI-99, Covance); SUN1 (ab74758), SYNE1 (ab24742), emerin (ab40688), MAG (with methanol pretreatment; MAB1567, Millipore Bioscience Research Reagents), cyclic nucleotide phosphodiesterase (CNP; SMI-91R, Covance), actin (clone AC40; A4700, Sigma-Aldrich), DAPI (D1306, ThermoFisher Scientific). The dilutions used for these reagents were based on the manufacturer's instructions.
Animals.
PDGFRα-H2B-EGFP mice were a generous gift of Dr. Philippe Soriano (Icahn School of Medicine at Mount Sinai) and were handled in accordance with Institutional Animal Care and Use Committee-approved protocols. Mice were maintained in a pathogen-free environment at the Icahn School of Medicine at Mount Sinai animal facility. All animal procedures were approved by the Institutional Animal Care and Use Committee at Icahn School of Medicine at Mount Sinai. Timed pregnancy mice were purchased from The Jackson Laboratory.
Oligodendrocyte cultures.
Oligodendrocyte progenitor cells (OPCs) were isolated from postnatal day (P) 6 male and female mice. After immunopanning for negative (BSL1 or RAN2/GC) and positive selection (PDGFRA or O4), cells were kept proliferating in PDGF (10 ng/ml) and FGF (20 ng/ml) for 5 d. Then mitogens were removed and T3 (30 ng/ml) was added to induce differentiation into oligodendrocytes (Emery and Dugas, 2013).
DRG neuron–oligodendrocyte cocultures.
Cocultures of DRG neurons and oligodendrocytes with and without microspheres were performed as described by Rosenberg et al. (2008).
Mechanostimulation using the cell-shortening device.
Before plating the cells, silicone sheets (0.20 inch NRV G/G 40D, SMI Silicone Sheeting) were washed, sterilized with UV light, assembled into a mechanical device built in-house, stretched, and coated with Matrigel (catalog #BD 356230, Corning) for 30 min at 37°C. As control, we used nonstretched silicon sheets on a culture plate. We seeded 250,000 cells per sheet (either stretched or not). Cells were allowed to attach at 37°C before adding medium to the culture plates. Mechanical stimulation was performed 24 h later by removing a spacer between the silicone membrane and the device, to deliver mechanical compression to the cells. Fixation was performed with 4% paraformaldehyde in phosphate buffer.
Immunocytochemistry.
After blocking nonspecific staining (phosphate buffer supplemented with gelatin, bovine albumin, and 10% normal goat serum and 0.05% Triton X-100), cells were incubated with primary antibodies overnight at 4°C, visualized by incubation with labeled secondary antibodies, and counterstained with 4′,6′-diamidino-2-phenylindole (DAPI; 1:2000) to identify nuclei. Monomeric G-actin was stained using Deoxyribonuclease I. F-actin was stained with phallodin. Labeled cells were imaged in five fields from three separate cultures using a confocal microscope (LSM710, Zeiss) or a fluorescence microscope with Apotome (Zeiss Axio Imager Z1). H3K9me3 and G-actin fluorescence staining was measured in arbitrary units, using National Institutes of Health ImageJ software.
Immunohistochemistry.
For immunohistochemistry, animals were perfused with 4% paraformaldehyde. After overnight postfixation in the same solution at 4°C, the tissue was cryopreserved in 30% (w/v) sucrose, embedded in optimal cutting temperature compound and sectioned (12 μm). Frozen sections were first permeabilized with blocking buffer [0.1 m phosphate buffer, 10% (v/v) normal goat serum (Vector Laboratories), and 0.5% (v/v) Triton X-100], followed by antigen retrieval using 10 mm citrate buffer, pH 6.0, for 1 h at 65°C. Incubation with primary antibodies was performed overnight at 4°C. The next day, tissue slices were incubated with the appropriate secondary antibody for 1 h. Monomeric G-actin was stained with Deoxyribonuclease I according to the vendor's instructions. Sections were then counterstained with DAPI (1:20,000) to visualize nuclei. Stained tissue was imaged in using a confocal microscope (LSM710, Zeiss).
Coimmunoprecipitation and Western blot.
Protein extracts were obtained by lysis in 10 nm HEPES, 5 mm EDTA, 50 mm NaCl, 1% Triton × 100 containing a protease inhibitor mixture (catalog #11836153001, Roche) for 20 min at 4 C. Immunoprecipitation of 2 mg of proteins was performed overnight using 2 μg of the corresponding antibody. Immunoprecipitated samples and whole-cell lysates were separated by SDS-PAGE and transferred to a PVDF (Millipore) membrane using a transfer buffer composed of 25 mm Tris base, pH 8.3, 192 mm glycine, 20% methanol for 75 min at 100 V at 4°C. Western blot analysis was performed using the appropriate dilution of primary antibodies. Immunoreactive bands were visualized using horseradish peroxidase-conjugated secondary antibodies (GE Healthcare).
Quantitative real-time PCR.
RNA was isolated following manufacturer's instruction and cleaned using RNeasy Mini kit (Qiagen) and amplified using the following primers for real-time PCR: for Syne1: GGGAGAGGCAGGAACGTCGC and GCGTTCGTGGCTGGAACCTG; for Sun1: TCAAGCACTGCGAAGGACTA and CTCCCTAGCCACCAGAAGGT: for GAPDH: ACCCAGAAGACTGTGGATGG and CACATTGGGGGTAGGAACAC.
Lentiviral infection.
Silencing of mouse primary OPCs was obtained using lentiviral transduction particles from Sigma-Aldrich (shRNAs targeting sequences for Syne1: TRCN0000239245; Emd: TRCN0000240585, and TurboGFPcontrol: SHC003V) in suspension at a multiplicity of infection of five with polybrene (2 μg/ml). After 2 h at 37°C, cells were plated at the desired density. For cocultures, OPCs were first seeded onto 10 cm, poly-lysine-coated dishes and incubated with shRNA lentiviruses overnight at 37°C. Infected OPCs were then trypsinized and seeded onto DRG cultures with or without beads at the desired density.
Statistical methods.
Student's t test was used to determine the statistical significance between two groups of independent variables.
Results
To define the molecular players transducing spatial constraints into epigenetic signals in OPCs, we compared the effect of three distinct stimuli: high cell density, presence of inert microspheres, and application of mechanostimulation using a mechanical device in oligodendrocyte cultures. We observed that in cocultures of OPCs with neurons, inert microspheres or high-density plating of OPCs resulted in increased myelination, as demonstrated by the staining of three myelin markers: MBP, CNP, and MAG (Fig. 1A,B; Rosenberg et al., 2008). Myelination was also accompanied by changes in nuclear structure, characterized by the presence of H3K9me3-immunoreactive heterochromatin (Fig. 1C) and a higher number of chromocenters (DAPI puncta underneath the nuclear membrane) in oligodendrocyte nuclei (Fig. 1E,F). Interestingly, we observed that G-actin subcellular distribution was also influenced by cell density. Nuclear and perinuclear G-actin was significantly increased in high-density conditions (Fig. 1D,G) and this was consistent with the redistribution of actin previously described in other cell types (Maniotis et al., 1997; Hirata et al., 2014; Li et al., 2014; Nagayama et al., 2014).
To directly address the effect of mechanical forces on nuclear heterochromatin in OPCs, we also plated them on a Matrigel-coated prestretched silicone substrate, mounted on a portable device (Fig. 2A) with the ability to impose uniaxial cell deformation upon mechanical release of the substrate (Costa et al., 2002). The mechanical stimulation resulted in the instantaneous transitioning of the cells from a spread state to a compressed configuration, which was accompanied by the reorganization of nuclear structure, as shown by the number of peripheral chromocenters (Fig. 2C) and the increased nuclear/perinuclear G-actin (Fig. 2B). Therefore, acute application of mechanical force to OPCs was sufficient to redistribute actin and transduce a signal to the nuclear compartment.
A potential connection between cell membrane and nuclear components is provided by components of the linker of nucleoskeleton and cytoskeleton (LINC) complex (Anno et al., 2012; Martins et al., 2012; Jahed et al., 2014; Kaminski et al., 2014), which is composed of SUN (Sad1, UNC-84) and KASH (Klarsicht, ANC-1, and Syne homology) domain-containing proteins (Starr and Fischer, 2005; Razafsky and Hodzic, 2009; Sosa et al., 2012). The expression of Sun1 and Syne1 was first evaluated during the differentiation of OPCs into oligodendrocytes in vitro, which revealed increased transcripts and protein levels (Fig. 3A, B). Similar results were obtained in vivo in the corpus callosum of mice expressing H2B-EGFP under the control of the Pdgfra promoter (Fig. 3C,D). Because SYNE1 has the ability to bind actin with its large N-terminal domain and to bind the nuclear protein SUN1 with its C-terminal domain (Ostlund et al., 2009), we conducted immunohistochemistry of the developing corpus callosum, using antibodies specific for SUN1 and actin at two developmental time points before (i.e., P1) and during myelination (i.e., P21; Fig. 3E). In agreement with the involvement of LINC complexes in myelination, nuclear colocalization of actin and SUN1 was only detected at P21 (Fig. 3E). Consistent with the immunohistochemical data, we detected the formation of the LINC complex in protein extracts from the CNS of P21 mice, which, in addition to the interaction between SYNE1 and SUN1 and actin, also revealed binding with nucleosomal histones characterized by the H3K9me3 mark (Fig. 3F). To further assess the specificity of complex detection, we showed that protein extracts immunoprecipitated with anti-SUN antibodies did not bind the unrelated nuclear protein emerin (Fig. 3F). Together these results identify the LINC complex as a potential mediator of mechanoinduced heterochromatin formation in oligodendrocytes during myelination. This hypothesis was tested by silencing Syne1 using lentiviral vectors and evaluating the response of OPCs to mechanostimulation induced by the device. Decreased Syne1 levels in OPCs resulted in impaired heterochromatin (Fig. 4A) and in chromocenter formation (Fig. 4B) and aberrant subcellular distribution of G-actin after mechanical stimulation (Fig. 4C). To further define whether pathways transducing spatial information were independent of those activated by chemical signals, we also assessed the ability of Syne1-silenced OPCs to differentiate in response to T3 thyroid hormone treatment and compared these results to those obtained in the presence of artificial microspheres. Syne1 silencing had no effect on T3-induced differentiation (Fig. 4D) while it profoundly affected the prodifferentiative effect of microsphere plating (Fig. 4E). The effect was specific for components of the LINC complex, since silencing the nonrelated nuclear envelope protein emerin (Emd) did not affect the number of myelinated fibers (Fig. 4D,E). We conclude that space limitations in OPCs promote heterochromatin formation and differentiation by activating a pathway involving the formation of SYNE–SUN–actin complexes.
Discussion
OPCs respond to conditions associated with limited space, as induced by plating of microspheres (Rosenberg et al., 2008) with enhanced differentiation and increased myelinated fibers in coculture. They can also respond to cell density by changing nuclear shape, and to stiffness of the extracellular matrix (Jagielska et al., 2012) by readjusting migration and differentiation rates. While the evidence linking physical stimulation to functional effect is clear, the molecules responsible for transducing the effect of an external force into a program of differentiation remain elusive.
In this study we have shown that distinct modalities of compressive force applied onto cultured OPCs favor their differentiation into myelinating oligodendrocytes. In addition to high-density plating, compressive forces were generated either by interspersing inert microspheres with cultured OPCs or by using a mechanical device. All these conditions resulted in the formation of nuclear heterochromatin, a hallmark of differentiation and increased myelination. Because the nuclear structure dynamic has been recently shown to modulate transcription, DNA replication and repair, and RNA processing (Schneider and Grosschedl, 2007), we reasoned that the rearrangements induced by compressive forces would be sufficient to initiate a differentiation program.
We have highlighted the existence of a cytosolic connection-mediating force transmission from the plasma membrane onto the nucleus, demonstrated by the redistribution of G-actin to the nuclear/perinuclear space only in stimulated cells, revealing the possibility that actin dynamics in regions with greater mechanical stress could modulate the mechanotransduction pathway. It has been recently shown that F-actin depolymerization is critical for myelination and necessary for the formation of myelin wraps along the axons in the CNS (Nawaz et al., 2015; Zuchero et al., 2015). However, the concept of G-actin-mediated modulation of a transcriptional program accompanying oligodendrocyte morphological changes has not been reported. In this context, our study identifies SYNE1, a component of the LINC complex, as an important molecular link coupling cytoskeletal changes with heterochromatin formation in cultured cells.
Because all the different modalities of mechanostimulation assessed in this study resulted in increased heterochromatin, as shown by the detection of repressive histone marks and chromocenters, it is remarkable to identify the KASH domain-containing protein SYNE1 as a key signaling molecule for mechanotransduction in oligodendrocytes. The ability of SYNE1's N-terminal cytosolic domain to interact with cytosolic actin and of its C-terminal domain to interact with the nuclear periphery (Lombardi et al., 2011) and with nucleosomal histone carrying repressive marks further highlight SYNE1's importance in transducing a mechanical signal into nuclear changes with likely transcriptional consequences. Although the identity of the genes affected by these repressive marks is not part of this study, it is conceivable that oligodendrocytes might rely on mechanosensitive transcription factors as reported in other cell types (Mammoto et al., 2012; Wang et al., 2015).
Decreased myelination was only detected in Syne1-silenced OPCs that were induced to differentiate by biophysical constraints (i.e., plating with microspheres), while differentiation induced by mitogen withdrawal and T3 supplementation were unaffected, thereby revealing the specificity of the SYNE–SUN–actin complex for mechanotransduction.
In conclusion, our study identifies signaling molecules relating spatial constraints to the activation of oligodendrocyte differentiation. Our results reveal SYNE1 as a key mediator of heterochromatin formation and differentiation induced by mechanical stimuli. Based on its distinct binding domain, SYNE1 has the ability to “sense” actin dynamics in the cytoplasm and “transfer” the signal to the inner nuclear membrane and chromatin, thereby linking external forces of compression to epigenetic changes associated with heterochromatin formation. We have previously shown that removal of mitogens from the culture medium also induces heterochromatin formation (Liu et al., 2015) but that pathways of “chemically induced” differentiation do not rely on SYNE1 function. Therefore, OPCs respond to multiple modalities of differentiation by activating distinct signaling pathways to integrate spatial and chemical signals and induce heterochromatin formation and oligodendrocyte identity.
Because the viscoelasticity of the brain has been documented to change in response to demyelinating lesions (Schregel et al., 2012) and because OPCs are mechanosensitive (Jagielska et al., 2012), a thorough analysis of mechanotransduction signals in conditions of failed remyelination is worth considering. This study, together with previous reports, further highlights the complexity of remyelination and suggests that a combination of mechanical and chemical stimuli might contribute to the optimization of therapeutic approaches and strategies to address demyelinating conditions and promote remyelination.
Footnotes
This work was supported by Grant F31NS083344 to M.H. and Grants R01NS52738 and R37NS42925 to P.C. from the National Institute of Neurological Disorders and Stroke at the National Institute of Health. We thank Dr. M. Wu for initial experiments.
The authors declare no competing financial interests.
- Correspondence should be addressed to Patrizia Casaccia, Icahn School of Medicine at Mount Sinai, One Gustave Levy Place, Box 1065, New York, NY 10029. patrizia.casaccia{at}mssm.edu