Nitric Oxide‐Induced Neuronal to Glial Lineage Fate‐Change Depends on NRSF/REST Function in Neural Progenitor Cells

Degeneration of central nervous system tissue commonly occurs during neuroinflammatory conditions, such as multiple sclerosis and neurotrauma. During such conditions, neural stem/progenitor cell (NPC) populations have been suggested to provide new cells to degenerated areas. In the normal brain, NPCs from the subventricular zone generate neurons that settle in the olfactory bulb or striatum. However, during neuroinflammatory conditions NPCs migrate toward the site of injury to form oligodendrocytes and astrocytes, whereas newly formed neurons are less abundant. Thus, the specific NPC lineage fate decisions appear to respond to signals from the local environment. The instructive signals from inflammation have been suggested to rely on excessive levels of the free radical nitric oxide (NO), which is an essential component of the innate immune response, as NO promotes neuronal to glial cell fate conversion of differentiating rat NPCs in vitro. Here, we demonstrate that the NO‐induced neuronal to glial fate conversion is dependent on the transcription factor neuron‐restrictive silencing factor‐1 (NRSF)/repressor element‐1 silencing transcription (REST). Chromatin modification status of a number of neuronal and glial lineage restricted genes was altered upon NO‐exposure. These changes coincided with gene expression alterations, demonstrating a global shift toward glial potential. Interestingly, by blocking the function of NRSF/REST, alterations in chromatin modifications were lost and the NO‐induced neuronal to glial switch was suppressed. This implicates NRSF/REST as a key factor in the NPC‐specific response to innate immunity and suggests a novel mechanism by which signaling from inflamed tissue promotes the formation of glial cells. Stem Cells 2014;32:2539–2549


INTRODUCTION
Neural stem/progenitor cells (NPCs) are characterized by their ability to divide and differentiate into the three major cell types that constitute the mammalian brain: astrocytes, oligodendrocytes, and neurons [1]. In the normal adult mammalian brain, NPCs originating from the subventricular zone (SVZ) give rise to neuroblasts, which then migrate to the olfactory bulb and striatum [2,3]. However, upon cell loss in central nervous system (CNS) tissue, NPCs migrate toward the site of injury. These areas have been shown to possess a substantial number of newly formed oligodendrocytes and astrocytes, whereas neurons are rarely generated at these sites [3,4]. This suggests that the local environment is influencing specific NPC lineage fate decisions during neuroinflammatory conditions. NPCs of the SVZ are located in close proximity to the ventricle system and are generally believed to be exposed to the cerebral spi-nal fluid (CSF) and its contents. Moreover, a substantial portion of CSF cycles through the brain interstitial space during normal conditions, indicating constant CSF-contact to all brain regions including other NPC populations [5]. During neuroinflammatory conditions, such as multiple sclerosis (MS) and neurotrauma, close access to CSF that contains inflammatory components will potentially influence NPClineage fate decisions.
The free radical nitric oxide (NO) is produced in various tissues and functions normally both as mediator of neural transmission as well as a regulator of other physiological events, such as vascular toning and adult neurogenesis [6][7][8]. However, during innate immune responses, NO exerts other functions as it is released in excessive amounts through induction of the inducible form of nitric oxide synthase (iNOS) [9]. In contrast to the other two isoforms of NOS (endothelial NOS; eNOS and neuronal NOS; nNOS), the production of NO via iNOS-synthase in various http://dx.doi.org/ 10.1002/stem.1749 This is an open access article under the terms of the Creative Commons Attribution Non-Commercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

TISSUE-SPECIFIC STEM CELLS
cells is induced by the JAK/STAT pathway and regulated at a transcriptional level [10,11]. This accounts for a relatively slow production, but in excessive amounts that can be 1,000-fold higher than the NO released under normal noninflammatory conditions [10]. While excessive amounts of NO exert antibacterial, antiparasitic, and tumoricidal effects, induction of iNOS in microglia and astrocytes within the CNS induces damage to neurons and oligodendrocytes [12][13][14][15]. Importantly, NO production is excessively increased in the CNS under various neuroinflammatory conditions, such as MS, meningitis, and neurotrauma [16][17][18][19][20]. For example, in MS lesions iNOS expression is increased in astrocytes and, interestingly, the level of NO production in the human CSF correlates with disease severity [19,21,22]. Accordingly, this raises the question of how pathologically relevant levels of NO affect the NPC population of the adult brain. Previously, we showed that adult rat NPCs that had been exposed to clinically relevant but pathological levels of NO in vitro were less prone to form neurons [23]. Relatively short exposure induced fate changes from proneuronal to proglial fate, as demonstrated by the downregulation of proneural gene Neurogenin2 (Ngn2) and the concentration-dependent reduction of neuron formation in the differentiated cultures [23]. The transcription factor NRSF/REST (neuron-restrictive silencing factor-1/repressor element-1 silencing transcription) has been shown to play an essential role in the regulation of neurogenesis [24,25]. NRSF/REST is expressed in NPCs and in non-neuronal tissues during embryonic development, whereby it acts as a repressor of neuronal genes [26]. As the NPCs differentiate toward the neuronal lineage, NRSF/REST is downregulated. Interestingly, recent studies in developing rodent brain demonstrate the requirement of NRSF/REST for oligodendrocyte differentiation and formation of the proper ratio of neurons and glial cells [27,28]. NRSF/REST regulates target genes through the binding to specific RE-1 (repressor element-1) sites and the recruitment of cofactors including histone deacetylases, demethylases, and methyltransferases, which induce changes in chromatin status and nucleosome repositioning [29]. A number of different histone modifications have been coupled to NRSF/REST function, both of which are associated with active and repressed chromatin such as acetylation and trimethylation of Lysine 27 on Histone 3 (H3K27Ac and H3K27me3) [29,30].
Here, we show that NRSF/REST is upregulated in rat and human primary NPC cultures that have been exposed to pathological levels of NO. This coincides with alterations in chromatin status and gene expression of a number of neuronal and glial lineage-specific genes, demonstrating a shift toward glial potential of the NPC cultures. The changes in histone modification status were reversed by the introduction of a dominant negative version of NRSF/REST (dnREST). We provide functional evidence that NRSF/REST is required for the NO-induced neuronal to glial fate change in differentiating rat NPCs, suggesting a role for NRSF/REST in the innate immunity-driven NPC response.

Rat Neural Progenitor Cell Cultures, NO-Exposure, and Transfection
Neural progenitor cells were isolated from adult Dark Agouti (DA) rats (Scanbur B&K, Sollentuna, Sweden, http://www.scan-bur.eu/) and cultured in Dulbecco's modified Eagle's medium (DMEM)/F-12 medium (GIBCO cat no 31331-028, Life Technologies, Carlsbad, CA, www.lifetechnologies.com) with B27 supplement without vitamin A (Life Technologies, cat no 12587-010), Penicillin-Streptomycin (100 U/ml), 20 ng/ml of epidermal growth factor (EGF), and 10 ng/ml of Basic fibroblast growth factor (FGF2) (FGF2 and EGF; added every second day). For all experiments, cells had undergone two passages before they were plated onto poly(D-lysine) (PDL) glass coverslips. For NO-exposure, we used the NO-donor DETA-NONO:ate (Alexis Biochemicals, Goteborg, Sweden, http://www.axxora.com/) in a total concentration of 0.1 mM (corresponding to approximately 500 nM of NO) for 4-20 hours (see reference for measured NO concentration) [23]. As control medium, we used 0.1 mM of depleted DETA-NONO:ate (incubated for 10 days in 14 C before usage). After NO-exposure, cells were either processed for analysis (immunohistochemistry, Western blot, ChIP analysis, or qPCR) or cultured for 5 days under differentiation conditions (DDC) (containing 1% Fetal calf serum (FCS) and EGF, FGF withdrawal) before being processed for immunohistochemistry. For transfection experiments, we used Lipofectamine 2000 (Invitrogen, Life Technologies) and 300 ng/30,000 cells of either pCAGG-dnRESTMyc, pCAGG-REST, or pCAGG-GFP expression constructs [31] for 12-15 hours directly after NOexposure but before initiation of differentiation conditions. For siRNA experiments, we used 20 pmol siRNA/10,000 cells of either two different NRSF/REST siRNA (Silencer Select Predesigned siRNA, siRNA id s136125, and siRNA id s136124, 10 pmol of each used/10,000 cells, cat no 4390771, Life Technologies) or control siRNA (Silencer Select Negative Control No. 1, cat no 4390843, Life Technologies). Transfected cells were analyzed by immunohistochemistry either directly after transfection or after 5 DDC. All animal experiments were carried out according to the Karolinska Institutet rules and guidelines for animal care. The protocol was approved by the Stockholm Animal Ethics committee (permit number 379/10).

Cell Death Assay
Cells were fixed with 2% paraformaldehyde for 20 minutes at RT and measured for apoptosis by Click-IT TUNEL Alexa Flour 488 Imaging Assay (Invitrogen, cat no C10245) according to the manufacture's recommendation. Apoptotic cells were imaged using florescence microscopy and cell-counting was performed from five individual experiments.

Human Cell Cultures
Tissue samples from two adult patients that had undergone temporal lobe resection for treatment of epilepsy were received and samples for experiments were restricted to the wall of the lateral ventricle (SVZ area). Both patients underwent MR scans and pathology screening to exclude tumors and were screened for infectious diseases. The used tissue was considered biological waste and would otherwise be discarded. The research protocol was approved by the local ethical committee (permit number 01-294). Biopsies were transported in Hibernate-A medium (GIBCO, cat no A12475) from operation theater. The tissue was dissociated mechanically with a scalpel followed by ready-to-use Accutase (Sigma-Aldrich, cat no A6964) in 37 C for 10 minutes. Propagation culture conditions used were DMEM/F-12 medium (GIBCO cat no 31331-028) with B27 supplement without vitamin A (Life Technologies, cat no 12587-010), HEPES buffer (1 M, 0.9%, Gibco), Penicillin-Streptomycin (100 U/ml), 20 ng/ml of EGF, and 10 ng/ml of FGF2 (added two to three times weekly) and have been described previously [32]. After 4-6 weeks in propagation conditions, human neurospheres were dissociated into single cells by Accutase, plated on PDL-coated coverslips in 50 ml droplets, and left for 24 hours to attach following NO-exposure via DETA-NONO:ate 0.1 mM according to description for rat cells. Cultures were either analyzed directly after 24 hours of NOexposure or left for 11 days under differentiation conditions (propagation medium containing 1% FCS and EGF, FGF2 withdrawal) before fixation with 2% paraformaldehyde for 20 minutes at RT and immunohistochemistry analysis.

NRSF/REST Is Upregulated in NO-Exposed NPCs
To better understand how NO can act as an instructive cue in causing inflammation-induced NPC fate-change, we aimed to characterize the mechanism for the reduced neurogenesis [23]. Since the transcription factor NRSF/REST is known to prevent premature expression of neuronal genes in NPCs and it has been shown to be required for oligodendrocyte development in neonatal rat CNS [25,27,34], we performed a series of experiments to investigate the involvement of NRSF/REST in the changed neuronal/glial ratio observed in the NO exposed NPC cultures. NPCs from SVZ of adult DA rats were exposed to the NO-donor DETA-NONO:ate 0.1 mM or control medium (containing 0.1 mM of depleted DETA-NONO:ate) for 20 hours followed by fixation and characterization. For the used cell culture conditions, 0.1 mM DETA-NONO:ate corresponds to approximately 500 nM NO that sub-sequentially decreases to undetectable levels over a period of 24 hours [23]. Interestingly, we detected a more than twofold increase in the expression of NRSF/REST already 4 hours after initiation of NO-  (Fig. 1A). Furthermore, Western blot analysis confirmed an increase in NRSF/REST protein levels compared to control cultures after 20 hours of NO-exposure (Fig. 1B). Exposure to NO for 20 hours did not significantly alter the morphology of the cells, the expression of the progenitor marker Sox2 or the cell-cycle marker Ki67 (Fig. 1C-1J). Together, these experiments indicate that rat NPCs upregulate the expression of NRSF/REST, but are maintained in a selfrenewing progenitor state after exposure to pathological NO concentrations. Importantly, the concentration of NO obtained from 0.1 mM DETA-NONO:ate after 20 hours did not induce cell death in NPCs, as the number of TUNEL positive cells was 16% and 17% in NO exposed and control cultures, respectively (Supporting Information Fig. S1A-S1G).

Increased Number of Oligodendrocytes After NO Exposure
We have previously reported that NO-exposure to NPCs leads to decreased neuronal differentiation paralleled by an increase in astroglial differentiation, demonstrated by an increase in GFAP protein levels [23]. Since the focus in the previous study was on measuring protein levels, we now investigated the number of astrocytes and oligodendrocytes within the cultures. Following 5 days of differentiation of the NO-exposed NPC cultures, the percentage of cells expressing the oligodendrocyte marker O4 was increased in NO-exposed cultures compared to control (18% 6 4% and 8% 6 3% O41 cells, respectively) ( Fig.  2B, 2E, 2N) Furthermore, also the proportion of cells expressing the oligodendrocyte precursor marker Olig2 was increased after NO-exposure (Supporting Information Fig. S2A-S2F). This suggests that NO-exposed NPCs are more prone to The percentage of cells expressing the astrocyte marker GFAP was 82% 6 7% and 78% 6 9% after NOexposure and control conditions, respectively (Fig. 2J, 2M, 2P). The reduction in the percentage of neurons after NO-exposure was confirmed by neuronal markers Tuj1 and NeuN (Fig 2F, 2I, 2O and Supporting Information Fig. S2G-S2L). Importantly, the decrease in generated neurons was not due an increased celldeath or an inability of NO-exposed NPCs to initiate the differentiation process, as the majority of NO-exposed and control cells had downregulated the progenitor marker Sox2 after 5 days of differentiation (Supporting Information Fig. S3A-S3F).

Blocking of NRSF/REST Function Rescues the NO-Induced Effect on NPC Differentiation
To determine the involvement of NRSF/REST in the NOinduced neuronal-to-glial fate change, we next examined if the suppression of NRSF/REST activity would restore the neuronal/ glial ratio. To examine this we used a dominant negative variant of NRSF/REST (dnREST), which includes the Zn-finger domain of NRSF/REST (aa 203-440) fused to a Myc-tag for detection. The N-and C-terminal repression domains of NRSF/ REST had been deleted to avoid repression of target genes, while it retains its DNA binding capacity to the RE-1 target   [31]. Thus, dnREST competes with endogenous NRSF/REST for the binding to the RE-1 sites and thereby impedes NRSF/ REST from carrying out its function. Following 20 hours of NOexposure, rat NPCs were transiently transfected with dnREST expression construct or a control green fluorescent protein (GFP)-vector. A transfection efficiency of 80% was detected 24 hours post-transfection (Supporting Information Fig. S4A-S4G). After transfection, EGF and FGF were withdrawn and the NPCs were cultured in serum-containing medium and differentiated for 5 days (Fig. 2A). Interestingly, under conditions when the function of NRSF/REST was blocked, the capacity of NOexposed rat NPCs to differentiate toward the neuronal lineage was restored. After 5 days of differentiation, the proportion of Tuj1 expressing cells in dnREST-transfected populations was comparable to cells cultured in the absence of NO (Fig. 2G, 2I, 2O). This event was not observed in the GFP-transfected control cultures (Fig. 2H, 2O) where the effect from NO-exposure was comparable to nontransfected NO-exposed cultures (Fig.  2F, 2O). Moreover, in comparison to nontransfected and GFPtransfected NO-exposed cultures, the percentage of O4positive oligodendrocytes was decreased after dnRESTtransfection and was more comparable to unexposed differentiated NPC cultures, 6% 6 4% and 8% 6 3%, respectively ( Fig.  2B-2E, 2N). No significant changes in the percentage of GFAP expressing astrocytes were found in dnREST-transfected cultures ( Fig. 2J-2M, 2P). In addition to blocking the NRSF/REST function, we also investigated the effects of lowering the protein levels of NRSF/ REST with siRNA in the NO-exposed cultures (Fig. 3A). In line with the results from dnREST transfection experiments, we found a significant increase in Tuj1 expressing cells in the siRNA transfected NO-exposed cultures after differentiation (8% 6 3%), compared to cultures transfected with control siRNA (2% 6 1%) (Fig. 3B, 3C, 3E). Furthermore, during conditions when NRSF/REST siRNA was cotransfected with a NRSF/REST expression vector (entitled "REST"), the proportion of Tuj1expressing cells was further decreased (4% 6 3%) and was lower than in those cultures transfected with NRSF/REST siRNA only (Fig. 3D, 3E). In addition, we observed a trend of increased percentage of O4-expressing oligodendrocytes in the siRNAtransfected cultures compared to control siRNA (9% 6 5% and 17% 6 7%; p-value 5 .0514) (Fig. 3F-3I), whereas the proportion of O4-positive cells in the REST cotransfected rescue experiment was 15% 6 4% (Fig. 3H, 3I). The number of GFAP expressing astrocytes remained unchanged (Fig. 3J-3M). Altogether, these results suggest that the effects achieved by decreasing the levels of NRSF/REST are similar to those when NRSF/REST activity was blocked and thus implicate NRSF/REST as a mediator of NO-induced NPC differentiation.

Altered Chromatin Modifications Are Reversed After Blocking of NRSF/REST Function
In order to understand the molecular mechanisms underlying the neuronal to glial fate-change after NO-exposure, we investigated chromatin modifications and expression levels of genes that are specifying the neuronal and glial lineages. The examined genes included the proneural factors Neurog2 (Ngn2), Ascl1, and NeuroD1 as well as the glial lineage genes Olig2 and Sox9 [35,37]. In the analysis, we also included Hes1, which blocks proneural gene expression, and thus plays an important role in the acquisition of the glial lineage [36]. Most of these genes (Ascl1, NeuroD1, Olig2, and Hes1) have previously been reported to be bound by NRSF/REST in mouse embryonic stem (ES) cells, where RE-1 sites can be found within the reported NRSF/REST binding areas for NeuroD1 and Hes1 [30,38]. Moreover, several of the genes have also been reported to be bound by NRSF/REST in NPCs (Ascl1, NeuroD1, and Hes1) [38,39]. In addition, expression levels of Hes1, Olig2, and Sox9 were shown to be upregulated following reduction in NRSF/REST levels in a NPC like cell line (NT2) [30]. Thus, previous studies indicate direct as well as indirect NRSF/REST involved regulation of these genes. We performed ChIP-qPCR experiments targeting the repressive histone modification mark H3K27me3, as well as the activating modification H3K27Ac, both of which previously have been connected to NRSF/REST activity [29,30]. Acetylation and methylation of H3K27 are mutually exclusive events at promoter sites of active and silent genes, respectively [40,41], and we therefore focused on the promoter regions of the investigated genes. Interestingly, within the promoter regions of the tested neuronal lineage restricted genes, Ngn2, NeuroD1, and Ascl1, and for the proneural blocking gene Hes1 we found H3K27me3 levels to be increased, whereas the levels of the active mark H3K27Ac were decreased after NO-exposure (Fig. 4A, 4B). Furthermore, in NO-exposed cultures where NRSF/REST function was blocked by dnREST transfection, the changes in H3K27me3 and H3K27Ac levels were less dramatic, resembling those levels observed in the control cultures (Fig. 4A,  4B). However, within the promoter regions of the tested glial restricted genes, Sox9 and Olig2, the situation was opposite and H3K27me3 levels were decreased, whereas H3K27Ac levels were found to be increased. Also for the glial-restricted genes these histone modification changes were reversed by transfection with dnREST. We did not identify any NO-induced histone modification changes within the negative control region; the promoter region of the cytokine gene TNF (tumor necrosis factor) (Fig. 4A,  4B). Furthermore, no significant changes in expression levels of these genes were observed. However, there was a trend that corresponded to increased/decreased levels of H3K27me3 and H3K27Ac; Ngn2, Ascl1, and Hes1 had decreased expression levels whereas the glial-restricted genes Sox9 and Olig2 had increased expression levels after NO-exposure (Fig. 4C). NeuroD1 levels could not be detected in undifferentiated cultures (Fig. 4C). In conclusion, the increased glial potential in NO-exposed NPCs involves NRSF/REST function at certain glial and neuronal gene promoter regions and suggests NRSF/REST as an important target for effects of innate immunity.

NRSF/REST Is Upregulated in NPCs from Adult Human After NO-Exposure
As the level of NO production in the human CNS correlates with disease activity in various neuroinflammatory conditions [16]- [20,42] it is reasonable to question whether adult human NPCs are affected by pathological levels of NO. To address this question we analyzed if NRSF/REST was upregulated after NO-exposure also in adult human NPCs and if the differentiation capacity of these cells was altered after NOexposure. Human tissue biopsies, restricted to the wall of the lateral ventricle (SVZ area), were dissociated and cells cultured as neurospheres during 4-6 weeks following dissociation to single cells and plating onto PDL-coated coverslips. After 24 hours, cells were fixed and immune-reactively labeled for NPC markers Sox2 and Pax6 (Fig. 5A-5D), or exposed to NO via DETA-NONO:ate 0.1 mM (or control medium) in a similar fashion as previously described experiments in rat NPCs. After 24 hours of NO-exposure, a 2.4-fold increase in NRSF/REST expression was detected compared to control cells (Fig. 5E), suggesting a possible function of NRSF/REST after NOexposure in human NPCs. Moreover, after 11 days of differentiation fewer Tuj1-positive neurons could be detected in the NO-exposed cultures compared to control cultures (2% 6 4% and 26% 6 10%, respectively) (Fig. 5F, 5I, 5H). Cells expressing the astrocyte marker GFAP occurred in both NO-exposed and control cultures (Fig. 5G, 5J, 5K). However, oligodendrocytes expressing O4 could not be detected in the control or NOexposed cultures after 11 days under differentiation conditions (data not shown). Taken together, these results demonstrate that pathological levels of NO increase the expression of NRSF/REST and affect lineage specification of the neural progenitor population from the adult human brain.

DISCUSSION
After CNS injury and in other neuroinflammatory conditions, the environment within the CNS changes rapidly. Innate immune responses are activated, releasing free radicals and reactive oxygen species in the surrounding tissue [16,43]. We have previously hypothesized that inflammatory cues may act to direct NPC-fate specification and part of the innate inflammatory response is the excess release of NO via iNOS synthase [9], which has been shown to affect NPCs in vitro by reducing their neurogenic capacity and promoting gliogenesis [23]. Here, we demonstrate that the NO-induced neuronal to glial lineage change in NPCs is regulated at the transcriptional level and depends on the function of the transcription factor NRSF/REST. Degeneration of CNS tissue commonly occurs during various neuroinflammatory conditions. In MS, for example, oligodendrocyte and myelin destruction lead to demyelination of neuronal axons and subsequent neuronal damage [44]. Moreover, the formation of glial scars after neurotrauma is important to limit the damage and inhibit axonal loss [45,46]. Glial scar tissue consists of several different cell types where astrocytes derived from resident NPCs constitute a considerable portion [46,47]. In MS, conversely, replacement of degenerated oligodendrocytes is evident within the MS lesions [48,49]. This suggests that the affected tissue requires a rapid contribution of glial cells and several studies demonstrate an increased formation of glial cells that migrate from proliferative areas toward the site Bergsland, Covacu, Estrada et al. of injury [3,47,50,51]. In addition, in mice induced with experimental autoimmune encephalomyelitis, progenitor cells from the SVZ showed less neurogenic capacity, but an increased capacity of generating Olig2-positive oligodendrocyte precursors [52], which is in line with the NPC fate-changes observed in this study after NO-exposure in vitro.
The transcription factor NRSF/REST is required for oligodendrocyte differentiation and the formation of proper ratio of neurons and glia in the developing rodent CNS [27,28]. Here, we show that NRSF/REST is upregulated in NPC cultures that have been exposed to pathological levels of NO. Furthermore, we present functional evidence that NRSF/REST is required for the NO-induced neuronal to glial fate-change in differentiating rat NPCs. Blocking of NRSF/REST function, as well as reducing the NRSF/REST levels with siRNA in NO-exposed NPCs, reversed the effect from NO, which suggests that NRSF/REST is a key factor in converting the NPC lineage specification from neuronal to glial.
NRSF/REST has an essential role during NPC differentiation in the adult CNS, where one of its functions is to repress proneural gene expression [39]. Here, we show that the chromatin at the promoter regions of Ngn2, NeuroD1, Ascl1, and Hes1 is set in a repressive state following NO-exposure, demonstrated by an increase in H3K27me3 and decrease in H3K27Ac modifications. In contrast, at the promoter regions of the glial lineage genes, Olig2 and Sox9, NO-exposure instead leads to the formation of an active chromatin state. We noticed a trend in gene expression changes that coincided with the changed levels of H3K27 methylation and acetylation. However, the changes were not significant and were possibly reflecting the parts of the NO-exposed : ChIP analysis on rat neural stem/progenitor cells (NPCs) exposed to NO (empty bars), nonexposed (checked bars), or exposed to NO and transfected with dnREST (black bars) was performed using antibodies against H3K27me3 (A) or H3K27Ac (B) for Ngn2, Ascl1, Hes1, NeuroD1, Sox9, and Olig2 promoter regions. The promoter region for the TNF gene has been included as a negative control. Results are presented as fold enrichment over IgG for the specific antibody, relative to the fold enrichment over IgG for H3 from each experiment. Error bars represent the SD of triplicate qPCR measurements from one representative experiment out of four, ***, p < .001; **, p < .01; *, p < .05 (Student's t test). (C): Gene expression analysis (Ngn2, Ascl1, Hes1, NeuroD1, Sox9, and Olig2) on rat NPCs exposed to NO, nonexposed, or exposed to NO and transfected with dnREST. NeuroD1 levels could not yet be detected in the NPC cultures. Results are presented as log scale mean-6 SEM relative to the control cultures. Abbreviations: REST, repressor element-1 silencing transcription; TNF, tumor necrosis factor.  (Fig. 2). We cannot exclude the possibility that alterations in gene expression and histone modifications at specific gene promoters may depend on the changes in expression of NPC genes rather than a direct effect of NRSF/REST. For instance, previous reports suggest that Sox9 is indirectly regulated by NRSF/REST via miR-124, whereby NRSF/REST repression of miR-124 allows Sox9 expression, which in turn promotes glial differentiation [53,54]. Thus, it is reasonable that the increased Sox9 levels after NO-exposure are mediated through miR-124 repression by NRSF/REST. Furthermore, the regulation of Ngn2 could also be an indirect event since the binding of NRSF/REST to this gene has not been identified in any cell-type as far as we know. However, a low affinity NRSF/REST binding motif has been identified within this gene, suggesting that NRSF/REST activity could include additional binding motifs and target genes [55]. Nevertheless, together our data provide evidence for the involvement of NRSF/REST in establishing glial commitment in NO-exposed NPCs.

CONCLUSION
Our results demonstrate that pathological levels of NO promote the expression of NRSF/REST, which affect lineage specification of the neural progenitor population from the adult brain. In conclusion, our findings implicate NRSF/REST as a key factor in the crosstalk between the NPC population and innate immunity and suggest a molecular mechanism by which signaling from inflamed tissue regulates the lineage fate of NPCs of the adult CNS.

ACKNOWLEDGMENTS
We are grateful to Britt Meijer for assistance with human NPC culturing. This work is supported by the Swedish Research Council (K2013-62X-20724-06-3); Torsten and Ragnar Soderberg Foundation; The Swedish Neuro Association; and Karolinska Institutet Theme-center for Regenerative Medicine.

AUTHOR CONTRIBUTION
M.B.: conception and design, financial support, provision of study material, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; R.C.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; C.P.E.: data analysis and interpretation and final approval of manuscript; M.S.: financial support, provision of study material, manuscript