Long noncoding RNA mediates stroke-induced neurogenesis

Abstract

Neurogenesis contributes to poststroke recovery. Long noncoding RNAs (lncRNAs) participate in the regulation of stem cell self-renewal and differentiation. However, the role of lncRNAs in stroke-induced neurogenesis remains unknown. In this study, we found that H19 was the most highly upregulated lncRNA in neural stem cells (NSCs) of the subventricular zone (SVZ) of rats subjected to focal cerebral ischemia. Deletion of H19 suppressed cell proliferation, promoted cell death, and blocked NSC differentiation. RNA sequencing analysis revealed that genes deregulated by H19 knockdown were those that are involved in transcription, apoptosis, proliferation, cell cycle, and response to hypoxia. H19 knockdown significantly increased the transcription of cell cycle-related genes including p27, whereas overexpression of H19 substantially reduced expression of these genes through the interaction with chromatin remodeling proteins EZH2 and SUZ12. Moreover, H19 regulated neurogenesis-related miRNAs. Inactivation of H19 in NSCs of ischemic rats attenuated spontaneous functional recovery after stroke. Collectively, our data provide novel insights into the epigenetic regulation of lncRNAs in stroke-induced neurogenesis.

1 |. INTRODUCTION

Stroke is a leading cause of disability worldwide with limited treatment options.¹ Neural stem cells (NSCs) are located in the subventricular zone (SVZ) of the lateral ventricle and the subgranular zone (SGZ) of the dentate gyrus (DG) in the hippocampus.² NSCs from the SVZ continually generate neuroblasts and oligodendrocyte progenitor cells. The neuroblasts migrate to the olfactory bulb where they differentiate into interneurons. Neurogenesis contributes to limited spontaneous recovery after stroke.^3,4 Elucidating the underlying molecular mechanisms in poststroke neurogenesis could provide new therapies to amplify endogenous neurogenesis and to improve neurological function during stroke recovery.

Long noncoding RNAs (lncRNAs) are a class of regulatory RNAs that are longer than 200 nucleotides, lack protein-coding potential, and are implicated in mediating physiologic and pathologic processes in mammals.^5,6 LncRNAs mediate gene expression via chromatin modification, translational and posttranslational regulation.⁷ Recent studies show that lncRNAs express in a lineage-specific manner, suggesting that lncRNAs may have a function in particular cell types, including cells in the brain.^8,9 Wang et al. reported that the lncRNA expression profiling was altered in the sera of stroke patients and the ischemic brains of experimental rodents.¹⁰ Moreover, the roles of lncRNAs in ischemic stroke have been recently studied in different brain cell types, including neurons, glial cells, and vascular cells. Among them, LncRNA-N1LR, SNHG1, C2dat1, and MALAT1 have been shown to facilitate neuroprotection after stroke, whereas TUG1, GAS5, FosDT, and MEG3 exacerbate ischemic brain injury.^11–18 SNHG12, MEG3, and MALAT1 enhance stroke-induced angiogenesis.^19–21 The functions of lncRNAs in neurogenesis are just beginning to be appreciated.^22,23

H19, encoding a polyadenylated lncRNA of ~2.6 kb, is located on 11p15.5 in humans and chromosome 7 in mice.²⁴ H19 is the earliest defined maternally imprinted lncRNA and it is involved in diverse biological processes and pathology as a multifunctional regulatory transcript.^25–27 Hypermethylation or hypomethylation of H19 during development induces human genetic syndrome, for example, Beckwith-Wiedemann syndrome and Silver-Russel syndrome with clinical features of asymmetry and growth retardation.²⁸ In adults, H19 maintains the character of hematopoietic stem cells and pluripotent stem cells and promotes skeletal muscle differentiation and regeneration.^27,29,30 Also, H19 regulates adipogenesis of bone marrow cells and embryonic stem cell differentiation.^31,32 Recently, six single-nucleotide polymorphisms (SNPs) of H19 were detected in ischemic stroke patients, whereas a preclinical study demonstrated that H19 is involved in acute stroke-induced neuroinflammation.¹⁰ These data suggest a critical role of H19 in the progression of stroke.

In the present study, we analyzed the lncRNA profiles alteration in the NSCs before and after stroke, investigated the role of H19 in mediating stroke-induced neurogenesis, and demonstrated that lncRNA H19 in the NSCs mediate stroke-augmented neurogenesis by potentially affecting chromatin remodeling proteins.

2 |. MATERIALS AND METHODS

2.1 |. Animals

All experimental procedures in this study were carried out according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee (IACUC) of Henry Ford Hospital.

2.2 |. Middle cerebral artery occlusion (MCAO)

Male Wistar rats (~10 weeks, purchased from Jackson Laboratory) were anesthetized with 2% isoflurane. Exposed the common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA) on the right side and gently insert a nylon filament (4 ± 0) from the ECA into the lumen of the ICA.

2.3 |. Neural stem cell culture

Neural stem cells (NSCs) were isolated from nonischemic rats and ischemic rats on day 7 after MCAO when stroke-increased neurogenesis reaches a peak, as previously described.^33,34 Briefly, the NSCs localized to the lateral walls of the lateral ventricles of the SVZ in sagittally dissected brain slices were isolated and digested in 0.05% trypsin (MilliporeSigma) at 37°C for 1 hour and the digestion was stopped with an equal volume of 0.1% Trypsin inhibitor (MilliporeSigma). For NSC proliferation, NSCs were cultured in a proliferation medium which consisted of Dulbecco’s modified Eagle’s medium (DMEM)/F-12 medium containing 1% B27 (Life Technologies), epidermal growth factor (EGF, 20 ng/mL, R&D System, USA), and basic fibroblast growth factor (bFGF, 20 ng/mL, R&D System). The formation of neurospheres was measured and characterized.³⁵ For NSC differentiation, the primary neurospheres were mechanically dissociated and seeded as single cells. The cells were incubated in the differentiation medium which consisted of DMEM/F-12 medium and 2% fetal bovine serum but without the growth factors. Passages 2 to 5 of the primary NSCs were used in the following experiments.

2.4 |. LncRNA array analysis

Total RNA was isolated from NSCs using TRIzol reagent and quantified using the NanoDrop ND-1000. Then, each RNA sample was amplified and transcribed into cDNA. The labeled cDNAs were then hybridized onto Rat LncRNA Array v2.0 (Arraystar). The arrays were scanned using the Agilent Scanner G2505C.

2.5 |. mRNA-sequencing analysis

Approximate 2 μg total RNA from each sample was used to prepare the sequencing. Agilent 2100 Bioanalyzer was used to qualify the completed libraries, and RNA content was quantified by real-time PCR absolute quantification method. The differentially expressed genes and transcripts were filtered using R package Ballgown. The novel genes were predicted by comparing to the reference annotation using StringTie and Ballgown.

2.6 |. Gene ontology (GO)

Gene ontology was used to describe gene and gene product properties. The ontology covers biological process, cellular component, and molecular function. The P-value denotes the significance of GO terms enrichment in the deregulated genes (a P-value ≤.05 is considered significant).

2.7 |. LncRNA-protein-coding gene coexpression analysis

The coexpression network for lncRNAs and the associated protein-coding genes was constructed.³⁶ For each pair of genes, Pearson correlation coefficients (PCCs) between the normalized data of provided lncRNAs and all coding gene expression were calculated. The coding genes with abs (PCC) ≥0.9 and FDR ≤0.05 were selected.

2.8 |. H19 knockdown with clustered regularly interspaced short palindromic repeat-associated system (CRISPR) associated 9 (Cas9) genome editing

CAG-mKate2-Cas9 expression plasmid DNA, trans-activating crRNA (tracrRNA), and single-guide RNAs (sgRNA nontargeting control [sgRNA-NC] and sgRNA-H19) were purchased from Dharmacon (Lafayette, Colorado). Four sgRNAs containing seed sequence targeting H19 were synthesized by Dharmacon. For each transfection, NSCs (~10⁶) were electroporated in electroporation buffer containing 10 μg CAG-mKate-Cas9 expression plasmid DNA using nucleofector electroporation with program A33 (Amaxa Biosystems, Germany). The cells were incubated for 48 hours and transfected with a sgRNA-tracrRNA mixture using the same electroporation program. After an additional 48 hours, cells were used for the following procedure.

2.9 |. Production of adeno-associated virus carrying H19 construction and delivery

Guide sequences and guide RNA scaffold were subcloned into the corresponding sites of pAV-FH (Vigene Biosciences, Maryland) containing the adeno-associated virus (AAV) vector backbone with Cre recombinase. Mouse Ascl1 clone (MPRM39894, GeneCopoeia) was subcloned into the pAV-FH plasmid. Infectious recombinant AAV vector particles were generated in HEK293T cells by a cross-packaging approach whereby the vector genome was packaged into AAV serotype-9. Viral stocks were obtained through PEG precipitation and CsCl₂ gradient centrifugation and stored at −80°C until use. The physical titer of recombinant AAVs was quantified using real-time PCR against a standard curve of a plasmid containing vector genome (vg).

For in vitro infection, NSCs were incubated with AAV-containing media (10⁴ vg per cell) for 12 hours and an equal volume of media was added to the cultures. Seven days after the infection, NSCs were lysed and gene expression was tested.

For in vivo ablation of H19, floxed Cas9 knockin mice were used. Briefly, AAV-H19 gRNA (2 μL, 2 × 10¹³ vg/mL) was injected into the ipsilateral SVZ and hippocampus of transgenic mice. Mice that received AAV-GFP injection were used as a control. One week later, all the animals were subjected to MCAO surgery. Three weeks after MCAO, mice were euthanized and their brains were processed for neurogenesis analysis according to our published protocol.³⁴

2.10 |. RNA immunoprecipitation (RIP)

RIP was performed using Magna RIP kit (MilliporeSigma). One hundred microliters of NSCs lysates were incubated with magnetic protein A/G beads precoated with SUZ12 or EZH2 antibody (Abcam) at 4°C overnight. Digested the protein with Proteinase K and isolated the immunoprecipitated RNA for quantitative RT-PCR (qRT-PCR) analysis. The presence of the binding targets was detected using the respective primers.

2.11 |. Chromatin Immunoprecipitation (ChIP) coupled with quantitative PCR

NSCs were cross-linked with 1% (v/v) formaldehyde for 10 minutes and sonicated until DNA was sheared to the lengths of 200 ~ 1000 bps. Chromatin extracts were immunoprecipitated with H3K27me3 antibody (Abcam) on Protein A Agarose/Salmon Sperm DNA. The ChIP DNA was isolated and the enrichment of H3K27me3 in the promoter of p27 was analyzed using real-time PCR and SYBR Green probes.

2.12 |. Chromatin isolation by RNA purification (ChIRP)

ChIRP was performed using Magna ChIRP RNA Interactome Kit (MilliporeSigma) following the manufacturer’s protocol. Biotinylated H19 probes were designed by Biosearch Technologies (Petaluma) and labeled as odd and even groups.

2.13 |. Statistical analysis

Data are presented as mean ± SE of the mean (SEM). The GraphPad Prism software (GraphPad Software Inc, California) was used for statistical analysis. Nonparametric one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was performed for multiple groups comparisons. Two-tailed Student’s t test was performed for two-group comparisons. The ImageJ software was used to quantify the band intensity in Western blot images. Quantification was performed from at least three independent experiments. A value of P < .05 was considered significant.

3 |. RESULTS

3.1 |. The alteration of lncRNAs and mRNAs profiles in adult neural stem cells after stroke

We first performed lncRNA profile analysis in NSCs isolated from nonischemic and ischemic rats. In total, 6101 lncRNAs were detected. Among the significantly and differentially expressed transcripts, 155 lncRNAs were downregulated (<0.33-fold), 86 lncRNAs were upregulated (>1.5-fold) after stroke (Figure 1A). The top three elevated lncRNAs in ischemic NSCs were lnc-NR_027324 (H19, 20-fold), lnc-EF094477 (6-fold), and lnc-BC090003 (6-fold), while top three decreased ones were lnc-M64384 (−5-fold), lnc-MRAK013682 (−4-fold), and lnc-MRAK051099 (−4-fold). mRNA-sequencing analysis showed a total of 11 591 mRNAs were detected in adult NSCs. 409 were significantly upregulated (>1.5-fold) and 255 were downregulated (<0.33-fold) in ischemic NSCs compared with nonischemic NSCs (Figure S1). GO analysis demonstrated that the stroke-induced upregulated mRNAs in NSCs were correlated with chemokine signaling and cell cycle pathways (Figure 1B), whereas the downregulated mRNAs were associated with calcium signaling, glutathione metabolism, and lysosome pathways (Figure 1C).

FIGURE 1.

Bioinformatic analysis of deregulated lncRNAs. A, Heatmap of lncRNA array analysis in control and ischemic neural stem cells (NSCs). B,C, Function gene ontology (GO) analysis of differentially upregulated (B) and downregulated (C) genes in stroke NSCs as compared with control NSCs

3.2 |. Networks of lncRNAs and protein-coding genes in NSCs

LncRNAs regulate the expression or function of genes in cis.³⁷ In ischemic NSCs, four upregulated and two downregulated lncRNAs were positively correlated with the genes in cis, but no negative correlation was detected. Gene coexpression networks were built according to PCCs between lncRNA and mRNA expression. The network consists of upregulated lncRNAs showed 1679 positive matched lncRNA-mRNA correlations (Figure S2A). The network consists of downregulated lncRNAs identified 1944 negative matched lncRNA-mRNA correlations (Figure S2B). Gene set enrichment analysis (GESA) was performed to predict the functions of coexpressed genes. The coexpressed genes positively modulated by lncRNAs were related to cognition, synapse plasticity, neuron development, mitochondrion, calcium channel regulator, neurotransmitter signal, etc. The coexpressed genes negatively regulated by lncRNAs were associated with cytokine production, DNA replication, cell adhesion, NF-κB signaling, metabolic process, T cell process, and so on.

3.3 |. Stroke upregulates lncRNA H19 expression in NSCs

H19 (NR_027324) was the most upregulated lncRNA after stroke. H19 expression was significantly increased by 19- and 9-fold in ischemic SVZ and DG, compared with the nonischemic tissues (Figure 2A). In situ hybridization (ISH) further confirmed that H19 expression was enhanced in ischemic SVZ (Figure 2B). In contrast, the content of H19 was not increased in cortex and striatum (Figure 2A). Using fluorescence-activated cell sorting (FACS), we isolated Ascl1 positive NSCs from ischemic Ascl1-CreER^T2;tomato reporter mice and DCX positive neuroblasts from and DCX-CreER^T2;tomato reporter mice. QRT-PCR analysis revealed that H19 was mainly expressed in Ascl1⁺ NSCs rather than DCX⁺ neuroblasts (Figure 2C).

FIGURE 2.

Stroke elevates lncRNA H19 expression in neural stem cells (NSCs). A, Expression of H19 in nonischemic and ischemic SVZ, dentate gyrus (DG), cortex, and striatum. (^***P < .001; n = 3). B, Representative images of ISH histochemistry using a digoxigenin-labeled H19 probe in SVZ (red outlines the NSCs, LV: lateral ventricle, Str: striatum). H19 content was enriched after stroke. C, Comparison of H19 expression in Ascl1 lineage NSCs and DCX lineage neuroblasts (^***P < .001; n = 3). D, Gene editing with Cas9 expression plasmid and sgRNA: tracrRNA was performed by cotransfecting all components into NSCs. E, We designed four paired of sgRNAs for H19 (sgH19) and observed over 70% knockdown of sgH19-1 in NSCs (F). (**P < .01, n = 3). G, Immunocytochemistry shows BrdU⁺, Tuj1⁺, in NSCs transfected by sgH19-1 or control nontargeting sgRNA. Bar = 50 μm. White arrows indicate positive cells. H,I, Quantitative data analysis shows the percentage of BrdU⁺ (H), Tuj1⁺ (I), NG2⁺ (I) and GFAP⁺ cells (I) cells (^**P < .01, ^***P < .001, n = 3). J, Caspase3/7 activity was substantially increased in NSCs transfected with sgH19-1 compared with cells transfected with control sgRNA (^***P < .001, n = 3)

3.4 |. Ablation of H19 suppresses the proliferation and neuronal differentiation of NSCs in vitro

We knocked down H19 in ischemic NSCs using CRISPR-Cas9 genome editing, and examined the alteration in the self-renewal and multipotency of ischemic NSCs (Figure 2D). Among the four sgRNAs designed to pair H19 (sgH19), sgH19-1 caused a 70% reduction in H19 expression compared with nontargeting control sgRNA (sgRNA-NC) in ischemic NSCs (Figure 2D–F). Immunostaining analysis demonstrated that the H19 knockdown significantly decreased the proliferation rate and increased the apoptosis of ischemic NSCs, as assayed by the percentage of BrdU⁺ cell (Figure 2G,H) and Caspase-3/7 activity (Figure 2J), compared with sgRNA-NC treated cells. Moreover, H19 knockdown significantly reduced the percentage of Tuj1⁺ neurons, NG2⁺ oligodendrocyte progenitor cells (OPCs), and GFAP⁺ astrocytes in ischemic NSCs (Figure 2G–I).

3.5 |. Effects of H19 on the key genes in the proliferation and survival of NSCs

To elucidate the effects of H19 in the regulation of NSCs growth and differentiation, RNA-seq analysis was performed to examine transcriptome alteration after H19 knockdown. Compared with ischemic NSCs treated with sgRNA-NC, 356 genes were upregulated (>2-fold), whereas 228 genes were downregulated (<0.5-fold) in H19 knockdown cells. GO analysis revealed that upregulated genes were closely related to oxidative response, NF-κβ signaling, cell cycle, translation, proliferation, apoptosis, response to hypoxia, and transcription (Figure 3A). Many of these upregulated genes located both in nucleus and cytoplasm (Figure 3B), and this result was further confirmed using ISH analysis (Figure 3C). Furthermore, qRT-PCR analysis revealed that H19 content was enhanced in both cytoplasmic and nuclear RNA extractions of ischemic NSCs (Figure 3D).

FIGURE 3.

H19 regulates key genes required for neural stem cell (NSC) proliferation and survival. Gene ontology (GO) analysis demonstrates the functional enrichment (A) and localization (B) of upregulated genes in NSCs upon H19 deletion relative to control. C, RNA-FISH showed that H19 is localized both in the nucleus and cytoplasm. Bar = 10 μm. D, QRT-PCR showed that nuclear H19 is significantly increased in ischemic NSCs after stroke (^***P < .001; n = 3). E, Western blot analysis of NSCs harvested from 7d poststroke shows attenuation of H19 (H19^−/−) by sgH19 induced expression of cell cycle and apoptosis-related proteins including p21, p27, PTEN, and cleaved Caspase-3, nevertheless reduced expression of CDK4 compared with control nontarget sgRNA (left). However, overexpression of H19 with AAV9-H19 cDNA viral vector (H19^+/+) has the opposite effects on the above proteins compared with cells transfected with control empty vector (right). F, The quantification data of protein expression when H19 was knocked down or overexpressed (^**P < .01, n = 3)

We then examined the impact of cytoplasmic and nuclear H19 on gene transcription. Western blot analysis revealed that the expression of cell cycle-related genes, as well as apoptosis-related genes, was increased upon H19 depletion (Figure 3E,F). These genes have been shown to inhibit NSC proliferation and survival.³⁸ In contrast, H19 overexpression (H19^+/+) in NSCs by AAV9-H19 cDNA infection led to the inverse expression pattern of the genes (Figure 3E,F). These data suggest that H19 mediates the transcriptional program that plays critical roles in regulating cell cycles and programmed cell death.

3.6 |. H19 improves repressive H3K27me3 histone modification

Using ChIP coupled with qPCR, we found that repressive H3K27me3 histone modification was highly enriched in the bp −2000 to −1000 from the transcription start site of p27 in ischemic NSCs (Figure 4A). Furthermore, knockdown of H19 directly reduced the enrichment of the H3K27me3 in the promoter of the p27 gene (Figure 4B), leading to the upregulation of p27 expression in NSCs (Figure 3E,F). These results indicate that repressive H3K27me3 histone modification in neurogenesis-related gene transcription is involved in the regulation of H19 in poststroke neurogenesis.

FIGURE 4.

H19 enhances H3K27me3 histone modification through interaction with chromatin remodeling factors. A, Chromatin immunoprecipitation (CHIP) combined qPCR demonstrates the enrichment of H3K27me3 histone modification in the bp −2000 to −1000 region relative to the transcription start sites of p27. B, In H19 knockdown neural stem cells (NSCs), the same strategy was used to test histone modification. No obvious enrichment was observed in any region of p27 promoter. C, RNA immunoprecipitation (RIP) was performed using antibodies against SUZ12, EZH2, or normal rabbit IgG. The association of H19 with these two chromatin proteins was enhanced in ischemic NSCs than in control NSCs (^**P < .01, ^***P < .001, n = 3). D, Chromatin isolation by RNA purification (ChIRP)-Western blot analysis of chromatin proteins immunoprecipitated by DNA probes against H19 in control and stroke NSCs. E,F, Representative Western blot images and quantitative data of proteins in SVZ and SGZ extracted from control and stroke half brain (^**P < .01, ^***P < .001, n = 3)

3.7 |. H19 interacts with SUZ12 and EZH2 chromatin remodeling factors

Polycomb repressive complex 2 (PRC2) induces gene silencing by methylation of histone H3 on K27 via its subunit EZH2.³⁹ We next performed RIP using antibodies against SUZ12 and EZH2 and observed a significant enrichment of H19 to chromatin proteins SUZ12 and EZH2 in NSCs after stroke (Figure 4C). Using ChIRP-Western blot, SUZ12, EZH2, H3K27me3 were observed significantly higher in the purified H19 regulatory complexes isolated from ischemic NSCs compared with those from nonischemic NSCs (Figure 4D). Also, stroke significantly increased SUZ12 and EZH2 in NSCs derived from SVZ and SGZ, whereas histone active marker H3K4me3 exhibited a significant decrease (Figure 4E,F). These data suggest that H19 promotes SUZ12 and EZH2 recruitment after stroke.

3.8 |. H19 regulates its encoding miR-675 expression in NSCs

H19 is the primary precursor of miR-675.⁴⁰ We found the expression of miR-675-5p and −3p was significantly increased after stroke (Figure 5A), which was positively correlated with the upregulation of H19. Furthermore, when the H19 expression was manipulated, the content of miR-675-3p and miR-675-5p showed an identical change trend (Figure 5B). Meanwhile, we observed that the expression of insulin-like growth factor-1 receptor (IGF1R) was significantly reduced in H19 knockdown NSCs (Figure 5C,D), indicating IGF1R was not a direct target gene of H19/miR-675. Further analysis revealed that TGF-β1, a target gene of miR-675, showed an inverse expression pattern of miR-675 (Figure 5C,D). These data suggest that H19/miR-675/TGF-β1 participates in the regulation of poststroke neurogenesis.

FIGURE 5.

H19 mediates poststroke neurogenesis via the regulation of miRNAs. A, Expression of miR-675-3p and miR-675-5p was evaluated by qRT-PCR in neural stem cells (NSCs) extracted from control and stroke half brain (^***P < .001, n = 3). B, Gain- and loss-of-function of H19 were obtained by overexpression (H19^+/+) and knockdown of H19 (H19^−/−) in NSCs extracted from stroke half brain. Content of H19, miR-675-3p and miR-675-5p was measured by qRT-PCR. Both miRNAs showed the same expression trend as H19 (^***P < .001, n = 3). C,D, In H19^−/− ischemic NSCs, the expression of TGF-β1 was increased compared with the nontarget control (NC) group, whereas the IGF1R expression was reduced (^***P < .001, n = 3). E, Expression of H19, miR-675-3p, and miR-675-5p was evaluated by qRT-PCR in NSCs extracted from wide-type (WT) and Ago2 knockout (Ago2-KO) mice (^***P < .001, n = 3). F,G, In Ago2-ablated NSCs, Ago2 expression was suppressed compared with the WT group, whereas the miR-675 target gene, TGF-β1, was increased (^***P < .001, n = 3). H) NSCs from WT and Ago2-KO mice were treated/un-treated with actinomycin D (AMD, 10 μg/mL) and content of H19 were tested 3 hours The degradation of H19 was accelerated in Ago2 ablated NSCs (*P < .05, ^##P < .01, ^ΔΔΔP < .001, ^{^^^}P < .001). I, The expression of miRNAs related to neurogenesis was substantially decreased in H19 knockdown ischemic NSCs compared with the control group measured by qRT-PCR (^***P < .001, n = 3). J,K, In H19 knockdown ischemic NSCs, expression of Dicer was decreased (^**P < .01, n = 3), whereas no obvious difference was observed in Ago2 expression (P = .672, n = 3)

3.9 |. Argonaute 2 is indispensable for H19-regulated miRNAs

Argonaute 2 (Ago2) is a key RNA binding protein in miRNA processing and function.⁴¹ We assessed whether Ago2 was required in the miRNA regulation of H19. NSCs were isolated from transgenic mice with conditional ablation of Ago2 in Ascl1 lineage cells. Western blot analysis showed that Ago2 expression was suppressed compared with wild-type mice (Figure 5F,G). QRT-PCR analysis showed that the ablation of Ago2 in Ascl1 lineage cells significantly reduced levels of H19 and encoding miR-675 compared with NSCs isolated from wild-type mice (Figure 5E). Meanwhile, TGF-β1 was significantly increased in Ago2-KO NSCs (Figure 5F,G). To determine the mechanism underlying Ago2-regulated H19 expression, we blocked the on-going transcription using actinomycin D (AMD) and observed that Ago2-KO enhanced H19 degradation (Figure 5H).

We and others previously demonstrated that miRNAs contribute to stroke-increased neurogenesis.^35,42 However, whether deregulated lncRNAs affect miRNAs that mediate stroke-increased neurogenesis remains unknown. When H19 was knocked down in ischemic NSCs by sgH19, miR-146a, miR-17-92 cluster, miR-21, miR-124a, and miR-23 were substantially suppressed (Figure 5I). Besides, H19 knockdown did not significantly alter Ago2 expression, but substantially reduced the expression of Dicer in NSCs (Figure 5J,K).

3.10 |. H19 inactivation in neural stem cells attenuates motor and cognitive function after stroke

To assess the impact of H19 on poststroke functional outcome, we selectively knocked down H19 in NSCs by injection of AAV-H19 gRNA into the ipsilateral SVZ of floxed Cas9 knockin mice. Compared with ischemic mice that received AAV-GFP, ischemic mice treated with AAV-H19 gRNA exhibited significant impairments of neurological function at 14 and 21 days after MCAO assayed by an array of behavioral tests (Figure 6A,B). Moreover, ischemic mice with H19 knockdown in NSCs significantly diminished their sociability and novel object recognition compared with the mice treated with AAV-GFP (Figure 6C,D). RT-PCR analysis showed the downregulation of H19 both in the SVZ and in the hippocampus (Figure S3). Double-immunofluorescent staining analysis demonstrated that knockdown of H19 in NSCs significantly reduced the numbers of GFP⁺/BrdU⁺ proliferating progenitor cells and GFP⁺/DCX⁺ neuroblasts in SVZ and DG of the hippocampus (Figure 6E–L), as well as reduced GFP⁺/calretinin⁺ newly born neurons in DG (Figure 6M,N).

FIGURE 6.

The deletion of H19 in neural stem cells (NSCs) attenuated neurological outcome and neurogenesis poststroke. Modified neurological severity score (mNSS) score (A) and Foot- fault test (B) results show that H19 knockdown attenuated poststroke motor functional recovery. (*P < .05, ^**P < .01, n = 6). C, Social recognition memory test shows that mice injected with AAV-GFP spent a significantly longer time around the stranger, but the mice injected with AAV-H19 spent an equivalent amount of time between the two cages (*P < .05, n = 6). D, Novel object recognition test shows that compared with AAV-GFP group, H19 knockout mice significantly lost their focused responses to a new bead (N2) (*P < .05, n = 6). Representative images and quantification of GFP⁺/BrdU⁺ proliferating neural progenitors in SVZ (E,F) and DG of hippocampus (G,H). Bar = 50 μm (^**P < .01, ^***P < .001, n = 6). Representative images and quantification of GFP⁺/DCX⁺ neuroblasts in SVZ (I,J) and DG of hippocampus (K,L). Bar = 50 μm (^***P < .001, n = 6). Representative images and quantification of differentiated GFP⁺/calretinin⁺ neurons in DG of hippocampus (M,N). Bar = 50 μm (^***P < .001, n = 6)

4 |. DISCUSSION

In this study, we for the first time provide a comprehensive lncRNA profile in NSCs and analyze the association of lncRNAs with corresponding protein-coding genes after stroke. Ischemic stroke robustly upregulated lncRNA H19 in NSCs. The deletion of H19 suppressed the proliferation and survival and blocked the neuronal differentiation of NSCs, leading to an exacerbation of motor and cognitive deficits after stroke. H19 physically interacted with and regulated levels of enriched SUZ12 and EZH2 chromatin proteins that catalyzed H3K27 trimethylation on p27. Besides, H19 regulated neurogenesis-related miRNAs in the NSCs. These data indicate that H19 contributes to stroke-increased neurogenesis.

The roles of lncRNAs in neurogenesis have been recently unveiled.⁴³ We systematically analyzed transcriptome profile alteration in ischemic and nonischemic NSCs in attempting to identify lncRNAs that are potentially involved in the regulation of neurogenesis. Six lncRNAs were discovered positively associated with their adjacent genes (Table. 1). Stat1 is reported to mediate neurogenesis when activated by interferon-γ.^44,45 Lamp2 is a lysosome marker gene and participates in ischemic neuronal cell damage.⁴⁶ GFAP is a widely used marker for NSCs.^35,47 Rt1.aa, an MHC class antigen, was increased after stroke and correlated with lncRNA NR_002597. The function of Zfhx2 and Cyld in adult neurogenesis has not been documented in adult NSCs after stroke, but both genes are involved in cell proliferation and differentiation.^48,49 PCC analysis revealed numerous integrative networks of deregulated lncRNAs with distant genes. Further investigation is needed to elucidate mechanisms underlying the effects of lncRNA poststroke.

TABLE 1.

The correlation between lncRNA and adjacent genes

LncRNAs	Fold changes	Associated Gene name	Associated_protein_name	Fold changes	Relationship
MRAK088874	2.49	Stat1	signal transducer and activator of transcription	4.41	sense_exon_overlap
MRAK080234	2.19	Lamp2	lysosomal-associated membrane protein 2	2.71	sense_intron_overlap
EF094477	6.38	GFAP	glial fibrillary acidic protein	3.57	sense_exon_overlap
NR_002597	2.29	Rt1.aa	MHC class I RT1.Aa alpha-chain	3.76	sense_intron_overlap
MRAK138051	−1.75	Zfhx2	zinc finger homeobox 2	−3.36	sense_exon_overlap
AY383658	−2.09	Cyld	ubiquitin carboxyl-terminal hydrolase CYLD	−2.23	sense_exon_overlap

The present study showed that H19 was among the most upregulated lncRNAs in ischemic NSCs and H19 knockdown in NSCs led to suppressed proliferation and neuronal differentiation, which was associated with the induction of cognitive deficits. Using a transgenic ablation of doublecortin-expressing cells, with doublecortin as a marker of neuroblasts, and nestin transgenic mice, recent studies have demonstrated that ablation of newly generated neuroblasts or NSCs significantly suppresses the neurogenesis in the SVZ and the SGZ of the hippocampus, leading to the impairment of cognition.^50,51 Moreover, the ablation of newly generated neuroblasts and NSCs in the ischemic brain reduces spontaneous functional recovery.⁵² It is reasonable to assume that H19 may improve neurological outcomes after stroke.

SUZ12 and EZH2 are polycomb repressive complex 2 (PRC2) core proteins that catalyze the di/trimethylation of H3K27.^53,54 We observed that upregulated H19 was physically associated with SUZ12 and EZH2 in NSCs. The absence of H19 markedly reduced levels of EZH2 and H3K27me3 at the promoters of cell cycle-related genes. Also, we demonstrated that H19 not only directly regulates its encoding miRNAs but also indirectly impacts its nonneighboring miRNAs, which is consistent with previous studies showing that lncRNAs can function as molecular decoys of miRNAs.⁶ Meanwhile, we found that ablation of Ago2 in Ascl1 lineage NSCs reduces the expression of H19 and its stability and related miRNAs. In comparison, knockdown of H19 has no significant effect on Ago2 expression, suggesting Ago2 is required for H19-mediated miRNA modulation but not vice versa. Interestingly, we revealed that knockdown of H19 reduces Dicer expression, another key factor in the miRNA biogenesis, which provides new insight into the role of lncRNA in miRNA biogenesis. Nevertheless, it remains to be determined how H19 regulates the transcription or translation of the Dicer gene in NSCs. Our novel observations suggest that H19 regulates stroke-induced neurogenesis by altering the epigenetic states to modulate the transcriptional silencing of genes as well as by the interaction with noncoding miRNAs, which underscore molecular mechanisms underlying the effect of H19 on adult neurogenesis.

Recently, it was reported that H19 is involved in the neuroinflammation and autophagy at the acute stage of stroke, indicating a detrimental effect of H19 on stroke.^10,55 The differences between our study and the prior reports may lie in the stages poststroke that were focused on. In this study, we mainly investigated the differential role of H19 at the neurorestorative stage. GESA analysis revealed that H19 was correlated with immune responses signaling pathways, such as Toll-like receptors (TLR). In the later stage poststroke, the immune response can promote NSCs proliferation and neuroblast migration.^56,57

Adult neurogenesis declines in aging rodents and primates. During aging, NSCs and their progenitors display reduced proliferation and new neuron generation, which contribute to age-related cognitive impairment and reduced neuroplasticity for brain repair.^58,59 Compared with young animals, aged animals subjected to stroke showed more severe impairment of neurogenesis and worse neurological recovery, suggesting that age affects the stroke-induced neurogenesis and neurological outcomes. Dysregulation of lncRNAs is associated with age-associated pathologies and related neurodegenerative disorders.^60,61 It remains unknown if lncRNAs will contribute to the stroke-induced neurogenesis in aging rodents. Further studies to investigate the function of H19 in neurogenesis using aging animals are warranted.

5 |. CONCLUSION

Therapies designed to boost neurogenesis following stroke offers a potentially effective approach to treatment aimed at improving recovery. Our study may set the groundwork for the development of a lncRNA-based therapy for improvement of neurological function after stroke.

Significance statement.

Adult neurogenesis contributes to neurological function. Elucidating the underlying molecular mechanisms in post-stroke neurogenesis could provide new therapies to amplify endogenous neurogenesis and to improve neurological function during stroke recovery. The present study for the first time reveals that stroke substantially changes lncRNA profiles and lncRNA-mRNA coexpression networks in the neural stem cells. Furthermore, it demonstrates that lncRNA H19 mediates stroke-augmented neurogenesis by recruiting chromatin remodeling proteins and regulating microRNA expression to modulate neurogenesis-related transcription. Our results provide new insights into the epigenetic control of adult neurogenesis after cerebral ischemia.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author.